Cycle Polyamide Motif for Recognition of the Minor Groove of DNA

Article abstract of DOI:10.1021/ja983206x

Motifs for covalent linkage of side-by-side complexes of pyrrole-imidazole (Py-Im) polyamides in the DNA minor groove provide for small molecules that specifically recognize predetermined sequences with subnanomolar affinity. Polyamide subunits linked by a turn-specific γ-aminobutyric acid (γ) residue form hairpin polyamide structures. Selective amino-substitution of the prochiral α-position of the γ-turn residue relocates the cationic charge from the hairpin C terminus. Here we report the synthesis of pyrrole resin as well as a solid-phase strategy for the preparation of cycle polyamides. The DNA binding properties of two eight-ring cycle polyamides were analyzed on a DNA restriction fragment containing six base pair match and mismatch binding sites. Quantitative footprint titrations demonstrate that a cycle polyamide of sequence composition cyclo-(γ-ImPyPyPy-(R)^H2Nγ-ImPyPyPy-) binds a 5′-AGTACT-3′ site with an equilibrium association constant K_a = 7.6 × 10¹⁰ M^-1, a 3600-fold enhancement relative to the unlinked homodimer (ImPyPyPy-β-Dp)₂·5′-AGTACT-3′, and an 8-fold enhancement relative to hairpin analogue ImPyPyPy-(R) ^H2Nγ-ImPyPyPy-C3-OH·5′-AGTACT-3′. Replacement of a single nitrogen atom with a C-H (Im→Py) regulates affinity and specificity of the cycle polyamide by 2 orders of magnitude. The results presented here suggest that addition of a chiral γ-turn combined with placement of a second γ-turn within the hairpin structure provides a cycle polyamide motif with favorable DNA binding properties.

Full text of DOI:10.1021/ja983206x

J. Am. Chem. Soc. 1999, 121, 1121-1129
1121
Cycle Polyamide Motif for Recognition of the Minor Groove of DNA
David M. Herman, James M. Turner, Eldon E. Baird, and Peter B. Dervan*
Contribution from the DiVision of Chemistry and Chemical Engineering, California Institute of
Technology, Pasadena, California, 91125
ReceiVed September 8, 1998
Abstract: Motifs for covalent linkage of side-by-side complexes of pyrrole-imidazole (Py-Im) polyamides
in the DNA minor groove provide for small molecules that specifically recognize predetermined sequences
with subnanomolar affinity. Polyamide subunits linked by a turn-specific γ-aminobutyric acid (γ) residue
form hairpin polyamide structures. Selective amino-substitution of the prochiral R-position of the γ-turn residue
relocates the cationic charge from the hairpin C terminus. Here we report the synthesis of pyrrole resin as well
as a solid-phase strategy for the preparation of cycle polyamides. The DNA binding properties of two eight-
ring cycle polyamides were analyzed on a DNA restriction fragment containing six base pair match and mismatch
binding sites. Quantitative footprint titrations demonstrate that a cycle polyamide of sequence composition
cyclo-(γ-ImPyPyPy-(R)H2Nγ-ImPyPyPy-) binds a 5′-AGTACT-3′ site with an equilibrium association constant
Ka ) 7.6 × 1010 M-1, a 3600-fold enhancement relative to the unlinked homodimer (ImPyPyPy-â-Dp)2‚5′-
AGTACT-3′, and an 8-fold enhancement relative to hairpin analogue ImPyPyPy-(R)H2Nγ-ImPyPyPy-C3-
OH‚5′-AGTACT-3′. Replacement of a single nitrogen atom with a C-H (ImfPy) regulates affinity and
specificity of the cycle polyamide by 2 orders of magnitude. The results presented here suggest that addition
of a chiral γ-turn combined with placement of a second γ-turn within the hairpin structure provides a cycle
polyamide motif with favorable DNA binding properties.
Introduction
completes the four base pair (bp) code.3 The linker amino acid
γ-aminobutyric acid (γ) connects polyamide subunits CfN in
a “hairpin motif”, and these ligands bind to predetermined target
sites with >100-fold enhanced affinity relative to dimers.2,9 In
both published and unpublished work, eight-ring hairpin polya-
mides have been found to regulate transcription and permeate
a variety of cell types in culture.1 Because topology could
Small molecules that target predetermined DNA sequences
have the potential to control gene expression.1 Polyamides
containing the three aromatic amino acids 3-hydroxypyrrole
(Hp), imidazole (Im), and pyrrole (Py) are synthetic ligands that
bind to predetermined DNA sequences with subnanomolar
affinity.2,3 DNA recognition depends on a code of side-by-side
amino acid pairings oriented N-C with respect to the 5′-3′
direction of the DNA helix in the minor groove.2-9 An
antiparallel pairing of Im opposite Py (Im/Py pair) distinguishes
G‚C from C‚G and both of these from A‚T/T‚A base pairs.4 A
Py/Py pair binds both A‚T and T‚A in preference to G‚C/C‚
G.4,5 The discrimination of T‚A from A‚T using Hp/Py pairs
(5) For Py/Py pairing, see: (a) Pelton, J. G.; Wemmer, D. E. Proc. Natl.
Acad. Sci. U.S.A. 1989, 86, 5723-5727. (b) Pelton, J. G.; Wemmer, D. E.
J. Am. Chem. Soc. 1990, 112, 1393-1399. (c) Chen, X.; Ramakrishnan,
B.; Rao, S. T.; Sundaralingham, M. Nat. Struct. Biol. 1994, 1, 169-175.
(d) White, S.; Baird, E. E.; Dervan, P. B. Biochemistry 1996, 35, 12532-
12537.
(6) For Im/Im pairing, see: (a) Geierstanger, B. H.; Dwyer, T. J.; Bathini,
Y.; Lown, J. W.; Wemmer, D. E. J. Am. Chem. Soc. 1993, 115, 4474-
4482. (b) Singh, S. B.; Ajay; Wemmer, D. E.; Kollman, P. A. Proc. Natl.
Acad. Sci. U.S.A. 1994, 91, 7673-7677. (c) White, S.; Baird, E. E.; Dervan,
P. B. Chem. Biol. 1997, 4, 569-578.
(7) For H-pin motif, see: (a) Mrksich, M.; Dervan, P. B. J. Am. Chem.
Soc. 1994, 116, 3663. (b) Dwyer, T. J.; Geierstanger, B. H.; Mrksich, M.;
Dervan, P. B.; Wemmer, D. E. J. Am. Chem. Soc. 1993, 115, 9900. (c)
Chen, Y. H.; Lown, J. W. J. Am. Chem. Soc. 1994, 116, 6995. (d) Greenberg,
W. A.; Baird, E. E.; Dervan, P. B. Chem. Eur. J. 1998, 4, 796-805.
(8) For hairpin motif, see: (a) Mrksich, M.; Parks, M. E.; Dervan, P. B.
J. Am. Chem. Soc. 1994, 116, 7983-7988. (b) Parks, M. E.; Baird, E. E.;
Dervan, P. B. J. Am. Chem. Soc. 1996, 118, 6147-6152. (c) Parks, M. E.;
Baird, E. E.; Dervan, P. B. J. Am. Chem. Soc. 1996, 118, 6153-6159. (d)
Swalley, S. E.; Baird, E. E.; Dervan, P. B. J. Am. Chem. Soc. 1996, 118,
8198-8206. (e) Pilch, D. S.; Poklar, N. A.; Gelfand, C. A.; Law, S. M.;
Breslauer, K. J.; Baird, E. E.; Dervan, P. B. Proc. Natl. Acad. Sci. U.S.A.
1996, 93, 8306-8311. (f) de Clairac, R. P. L.; Geierstanger, B. H.; Mrksich,
M.; Dervan, P. B.; Wemmer, D. E. J. Am. Chem. Soc. 1997, 119, 7909-
7916.
(9) For polyamide dimers, see: (a) Kelly, J. J.; Baird, E. E.; Dervan, P.
B. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 6981-6985. (b) Trauger, J. W.;
Baird, E. E.; Mrksich, M.; Dervan, P. B. J. Am. Chem. Soc. 1996, 118,
6160-6166. (c) Geierstanger, B. H.; Mrksich M.; Dervan, P. B.; Wemmer,
D. E. Nat. Struct. Biol. 1996, 3, 321-324. (d) Swalley, S. E.; Baird, E. E.;
Dervan, P. B. Chem. Eur. J. 1997, 3, 1600-1607. (e) Trauger, J. W.; Baird,
E. E.; Dervan, P. B. J. Am. Chem. Soc. 1998, 120, 3534-3535.
(1) (a) Gottesfield, J. M.; Nealy, L.; Trauger, J. W.; Baird, E. E.; Dervan,
P. B. Nature 1997, 387, 202-205. (b) Dickenson; L. A., Guzilia; P. Trauger,
J. W.; Baird, E. E.; Mosier, D. M.; Gottesfeld, J. M.; Dervan, P. B. Proc.
Natl. Acad. Sci. U.S.A. 1998, 95, 12890.
(2) For subnanomolar binding, see: (a) Trauger, J. W.; Baird, E. E.;
Dervan, P. B. Nature 1996, 382, 559-561. (b) Swalley, S. E.; Baird, E.
E.; Dervan, P. B. J. Am. Chem. Soc. 1997, 119, 6953-6961. (c) Turner, J.
M.; Baird, E. E.; Dervan, P. B. J. Am. Chem. Soc. 1997, 119, 7636-7644.
(d) Trauger, J. W.; Baird, E. E.; Dervan, P. B. Angew. Chem., Int. Ed.
1998, 37, 1421-1423. (e) Turner, J. M.; Swalley, S. E.; Baird, E. E.; Dervan,
P. B. J. Am. Chem. Soc. 1998, 120, 6219-6226.
(3) For Hp/Py pairings, see: (a) White, S. E.; Szewczyk, J. W.; Turner,
J. M.; Baird, E. E.; Dervan, P. B. Nature 1998, 391, 468-471. (b) Kielkopf,
C. L.; White, S. E.; Szewczyk, J. W.; Turner, J. M.; Baird, E. E.; Dervan,
P. B.; Rees, D. C. Science 1998, 282, 111.
(4) For specificity of Im/Py pairings, see: (a) Wade, W. S.; Mrksich,
M.; Dervan, P. B. J. Am. Chem. Soc. 1992, 114, 8783-8794. (b) Mrksich,
M.; Wade, W. S.; Dwyer, T. J.; Geierstanger, B. H.; Wemmer, D. E.;
Dervan, P. B. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 7586-7590. (c) Wade,
W. S.; Mrksich, M.; Dervan, P. B. Biochemistry 1993, 32, 11385-11389.
(d) Mrksich, M.; Dervan, P. B. J. Am. Chem. Soc. 1993, 115, 2572-2576.
(e) Geierstanger, B. H.; Mrksich, M.; Dervan, P. B.; Wemmer, D. E. Science
1994, 266, 646-650. (f) Kielkopf, C. L.; Baird, E. E.; Dervan, P. B.; Rees,
D. C. Nat. Struct. Biol. 1998, 5, 104-109. (g) White, S.; Baird, E. E.;
Dervan, P. B. J. Am. Chem. Soc. 1997, 119, 8756-8765.
10.1021/ja983206x CCC: $18.00 © 1999 American Chemical Society
Published on Web 01/27/1999

1122 J. Am. Chem. Soc., Vol. 121, No. 6, 1999
Herman et al.
Figure 1. (Top a-c) Hydrogen bond models of the polyamide-DNA complexes formed between the 2:1 dimer ImPyPyPy-â-Dp (1), the 1:1
hairpin ImPyPyPy-(R)H2Nγ-ImPyPyPy-OH (2), and the 1:1 cyclo-(γ-ImPyPyPy-(R)H2Nγ-ImPyPyPy) (3) with the six base pair 5′-AGTACT-3′ match
site. Circles with dots represent lone pairs of N3 of purines and O2 of pyrimidines. Circles containing an H represent the N2 hydrogen of G.
Putative hydrogen bonds are illustrated by dotted lines. (Bottom a-c) Schematic binding models for 1, 2, and 3. Im and Py rings are represented
as shaded and unshaded spheres, respectively.
potentially regulate cell-permeation properties, discovery of new
motifs for covalent linkage that provide polyamides with
affinities and specificities comparable to naturally occurring
DNA binding proteins remains a high priority.
increased from three to four. Polyamide-DNA binding affinity
is predicted to increase as the number of consecutive ring
pairings increases from three to four.9a In addition, the charge
has been moved from a Py-N-methyl group to a γ-turn. Although
the detailed effects of the placement of charge remain to be
determined, substitution of the prochiral R-position of the γ-turn
residue to provide (R)-2,4,-diaminobutyric acid has been previ-
ously found to yield chiral linked hairpins with enhanced DNA
binding sequence specificity and orientation preference.12 Cycle
polyamides must be capped at the N terminus in order to
complete the cycle; however, N-terminal acetylation has been
found to reduce hairpin-binding specificity and orientation
preference.4g,8b Analogous to hairpin polyamides, cycle polya-
mides are capable of forming two mirror image folded structures,
one of which is responsible for 5′-3′, N-C match DNA
binding4g (see Figure 1). It is likely that the chiral amine group
could offset the “acetylation effect” by controlling cycle
polyamide binding orientation preference and hence binding
specificity.12
Here we report the DNA binding affinities and sequence
specificities of the eight-ring cycle polyamides, cyclo-(γ-
ImPyPyPy-(R)H2Nγ-ImPyPyPy-) (3) and cyclo-(γ-ImPyPyPy-
(R)H2Nγ-PyPyPyPy-) (4) that differ by a single amino acid
substitution (underlined), for their respective six base pair match
sites, 5′-AGTACT-3′ and 5′-AGTATT-3, which differ by a
single base pair (underlined). In control experiments, the binding
affinity and sequence specificity of the ImPyPyPy-â-Dp (1)
dimer, and a hairpin analogue ImPyPyPy-(R)H2Nγ-ImPyPyPy-
C3-OH (2) were also studied (see Figure 2). An EDTA
Design of Cycle Polyamides. In a formal sense, addition of
a second γ-turn at the C and N termini of a hairpin polyamide
allows covalent closure to form a cycle.10An initial report
described a six-ring (-γ-3-γ-3-) cycle polyamide which bound
to a five base pair DNA sequence with higher affinity than a
corresponding hairpin polyamide;10 however, sequence specific-
ity versus mismatch DNA sequences was extremely poor (∼3-
fold), compared to 40-fold observed for the hairpin.11 It was
initially thought that the cycle restricted polyamide flexibility,
limiting the available conformers to prevent formation of a
specific recognition complex. It remained to be determined if
γ-turn cycle polyamides could be designed that have comparable
DNA binding properties to hairpin-polyamides.
Because of the labor intensive solution-phase cycle polyamide
synthesis and the initial discouraging thermodynamics with
regard to sequence specificity, the cycle polyamides have not
been investigated further until this report. We describe here a
pyrrole resin which enables cycle polyamide linear precursors
to be synthesized by solid-phase methods, reducing the synthetic
effort from weeks to days. Two eight-ring (-γ-4-γ-4-) cycle
polyamides have been prepared for the studies described here.
Two key design elements from the original six-ring cycle have
been altered. First, the number of ring pairings has been
(10) For cycle polyamides, see: Cho, J. Y.; Parks, M. P.; Dervan, P. B.
Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 10389.
(11) For a review, see: Wemmer, D. E.; Dervan, P. B. Curr. Opin. Struct.
Biol. 1997, 7, 355.
(12) Herman, D. M.; Baird, E. E.; Dervan, P. B. J. Am. Chem. Soc. 1998,
120, 1382.

DNA Binding Properties of Cycle Polyamide Motifs
J. Am. Chem. Soc., Vol. 121, No. 6, 1999 1123
Figure 3. Synthesis of Boc-Py-PAM resin (7). (i) K2CO3, DMF, (ii)
Zn/AcOOH (iii) DCC/HOBT, DMF; (iv) aminomethylated-polystyrene;
DIEA.
(Figure 3). The phenacyl ester (5) was selectively cleaved (Zn,
AcOH), and the resultant acid was (6) activated (DCC, HOBT)
followed by reaction of the activated ester with an excess of
0.7 mmol/g aminomethylated polystyrene for 24 h (DIEA, DMF)
to give Boc-Py-Pam resin (7). Reactions were stopped at 0.1
mmol/g substitution as determined by quantitative ninhydrin
analysis of free amine groups.18 Unreacted amine groups were
capped by acetylation (Ac2O, DIEA, DMF). Picric acid titra-
tion19 of Py-amino groups was used to verify resin loading of
0.1 mmol/g.
Figure 2. Structures of polyamides ImPyPyPy-â-Dp (1), ImPyPyPy-
(R)H2Nγ-ImPyPyPy-â-Dp (2), ImPyPyPy-(R)H2Nγ-ImPyPyPy-OH (2),
cyclo-(γ-ImPyPyPy-(R)H2Nγ-ImPyPyPy) (3) and cyclo-(γ-ImPyPyPy-
(R)H2Nγ-PyPyPyPy) (4) as synthesized by solid-phase methods.
Synthesis of Hairpin Control 2. ImPyPyPy-(R)H2Nγ-ImPy-
PyPy-â-Pam resin was synthesized by machine-assisted proto-
cols in 18 steps from commercially available Boc-â-Ala-Pam
resin (Figure 4a).13 The polyamide was cleaved from the resin
by a single-step reduction with lithium borohydride (EtOH, 60
°C), followed by reversed phase HPLC purification to yield the
hairpin polyamide ImPyPyPy-(R)H2Nγ-ImPyPyPy-OH (2).
Cycle Polyamide Synthesis. Two polyamide resins, Cbzγ-
ImPyPyPy-(R)Fmocγ-PyPyPyPy-Pam resin, and Cbzγ-ImPy-
PyPy-(R)Fmocγ-ImPyPyPy-Pam resin were synthesized in 18
steps from Boc-Py-Pam resin (600 mg of resin, 0.1 mmol/g of
substitution) using Boc-chemistry machine-assisted protocols
(Figure 4b).13 The (R)-2,4-diaminobutyric acid residue was
introduced as an orthogonally protected N-R-Fmoc-N-γ-Boc
derivative 10 (HBTU, DIEA). The final step was introduction
of Cbzγ-Im acid 11 as a dimer block (HBTU, DIEA).13 Fmoc-
protected polyamide resins Cbzγ-ImPyPyPy-(R)Fmocγ-PyPy-
PyPy-Pam resin, and Cbzγ-ImPyPyPy-(R)Fmocγ-ImPyPyPy-Pam
resin were treated with 1:4 DMF:piperidine (22 °C, 30 min) to
provide Cbzγ-ImPyPyPy-(R)H2Nγ-PyPyPyPy-Pam resin and
Cbzγ-ImPyPyPy-(R)H2Nγ-ImPyPyPy-Pam resin, respectively.
The amine resins were then treated with Boc-anhydride (DIEA,
DMF, 55 °C, 30 min) providing Cbzγ-ImPyPyPy-(R)Bocγ-
PyPyPyPy-Pam resin and Cbzγ-ImPyPyPy-(R)Bocγ-ImPyPyPy-
Pam resin. A single step catalytic transfer hydrogenolysis was
used to cleave the polyamide from the solid support and remove
the Cbz-protecting group from the N-terminal γ residue. A
sample of the resin (240 mg) was treated with palladium acetate
(2 mL DMF, 240 mg Pd(OAc)2, 37 °C, 10 min). Ammonium
analogue cyclo-(γ-ImPyPyPy-(R)EDTA‚Fe(II)γ-PyPyPyPy-) (4-E‚
Fe(II)) was constructed to confirm the binding orientation of
cycle polyamide 4 at its 5′-AGTATT-3′ match and 5′-AGTACT-
3′ mismatch site. All polyamides were synthesized by solid-
phase methods13 and their purity and identity confirmed by 1H
NMR, MALDI-TOF-MS, and analytical HPLC. Precise binding
site sizes were determined by MPE‚Fe(II) footprinting,14 and
binding orientation and stoichiometry were confirmed by affinity
cleaving experiments.15 Equilibrium association constants (Ka)
of the polyamides for respective match and mismatch binding
sites were determined by quantitative DNase I footprint titra-
tion.16
Results and Discussion
Resin Synthesis. The Py-Pam ester (5) was prepared accord-
ing to the published procedures of Merrifield,17 with Boc-Py
acid substituted for the standard Boc protected R-amino acid
(13) Baird, E. E.; Dervan, P. B. J. Am. Chem. Soc. 1996, 118, 6141.
(14) (a) Van Dyke, M. W.; Hertzberg, R. P.; Dervan, P. B. Proc. Natl.
Acad. Sci. U.S.A. 1982, 79, 5470. (b) Van Dyke, M. W.; Dervan, P. B.
Science 1984, 225, 1122.
(15) (a) Taylor, J. S.; Schultz, P. G.; Dervan, P. B. Tetrahedron 1984,
40, 457. (b) Dervan, P. B. Science 1986, 232, 464.
(16) (a) Brenowitz, M.; Senear, D. F.; Shea, M. A.; Ackers, G. K.
Methods Enzymol. 1986, 130, 132. (b) Brenowitz, M.; Senear, D. F.; Shea,
M. A.; Ackers, G. K. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 8462. (c)
Senear, D. F.; Brenowitz, M.; Shea, M. A.; Ackers, G. K. Biochemistry
1986, 25, 7344.
(17) Mitchell, A. R.; Kent, S. B. H.; Engelhard, M.; Merrifield, R. B. J.
Org. Chem. 1978, 43, 2845.
(18) Sarin, V. K.; Kent, S. B. H.; Tam, J. P.; Merrifield, R. B. Anal.
Biochem. 1981, 117, 147.
(19) Gisin, B. F. Anal. Chim. Acta 1972, 58, 248.

1124 J. Am. Chem. Soc., Vol. 121, No. 6, 1999
Herman et al.
Figure 4. (Box) Monomers for synthesis of compounds described here; Boc-Py-OBt ester 8, Boc-Im acid 9, (R)-Fmoc-R-Boc-γ-diaminobutyric
acid 10, and Cbz-γ-Im acid 11. (i) 1:4 DMF/piperidine (22 °C, 30 min): (ii) LiBH4, EtOH, reflux 16 h. (iii) 80% TFA/DCM, 0.4 M PhSH; (iv)
Boc-Py-OBt, DIEA, DMF; (v) 80% TFA/DCM, 0.4 M PhSH; (vi) Boc-Py-OBt, DIEA, DMF; (vii) 80% TFA/DCM, 0.4 M PhSH; (viii) Boc-Im
acid (DCC, HOBT) for 3, Boc-Py-OBt, DIEA, DMF for 4; (ix) 80% TFA/DCM, 0.4 M PhSH; (x) (R)-Fmoc-R-Boc-γ-diaminobutyric acid (HBTU,
DIEA); (xi) 80% TFA/DCM, 0.4 M PhSH; (xii) Boc-Py-OBt, DIEA, DMF; (xiii) 80% TFA/DCM, 0.4 M PhSH; (xiv) Boc-Py-OBt, DIEA, DMF;
(xv) 80% TFA/DCM, 0.4 M PhSH; (xvi) Boc-Py-OBt, DIEA, DMF; (xvii) 80% TFA/DCM, 0.4 M PhSH; (xviii) Cbz-γ-Im acid (HBTU, DIEA);
(xix) 80% Piperdine/DMF (25 °C, 30 min); (xx) Boc-anhydride, DIEA, DMF; (xxi) Pd(OAc)2, HCO2NH4, DMF (37 °C, 8 h); (xxii) DPPA, K2CO3
(xxiii) TFA (1 h). (Inset) Pyrrole, imidazole, and diaminobutyric acid monomers for solid-phase synthesis: Boc-Pyrrole-OBt ester13 (Boc-Py-OBt)
(8), imidazole-2-carboxylic acid4a (Im-OH) (9), (R)-Fmoc-R-Boc-γ-diaminobutyric acid (10), and CBZ-γ-aminobutyric acid-imidazole dimer (11).
formate was added (500 mg, 8 h) and the reaction mixture
purified by reversed phase HPLC to provide H2N-γ-ImPyPyPy-
(R)Bocγ-PyPyPyPy-COOH (13) and H2N-γ-ImPyPyPy-(R)Bocγ-
ImPyPyPy-COOH (12). Cyclization of H2N-γ-ImPyPyPy-
(R)Bocγ-PyPyPyPy-COOH (13) and H2N-γ-ImPyPyPy-(R)Bocγ-
ImPyPyPy-COOH (12) was achieved with DPPA and potassium

DNA Binding Properties of Cycle Polyamide Motifs
J. Am. Chem. Soc., Vol. 121, No. 6, 1999 1125
Figure 5. (Middle) Illustration of the 229 base pair restriction fragment with the position of the sequence indicated. MPE‚Fe(II) protection patterns
of 10 nM cyclo-(γ-ImPyPyPy-(R)H2Nγ-ImPyPyPy) (3) and 10 nM cyclo-(γ-ImPyPyPy-(R)H2Nγ-PyPyPyPy) (4). Bar heights are proportional to the
relative protection from cleavage at each band. Binding sites determined by MPE‚Fe(II) footprinting are boxed. MPE‚Fe(II) footprinting experiments
of cyclo-(γ-ImPyPyPy-(R)H2Nγ-ImPyPyPy) (3) (left) and cyclo-(γ-ImPyPyPy-(R)H2Nγ-PyPyPyPy) (4) (right) on the 3′ 32P-labeled 229 bp restriction
fragment of pJT8. 5′-AGTATT-3′ and 5′-AGTACT-3′ sites are shown adjacent to the autoradiograms. Additional 5′-TGTAAA-3′, 5′-TGTGCT-3′,
and 5′-TTAAGT-3′ mismatch sites are not analyzed. Lane 1, intact; lane 2, A reaction; lane 3, MPE‚Fe(II) standard; lanes 4-8, 100 pM, 1 nM, 10
nM, 100 nM and 1 µM polyamide. All lanes contain 15 kcpm of either 3′- or 5′-radiolabeled DNA, 25 mM HEPES buffer (pH 7.3), 200 mM NaCl,
50 mg/mL glycogen, 5 mM DTT, and 0.5 mM MPE‚Fe(II), 22 °C.
carbonate, as described previously.10 The Boc-protecting group
was then removed in situ by treatment with neat TFA to yield
the cyclic compounds cyclo-(γ-ImPyPyPy-(R)H2Nγ-ImPyPyPy-)
(3) and cyclo-(γ-ImPyPyPy-(R)H2Nγ-PyPyPyPy-) (4) after sub-
sequent purification by reversed phase HPLC. The cycle
polyamides are obtained with similar yield and purity and have
similar solubility as their hairpin counterparts.
anhydride was hydrolyzed (0.1 M NaOH, 55 °C, 10 min). Cyclo-
(γ-ImPyPyPy-(R)EDTAγ-PyPyPyPy-) (4-E‚Fe(II)) was then iso-
lated by reversed phase HPLC (Figure 6a). Affinity cleavage
experiments were performed on the same 3′- or 5′-32P end-
labeled 229 base pair DNA restriction fragment from the plasmid
pJT8 (20 mM HEPES buffer (pH 7.3), 200 mM NaCl, 50 µg/
mL glycogen, 5 mM DTT, 1 µM Fe(II), pH 7.0 and 22 °C).
The observed cleavage pattern for cyclo-(γ-ImPyPyPy-(R)EDTA‚
Fe(II)γ-PyPyPyPy) (4-E‚Fe(II)) (Figure 6b and c) are 3′-shifted,
consistent with minor groove occupancy. In the presence of 100
nM 4-E‚Fe(II), a major cleavage locus proximal to the 3′-side
of the 5′-AGTATT-3′ match sequence is revealed, consistent
with formation of an oriented 1:1 cycle polyamide-DNA
complex. At the same ligand concentration, minor cleavage loci
located 3′ and 5′ adjacent to the single base pair mismatch 5′-
AGTACT-3′ site appear, consistent with dual binding orienta-
tions at this symmetrical binding site. The cyclo-polyamide
binding model is further supported by the location of cleavage
loci at the 5′-side of the 5′-AGTATT-3′ match site (Figure 6c),
and at the 5′- and 3′- sides of the 5′-AGTACT-3′ mismatch
site corresponding to the EDTA‚Fe(II) moiety placement off
the ((R)H2Nγ)-turn residue.
Binding Site Size. MPE‚Fe(II) footprinting14 on 3′- or 5′-
32P end-labeled 229 base pair restriction fragments reveals that
each cycle polyamide, at 10 nM concentration, binds to its
designated six base pair match sites (25 mM HEPES buffer (pH
7.3), 200 mM NaCl, 50 µg/mL glycogen, 5 mM DTT, 0.5 µM
MPE‚Fe(II), and 22 °C) (Figure 5). The polyamide cyclo-(γ-
ImPyPyPy-(R)H2Nγ-ImPyPyPy-) (3) which contains an Im/Py
and a Py/Im pair, protects the cognate 5′-AGTACT-3′ match
site. Binding of the single base pair mismatch site 5′-AGTATT-
3′ is only seen at much higher polyamide concentrations. The
polyamide cyclo-(γ-ImPyPyPy-(R)H2Nγ-PyPyPyPy-) (4) which
contains a single Im/Py pair protects the designated match site
5′-AGTATT-3′ and the single base pair mismatch site 5′-
AGTACT-3′. The sizes of the asymmetrically 3′-shifted footprint
cleavage patterns are consistent with 1:1 cycle polyamide/DNA
complex formation at six base pair binding sites.
Binding Energetics. Quantitative DNase I footprint titra-
tions16 (10 mM Tris‚HCl, 10 mM KCl, 10 mM MgCl2, and 5
mM CaCl2, pH 7.0, and 22 °C) were performed to determine
the equilibrium association constants (Ka) of ImPyPyPy-â-Dp
(1), ImPyPyPy-(R)H2Nγ-ImPyPyPy-OH (2), cyclo-(γ-ImPyPyPy-
(R)H2Nγ-ImPyPyPy) (3), and cyclo-(γ-ImPyPyPy-(R)H2Nγ-Py-
PyPyPy) (4) for the six base pair match and mismatch sites
(Table 1 and Figure 7). Polyamide 1 binds the respective match
Binding Orientation. Affinity cleavage experiments15 using
cyclo-(γ-ImPyPyPy-(R)EDTA‚Fe(II)γ-PyPyPyPy-) (4-E‚Fe(II)), which
has an EDTA‚Fe(II) moiety appended to the γ-turn, were used
to confirm polyamide binding orientation and stoichiometry. For
synthesis of the EDTA analogue, cyclo-(γ-ImPyPyPy-(R)H2Nγ-
PyPyPyPy-) (4) was treated with an excess of EDTA-dianhy-
dride (DMSO/NMP, DIEA, 55 °C, 15 min) and the remaining

1126 J. Am. Chem. Soc., Vol. 121, No. 6, 1999
Herman et al.
Table 1. Association Constants (M-1) for Polyamides 1-4a-d
a The reported equilibrium association constants are the mean values
obtained from three DNase I footprint titration experiments. b The assays
were carried out at 22 °C, pH 7.0 in the presence of 10 mM Tris‚HCl,
10 mM KCl, 10 mM MgCl2, and 5 mM CaCl2. c Apparent monomer
association constants were determined for polyamide homodimers.23
d Specificity calculated as Ka(5′-AGTACT-3′)/Ka(5′-AGTATT-3′).
Figure 7. Quantitative DNase I footprint titration experiments with
(a) cyclo-(γ-ImPyPyPy-(R)H2Nγ-ImPyPyPy) (3) and (b) cyclo-(γ-
ImPyPyPy-(R)H2Nγ-PyPyPyPy) (4) on the 3′-end labeled 229 base pair
restriction fragment: lane 1, intact; lane 2, A reaction; lane 3, DNase
I standard; lanes 4-14, 5 pM, 10 pM, 20 pM, 50 pM, 100 pM, 200
pM, 500 pM, 1 nM, 2 nM, 5 nM, and 10 nM. The 5′-AGTACT-3′ and
5′-AGTATT-3′ sites were analyzed and are shown on the right side of
the autoradiogram. Additional sites not analyzed are 5′-TGTAAA-3′,
5′-TGTGCT-3′, and 5′-TTAAGT-3′. All reactions contain 20 kcpm
restriction fragment, 10 mM Tris‚HCl (pH 7.0), 10 mM KCl, 10 mM
MgCl2, and 5 mM CaCl2.
Figure 6. (a) Synthesis of 4-E‚Fe(II): (i) EDTA dianhydride (DMSO/
NMP, DIEA, 55 °C, 15 min); (ii) 0.1 M NaOH (55 °C, 10 min). (b)
Affinity cleavage pattern for cyclo-(γ-ImPyPyPy-(R)EDTA‚Fe(II)γ-PyPy-
PyPy) (4-E‚Fe(II)) at 100 nM concentration depicting a single binding
orientation at the 5′-AGTATT-3 match site and no orientational
preference at the 5′-AGTACT-3′ mismatch site. (c) Ball-and-stick model
of 4-E‚Fe(II)‚5′AGTATT-3′ complex. Bar heights are proportional to
the relative cleavage intensities at each base pair. Shaded and nonshaded
circles denote imidazole and pyrrole carboxamides, respectively. The
boxed Fe denotes the EDTA‚Fe(II) cleavage moiety. See Supporting
Information for autoradiograms.
1:1 polyamide-DNA complex.12,23 Polyamides bind the 5′-
AGTACT-3′ site with decreasing affinity; match cycle (3) >
match hairpin (2) > mismatch cycle (4) > match dimer (1).
Polyamides bind the 5′-AGTATT-3′ sequence with decreasing
affinity; match cycle (4) > mismatch cycle (3) > mismatch
hairpin (2) > mismatch dimer (1).
Covalent coupling of dimer (1) to form hairpin polyamide
(2) results in a 428-fold increase in the DNA binding affinity
and comparable DNA binding sequence specificity. It is
interesting to compare hairpin polyamide (2), ImPyPyPy-
(R)H2Nγ-ImPyPyPy-OH, to the previously reported hairpin
ImPyPyPy-γ-ImPyPyPy-â-Dp.2a Each hairpin contains eight
aromatic rings and a single charge located either on the γ-turn
or a C-terminal â-Dp group. Although hairpin polyamide (2)
binds to DNA with affinity and specificity comparable to DNA
and mismatch sites with binding isotherms consistent with 2:1
dimer formation.12,23 Hairpin polyamide (2) and cycle polya-
mides (3) and (4) bind their respective match and mismatch
sequences with binding isotherms consistent with binding in a
(20) Kent, S. B. H. Annu. ReV. Biochem. 1988, 57, 957.
(21) Barlos, K.; Chatzi, O.; Gatos, D.; Stravropoulos, G. Int. J. Pept.
Protein Res. 1991, 37, 513.
(22) Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular Cloning; Cold
Spring Harbor Laboratory: Cold Spring Harbor, NY, 1989.
(23) For the treatment of data on cooperative association of ligands,
see: Cantor, C. R.; Schimmel, P. R. Biophysical Chemistry, Part III: the
BehaVior of Biological Macromolecules; W. H. Freeman and Company:
New York, 1980; p 863.

DNA Binding Properties of Cycle Polyamide Motifs
J. Am. Chem. Soc., Vol. 121, No. 6, 1999 1127
(Boc-â-Pam resin) was purchased from Peptides International (0.2
mmol/gram) (R)-2-Fmoc-4-Boc-diaminobutyric acid, and (R)-2-amino-
4-Boc-diaminobutyric acid were from Bachem. N,N-diisopropylethy-
lamine (DIEA), N,N-dimethylformamide (DMF), N-methylpyrrolidone
(NMP), DMSO/NMP, acetic anhydride (Ac2O), and 0.0002 M potas-
sium cyanide/pyridine were purchased from Applied Biosystems.
Dichloromethane (DCM) and triethylamine (TEA) were reagent grade
from EM, thiophenol (PhSH) and dimethylaminopropylamine (Dp) were
from Aldrich, trifluoroacetic acid (TFA) Biograde was from Halocarbon,
phenol was from Fisher, and ninhydrin was from Pierce. All reagents
were used without further purification. Quik-Sep polypropylene dispos-
able filters were purchased from Isolab, Inc. A shaker for manual solid-
phase synthesis was obtained from St. John Associates, Inc. Screw-
cap glass peptide synthesis reaction vessels (5 mL and 20 mL) with a
no. 2 sintered glass frit were made as described by Kent.20 1H NMR
spectra were recorded on a General Electric-QE NMR spectrometer at
300 MHz with chemical shifts reported in parts per million relative to
residual solvent. UV spectra were measured in water on a Hewlett-
Packard model 8452A diode array spectrophotometer. Optical rotations
were recorded on a JASCO Dip 1000 digital polarimeter. Matrix-
assisted, laser desorption/ionization time-of-flight mass spectrometry
(MALDI-TOF) was performed at the Protein and Peptide Microana-
lytical Facility at the California Institute of Technology. HPLC analysis
was performed on either a HP 1090M analytical HPLC or a Beckman
Gold system using a RAINEN C18, Microsorb MV, 5µm, 300 × 4.6
mm reversed phase column in 0.1% (wt/v) TFA with acetonitrile as
eluent and a flow rate of 1.0 mL/min, gradient elution 1.25%
acetonitrile/min. Preparatory reversed phase HPLC was performed on
a Beckman HPLC with a Waters DeltaPak 25 × 100 mm, 100 µm
C18 column equipped with a guard, 0.1% (wt/v) TFA, 0.25%
acetonitrile/min. Water (18Ω) was obtained from a Millipore MilliQ
water purification system, and all buffers were 0.2 µm filtered.
Figure 8. Comparison of first-generation six-ring cycle polyamide,10
and second generation eight-ring cycle polyamide described here.
Schematic binding model is as described in Figure 1.
binding proteins, it does bind with 4-fold lower affinity and
5-lower sequence specificity than the previously described
hairpin ImPyPyPy-γ-ImPyPyPy-â-Dp. This probably results
from loss of favorable interactions between the â-Dp group and
A,T rich flanking sequences.8b Since cycle polyamides have no
C-terminal â-Dp group, hairpin polyamide (2) is a more
applicable control for the study described here.
On the basis of the pairing rules for polyamide-DNA
complexes, the 5′-AGTACT-3′ and 5′-AGTATT-3′ sites rep-
resent “match” and “single base pair mismatch” sites for cycle-
3, respectively, and “single base pair mismatch” and “match”
sites for cycle-4, respectively. Cycle polyamide (3), cyclo-(γ-
ImPyPyPy-(R)H2Nγ-ImPyPyPy), binds the six base pair 5′-
AGTACT-3′ target sequence with an equilibrium association
constant, Ka ) 7.6 × 1010 M-1, and 55-fold specificity over
the single base pair mismatch 5′-AGTATT-3′ site (Ka ) 1.3 ×
109 M-1). These affinities represent a 3600-fold increase relative
to dimer (1) and an 8-fold enhancement relative to hairpin
polyamide (2). Furthermore, the affinity and specificity of (3)
are comparable to those of the previously described hairpin
ImPyPyPy-γ-ImPyPyPy-â-Dp. The cycle polyamide (4), cyclo-
(γ-ImPyPyPy-(R)H2Nγ-PyPyPyPy), which contains a single Im/
Py pair, preferentially binds the 5′-AGTATT-3′ match site (Ka
) 3.1 × 109 M-1) versus the single base pair mismatch
5-AGTACT-3′ (Ka ) 4.2 × 108 M-1) with a 7-fold preference.
Therefore, replacing a single pyrrole amino acid in cyclo-(γ-
ImPyPyPy-(R)H2Nγ-PyPyPyPy) (4) with an imidazole residue
in cyclo-(γ-ImPyPyPy-(R)H2Nγ-ImPyPyPy) (3), regulates cycle
polyamide specificity and affinity by 2 orders of magnitude (see
Figure 7).
Conclusions. A second generation cycle polyamide motif has
been designed and characterized. Two eight-ring (-γ-4-γ-4-)
cycle polyamides were found here to bind to DNA with affinity
and specificity comparable to naturally occurring DNA binding
proteins. The polyamide DNA binding affinity increases as
expected as the number of consecutive ring pairings increases
from 3 to 4.9a Important key design factors likely contributed
to the improved specificity of the 8-ring cycles compared with
the original design (Figure 8).10 Moving the charge to the γ-turn
may enhance the cycle polyamide binding orientation preference
and hence binding-specificity. Determination of the exact
molecular basis for the optimization of the cycle polyamides
awaits further footprinting efforts as well as high resolution
structure studies. However, it is clear from comparison of first
and second generation cycles that polyamide design for DNA
recognition can be continually optimized using affinity and
specificity as two key criteria in parallel with continued
investment in synthetic methodology.
Resin Substitution. Resin substitution can be calculated as Lnew
-
(mmol/g) ) Lold/(1 + Lold(Wnew - Wold) × 10-3) (eq 2), where L is the
loading (mmol of amine per gram of resin), and W is the weight
(gmol-1) of the growing polyamide attached to the resin.21
Control Hairpin ImPyPyPy-(R)H2Nγ-ImPyPyPy-C3-OH (2). Im-
PyPyPy-(R)Fmocγ-ImPyPyPy-â-Pam resin was synthesized in a stepwise
fashion by machine-assisted solid-phase methods from Boc-â-alanine-
Pam resin (0.6 mmol/g).13 (R)-2-Fmoc-4-Boc-diaminobutyric acid (0.7
mmol) was incorporated as previously described for Boc-γ-aminobutyric
acid. ImPyPyPy-(R)Fmocγ-ImPyPyPy-â-Pam resin was placed in a glass
20-mL peptide synthesis vessel and treated with DMF (2 mL), and
then with piperidine (8 mL) and agitated (22 °C, 30 min). ImPyPyPy-
(R)H2Nγ-ImPyPyPy-â-Pam resin was isolated by filtration and washed
sequentially with an excess of DMF, DCM, MeOH, and ethyl ether,
and the amine resin was dried in vacuo. A sample of resin (200 mg
0.40 mmol/gram) was suspended in absolute ethanol (25 mL). LiBH4
(200 mg) was added and the mixture refluxed for 16 h. The reaction
mixture was then filtered to remove resin, neat TFA was added (6 mL),
and the resulting solution was concentrated in vacuo, resuspended in
0.1% (wt/v) TFA (8 mL), and purified twice by reversed phase HPLC
to provide the trifluoroacetate salt of ImPyPyPy-(R)H2Nγ-ImPyPyPy-
C3-OH (2) as a white powder upon lyophilization of the appropriate
1
fractions: 2.5 mg, 3% recovery; H NMR (DMSO-d6) δ 11.03 (s, 1
H), 10.47 (s, 1 H), 10.13 (s, 1 H), 9.97 (s, 2 H), 9.94 (s, 1 H), 9.90 (s,
1 H), 8.35 (br s, 3 H), 8.29 (m 1 H), 8.01 (m, 1 H), 7.39 (s, 2 H), 7.32
(s, 1 H), 7.26 (m, 2 H), 7.22 (s, 1 H), 7.17 (m, 2 H), 7.08 (s, 2 H), 7.05
(s, 1 H), 7.02 (s, 2 H), 6.93 (s, 1 H), 6.84 (s, 1 H), 5.32 (m, 1 H), 3.98
(s, 6 H), 3.85 (s, 3 H), 3.83 (m, 9 H), 3.80 (s, 3 H), 3.78 (s, 3 H), 2.73
(m, 2 H), 2.35 (m, 2 H), 1.95 (m, 4 H), 1.73 (m, 2 H), MALDI-TOF-
MS (monoisotopic) [M + H] 1139.5 (1139.5 calcd for C53H63N20O10).
Boc-Pyrrolyl-4-(oxymethyl)phenylacetic Acid Phenacyl Ester (5).
A solution of Boc-Py acid (6.9 g, 29 mmol), 4-(bromomethyl)-
phenylacetic acid phenacyl ester (10 g, 29 mmol), and DIEA (7.2 mL,
41 mmol) in 60 mL of DMF were stirred at 50 °C for 6 h. The solution
was cooled and partitioned between 400 mL of water and 400 mL of
ethyl ether. The ether layer was washed sequentially (2 × 200 mL
each) with 10% citric acid, brine, saturated NaHCO3, and brine. The
organic phase was dried (sodium sulfate) and concentrated in vacuo.
Experimental Section
Dicyclohexylcarbodiimide (DCC), Hydroxybenzotriazole (HOBt),
2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophos-
phate (HBTU); aminomethylated polystyrene, 4-(bromomethyl)pheny-
lacetic acid phenacyl ester, and 0.6 mmol/gram Boc-â-alanine-(4-
carboxamidomethyl)-benzyl-ester-copoly(styrene-divinylbenzene) resin

1128 J. Am. Chem. Soc., Vol. 121, No. 6, 1999
Herman et al.
The crude product was recrystallized from 3:1 ethyl acetate/hexanes
to yield 5 as a fluffy white foam: 6.1 g, 42% yield; TLC (2:3 hexanes/
ethyl acetate v/v) Rf 0.6 1H NMR (DMSO-d6) δ 9.07 (s, 1 H), 7.90 (d,
2 H, J ) 7.6 Hz), 7.61 (t, 1 H, J ) 7.2 Hz), 7.50 (t, 2 H, J ) 7.5 Hz),
7.31 (m, 4 H), 7.06 (s, 1 H), 6.62 (s, 1 H), 5.47 (s, 2 H), 5.16 (s, 2 H),
3.80 (s, 2 H), 3.76 (s, 3 H), 1.42 (s, 9H) 13C NMR (DMSO-d6) δ 193.1,
171.2, 160.6, 153.2, 135.7, 134.4, 130.1, 129.4, 128.5, 128.3, 123.7,
120.0, 119.0, 108.0, 79.0, 67.4, 65.1, 36.7, 28.6; FABMS m/e 506.205
(506.205 calcd for C28H30N2O7).
palladium acetate (330 mg, 37 °C, 10 min) followed by the addition of
ammonium formate (500 mg, 37 °C, 8 h). The reaction mixture was
then filtered to remove resin and diluted with 0.1% (wt/v) TFA (8 mL),
and the resulting solution was purified by reversed phase HPLC to
yield H2N-γ-ImPyPyPy-(R)Bocγ-ImPyPyPy-COOH as the trifluoroac-
1
etate salt: 4.2 mg, 10% recovery; H NMR (DMSO-d6) δ 10.36 (s, 1
H), 10.22 (s, 1 H), 10.14 (s, 1 H), 9.99 (s, 1 H), 9.97 (s, 2 H), 9.92 (s,
2 H), 7.97 (m, 1 H), 7.74 (br s 3 H), 7.45 (m, 1 H), 7.43 (s, 1 H), 7.42
(s, 1 H), 7.29 (s, 1 H), 7.26 (d, 1 H, J ) 1.7 Hz), 7.23 (m, 2 H), 7.17
(s, 1 H), 7.14 (s, 1 H), 7.12 (s, 1 H), 7.06 (m, 2 H), 6.95 (s, 1 H), 6.90
(s, 1 H), 6.83 (d, 1 H, J ) 1.3 Hz), 4.20 (m, 1 H), 3.94 (s, 6 H), 3.84
(m, 12 H), 3.81 (s, 3 H), 3.78 (s, 3 H), 3.22 (m, 2 H), 2.83 (quintet, 2
H, J ) 6.2 Hz), 2.41 (t, 2 H, J ) 6.9 Hz), 1.83 (m, 4 H), 1.38 (s, 9 H),
MALDI-TOF-MS (monoisotopic) [M + H] 1282.6 (1282.6 calcd for
C59H72N21O13).
Boc-Pyrrolyl-4-(oxymethyl)phenylacetic Acid (6) Zinc dust was
activated with 1 M HCl (aqueous) as described.17 Boc-Pyrrolyl-4-
(oxymethyl)phenylacetic acid phenacyl ester (3 g, 5.9 mmol) was
dissolved in 90 mL 4:1 acetic acid/water (v/v). Zinc dust (9.6 g, 147
mmol) was added and the reaction stirred for 18 h at room temperature.
The zinc was removed by filtration and the reaction mixture partitioned
between 200 mL of ethyl ether and 200 mL of water. The layers were
separated, the aqueous layer was extracted (ethyl ether, 1 × 200 mL),
and the combined ether layers were washed (water, 5 × 100 mL), dried
(sodium sulfate), concentrated in vacuo, and azeotroped (benzene, 6
× 100 mL). The crude acid product was purified by flash chromatog-
raphy (2:1 hexanes/ethyl acetate) to yield a yellow oil: 2.0 g, 86%
Cyclo-(γ-ImPyPyPy-(R)H2Nγ-ImPyPyPy-) (3). The amine-polya-
mide 12 (2.8 mg, 2.0 µmol) was dissolved in DMF (7 mL), and treated
with DPPA (12.5 µL) and K2CO3 (100 mg) for 3 h. The reaction mixture
was concentrated in vacuo, treated with TFA (3 mL, 1 h), and purified
by reversed phase HPLC to provide the trifluoroacetate salt of cyclo-
(γ-ImPyPyPy-(R)H2Nγ-ImPyPyPy) (3). Cyclo-(γ-ImPyPyPy-(R)H2Nγ-
ImPyPyPy) was recovered as a white powder upon lyophilization of
the appropriate fractions (1.0 mg, 38% recovery). MALDI-TOF-MS
(monoisotopic) [M + H] 1164.5 (1164.5 calcd for C54H62N21O10).
Cyclo-(γ-ImPyPyPy-(R)H2Nγ-PyPyPyPy) (4). The amine-polyamide
13 (7 mg, 5.0 µmol) was dissolved in DMF (7 mL), and treated with
DPPA (12.5 µL) and K2CO3 (100 mg) for 3 h. The reaction mixture
was concentrated in vacuo, treated with TFA (3 mL, 1 h), diluted to 8
mL with 0.1% (wt/v) TFA, and purified by reversed phase HPLC to
provide cyclo-(γ-ImPyPyPy-(R)H2Nγ-PyPyPyPy) (4). Cyclo-(γ-ImPy-
PyPy-(R)H2Nγ-PyPyPyPy) was recovered as a white powder upon
lyophilization of the appropriate fractions as the trifluoroacetate salt
(2.6 mg, 41% recovery). 1H NMR (DMSO-d6) δ 10.56 (s, 1 H), 10.26
(s, 1 H), 9.96 (s, 2 H), 9.95 (s, 3 H), 9.92 (s, 1 H), 8.27 (m, 4 H), 9.00
(m, 1 H), 7.45 (s, 1 H), 7.39 (s, 1 H), 7.38 (s, 1 H), 7.34 (s, 1 H), 7.29
(s, 1 H), 7.26 (s, 1 H), 7.25 (s, 1 H), 7.18 (s, 1 H), 7.08 (s, 1 H), 6.93
(m, 1 H), 6.91 (s, 1 H), 6.89 (s, 1 H), 6.85 (s, 2 H), 6.82 (s, 1 H), 5.31
(m, 1 H), 3.93 (s, 3 H), 3.85 (s, 3 H), 3.83 (m, 9 H), 3.80 (s, 3 H), 3.78
(s, 3 H), 3.76 (s, 3 H), 3.23 (m, 2 H), 2.72 (m, 2 H), 2.36 (m, 2 H),
1.97 (m, 4 H), 1.38 (s, 9 H), MALDI-TOF-MS (monoisotopic) [M +
H] 1163.4 (1163.5 calcd for C55H63N20O10).
Cyclo-(γ-ImPyPyPy-(R)EDTA‚Fe(II)γ-PyPyPyPy-) (4-E‚Fe(II)). Ex-
cess EDTA-dianhydride (180 mg) was dissolved in 1:1 DMSO/NMP
(1 mL) and DIEA (1 mL) by heating at 60 °C for 10 min. The
dianhydride solution was added to cyclo(γ-ImPyPyPy-(R)H2Nγ-PyPy-
PyPy-) (4) (1.3 mg, 1.0 µmol) dissolved in DMSO (1 mL). The mixture
was heated (60 °C, 2 h) and the remaining EDTA-dianhydride
hydrolyzed with 0.1M NaOH (2 mL, 60 °C, 15 min). Aqueous TFA
(0.1wt %/v) was added to adjust the total volume to 8 mL and the
solution purified directly by reversed phase HPLC to provide to cyclo-
(γ-ImPyPyPy-(R)EDTA‚Fe(II)γ-PyPyPyPy-) (4-E‚Fe(II)) as a white powder
upon lyophilization of the appropriate fractions: 0.25 mg, 18%
recovery; MALDI-TOF-MS (monoisotopic) [M + H] 1438.1 (1437.6
calcd for C65H77N22O17).
DNA Reagents and Materials. Enzymes were purchased from
Boehringer-Mannheim and used with their supplied buffers. Deoxy-
adenosine and thymidine 5′-[R-32P] triphosphates were obtained from
Amersham, and deoxyadenosine 5′-[γ-32P]triphosphate was purchased
from I. C. N. Sonicated, deproteinized calf thymus DNA was acquired
from Pharmacia. RNase free water was obtained from USB and used
for all footprinting reactions. All other reagents and materials were
used as received. All DNA manipulations were performed according
to standard protocols.22
Preparation of 3′- and 5′-End-Labeled Restriction Fragments.
The plasmid pJT8 was constructed as previously reported. pJT8 was
linearized with AflII and FspI restriction enzymes, then treated with
the Sequenase enzyme, deoxyadenosine 5′-R-32P]triphosphate, and
thymidine 5′-[R-32P]triphosphate for 3′ labeling. Alternatively, these
plasmids were linearized with AflII, treated with calf alkaline phos-
phatase, and then 5′-labeled with T4 polynucleotide kinase and
deoxyadenosine 5′-[γ-32P]triphosphate. The 5′-labeled fragment was
1
yield; TLC (ethyl acetate) Rf 0.7; H NMR (DMSO-d6) δ 9.06 (s, 1
H), 7.29 (d, 2 H, J ) 7.8 Hz), 7.21 (d, 2 H. J ) 7.8 Hz), 7.06 (s, 1 H),
6.6 (s, 1 H), 5.14 (s, 2 H), 3.74 (s, 3 H), 3.52 (s, 2 H), 1.38 (s, 9 H).
Boc-aminoacyl-Pyrrolyl-4-(oxymethyl)-Pam Resin (7). Boc-Pyr-
rolyl-4-(oxymethyl)phenylacetic acid (6) (1 g, 2.57 mmol) was dissolved
in 6.5 mL DMF. HOBt (382 mg, 2.8 mmol) followed by DCC (735
mg, 3.2 mmol) was added and the reaction mixture shaken for 4 h at
room temperature. The precipitated DCU was filtered and the reaction
mixture was added to 10 g aminomethyl-polystyrene resin (0.7 mmol/g
substitution) previously swollen for 30 min in DMF. DIEA (1 mL)
was added and the reaction shaken until the resin was determined by
the ninhydrin test and picric acid titration to be approximately 0.1
mmol/g substituted. The resin was washed with DMF, and the remaining
amine groups were capped by acetylation (2×) with excess acetic
anhyride capping solution (2:2:1 DMF/Ac2O/DIEA). The resin was
washed with DMF (1 × 20 mL), DCM (1 × 20 mL), and MeOH (1 ×
20 mL) and dried in vacuo.
H2N-γ-ImPyPyPy-(R)Bocγ-PyPyPyPy-COOH (13). Cbz-γ-ImPy-
PyPy-(R)Fmocγ-PyPyPyPy-Pam resin was synthesized in a stepwise
fashion by machine-assisted solid-phase methods from Boc-Py-Pam
resin (0.1 mmol/g). (R)-2-Fmoc-4-Boc-diaminobutyric acid (0.7 mmol)
was incorporated as previously described for Boc-γ-aminobutyric acid.
Cbzγ-ImPyPyPy-(R)Fmocγ-PyPyPyPy-Pam resin was placed in a glass
20-mL peptide synthesis vessel and treated with DMF (2 mL) followed
by piperidine (8 mL) and agitated (22 °C, 30 min). Cbzγ-ImPyPyPy-
(R)H2Nγ-PyPyPyPy-Pam resin was isolated by filtration and washed
sequentially with an excess of DMF, DCM, MeOH, and ethyl ether,
and the amine resin was dried in vacuo. Cbzγ-ImPyPyPy-(R)H2Nγ-
PyPyPyPy-Pam resin was then treated with 300 mg of Boc-anhydride
(2 mL DMF, 1 mL DIEA, 55 °C, 30 min) to provide Cbzγ-ImPyPyPy-
(R)Bocγ-PyPyPyPy-Pam resin. Palladium acetate (200 mg, 37 °C, 10
min) in 2 mL of DMF was added to a sample of the Boc-resin (200
mg, 0.09 mmol/gram), ammonium formate (500 mg) was then added,
and the reaction was allowed to shake for 8 h at 37 °C. The reaction
mixture was then filtered to remove resin and precipitated palladium
metal and diluted with 0.1% (wt/v) TFA (8 mL), and the resulting
solution was purified by reversed phase HPLC to yield the trifluoro-
acetate salt of H2N-γ-ImPyPyPy-(R)Bocγ-PyPyPyPy-COOH: 17 mg,
1
66% recovery; H NMR (DMSO-d6) δ 10.35 (s, 1 H), 9.98 (s, 2 H),
9.96 (s, 3 H), 9.92 (s, 2 H), 7.98 (m, 1 H), 7.75 (br s 3 H), 7.44 (m, 1
H), 7.42 (s, 1 H), 7.30 (s, 1 H), 7.26 (s, 1 H), 7.25 (m, 2 H), 7.18 (m,
2 H), 7.15 (d, 1 H, J ) 1.3 Hz), 7.13 (s, 1 H), 7.06 (m, 3 H), 6.96 (s,
1 H), 6.93 (s, 1 H), 6.90 (s, 1 H), 6.84 (d, 1 H, J ) 1.8 Hz), 4.07 (q,
1 H, J ) 6.0 Hz), 3.95 (s, 3 H), 3.84 (m, 15 H), 3.82 (s, 3 H), 3.79 (s,
3 H), 3.22 (m, 2 H), 2.81 (m, 2 H), 2.39 (t, 2 H, J ) 6.9 Hz), 1.83 (m,
4 H), 1.39 (s, 9 H), MALDI-TOF-MS (monoisotopic) [M + H] 1281.5
(1281.6 calcd for C60H73N20O13).
H2N-γ-ImPyPyPy-(R)Bocγ-ImPyPyPy-COOH (12). Cbz-γ-ImPy-
PyPy-(R)Boc-γ-ImPyPyPy-Pam resin was prepared as described for (13).
A sample of the Boc-resin (330 mg, 0.09 mmol/gram) was treated with

DNA Binding Properties of Cycle Polyamide Motifs
J. Am. Chem. Soc., Vol. 121, No. 6, 1999 1129
then digested with FspI. The labeled fragment (3′ or 5′) was loaded
onto a 6% nondenaturing polyacrylamide gel, and the desired 229 base
pair band was visualized by autoradiography and isolated.
DNase I Footprinting.16 All reactions were carried out in a volume
of 400 µL. We note explicitly that no carrier DNA was used in these
reactions until after DNase I cleavage. A polyamide stock solution or
water (for reference lanes) was added to an assay buffer where the
final concentrations were 10 mM Tris‚HCl buffer (pH 7.0), 10 mM
KCl, 10 mM MgCl2, 5 mM CaCl2, and 30 kcpm 3′-radiolabeled DNA.
The solutions were allowed to equilibrate for a minimum of 12 h at 22
°C. Cleavage was initiated by the addition of 10 µL of a DNase I stock
solution (diluted with 1 mM DTT to give a stock concentration of 1.875
u/mL) and was allowed to proceed for 7 min at 22 °C. The reactions
were stopped by adding 50 µL of a solution containing 2.25 M NaCl,
150 mM EDTA, 0.6 mg/mL glycogen, and 30 µM base pair calf thymus
DNA and then ethanol-precipitated. The cleavage products were
resuspended in 100 mM Tris-borate-EDTA/80% formamide loading
buffer, denatured at 85 °C for 6 min, and immediately loaded onto an
8% denaturing polyacrylamide gel (5% cross-link, 7 M urea) at 2000
V for 1 h. The gels were dried under vacuum at 80 °C and then
quantitated using storage phosphor technology. Equilibrium association
constants were determined as previously described.12
MPE‚Fe(II) Footprinting.14 All reactions were carried out in a
volume of 400 µL. A polyamide stock solution or water (for reference
lanes) was added to an assay buffer where the final concentrations were
20 mM HEPES buffer (pH 7.0), 10 mM NaCl, 100 µM/base pair calf
thymus DNA, and 30 kcpm 3′- or 5′-radiolabeled DNA. The solutions
were allowed to equilibrate for 4 h. A fresh 50 µM MPE‚Fe(II) solution
was prepared from 100 µL of a 100 µM MPE solution and 100 µL of
a 100 µM ferrous ammonium sulfate (Fe(NH4)2(SO4)2‚6H2O) solution.
MPE‚Fe(II) solution (5 µM) was added to the equilibrated DNA, and
the reactions were allowed to equilibrate for 5 min. Cleavage was
initiated by the addition of dithiothreitol (5 mM) and allowed to proceed
for 14 min. Reactions were stopped by ethanol precipitation, resus-
pended in 100 mM Tris-borate-EDTA/80% formamide loading buffer,
and denatured at 85 °C for 6 min, and a 5-µL sample (∼15 kcpm) was
immediately loaded onto an 8% denaturing polyacrylamide gel (5%
cross-link, 7 M urea) at 2000 V.
Affinity Cleaving.15 All reactions were carried out in a volume of
400 µL. A polyamide stock solution or water (for reference lanes) was
added to an assay buffer where the final concentrations were 20 mM
HEPES buffer (pH 7.0), 20 mM NaCl, 100 µM/base pair calf thymus
DNA, and 20 kcpm 3′- or 5′-radiolabeled DNA. The solutions were
allowed to equilibrate for 8 h. A fresh solution of ferrous ammonium
sulfate (Fe(NH4)2(SO4)2‚6H2O) (10 µM) was added to the equilibrated
DNA, and the reactions were allowed to equilibrate for 15 min.
Cleavage was initiated by the addition of dithiothreitol (10 mM) and
allowed to proceed for 30 min. Reactions were stopped by ethanol
precipitation, resuspended in 100 mM tris-borate-EDTA/80% forma-
mide loading buffer, and denatured at 85 °C for 6 min, and the entire
sample was immediately loaded onto an 8% denaturing polyacrylamide
gel (5% cross-link, 7 M urea) at 2000 V.
Acknowledgment. We are grateful to the National Institutes
of Health (GM-27681) for research support, the National
Institutes of Health for a research service award to D.M.H., J.
Edward Richter for an undergraduate fellowship to J.M.T., and
the Howard Hughes Medical Institute for a predoctoral fellow-
ship to E.E.B. We thank G.M. Hathaway for MALDI-TOF mass
spectrometry.
Supporting Information Available: Affinity-cleaving ex-
periments of 4-E‚Fe(III) (PDF). This material is available free
of charge via the Internet at http://pubs.acs.org.
JA983206X