Programmable oligomers for minor groove DNA recognition

Article abstract of DOI:10.1021/ja0621795

The four Watson-Crick base pairs of DNA can be distinguished in the minor groove by pairing side-by-side three five-membered aromatic carboxamides, imidazole (Im), pyrrole (Py), and hydroxypyrrole (Hp), four different ways. On the basis of the paradigm of

Full text of DOI:10.1021/ja0621795

Published on Web 06/20/2006
Programmable Oligomers for Minor Groove DNA Recognition
Raymond M. Doss, Michael A. Marques, Shane Foister, David M. Chenoweth, and
Peter B. Dervan*
Contribution from The DiVision of Chemistry and Chemical Engineering, California Institute of
Technology, Pasadena, California 91125
Received March 30, 2006; E-mail: dervan@caltech.edu
Abstract: The four Watson-Crick base pairs of DNA can be distinguished in the minor groove by pairing
side-by-side three five-membered aromatic carboxamides, imidazole (Im), pyrrole (Py), and hydroxypyrrole
(Hp), four different ways. On the basis of the paradigm of unsymmetrical paired edges of aromatic rings
for minor groove recognition, a second generation set of heterocycle pairs, imidazopyridine/pyrrole (Ip/Py)
and hydroxybenzimidazole/pyrrole (Hz/Py), revealed that recognition elements not based on analogues of
distamycin could be realized. A new set of end-cap heterocycle dimers, oxazole-hydroxybenzimidazole
(No-Hz) and chlorothiophene-hydroxybenzimidazole (Ct-Hz), paired with Py-Py are shown to bind
contiguous base pairs of DNA in the minor groove, specifically 5′-GT-3′ and 5′-TT-3′, with high affinity and
selectivity. Utilizing this technology, we have developed a new class of oligomers for sequence-specific
DNA minor groove recognition no longer based on the N-methyl pyrrole carboxamides of distamycin.
Introduction
Aberrant gene expression is the cause of many diseases, and
the ability to reprogram transcriptional pathways using cell-
permeable small molecules may, one day, have an impact on
human medicine.1 DNA-binding polyamides, which are based
on the architecture of the natural products netropsin and
distamycin A,2,3a,b are capable of distinguishing all four Wat-
son-Crick base pairs in the DNA minor groove and have been
the subject of intense study along with many other classes of
minor groove binders.3c-f,4,5 Sequence-specific recognition of
the minor groove of DNA by polyamides arises from the pairing
of three different antiparallel five-membered heterocyclic amino
acids, pyrrole (Py), imidazole (Im), and hydroxypyrrole (Hp).4,5
The direct read out, or information face, on the inside of the
crescent-shaped polyamide may be programmed by the incre-
mental change of atoms on the corners of the ring pairs presented
to the DNA minor groove floor. Stabilizing and, importantly,
destabilizing interactions with the different edges of the four
Figure 1. Structures of dimers (a) imidazole-hydroxybenzimidazole (Im-
Hz), (b) oxazole-hydroxybenzimidazole (No-Hz), and (c) chlorothiophene-
hydroxybenzimidazole (Ct-Hz) dimer caps in comparison with their
respective five membered ring systems. Hydrogen-bonding surfaces to the
DNA minor-groove floor are bolded.
(1) (a) Darnell, J. E. Nat. ReV. Cancer 2002, 2, 740-748. (b) Pandolfi, P. P.
Oncogene 2001, 20, 3116-3127. (c) Dervan, P. B.; Edelson, B. S. Curr.
Opin. Struct. Biol. 2003, 13, 284-299.
(2) (a) Finlay, A. C.; Hochstein, F. A.; Sobin, B. A.; Murphy, F. X. J. Am.
Chem. Soc. 1951, 73, 341-343. (b) Arcamone, F. N. V.; Penco, S.; Orezzi,
P.; Nicolella, V.; Pirelli, A. Nature 1964, 203, 1064-1065.
(3) (a) Kopka, M. L.; Yoon, C.; Goodsell, D.; Pjura, P.; Dickerson, R. E. Proc.
Natl. Acad. Sci. U.S.A. 1985, 82, 1376-1380. (b) Pelton, J. G.; Wemmer,
D. E. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 5723-5727. (c) Wemmer,
D. E. Annu. ReV. Biomol. Struct. 2000, 29, 439-461. (d) Buchmueller, K.
L.; Staples, A. M.; Howard, C. M.; Horick, S. M.; Uthe, P. B.; Minh Le,
N.; Cox, K. K.; Nguyen, B.; Pacheco, K. A. O.; Wilson, D. W.; Lee, M.
J. Am. Chem. Soc. 2004, 127, 742-750. (e) Baraldi, P. G.; Bovero, A.;
Fruttarolo, F.; Preti, D.; Tabrizi, M. A.; Pavani, M. G.; Romagnoli, R. Med.
Res. ReV. 2004, 24, 475-528. (f) Reddy, P. M.; Toporowski, J. W.; Kahane,
A. L.; Bruice, T. C. Bioorg. Med. Chem. Lett. 2005, 15, 5531-5536.
(4) White, S.; Szewczyk, J. W.; Turner, J. M.; Baird, E. E.; Dervan, P. B.
Nature 1998, 391, 468-471.
Watson-Crick bases are modulated by shape complementarity
and specific hydrogen bonds.4-6,7l For example, the Im/Py pair
distinguishes G‚C from C‚G, T‚A, and A‚T. Im presents a lone
pair of electrons to the DNA minor groove and can accept a
hydrogen bond from the exocyclic amine of guanine.5 Addition-
ally, the Hp/Py pair distinguishes T‚A from A‚T, G‚C, and
C‚G.4-6 Hp projects an exocyclic OH group toward the minor
(5) (a) Kielkopf, C. L.; White, S.; Szewcyzk, J. W.; Turner, J. M.; Baird, E.
E.; Dervan, P. B.; Rees, D. C. Science 1998, 282, 111-115. (b) 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) Urbach, A. R.; Szewczyk, J. W.; White, S.; Turner, J. M.; Baird, E. E.;
Dervan, P. B. J. Am. Chem. Soc. 1999, 121, 11621-11629. (b) White, S.;
Turner, J. M.; Szewezyk, J. W.; Baird, E. E.; Dervan, P. B. J. Am. Chem.
Soc. 1999, 121, 260-261.
9
9074
J. AM. CHEM. SOC. 2006, 128, 9074-9079
10.1021/ja0621795 CCC: $33.50 © 2006 American Chemical Society

Programmable Oligomers for Minor Groove DNA Recognition
A R T I C L E S
Figure 2. Postulated hydrogen-bonding models for the 1:1 polyamide-DNA complexes with their matched sequence and their ball-and-stick representations.
(a) Im-Hz-Py-Py-γ-Im-Py-Py-Py-â-Dp (1), (b) No-Hz-Py-Py-γ-Im-Py-Py-Py-â-Dp (2), and (c) Ct-Hz-Py-Py-γ-Im-Py-Py-Py-â-Dp (3).
groove floor that is sterically accommodated in the cleft of the
T‚A base pair, preferring to lie over T, not A.5 These pairing
rules have proven useful for programmed recognition of a broad
repertoire of DNA sequences; however, the hydroxypyrrole ring
system has proven to be unstable over time and in the presence
of acid, further prompting our search for new T‚A/A‚T
recognition elements. In addition, sequence-dependent changes
in the microstructure of DNA (intrinsic minor groove width,
minor groove flexibility, and inherent DNA curvature)7a-k
combined with structural and conformational changes among
polyamides make the targeting of certain sequences less than
optimal, leading us to explore whether other novel heterocyclic
recognition elements could be discovered for use in DNA groove
recognition within the unsymmetrical pairing paradigm.7l,8-10
Furthermore, from a medicinal chemistry point of view, a
broader tool kit of sequence-specific recognition elements for
DNA beyond polyamides would be useful as our artificial
transcription factor program moves from cell culture11 to small
animal studies.
We recently reported that the benzimidazole ring can be an
effective platform for the development of modular paired
recognition elements for the minor groove of DNA.9,10 The
benzimidazole 6-5 bicyclic ring structure, though having
slightly different curvature from the classic five-membered
pyrrole-carboxamides, presents an “inside edge” with a similar
atomic readout to the DNA minor groove floor, effectively
mimicking Py, Im, and Hp. We demonstrated that the imida-
zopyridine/pyrrole pair Ip/Py distinguishes G‚C from C‚G and
the hydroxybenzimidazole/pyrrole pair Hz/Py distinguishes T‚
A from A‚T, providing a solution to the unanticipated hydroxy-
pyrrole instability limitation.9,10 The question arises whether this
second generation solution to DNA recognition can be elabo-
rated further, deleting incrementally almost all carboxamide
linkages in the backbone of the hairpin motif.12
We report here a new set of heterocycle dimer pairs,12 which
represents a step from single base-pair recognition toward a two
letter approach to molecular recognition of the minor groove
of DNA (Figure 1). We move from single letters to syllables.
New heterocycles were designed by combining the T-specific
hydroxybenzimidazole (Hz) with oxazole (No) rings and
chlorothiophene (Ct) caps8 to afford the recognition elements
No-Hz and Ct-Hz, respectively (Figure 1). Quantitative
DNase I footprinting titrations were used to determine DNA
binding affinities of hairpin oligomers containing the No-Hz
and Ct-Hz dimers paired with Py-Py dimer for each of the
four Watson-Crick bases (Figure 2). When positioned at the
termini of hairpin polyamides, the No-Hz/Py-Py and Ct-
(7) (a) Hays, F. A.; Teegarden, A.; Jones, Z. J. R.; Harms, M.; Raup, D.;
Watson, J.; Cavaliere, E.; Shing Ho, P. Proc. Natl. Acad. Sci. U.S.A. 2005,
102, 7157-7162. (b) Beveridge, D. L.; Barreiro, G.; Byun, K. S.; Case,
D. A.; Cheatham, T. E., 3rd; Dixit, S. B.; Giudice, E.; Lankas, F.; Lavery,
R.; Maddocks, J. H.; Osman, R.; Seibert, E.; Sklenar, H.; Stoll, G.; Thayer,
K. M.; Varnai, P.; Young, M. A. Biophys. J. 2004, 87, 3799-3813. (c)
Dixit, S. B.; Beveridge, D. L.; Case, D. A.; Cheatham, T. E., 3rd; Giudice,
E.; Lankas, F.; Lavery, R.; Maddocks, J. H.; Osman, R.; Sklenar, H.; Thayer,
K. M.; Varnai, P. Biophys. J. 2005, 89, 3721-3740. (d) Wu, H.; Crothers,
D. M. Nature 1984, 308, 509. (e) Steitz, T. A. Annu. ReV. Biophys. 1990,
23, 205. (f) Goodsell, D. S.; Kopka, M. L.; Cascio, D.; Dickerson, R. E.
Proc. Natl. Sci. U.S.A. 1993, 90, 2930. (g) Paolella, D. N.; Palmer, R.;
Schepartz, A. Science 1994, 264, 1130. (h) Kahn, J. D.; Yun, E.; Crothers,
D. M. Nature 1994, 368 163. (i) Geierstanger, B. H.; Wemmer, D. E. Annu.
ReV. Biochem. 1995, 24, 463. (j) Hansen, M. R.; Hurley, L. H. Acc. Chem.
Res. 1996, 29, 249. (k) Turner, J. M.; Swalley, S. E.; Baird, E. E.; Dervan,
P. B. J. Am. Chem. Soc. 1998, 120, 6219-6226. (l) Marques, M. A.; Doss,
R. M.; Urbach, A. R.; Dervan, P. B. HelV. Chim. Acta 2002, 85, 4485-
4517.
(8) Foister, S.; Marques, M. A.; Doss, R. M.; Dervan, P. B. Bioorg. Med. Chem.
2003, 11, 4333-4340.
(11) For downregulation of an endogenous gene in cell culture by polyamides
see: Olenyuk, B. Z.; Zhang, G.; Klco, J. M.; Nickols, N. G.; Kaelin, W.
G., Jr.; Dervan, P. B. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 16768-
16773.
(12) Heterocycle pair refers to cofacial stack (noncovalent interaction), whereas
dimer refers to two covalently attached heterocycles.
(9) Briehn, C. A.; Weyermann, P.; Dervan, P. B. Chem.-Eur. J. 2003, 9, 2110-
2122.
(10) (a) Renneberg, D.; Dervan, P. B. J. Am. Chem. Soc. 2003, 125, 5707-
5716. (b) Marques, M. A.; Doss, R. M.; Foister, S.; Dervan, P. B. J. Am.
Chem. Soc. 2004, 126, 10339-10349.
9
J. AM. CHEM. SOC. VOL. 128, NO. 28, 2006 9075

A R T I C L E S
Doss et al.
Scheme 1. Representative Solid-Phase Synthesis of Polyamide 2 and 3 along with a Table of the Amino Acid Building Blocks Useda
a Reaction conditions: (i) 80% TFA/DCM; (ii) Boc-Py-OBt, DIEA, DMF; (iii) Ac2O, DIEA, DMF; (iv) repeat i-iii × 2; (v) 80% TFA/DCM; (vi)
Boc-Im-OH, HBTU, DIEA, DMF; (vii) Ac2O, DIEA, DMF; (viii) 80% TFA/DCM; (ix) Boc-γ-OH, HBTU, DIEA, DMF; (x) Ac2O, DIEA, DMF; (xi) repeat
i-iii × 2; (xii) 80% TFA/DCM; (xiii) No-HzOMe-OH, HBTU, DIEA, DMF; (xiv) 3-(dimethylamino)-1-propylamine (Dp), 80 °C, 2 h; (xv) preparative
HPLC; (xvi) thiophenol, NaH, DMF; (xvii) preparative HPLC; (xviii) Ct-HzOMe-OH, HBTU, DIEA, DMF; (xix) same as steps xiv-xvii.
amino acid building blocks: Boc-Py-OBt (4), Boc-Im-OH (5), and
Hz/Py-Py dimer pairs are found to target 5′-GT-3′ and 5′-TT-
Boc-γ-OH (6) (Scheme 1). The base resins were then split into smaller
batches for coupling to the final dimeric caps. Boc-protected amino
acid monomers for Boc-Py-OBt (4) and Boc-Im-OH (5) were synthe-
sized according to previously reported procedures.8,13,14 Dimeric cap
synthesis for No-HzOMe-OH (7) and Ct-HzOMe-OH (8) are detailed
in the Supporting Information. Couplings were achieved using preac-
tivated monomers (Boc-Py-OBt) or HBTU activation in a DIEA and
DMF mixture. Coupling times ran from 3 to 24 h at 25-40 °C.
Deprotection of the growing polyamide was accomplished using 80%
TFA/DCM. Polyamides were cleaved from the resin by treatment with
neat 3-(dimethylamino)-1-propylamine (Dp) at 80 °C for 2 h and
purified by preparatory reverse phase HPLC. Deprotection of the
methoxy-protected polyamides was done using a mix of thiophenoxide
in DMF at 80 °C to provide the free hydroxy derivatives after a second
HPLC purification: Im-Hz-Py-Py-γ-Im-Py-Py-Py-â-Dp (1),
No-Hz-Py-Py-γ-Im-Py-Py-Py-â-Dp (2), Ct-Hz-Py-Py-γ-Im-
Py-Py-Py-â-Dp (3). (See Supporting Information for full experimental
details.)
3′ sequences, respectively, with high affinity and good speci-
ficity. With the development of dimer pairs capable of recognizing
a 5′-GT-3′ sequence of DNA, we could address the question
whether a hairpin oligomer comprised of four dimer units will
bind the site 5′-GTAC-3′, a sequence formally containing all
four Watson-Crick base pairs. Such a molecule represents our
first programmable oligomer, which demonstrates excellent
DNA binding properties without containing a single pyrrole-
or imidazole-carboxamide based on the natural product dista-
mycin, a design benchmark for biomimetic chemistry and the
field of DNA recognition.
Experimental Methods
Polyamide Synthesis. Hairpin polyamides were synthesized manu-
ally from Boc-â-PAM resin in a stepwise fashion using Boc-protected
monomeric and dimeric amino acids according to established solid-
phase protocols.13 Base Resin 1 (BR1) (H2N-Py-Py-γ-Im-Py-Py-
Py-â-Pam) was synthesized in gram quantities using the following
(14) Urbach, A. R.; Dervan, P. B. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 4343-
(13) Baird, E. E.; Dervan, P. B. J. Am. Chem. Soc. 1996, 118, 6141-6146.
4348.
9
9076 J. AM. CHEM. SOC. VOL. 128, NO. 28, 2006

Programmable Oligomers for Minor Groove DNA Recognition
A R T I C L E S
Figure 3. Quantitative DNase I footprinting experiments in the hairpin motif for polyamides 1, 2, and 3, respectively, on the 278 bp, 5′-end-labeled PCR
product of plasmid CW15: (lane 1) intact DNA; (lane 2) A reaction; (lane 3) DNase I standard; (lanes 4-14) 1 pM, 3 pM, 10 pM, 30 pM, 100 pM, 300
pM, 1 nM, 3 nM, 10 nM, 30 nM, 100 nM polyamide, respectively. Each footprinting gel is accompanied by the following: (top) Chemical structure of the
pairing of interest; (bottom) binding isotherms for the four designed sites. θnorm values were obtained according to published methods.15 A binding model
for the hairpin motif is shown centered at the top as a ball-and-stick model with the polyamide bound to its target DNA sequence. Imidazoles and pyrroles
are shown as filled and nonfilled circles, respectively; â-alanine is shown as a diamond; the γ-aminobutyric acid turn residue is shown as a semicircle
connecting the two subunits.
site 5′-TTTACA-3′ (Ka ) 2.4 × 109 M-1) (Figure 3, Table 1)
Results
with both the Ct and the Hz halves of the dimer preferring to
DNA Affinity and Sequence Specificity of Dimer Caps.
Quantitative DNase-I footprinting titrations were carried out for
polyamides 1-3. All polyamides were footprinted on the 285-
base-pair PCR product of plasmid pCW15. In all cases, the
DNA-sequence specificity at the cap position (in bold) was
determined by varying a single DNA base pair within the
sequence, 5′-TXTACA-3′, to all four Watson-Crick base pairs
(X ) A, T, G, C) and comparing the relative affinities of the
rest over the less bulky T in the asymmetric cleft of a T‚A base
pair. Placing the Ct ring adjacent to the Hz resulted in a 10-
fold specificity for T > A using the Ct-Hz system. Polyamide
1, which contains the Im-Hz dimer, did not bind its designed
match site 5′-TGTACA-3′ with any appreciable level of
specificity exhibiting affinities of Ka ) 1.6 × 108 and 4.0 ×
108 M-1 for the GT and AT sites, respectively.
Oxazole cap (polyamide 2) was incorporated into the dimer
cap system, and the affinity for its designed match site,
5′-TGTACA-3′, was examined. Polyamide 2 successfully
targeted its designed match site with an appreciable level of
specificity (25-fold) and a match site affinity of Ka ) 6.8 ×
109 M-1 (Figure 3, Table 1). With the development of the
chlorothiophene and oxazole dimer caps, the range of targetable
sequences by polyamides has been expanded (Table 1).
resulting complexes. The variable base-pair position was
designed to be adjacent to the Hz ring, which has been shown
to specify for T when paired across from Py, so as to be able
to determine the binding properties of each compound to the
following two base-pair sequences: AT, TT, GT, and CT.
The sequence specificity of the Im-Hz and Ct-Hz dimers
for 5′-TXTACA-3′ were evaluated in polyamides 1 and 3,
respectively. As expected, polyamide 3 bound its designed match
9
J. AM. CHEM. SOC. VOL. 128, NO. 28, 2006 9077

A R T I C L E S
Table 1. Affinities of X/Py Ring Pairs Proximal to a Hydroxybenzimidazole Bicycle Ka [M-1 a,b
Doss et al.
]
a Values reported are the mean values from at least three DNase I footprinting titration experiments, with the standard deviation given in parentheses.
b Assays were performed at 22 °C in a buffer of 10 mM Tris‚HCl, 10 mM KCl, 10 mM MgCl2, and 5 mM CaCl2 at pH 7.0.
Figure 4. (a) Postulated hydrogen-bonding model and structure of oligomer 9. (b) Ball and stick representation of 9 and the 6-base-pair binding site with
variable region (W ) A or T) shown. (c) Quantitative DNase I footprint titration experiment on the 5′-32P-labeled PCR product shown with an illustration
and complete sequence of the 285 bp EcoRI/PυuII restriction fragment from plasmid pDEH10. Binding affinities are shown next to their respective binding
sites and the match site is designated.
Design of a Programmable Oligomer for 5′-GTAC-3′. The
synthesis of oligomer 9 containing four dimer units was achieved
via the stepwise addition of Boc-amino acid dimers in the same
manner as previously described polyamide syntheses. This “third
generation” oligomer’s binding properties were assessed in the
same context16 as previously reported for first and second
generation hairpin polyamides targeting the sequence, 5′-GTAC-
3′, containing the four Watson-Crick base pairs.4,10 Footprinting
of the oligomer on the previously characterized plasmid DEH10
showed a binding affinity of Ka ) 2.3 × 1010 M-1 for the match
site 5′-GTAC-3′ and affinities of Ka ) 3.5 × 109 and 9.8 ×
108 M-1 for the mismatch sites 5′-GAAC-3′ and 5′-GATC-3′,
respectively (Figure 4). Such a result demonstrates that a
compound consisting exclusively of 6-5 fused ring systems
and minimal carboxamide linkages is able to maintain good
levels of specificity and excellent binding affinity.
Discussion. Recent advances in hairpin polyamide designs
have traditionally focused on developing new modes of single
base-pair recognition by heterocyclic ring pairings. Previous
studies, however, have highlighted the fact that the microstruc-
(15) Trauger, J. W.; Dervan, P. B. J. Methods Enzymol. 2001, 340, 450-466.
(16) Im-Hp-Py-Py-γ-Im-Hp-Py-Py-â-Dp targets 5′-WGTACW-3′ with Ka
) 7.0 × 108 M-1 and Im-Hz-Py-Py-γ-Im-Hz-Py-Py-â-Dp targets
5′-WGTACW-3′ with Ka ) 4.6 × 108 M-1 where W ) A or T (see ref
10b).
9
9078 J. AM. CHEM. SOC. VOL. 128, NO. 28, 2006

Programmable Oligomers for Minor Groove DNA Recognition
A R T I C L E S
ture of DNA depends on the sequence in question.17 In addition,
structural and conformational changes among polyamides are
thought to have an impact on DNA affinity and sequence
specificity. Thus, we have taken a more global view of
molecular recognition, where our efforts have expanded from
designing modules that distinguish the four DNA base-pairs (i.e.,
pairing rules) to designing those that target short, discrete DNA
sequences.
Upon incorporation and evaluation of the G specific Im ring
into the Im-Hz dimer cap (polyamide 1) we were surprised to
find that it failed to demonstrate any preference for its designed
site in addition to displaying a significantly decreased affinity.
The shortcomings of the Im-Hz dimer prompted a search for
a ring system that was capable of specifying for G > C within
the X-Hz context. The oxazole (No) cap (Figure 1) was
considered because of its structural resemblance to Imsboth
rings present a nitrogen atom capable of hydrogen bonding to
the minor groove. When the No-Hz dimer was incorporated
into polyamide 2, it was found to be specific for its designed
sequence of 5′-TGTACA-3′ with a 25-fold preference for G >
C and an affinity of Ka ) 6.8 × 109 M-1 at its match site (Figure
3). The No-Hz dimer presents the same functionality to the
minor groove as the Im-Hz dimer but with an enhanced ability
to target a 5′-GT-3′ site, which could be due to a combination
oxazole lone-pair basicity and differential solvation/desolvation
effects between oxazole and imidazole.
The Ct-Hz dimer cap represents our first effort to target a
short sequence of DNA using sequence-inspired recognition
elements. Studies have shown that Hz exhibits specificity for
T > A at the N-1 positionsrelative to the polyamide
N-terminussand that Ct polyamides exhibited specificity for
T > A at the cap position with excellent polyamide affinities.
We hoped that a hybrid dimer would impart specificity for the
TT sequence while maintaining a biologically relevant affinity.
Polyamide 3 bound its designed match site 5′-TTTACA-3′ with
an affinity of Ka ) 2.4 × 109 M-1 and a specificity of 10-fold
for T‚A over A‚T. This result is attributed to the fact that both
the sulfur and hydroxyl groups prefer to lie over the less-bulky
thymine base in a T‚A base pair and that the -OH of the Hz
ring is able to form an energetically favorable hydrogen bond
with the O(2) carbonyl of thymine.10 Combined, these attributes
make this dimer the preferred solution for targeting consecutive
thymine residues.
As a first step in the design of programmable oligomers
devoid of directly linked pyrrole- or imidazole-carboxamides,
we incorporated the new No-Hz dimer into a hairpin structure,
oligomer 9, consisting only of 6-5 fused ring systems (Figure
2). The oligomer was designed to target the site 5′-GTAC-3′
and is a third generation molecule from our previously reported
hairpin polyamides, which were shown to code for the four
Watson-Crick base pairs in a sequence specific manner.16 To
evaluate the impact of removing four carboxamide linkages and
moving to a system consisting of only 6-5 fused recognition
elements, binding properties were evaluated using quantitative
DNase I footprinting titrations. Oligomer 9 was found to bind
its match site with an impressive affinity of Ka ) 2.3 × 1010
M-1 while discriminating against its mismatch sites of 5′-
GAAC-3′ and 5′-GATC-3′ with specificities (Ka match/Ka
mismatch) of ∼7- and ∼ 23-fold, respectively (Figure 4). The
four 6-5 fused rings of oligomer 9 present an “inside edge”
with complimentary shape to the minor groove floor as the
carboxamide linkages of traditional Py-Im-Hp polyamides.
The complementary bumps and holes fit together between the
oligomer and the DNA surface, which is the key to specificity.
The dimer recognition elements in oligomer 9 are linked by a
single carbon-carbon bond, resulting in fewer degrees of
rotational freedom, which may result in a reduced entropic
penalty for minor-groove binding. Undoubtedly, much of the
favorable energetics for complexation with DNA for all these
molecules is a result of differential solvation. The oligomer’s
large hydrophobic surface may result in increased favorable van
der Waals interactions with the walls of the minor groove.
Conclusion
Hairpin polyamides containing the No-Hz and Ct-Hz dimer
caps at the polyamide N-terminus are able to target 5′-GT-3′
and 5′-TT-3′ sequences with good affinity and specificity and
represent new recognition elements for the minor groove of
DNA. The No-Hz and Ct-Hz dimer caps represent attempts
to broaden heterocycle designs beyond single base pair interac-
tions. In addition, the development of the No-Hz cap has
allowed for the design of a DNA binding molecule, which in a
formal sense is no longer a polyamide, hence the term
programmable oligomer. We are encouraged by the fact that
this oligomer demonstrates excellent affinity for DNA while
exhibiting good levels of specificity. We hope to apply these
new heterocycles to the targeting of biologically relevant
sequences in the context of integrating artificial transcription
factors with living biological systems.
Acknowledgment. We thank The National Institutes of
Health for grant support, Caltech for a James Irvine Fellowship
to R.M.D., the Parsons Foundation Fellowship to M.A.M, and
the NSF for a fellowship to S.F.
Supporting Information Available: Complete experimental
details for the synthesis of building blocks (7 and 8) and
polyamides (1-3 and 9) along with compound characterization
data. This material is available free of charge via the Internet
at http://pubs.acs.org.
(17) Urbach, A. R.; Love, J. J.; Ross, S. A.; Dervan, P. B. J. Mol. Biol. 2002,
320, 55-71.
JA0621795
9
J. AM. CHEM. SOC. VOL. 128, NO. 28, 2006 9079