Discrimination of hairpin polyamides with an α-substituted-γ- aminobutyric acid as a 5′-TG-3′ reader in DNA minor groove

Article abstract of DOI:10.1021/ja0580587

Pyrrole-imidazole (Py-lm) polyamides containing stereospecifically α-amino- or α-hydroxyl-substituted γ-aminobutyric acid as a 5′-TG-3′ recognition element were synthesized by machine-assisted Fmoc solid-phase synthesis. Their binding properties to predetermined DNA sequences containing a core binding site of 5′-TGCNCA-3′/3′- ACGN′GT-5′ (N·N′ = A·T, T·A, G·C, and C·G) were then systematically studied by surface plasmon resonance (SPR). SPR results revealed that the pairing of stereospecifically α-amino-/α-hydroxyl-substituted γ-aminobutyric acids, (R or S)-α,γ-diaminobutyric acid (γRN or γSN) and (R or S)-α-hydroxyl-γ-aminobutyric acid (γRO or γSO), side-by-side with β-alanine (β) in such polyamides significantly influenced the DNA binding affinity and recognition specificity of hairpin polyamides in the DNA minor groove compared with β/β, β/γ, and γ/β pairings. More importantly, the polyamide Ac-lm-γSO-lmPy-γ-lmPyβPy-β-Dp (β/γSO) favorably binds to a hairpin DNA containing a core binding site of 5′-TGCNCA- 3′/3′-ACGN′GT-5′ (N·N′ = A·T) with dissociation equilibrium constant (K_D) of 1.9 × 10^-7 M over N·N′ = T·A with K_D = 3.7 × 10 ^-6 M, with a 19-fold specificity. By contrast, Ac-lm-γSN-lmPy- γ-lmPyβPy-β-Dp (β/γSN) binds to the above sequence with N·N′ = A·T with K_D = 8.7 × 10 ^-7 M over N·N′ = T·A with K_D = 8.4 × 10^-6 M, with a 9.6-fold specificity. The results also show that the stereochemistry of the α-substituent, as well as the α-substituent itself may greatly alter binding affinity and recognition selectivity of hairpin polyamides to different DNA sequences. Further, we carried out molecular modeling studies on the binding by an energy minimization method, suggesting that α-hydroxyl is very close to N3 of the 3′-terminal G to induce the formation of hydrogen bonding between hydroxyl and N3 in the recognition event of the polyamide Ac-lm-γSO-lmPy-γ- lmPyβPy-β-Dp (β/γSO) to 5′-TGCNCA-3′/3′- ACGN′GT-5′ (N·N′ = A·T). Therefore, SPR assays and molecular modeling studies collectively suggest that the (S)-α-hydroxyl-γ-aminobutyric acid (γSO) may act as a 5′-TG-3′ recognition unit.

Full text of DOI:10.1021/ja0580587

Published on Web 06/20/2006
Discrimination of Hairpin Polyamides with an
r-Substituted-γ-aminobutyric Acid as a 5′-TG-3′ Reader in
DNA Minor Groove
Wen Zhang, Toshikazu Bando, and Hiroshi Sugiyama*
Contribution from the Department of Chemistry, Graduate School of Science, Kyoto UniVersity,
Sakyo-ku, Kyoto 606-8502, Japan
Received November 28, 2005; E-mail: hs@kuchem.kyoto-u.ac.jp
Abstract: Pyrrole-imidazole (Py-Im) polyamides containing stereospecifically R-amino- or R-hydroxyl-
substituted γ-aminobutyric acid as a 5′-TG-3′ recognition element were synthesized by machine-assisted
Fmoc solid-phase synthesis. Their binding properties to predetermined DNA sequences containing a core
binding site of 5′-TGCNCA-3′/3′-ACGN′GT-5′ (N‚N′ ) A‚T, T‚A, G‚C, and C‚G) were then systematically
studied by surface plasmon resonance (SPR). SPR results revealed that the pairing of stereospecifically
R-amino-/R-hydroxyl-substituted γ-aminobutyric acids, (R or S)-R,γ-diaminobutyric acid (γRN or γSN) and
(R or S)-R-hydroxyl-γ-aminobutyric acid (γRO or γSO), side-by-side with â-alanine (â) in such polyamides
significantly influenced the DNA binding affinity and recognition specificity of hairpin polyamides in the
DNA minor groove compared with â/â, â/γ, and γ/â pairings. More importantly, the polyamide Ac-Im-γSO-
ImPy-γ-ImPyâPy-â-Dp (â/γSO) favorably binds to a hairpin DNA containing a core binding site of
5′-TGCNCA-3′/3′-ACGN′GT-5′ (N‚N′ ) A‚T) with dissociation equilibrium constant (KD) of 1.9 × 10-7
M
over N‚N′ ) T‚A with KD ) 3.7 × 10-6 M, with a 19-fold specificity. By contrast, Ac-Im-γSN-ImPy-γ-ImPyâPy-
â-Dp (â/γSN) binds to the above sequence with N‚N′ ) A‚T with KD ) 8.7 × 10-7 M over N‚N′ ) T‚A with
KD ) 8.4 × 10-6 M, with a 9.6-fold specificity. The results also show that the stereochemistry of the
R-substituent, as well as the R-substituent itself may greatly alter binding affinity and recognition selectivity
of hairpin polyamides to different DNA sequences. Further, we carried out molecular modeling studies on
the binding by an energy minimization method, suggesting that R-hydroxyl is very close to N3 of the 3′-
terminal G to induce the formation of hydrogen bonding between hydroxyl and N3 in the recognition event
of the polyamide Ac-Im-γSO-ImPy-γ-ImPyâPy-â-Dp (â/γSO) to 5′-TGCNCA-3′/3′-ACGN′GT-5′ (N‚N′ ) A‚
T). Therefore, SPR assays and molecular modeling studies collectively suggest that the (S)-R-hydroxyl-
γ-aminobutyric acid (γSO) may act as a 5′-TG-3′ recognition unit.
at a specific site when conjugated with an alkylating agent,4,5
and be introduced into living cells with the aim of possible
Introduction
Py-Im polyamides composed of two or three different types
of aromatic heterocyclic rings, analogous to the composition
of the antibiotics, netropsin and distamycin, are currently among
the most promising synthetic compounds for use in chemo-
therapy and biosensing.1 They can recognize a predetermined
DNA sequence in the minor groove of B-DNA,2,3 alkylate DNA
chemical regulation of gene expression.6 In their hairpin
structure, antiparallel side-by-side pairings of two aromatic
amino acids bind to DNA sequences, with a polyamide ring
packed specifically against each DNA base. N-Methylpyrrole
(Py) favors T, A, and C bases, excluding G; N-methylimidazole
(Im) is a G-reader; and 3-hydroxyl-N-methylpyrrol (Hp) is
specific for thymine base. To date, all four Watson-Crick base
pairs can be recognized using different pairings of three aromatic
amino acids, with an Im/Py pairing reading G‚C by symmetry,
a Py/Im pairing reading C‚G, an Hp/Py pairing being able to
distinguish T‚A from A‚T, G‚C, and C‚G, and a Py/Py pairing
nonspecifically discriminating both A‚T and T‚A from G‚C and
C‚G.7
(1) (a) Ehley, J. A.; Melander, C.; Herman, D.; Baird, E. E.; Ferguson, H. A.;
Goodrich, J. A.; Dervan, P. B.; Gottesfeld, J. M. Mol. Cell. Biol. 2002, 22,
1723-1733. (b) Wang Y.-D.; Dziegielewski, J.; Wurtz, N. R.; Dzi-
egielewski, B.; Dervan, P. B.; Beerman, T. Nucleic Acids Res. 2003, 31,
1208-1215. (c) Edayathumangalam, R. S.; Weyermann, P.; Gottesfeld, J.
M.; Dervan, P. B.; Luger, K. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 6864-
6869. (d) Philips, B. J.; Chang, A. Y.; Dervan, P. B.; Beerman, T. A. Mol.
Pharmacol. 2005, 67, 877-882. (e) Warren, C. L.; Kratochvil, N. C. S.;
Hauschild, K. E.; Foister, S.; Brezinski, M. L.; Dervan, P. B.; Phillips, G.
N., Jr.; Ansari, A. Z. Proc. Natl. Acad. Sci. U.S.A. 2005, 103, 867-872.
(f) Lai, Y.-M.; Fukuda, N.; Ueno, T.; Kishioka, H.; Matsuda, H.; Saito, S.;
Matsumoto, K.; Ayame, H.; Bando, T.; Sugiyama, H.; Mugishima, H.; Kim,
S.-J. J. Pharmacol. Exp. Ther. 2005, 315, 5-10. (g) Matsuda, H.; Fukuda,
N.; Ueno, T.; Tahira, Y.; Ayame, H.; Zhang, W.; Bando, T.; Sugiyama,
H.; Saito, S.; Matsumoto, K.; Mugishima, H.; Serie, K. J. Am. Soc. Nephrol.
2006, 17, 422-432. (h) Shinohara, K.; Sasaki, S.; Minoshima, M.; Bando,
T.; Sugiyama, H. Nucleic Acids Res. 2006, 34, 1189-1195. (i) Shinohara,
K.; Bando, T.; Sasaki, S.; Sakakibara, Y.; Minoshima, M.; Sugiyama, H.
Cancer Sci. 2006, 97, 219-225.
(4) Chang, A. Y,; Dervan, P. B. J. Am. Chem. Soc. 2000, 122, 4856-4864.
(5) (a) Tao, Z.-F.; Saito, I.; Sugiyama, H. J. Am. Chem. Soc. 2000, 122, 1602-
1608. (b) Bando, T.; Narita, A.; Saito, I.; Sugiyama, H. J. Am. Chem. Soc.
2003, 125, 3471-3485. (c) Bando, T.; Iida, H.; Tao, Z.-F.; Narita, A.;
Fukuda, N.; Yamori, T.; Sugiyama, H. Chem. Biol. 2003, 10, 751-758.
(d) Bando, T.; Narita, A.; Iwai, A.; Kihara, K.; Sugiyama, H. J. Am. Chem.
Soc. 2004, 126, 3406-3407.
(2) Dervan, P. B. Bioorg. Med. Chem. 2001, 9, 2215-2235.
(3) Dervan, P. B.; Edelson, B. S. Curr. Opin. Struct. Biol. 1999, 13, 284-
299.
(6) Best, T. P.; Edelson, B. S.; Nickols, N. G.; Dervan, P. B. Proc. Natl. Acad.
Sci. U.S.A. 2003, 100, 12063-12068.
9
8766
J. AM. CHEM. SOC. 2006, 128, 8766-8776
10.1021/ja0580587 CCC: $33.50 © 2006 American Chemical Society

Discrimination of Hairpin Polyamides
A R T I C L E S
There are disadvantages in using Hp as a T-specific recogni-
tion element. These disadvantages include:8 (1) the Hp residue
can degrade in solution over time in the presence of acid or
free radicals; (2) it has a lower affinity for T bases relative to
Py; and (3) Hp at the N-terminus of polyamides has almost no
specificity for any base of the Watson-Crick base pairs.
For these reasons, researchers have been investigating other
T-specific recognition elements as substitutes for Hp. Recently,
Dervan’s group has found that hydroxybenzimidazole (Hz) may
replace Hp as a T-reader.9 Although an Hz/Py pairing has an
association equilibrium constant (KA) of 5.7 × 108 M-1 for T‚
A, close to that of Hp/Py, and a specificity approximately 18-
fold greater for T‚A over A‚T, and the Hz-containing polyamide
is stable under acidic conditions, it only recognizes T base at
the 3′ end of contiguous 5′-TT-3′ due to structural limitations
of conjugates of Hz directly linked to Py.
In addition, it has been reported that the polyamide rise per
aromatic amino acid residue basically matches the pitch of the
B-DNA helix in the polyamide composed of not more than five
contiguous aromatic rings, but the polyamides are overcurved
compared to the curvature and twist of the DNA helix,
suggesting that not more than five contiguous rings are better
bound along the minor groove of B-DNA without loss of affinity
and specificity.10,11 Despite introducing â-alanine (â) as an
aliphatic substitute for a Py ring in the construction of the
polyamides, with the â residue taking advantage of its better
flexibility to act like a “spring” to allow the crescent-shaped
polyamide to match the curvature and twist of B-DNA, the â/â
pairing still nonspecifically reads T‚A and A‚T pairs in the same
way as Py/Py pairs.12 This has led to attempts to identify an
aliphatic monomer that can function as a base recognition
element, as well as a “spring” to adjust the curvature and length
of polyamides to fit to twist and pitch of B-DNA as a substitute
for the aromatic ring. Floreancig and colleagues have shown
that (S)-isoserine could serve as a T-recognition unit able to
discriminate T specifically over A with 7.2-fold increased
specificity.13
Figure 1. Molecular interaction model of polyamides with biotinylated
hairpin DNAs in SPR assays. X/Y represents different aliphatic monomer
pairings: â/â, â/γ, γ/â, γ/γ, â/γSO, â/γRO, γSO/â, γRO/â, â/γSN, â/γRN,
γSN/â, γRN/â; N‚N′ represents the four base pairs: T‚A, A‚T, C‚G, G‚C.
The biotinylated hairpin DNA was incorporated by the 5′ end onto precoated
streptavidin sensor chips. Im is indicated by the black circles, and Py the
open circles; â represents â-alanine residue, Dp is 3-(dimethylamino)-
propylamide, ⊂ represents the γ-aminobutanoic acid turn residue connecting
the two subunits, and Ac is acetyl.
amides: (R)-R,γ-diaminobutyric acid (γRN), (S)-R,γ-diami-
nobutyric acid (γSN), (R)-R-hydroxyl-γ-aminobutyric acid
(γRO), and (S)-R-hydroxyl-γ-aminobutyric acid (γSO). These
residues were based on the flexibility of the aliphatic γ-ami-
nobutyric acid, the same covalence number as that between
adjacent Py-Im aromatic ring carboxamides, and different
stereochemical effects of substituents of functionalized â-alanine
on the recognition of polyamides.13 The binding properties of
the designed polyamides to selected hairpin DNAs, including
their binding affinity, specificity, stoichiometry, kinetics, and
thermodynamics studied through equilibrium analysis [(KA) or
dissociation equilibrium constant (KD)] and kinetic analysis
[association rate constant (ka) or dissociation rate constant (kd)],
were investigated using the Biacore SPR technique (Figures1
and 2).14
Results and Discussion
Monomer Synthesis. The protected optically active mono-
mers 17a (S) and 17b (R) were synthesized according to the
procedure in Scheme 1. (2S)-4-[(9-Fluorenylmethoxy)carbonyl-
amino]-2-hydroxybutyric acid (14a) was prepared from the
corresponding optically active starting material 13a in a mixture
of water and 1,2-dimethoxyethane in the presence of a small
amount of sodium bicarbonate. The resultant Fmoc-protected
2-hydroxybutyric acid 14a was suspended in DCM, and an
excess of isobutene was introduced at a lower temperature. The
reaction was carried out until a clear solution was formed, and
after the mixture was worked up, the tert-butyl ester of the (2S)-
4-[(9-fluorenylmethoxy)carbonylamino]-2-tert-butoxybutyric acid
15a was obtained.15 The cleavage of the tert-butyl ester from
15a was accomplished in the presence of the high-activity
silylating agent TMSOTf to give (2S)-2-tert-butyloxy-4-(9-
fluorenylmethoxy) carbonylaminobutyric acid (17a) through a
very labile intermediate 16a.16 17b was also synthesized using
the above approach. The enantiomeric purity of these com-
Subsequent attempts have been made to explore ideal
aromatic B-DNA base binders binding in the minor groove,
including T-specific binders, to adjust the curvature and length
of the polyamide to fit the twist of the B-DNA helix by
examining other types of single five-membered heterocyclic
residues, including furan, thiazole, thiophene, 3-methyl-
thiophene, 4-methylthiazole, 3-hydroxythiophene, 1H-pyrrol,
and 1H-pyrazole; unfortunately, however, such hyperspecific
rings have not yet been discovered.8a
To address the above problems, as shown in Figure 1, we
designed four R-substituted γ-aminobutyric acid residues as
T-specific recognition elements in the construction of poly-
1
pounds based on H NMR analysis of (S)-(-)-1-(1-naphthyl)-
ethyl amide derivatives of 17a and 17b was shown to be more
than 97%.
It is worth noting that it was difficult to optimize the synthesis
conditions to give rise to the optically active compound 17a or
17b, without side reactions and racemization, as a base-labile
amino protective group, Fmoc, and an acid-labile hydroxy
protective group, tert-butyl, are simultaneously present in the
17a or 17b monomer. Two key steps are necessary to enable
(7) (a) Pelton, J. G.; Wemmer, D. E. Proc. Natl. Acad. Sci. U.S.A. 1989, 86,
5723-5727. (b) Wade, W. S.; Mrksich, M.; Dervan, P, B. J. Am. Chem.
Soc. 1992, 114, 8783-8794. (c) White, S.; Szewczyk, J. W.; Turner, J.
M.; Baird, E. E.; Dervan, P. B. Nature 1998, 391, 468-471.
(8) (a) Marques, M. A.; Doss, R. M.; Urbach, A. R.; Dervan, P. B. HelV. Chim.
Acta 2002, 85, 4485-4517. (b) Walker, W. L.; Kopka, M. L.; Goodsell,
D. S. Biopolymers. 1997, 44, 323-334.
(9) Renneberg, D.; Dervan, P. B. J. Am. Chem. Soc. 2003, 125, 5707-5716.
(10) Kielkopf, C. L.; Baird, E. E.; Dervan, P. B.; Rees, D. C. Nat. Struct. Biol.
1998, 5, 104-109.
(11) Kelly, J. J.; Baird, E. E.; Dervan, P. B. Proc. Natl. Acad. Sci. U.S.A. 1996,
93, 6981-6985.
(14) Supekova, L.; Pezacki, J. P.; Su, A. I.; Loweth, C. J.; Riedl, R.; Geierstanger,
B.; Schultu, P. G.; Wemmer, D. E. Chem. Biol. 2002, 9, 821-827.
(15) (a) Beyerman, H. C.; Bontekoe, J. S. Recl. TraV. Chim. Pays-Bas 1962,
81, 691-697. (b) Adamson, J. G.; Blaskovich, M. A.; Groenevelt, H.;
Lajoie, G. A. J. Org. Chem. 1991, 56, 3447-3449.
(12) Trauger, J. W.; Baird, E. E.; Mrksich, M.; Dervan, P. B. J. Am. Chem.
Soc. 1996, 118, 6160-6166.
(13) Floreancig, P. E.; Swalley, S. E.; Trauger, J. W.; Dervan, P. B. J. Am.
Chem. Soc. 2000, 122, 6342-6350.
(16) Trzeciak, A.; Bannwarth, W. Synthesis 1996, 1434.
9
J. AM. CHEM. SOC. VOL. 128, NO. 27, 2006 8767

A R T I C L E S
Zhang et al.
Figure 2. Typical SPR sensorgrams for the interaction of polyamides with hairpin DNAs immobilized on the surface of sensor chips. The binding response
is changed over time. Experiments at each concentration were repeated three times, and the data were globally fitted to a 1:1 interaction model with mass
transport. The association phase was allowed to come to equilibrium at the different concentrations of the polyamides. All experiments were done using the
same DNA immobilization level sensor chips in HBS-EP buffer (0.01 M HEPES, pH 7.4, 0.15 M NaCl, 3 mM EDTA, 0.005% surfactant P20) with 0.1%
DMSO at 25 °C. (A) Polyamide 1 with â/â pairing for the recognition of A‚T base pairs at a concentration range from 3.75 × 10-8 M (lowest curve) to 2.00
× 10-7 M (highest curve). (B) Polyamide 5 with γ/γSO pairing for the recognition of A‚T base pairs at a concentration range from 2.00 × 10-8 M (lowest
curve) to 5.00 × 10-7 M (highest curve). (C) Polyamide 5 with γ/γSO pairing for the recognition of T‚A base pairs at a concentration range from 1.00 ×
10-7 M (lowest curve) to 1.00 × 10-6 M (highest curve). (D) Polyamide 9 with γ/γSN pairing for the recognition of A‚T base pair at a concentration range
from 1.00 × 10-8 M (lowest curve) to 5.00 × 10-7 M (highest curve). (E) Polyamide 9 with γ/γSN pairing for the recognition of T‚A base pair at a
concentration range from 1.50 × 10-7 M (lowest curve) to 1.25 × 10-6 M (highest curve).
our method to synthesize such optically active targets and to
avoid side reactions and the concomitant danger of racemization.
The first is the simultaneous formation of tert-butyl-ester and
-ether and the second is the conversion of tert-butyl ester to
the corresponding trimethylsilyl ester. The selective cleavage
of tert-butyl ester in the presence of tert-butyl ether has rarely
been reported.16 Less than 3% racemization (see Experimental
Section) was observed, and the final products 17a and 17b were
afforded at overall yields of 69.5% and 67.2%, respectively,
by the four steps in our experimental conditions, as shown in
the Experimental Section. More importantly, our experiments
showed that selective protection of the hydroxyl group by tert-
9
8768 J. AM. CHEM. SOC. VOL. 128, NO. 27, 2006

Discrimination of Hairpin Polyamides
A R T I C L E S
Scheme 1 . Synthesis of Two New Monomersa
a Reagents (i) (S)-(-)-4-amino-2-hydroxybutyric acid (13a) or (R)-(+)-amino-2-hydroxybutyric acid (13b), H2O, NaHCO3, (ii) FmocONSu, 1,2-
dimethoxyethane. (iii) Concentrated hydrochloric acid. (iv) 98% H2SO4, DCM, then CH2dCMe2. (v) 2 M aqueous NaOH. (vi) TMSOTf, DIEA, dioxane.
(vii) EtOAc/H2O. (viii) 10% aqueous hydrochloric acid. (ix) 5% aqueous KHSO4, and 10% aqueous K2SO4.
butyl and selective deprotection for tert-butyl ester with optically
active hydroxyl amino acids are possible. Our method has
potential for the synthesis of tert-butyl ethers of amino-protected
hydroxyl amino acids, which are valuable building blocks for
peptide synthesis.
aminolysis with 3-(dimethylamino)propylamine (Dp), to produce
the desired polyamides 1, 2, 3, 4, and tert-butyl-protected 5, 6,
7, 8, and Boc-protected 9, 10, 11, 12. The deprotection of
protected 5, 6, 7, 8, 9, 10, 11, and 12 was readily achieved
using an optimal deprotecting agent, TFA/TMS/TIS/H2O (91/
3/3/3, v/v), to give the final products 5, 6, 7, 8, 9, 10, 11, 12.
No racemization occurred as the removal of tert-butyloxy was
realized by breaking the oxygen-tert-butyl bond.15a A final purity
for all synthesized polyamides greater than 98% was determined
Polyamide Synthesis. To examine the ability of R-substituted
γ-butyric acid to recognize T bases, 12 polyamides were
synthesized by the Fmoc-chemistry solid-phase synthesis method,
starting with Fmoc-â-Ala-clear-acid resin and using the mono-
mers shown in Scheme 2. Wurtz et al. first reported general
protocols for machine-assisted Fmoc-chemistry solid-phase
synthesis of Py-Im polyamides starting from Fmoc-â-alanine-
Wang resin.17 Here, N-methylpyrrolidone (NMP) was replaced
with DMF as solvent, and HBTU, with HATU as activator. To
minimize side products in the synthesis of our polyamides,
capping was used for each cycle after coupling.18 After the first
Fmoc-Py acid 19 was activated by HATU in the presence of
DIEA, it was readily coupled to swollen resin 18 in the
synthesizer column for 60 min, in accordance with the estab-
lished protocols shown in the Experimental Section. Coupling
of the remaining aromatic monomers 19 and 20 was also
completed using this procedure. Coupling of modified and
unmodified aliphalic monomers 17a, 17b, 21, 22, 23, and 24
was accomplished with HATU using conditions the same as
those in the coupling of aromatic monomers. No effects were
found in the polyamide solid-phase synthesis by the respective
tert-butyloxy and tert-butyloxycarbonylamino groups in the
aliphatic monomers 17a, 17b and 23, 24. Interestingly, to obtain
higher recoveries of polyamides successfully, a double-couple
cycle must be employed upon coupling of any monomer to
imidazole and the coupling of imidazole itself, but their
mechanism is not well understood.
1
by a combination of analytical HPLC, H NMR, and mass
spectroscopy.
SPR Assay. By designing and using a self-complementary
biotinylated oligonucleotide, 5′-biotin-GGCCGATGN′GCAT-
GCTATAGCATGCNCATCGGCC-3′ (34 bases, N‚N′: T‚A,
A‚T, C‚G, G‚C) that was noncovalently bound to streptavidin-
functionalized Biacore sensor chips and that formed double-
stranded hairpin DNA including the target binding sites of the
polyamides with a T-specific recognition unit (Figure 1), the
kinetic and thermodynamic properties of polyamide binding to
DNA, evaluated from binding affinity, specificity, and stoichi-
ometry, can be obtained using SPR assays (Figure 2).
All SPR experiments for polyamides 1-12 were performed
using a Biacore X instrument with a sensor chip SA in HBS-
EP buffer (0.01 M HEPES, pH 7.4, 0.15 M NaCl, 3 mM EDTA,
0.005% surfactant P20) with 0.1% DMSO at 25 °C. The
dissociation equilibrium constants (KD, M) were obtained by
fitting the resulting sensorgrams (Figure 2) to a theoretical
model, providing the affinity and specificity of binding for the
functionalized γ-monomer/â against all four base pairs: T‚A,
A‚T, G‚C, C‚G. In addition, standard free energy change (∆G°,
kcal/mol-1) for binding was calculated from KD.
To obtain rational and accurate binding constants, varied
concentration ranges were used for different analytes in the SPR
experiments. The stoichiometry for polyamide binding to hairpin
DNA was determined by comparison of the maximum response
After the completion of machine-assisted solid-phase syn-
thesis, the polyamides were cleaved from resins 1a-12a by
(17) Wurtz, N. R.; Turner, J. M.; Baird, E. E.; Dervan, P. B. Org. Lett. 2001,
3, 1201-1203.
(18) Albericio, F. Curr. Opin. Chem. Biol. 2004, 8, 211-221.
for binding one molecule of polyamide per binding site, RUmax,
9
J. AM. CHEM. SOC. VOL. 128, NO. 27, 2006 8769

A R T I C L E S
Zhang et al.
Scheme 2 . Fmoc Solid-Phase Synthesis of Polyamides Starting from Commercially Available Fmoc-â-Ala-Clear-Acid Resin 18a
a Reaction conditions: (i) 20% piperidine/DMF; (ii) FmocPyCOOH, HATU, DIEA; (iii) 5% acetic anhydride + 5% pyridine/DMF; (iv) 20% piperidine/
DMF; (v) Fmoc-γ-Ala-OH for 1a, 2a, 5a, 6a, 9a, 10a, HATU, DIEA; Fmoc-γ-Abu-COOH for 3a, 4a, HATU, DIEA; 17a for 7a, HATU, DIEA; 17b for
8a, HATU, DIEA; 23 for 11a, HATU, DIEA; 24 for 12a, HATU, DIEA; (vi) 5% acetic anhydride + 5% pyridine/DMF; (vii) 20% piperidine/DMF; (viii)
FmocPyCOOH, HATU, DIEA; (ix) 5% acetic anhydride + 5% pyridine/DMF; (x) 20% piperidine/DMF; (xi) FmocImCOOH, HATU, DIEA; (xii) 5% acetic
anhydride + 5% pyridine/DMF; (xiii) 20% piperidine/DMF; (xiv) Fmoc-γ-Abu-COOH, HATU, DIEA; (xv) 5% acetic anhydride + 5% pyridine/DMF; (xvi)
20% piperidine/DMF; (xvii) FmocPyCOOH, HATU, DIEA; (xviii) 5% acetic anhydride + 5% pyridine/DMF; (xix) 20% piperidine/DMF; (xx) FmocImCOOH,
HATU, DIEA; (xxi) 5% acetic anhydride + 5% pyridine/DMF; (xxii) 20% piperidine/DMF; (xxiii) Fmoc-γ-Ala-OH for 1a, 3a, 7a, 8a, 11a, 12a, HATU,
DIEA; Fmoc-γ-Abu-COOH for 2a, 4a, HATU, DIEA; 17a for 5a, HATU, DIEA; 17b for 6a, HATU, DIEA; 23 for 9a, HATU, DIEA; 24 for 10a, HATU,
DIEA; (xxiv) 5% acetic anhydride + 5% pyridine/DMF; (xxv) 20% piperidine/DMF; (xxvi) FmocImCOOH, HATU, DIEA; (xxvii) 5% acetic anhydride +
5% pyridine/DMF; (xxviii) N,N-dimethylaminopropylamine for 1-4 and Boc and tert-butyl-protected 5-12, 55 °C, overnight; (xxix) TFA/TMS/TIS/H2O
(91/3/3/23, v/v) for 5-12
with the RU value at saturation. For a single-site model, RUmax
can be calculated using the following equation:
were examined by SPR (Figure 2, Table 1 and Supporting Infor-
mation Figures 1 and 2). It is noteworthy that almost no binding
of all examined polyamides 1-12 to 5′-TGCNCA-3′ (N ) C,
G) was observed at low concentrations of polyamides, but there
was large nonspecific binding at higher concentrations, resulting
in the inability to obtain precise dissociation equilibrium con-
stants. For polyamides 1, 2, 3, and 4 with an unsubstituted
aliphatic amino acid pairing, almost no specificity for 5′-
TGCNCA-3′/3′-ACGN′GT-5′ (N‚N′ ) T‚A and A‚T) was found
under our experimental conditions. However, a phenomenon
that should be noted is that the introduction of γ into 2 and 3
to pair with â reduces the binding affinity to A‚T by 35-fold
and 4.7-fold, respectively, and to T‚A by 30-fold and 4.5-fold,
respectively, when compared to 1 with a â/â pairing. In addition,
the substitution of two γ residues for the two â monomers
significantly decreased the binding affinity to the sequence 5′-
TGCACA-3′/3′-ACGTGT-5′ by approximately 846-fold and to
5′-TGCTCA-3′/3′-ACGAGT-5′ by 724-fold. These observations
indicate that introduction of a single γ to pair with â to allow
RUmax ) (RUDNA/MWDNA) × MWpolyamide × RII
where RUDNA is the amount of hairpin DNA immobilized on
the sensor chip in response units (RU), MW is the molecular
weight of the polyamide and the hairpin DNA, and RII is the
refractive index increment ratio of the compound to the
refractive index of the DNA. Here, RII values for these
compounds should range from 1.25 to 1.40.19-21
Effects of Stereochemistry of the Hydroxy Substituent on
Recognition and Binding Stoichiometry. To evaluate the
recognition ability of polyamides 1-12 with an aliphatic
monomer pairing to all four possible base pairs T‚A, A‚T, C‚
G, and G‚C in the DNA minor groove, their binding properties
(19) Davis, T. M.; Wilson, W. D. Anal. Biochem. 2000, 284, 348-353.
(20) Tanious, F. A.; Hamelberg, D.; Baillly, C.; Czarny, A.; Boykin, D. W.;
Wison, W. D. J. Am. Chem. Soc. 2004, 126, 143-153.
(21) Biacore, A. B. In BIAapplications Handbook; Biacore AB: Rapsgatan 7
S-754 50 Uppsala, Sweden, 1998; p 3-2.
9
8770 J. AM. CHEM. SOC. VOL. 128, NO. 27, 2006

Discrimination of Hairpin Polyamides
A R T I C L E S
Table 1. Dissociation Equilibrium Constants for Polyamides 1-12
recognition of A‚T and T‚A did reduce the binding affinity,
but not greatly.
In this case, when polyamide 5 with a â/γSO pairing binds
to 5′-TGCTCA-3′/3′-ACGAGT-5′, great steric hindrance by the
A base makes the additional hydrogen bond between N3 of 3′-
terminal guanine and R-hydroxyl group weak and leads in turn
to a weaker binding affinity. Therefore, monomer γSO actually
acts as a 5′-TG-3′ reader in the recognition events.
In this study, polyamide 5, with a â/γSO, showed a strong
binding preference for 5′-TGCACA-3′/3′-ACGTGT-5′ (N‚N′
) A‚T) over 5′-TGCTCA-3′/3′-ACGAGT-5′ (N‚N′ ) T‚A) with
a 19-fold higher specificity (Table 1). Comparison of the binding
affinities of polyamides with opposing paired aliphatic mono-
mers, such as â/γSO and γSO/â in polyamides 5 and 7,
respectively, to the above sequence with A‚T by polyamides
5-12, showed that the R-hydroxy-/R-amino-substituted γ
monomer is not able to simply recognize T or A by hydrogen
bonding between the OH/NH2 and O2 of T, but quite probably
simultaneously recognizes two bases, that is, 5′-TG-3′, through
strong hydrogen bonding between the OH/NH2 and N3 of the
3′-G base adjacent to the T (Table 1). Furthermore, the
R-position of the hydroxy substituent in the γ-monomer leads
us to assume that an additional hydrogen bond occurs between
the N3 of the 3′-terminal guanine adjacent to T and the hydroxy
group, which may be responsible for the binding specificity of
polyamide 5 to 5′-TGCNCA-3′/3′-ACGN′GT-5′ (N‚N′ ) T‚A
and A‚T) (Figure 3). In addition, van der Waals contacts exist
between the protons at the R-position of the γ residue and H1′
and/or H4′ of the deoxyribose of guanine.22,23 These combined
factors imply that binding affinity and specificity might be
directly related to adjacent bases.
To characterize the binding differences quantitatively among
these polyamides, we obtained the thermodynamic parameter,
standard free energy change (∆G°) upon the formation of the
complexes, as derived from the SPR results (Supporting
Information Table 1). The binding of 5 to the sequence with
A‚T base pair is ∼2 kcal/mol more favorable than binding to
the sequence with a T‚A base pair. The improved binding
stability to sequence 5′-TGCACA-3′/3′-ACGTGT-5′ is cor-
related with much slower dissociation kinetics relative to the
sequence 5′-TGCTCA-3′/3′-ACGAGT-5′ (B and C of Figure
2; data for dissociation rate constants not shown). The slow
dissociation kinetics for the complex of polyamide 5-5′-
TGCACA-3′/3′-ACGTGT-5′ could be governed mainly by the
breakage of the relatively strong additional H-bond between the
OH and N3 of the G and maybe a possible factor for the release/
reuptake of water molecules in the DNA minor groove. In
contrast, only a 4.8-fold increase in specificity was observed in
the recognition of the polyamide 6 with an γRO monomer to
5′-TGCACA-3′/3′-ACGTGT-5′ and to 5′-TGCTCA-3′/3′-AC-
GAGT-5′ (Table 1 and Supporting Information Figure 1, A and
B) and replacing γSO of 5 by γRO at the same position to give
6 increases destabilization of binding to 5′-TGCACA-3′/3′-
ACGTGT-5′ and 5′-TGCTCA-3′/3′-ACGAGT-5′, particularly
with γRO opposite T, with a 4.3-fold reduction in affinity. One
(22) (a) Lamamie de Clairac, R. P.; Geierstanger, B. H.; Mrksich, M.; Dervan,
P. B.; Wemmer, D. E. J. Am. Chem. Soc. 1997, 119, 7909-7916. (b)
Lamamie de Clairac, R. P.; Seel, C. J.; Geierstanger, B. H.; Mrksich, M.;
Baird, E. E.; Dervan, P. B.; Wemmer, D. E. J. Am. Chem. Soc. 1999, 121,
2956-2964.
(23) 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.
9
J. AM. CHEM. SOC. VOL. 128, NO. 27, 2006 8771

A R T I C L E S
Zhang et al.
Figure 3. Molecular modeling study of the recognition by the polyamide with â/γSO of the sequence 5′-GCATGCA/TCATCG-3′/5′-CGATGA/TGCATGC-
3′. (a) Molecular model of the 1:1 complex of polyamide 5 with the sequence 5′-GCATGCACATCG-3′/5′-CGATGTGCATGC-3′ using a computer-assisted
molecular simulation was obtained by energy minimization with the use of semiquantitative distance restraints derived from NOESY. The polyamide is
shown as a stick model, with the A base in blue, T base in red, G base in yellow, but N3 in blue; in the polyamide, oxygen is represented in red, hydrogen
is in gray, nitrogen is in blue, and covalent bonds are green. (b) Partial intermolecular contacts between the hairpin polyamide 5 with a â/γSO pairing and
the sequence with an A‚T base pair. (c) Schematic for (b) with hydrogen bond between OH and N3 of G shown as red dashed lines. (d) Partial intermolecular
contacts between the hairpin polyamide 5 with a â/γSO pairing and the sequence with a T‚A base pair. (e) Schematic for (d) with a hydrogen bond between
OH and N3 of G shown as red dashed line.
explanation for this result is that the (R)-R-hydroxy substituent
points toward the wall of the minor groove of B-DNA, causing
a steric clash with the minor groove, leading to increased
destabilization, whereas the (S)-R-hydroxy is in a favorable
orientation, increasing binding affinity. Conversely, polyamides
7 and 8 with a γSO/â and a γRO/â pairing, respectively, almost
nonspecifically bind to 5′-TGCACA-3′/3′-ACGTGT-5′ and to
5′-TGCTCA-3′/3′-ACGAGT-5′. However, their binding affini-
ties all increase relative to the corresponding polyamides 5, with
a â/γSO pairing, and 6, with a â/γRO pairing. This implies
that van der Waals interactions between the hydroxyl-modified
γ-monomer and flanking bases were also playing an important
role in the discrimination events.
In addition, destabilization of polyamides 5-12 with a
substituted γ residue increases for recognition of 5′-TGCACA-
3′/3′-ACGTGT-5′ and 5′-TGCTCA-3′/3′-ACGAGT-5′ relative
to binding of the polyamide 1 with a â/â pairing to the above
sequences, but compared with their parent polyamides desta-
bilization of some polyamides enhances while that of others
decreases destabilization (Table 1). This result illustrates that
in some cases specificity in the recognition of sequence-specific
DNAs by polyamides may indeed come at the cost of binding
affinity.7c,13
Effects of the Amino Substituent on Binding. Polyamide
9, with a â/γSN pairing, prefers to bind the sequence 5′-
TGCACA-3′/3′-ACGTGT-5′ over 5′-TGCTCA-3′/3′-ACGAGT-
5′ by a factor of 9.6, while 10, with a â/γRN pairing, displays
a 2.3-fold preference for the sequence 5′-TGCACA-3′/3′-
ACGTGT-5′ (Table 1). This difference in specificity may result
from the reasons indicated above, that is, the different binding
orientation of the S- and R-amino substituents could explain
the differences in recognition. In addition, the SPR results shown
in D and E of Figure 2 indicate that the faster dissociation
kinetics for the complex of 9 and the sequence with T‚A could
be responsible for the difference in binding affinity between
sequences with A‚T and T‚A (data for dissociation rate constants
not shown). Likewise, ∆G values may also account for the
binding preference of polyamide 9 to A‚T over T‚A (Supporting
Information Table 1).
Similarly, almost equal binding affinities for 11, with an γSN/
â, and 12, with an γRN/â, to A‚T and T‚A were found, but
their binding stability increases relative to the corresponding 9,
with an γSN/â, and 10, with an γRN/â. Furthermore, it was
significant that polyamide 5, with a â/γSO pairing, has a higher
binding affinity to both 5′-TGCACA-3′/3′-ACGTGT-5′ and to
5′-TGCTCA-3′/3′-ACGAGT-5′ than 9, with a â/γSN, in
particular, with the γSO monomer on the T side giving an
approximately 4.6-fold increase. This difference may be caused
by at least two factors. One is that (S)-R-hydroxy has a stronger
ability to form hydrogen bonds with the N3 of guanine than
the protonated (S)-R-amino; the other is that the bulkier
As expected, all results were a better fit to a model with 1:1
polyamide to DNA stoichiometry, except for the recognition
of G‚C and C‚G. No 2:1 binding motif of polyamides to hairpin
DNAs was found despite high concentrations being used (Figure
2, Supporting Information Figures 1 and 2).
9
8772 J. AM. CHEM. SOC. VOL. 128, NO. 27, 2006

Discrimination of Hairpin Polyamides
A R T I C L E S
protonated amino has a steric clash with the minor groove, such
as the NH of guanine, or induces serious distortion of the minor
groove of the DNA. In short, these observations suggest that,
although hydrogen-bonding ability is a key factor, other factors
such as the surface complementarity of the DNA minor groove
and the hairpin polyamide in the complex and van der Waals
interactions may also play a role in DNA minor groove
recognition by polyamides.
examination of the recognition properties of longer polyamides
containing the γSO monomer(s) as a “spring” to target longer
specific DNA sequences are in progress.
Experimental Section
General. The following abbreviations are used: Fmoc-ONSu:
9-fluorenyl succinimidyl carbonate, HOBt: 1-hydroxybenzotriazole,
DCC: dicyclohexylcarbodiimide, DCM: dichloromethane, DMF: N,N-
dimethylformamide, DIEA: N,N-diisopropylethylamine, TFA: trifluo-
roacetic acid, HATU: N-{(dimethylamino)-1H-1,2,3-triazolo[4,5-b]-
pyridino-1-ylmethylene}-N-methylmethanaminium hexafluorophosphate,
HOAc: acetic acid, Boc: tert-butoxycarbonyl, Fmoc: fluorenyl-
methoxycarbonyl, DMSO: dimethyl sulfoxide, â or â-head: â-alanine,
γ or γ-turn: γ-aminobutyric acid, Hp: N-methyl-3-hydroxypyrrole,
Py: N-methylpyrrole, Im: N-methylimidazole, Dp: 3-(dimethylamino)-
propylamine, bp: base pairs, Hz: hydroxybenzimidazole, Ac: acetyl,
γSN: (S)-R, γ-diaminobutyric acid, γRN: (R)-R, γ-diaminobutyric acid,
γSO: (S)-R-hydroxyl-γ-aminobutyric acid, γRO: (R)-R-hydroxyl-γ-
aminobutyric acid, TMSOTf: trimethylsilyl trifluoromethanesulfonate,
TIS: triisopropylsilane, and DMS: dimethyl sulfide.
Molecular Modeling. To understand more clearly the mech-
anism for the strong and preferential binding of 5 to A‚T over
T‚A, a plausible molecular model of the 5-DNA complexes
was estimated by energy minimization (Figure 3).22a,24 As shown
in Figure 3a, the hairpin polyamide 5 is significantly better
docked in an antiparallel manner into the sequence 5′-GCAT-
GCACATCG-3′/5′-CGATGTGCATGC-3′, with A‚T in the
DNA minor groove, to make close contact with the bases
recognized. As expected, b and d of Figure 3 both show that
the R-hydroxy of the γ-monomer is indeed close to the N3 of
the 3′-G and the proximity and the match between the γ residue
and the G base in this conformation may induce the formation
of a hydrogen bond between the OH and N3 of the G with
bond lengths of 2.53 and 4.11 Å in the recognition by 5 of the
sequences with A‚T and T‚A, respectively (Figure 3, c and e).
However, as can be seen in Figure 3d, the bulkier steric
structure of the A base pushes the γ-monomer away from the
floor of the minor groove to prevent the hydroxyl from forming
a strong H-bonding interaction of the type seen in the 5-A‚T
complex (Figure 3b) and/or influences more favorable aromatic
overlap between polyamide monomers and the DNA bases in
the 5-T‚A complex. These combined factors result in weaker
binding affinity of 5 to the sequence with T‚A. In summary,
molecular modeling studies are in agreement with the SPR
experimental results and provide a molecular rationale for the
thermodynamic observations.
In conclusion, 12 new polyamides with different aliphatic
amino acid pairings, â/â, â/γ, γ/â, γ/γ, â/γSO, â/γRO, γSO/
â, γRO/â, â/γSN, â/γRN, γSN/â, γRN/â, have been designed
and synthesized and their binding kinetic properties to hairpin
DNA examined by SPR. Importantly, we found that the steric
configuration of the R-substituent of the γ-aminobutyric acid
residue in the polyamides could lead to alteration of their binding
properties to predetermined DNA sequences relative to unmodi-
fied â-alanine or γ-aminobutyric acid. More importantly, the
polyamide with the â/γSO pairing preferably bound to 5′-
TGCACA-3′/3′-ACGTGT-5′ over 5′-TGCTCA-3′/3′-ACGAGT-
5′ with a 19-fold higher specificity, while the polyamide with
the â/γSN has a moderate selectivity of 9.6-fold in the
discrimination of these sequences. We found that the recognition
by hairpin polyamides with an R-substituted-γ-aminobutyric acid
in the DNA minor groove was significantly affected by the
environment, here mainly the flanking bases, as well as the
stereochemistry, steric bulk, electronic and hydrogen-binding
ability of the substituent at the R-position of γ-aminobutyric
acid. The combined results from the SPR and molecular
modeling indicated that (S)-R-hydroxyl-γ-aminobutyric acid
(γSO) may not only serve as a flexible subunit in the
construction of polyamides but also as a 5′-TG-3′ recognition
unit in the hairpin polyamides examined. The synthesis and
1H NMR spectra were recorded on a JEOL JNM-FX400 Supercon-
ducting NMR spectrometer 400 MHz, with chemical shifts reported in
parts per million relative to residual solvent and coupling constants in
Hertz. The following abbreviations were applied to spin multiplicity:
s (singlet), d (doublet), t (triplet), q (quartet), qu (quintet), m (multiplet),
and br (broad). UV spectra were measured in HBS-EP buffer on a
Beckman Coulter DU 650 diode array spectrophotometer. Mass spectra
were recorded on the following mass spectrometers using a positive
ionization mode (unless otherwise stated): matrix-assisted laser de-
sorption/ionization time-of-flight (MALDI-TOF) with a BioTOF II from
Bruker Daltonics and electrospray ionization mass spectrometry (ESI-
Mass) with an API 165 from PE SCIEX (APPLIED BIOSYSTEMS).
Biacore assays were performed on a Biacore X system (Sweden), and
processing of data was carried out using BIAevaluation version 4.1.
Glass plates precoated with silica gel 60F254 (Merck) were used for
thin-layer chromatography (TLC) and silica gel 60 (63-200 µm) for
column chromatography. Visualization was realized by UV exposure
and/or with ninhydrin spray. High-performance liquid chromatography
(HPLC) analysis was performed on a Tosoh CCP & 8020 series system
using a 150 × 10 mm ChemcoPak Chemcoband 5-ODS-H reversed
phase column in 0.1% AcOH in water with acetonitrile as eluent at a
flow rate of 3.0 mL/min, and a linear gradient elution of 0 to 100%
acetonitrile over 30 min with detection at 254 nm (unless otherwise
stated). Preparatory reversed phase HPLC was performed on Tosoh
CCP & 8020 series system using a 150 mm × 20 mm YMC-Pack R
& D ODS D-ODS-5-ST S-5 µm reversed phase column in 0.1% AcOH
in water with acetonitrile as the eluent at a flow rate of 6.0 mL/min,
appropriate gradient elution conditions, and detection at 254 or 350
nm. Collected fractions were analyzed by ESI-Mass (PE SCIEX) or
TOF-Mass (Bruker). All HPLC retention times quoted refer to the
analytical column. A Millipore MilliQ water purification system was
used to produce 18 MΩ water, and all buffers for Biacore assays were
filtered at 0.22 µm and degassed. Hairpin DNAs were obtained from
Proligo and used without further purification. HBS-EP buffer (0.01 M
HEPES, pH 7.4, containing 0.15 M NaCl, 3 mM EDTA, and 0.005%
surfactant P20) and sensor chip SA were purchased from Biacore AB
(Sweden). Clear resin (0.52 mequiv/g), (S)-R-tert-butyloxycarbonyl-
amino-γ-(9-fluorenylmethoxy)carbonylaminobutyric acid and (R)-R-
tert-butyloxycarbonylamino-γ-(9-fluorenylmethoxy)carbonylamino-
butyric acid were purchased from Peptides International. Fmoc-
ImCOOH, Fmoc-PyCOOH, DMF, DMSO, piperidine, and pyridine
were purchased from Wako, and Fmoc-â-Ala-OH, Fmoc-γ-Abu-COOH,
and HOBt were from Novabiochem. DIEA, acetic anhydride, and N,N-
dimethyl-1,3-propanediamine were from Nacalai Tesque, Inc. TFA was
from Kanto Chemical Co., Inc. HATU was purchased from Peptide
Institute, Inc. (S)-R-Hydroxyl-γ-aminobutyric acid (γSO) was from
(24) Bando, T.; Narita, A.; Sasaki, S.; Sugiyama, H. J. Am. Chem. Soc. 2005,
127, 13890-13895.
9
J. AM. CHEM. SOC. VOL. 128, NO. 27, 2006 8773

A R T I C L E S
Zhang et al.
Aldrich and (R)-R-hydroxyl-γ-aminobutyric acid (γRO) from Xiamen
Forever Green Source Bio-Chem Technology Co., Ltd. DCC was
fromTIC of Japan. TMSOTf was from Acros Organics. All other
reagents and materials were from standard suppliers (highest quality
available).
3JH,H ) 7.2 Hz, 2H, Fluoren. H), 7.67 (d, 3JH,H ) 7.6 Hz, 2H, Fluoren.
H), 7.40 (m, 2H, Fluoren. H), 7.31 (m, 2H, Fluoren. H), 7.24 (t, JH,H
3
) 4.8 Hz, 1H, NH), 4.28 (m, 2H, OCH2), 4.20 (m, 1H, Fluoren. H),
3
3
3.88 (dd, JH,H ) 8.8 Hz, JH,H ) 3.6 Hz, 1H, CH2CH), 3.04 (m, 2H,
NHCH2), 1.67 and 1.58 (m, 2H, CH2CH), 1.39 (s, 9H, COOCMe3),
1.09 (s, 9H, OCMe3). ESI-Mass: m/z calculated for C27H35NO5: 453.25;
found: 454.26 [M + H]+.
Monomer Synthesis. (2S)-4-[(9-Fluorenylmethoxy)carbonylamino]-
2-hydroxybutyric Acid (14a) and (2R)-4-[(9-Fluorenylmethoxy)-
carbonylamino]-2-hydroxybutyric Acid (14b). A solution of Fmoc-
ONSu (6.72 g, 24 mmol) in 90 mL of 1,2-dimethyoxyethane was added
dropwise to a solution of (S)-(-)-4-amino-2-hydroxybutyric acid (13a)
(2.38 g, 20 mmol) and sodium hydrogen carbonate (2.0 g) in 30 mL of
water with stirring at room temperature over 45 min, and the resultant
solution was stirred for a further 15 h. The reaction was monitored by
TLC with a mixture of DCM/methanol/acetic acid (4/1/0.2, v/v) as the
developing solvent. The solution was directly acidified with concen-
trated hydrochloric acid and concentrated on a rotary evaporator. After
the addition of 20 mL of ice water to the residue, the suspension was
filtered, and then washed with a small amount of cold water to obtain
a white solid. The filtrate was continuously concentrated, cold water
was added, and the precipitate again was collected by filtration. The
combined products were dried to remove traces of water to afford a
white solid at a yield of 98.2%. The product was analyzed and shown
to be 98% pure by HPLC and confirmed by ESI-Mass to give 14a.
Data for 14a: Rf 0.70 (DCM/methanol/acetic acid 4/1/0.2, v/v).
Analytical HPLC: tR ) 19.7 min. 1H NMR (400 MHz, DMSO-d6): δ
7.87 (d, 3JH,H ) 7.6 Hz, 2H, Fluoren. H), 7.67 (d, 3JH,H ) 7.2 Hz, 2H,
Fluoren. H), 7.40 (m, 2H, Fluoren. H), 7.32 (m, 2H, Fluoren. H), 4.25
15b was synthesized according to the above procedure with a yield
of 79.6% at 98% minimum purity by HPLC. Data for 15b: Rf 0.53
(ethyl acetate/hexane, 1/2, v/v). Analytical HPLC: tR ) 29.70 min. 1H
NMR (400 MHz, DMSO-d6): δ 7.88 (d, 3JH,H ) 7.6 Hz, 2H, Fluoren.
3
3
H), 7.66 (d, JH,H ) 7.6 Hz, 2H, Fluoren. H), 7.40 (t, JH,H ) 7.6 Hz,
3
2H, Fluoren. H), 7.31 (m, JH,H ) 7.6 Hz, 2H, Fluoren. H), 7.23 (t,
3JH,H ) 4.8 Hz,1H, NH), 4.28 (m, 2H, OCH2), 4.19 (t, 3JH,H ) 6.8 Hz,
1H, Fluoren. H), 3.89 (dd, 3JH,H ) 8.8 Hz, 3JH,H ) 4.0 Hz, 1H, CH2CH),
3.03 (m, 2H, NHCH2), 1.66 & 1.57 (m, 2H, CH2CH), 1.39 (s, 9H,
COOCMe3), 1.08 (s, 9H, OCMe3). ESI-Mass: m/z calculated for C27H35-
NO5: 453.25; found: 454.27 [M + H]+.
(2S)-2-tert-Butyloxy-4-(9-fluorenylmethoxy)carbonylaminobu-
tyric Acid (17a) and (2R)-2-tert-Butyloxy-4-(9-fluorennylmethoxy)-
carbonylaminobutyric Acid (17b). Trimethylsilyl trifluoromethane-
sulfonate (TMSOTf, 215.6 mg, 0.97 mmol) was added dropwise with
stirring to a solution of 15a (295.5 mg, 0.65 mmol) in 6.5 mL of dioxane
in the presence of 150 µL of DIEA over 25 min at room temperature,
and the resultant solution was continuously stirred for a further 50 min.
The reaction was monitored by TLC with a mixture of ethyl acetate/
hexane/acetic acid (2/1/0.01, v/v). After completion of the reaction,
the reaction mix was cooled to room temperature; 5 mL of water and
8 mL of ethyl acetate were added promptly to the reaction mixture in
an ice-water bath, and the solution was transferred to a separatory
funnel. The organic layer was separated and washed with 5 mL of 5%
aqueous sodium hydrogen carbonate. Because intermediate 16a is very
unstable and hydrolyzes easily to the target compound 17a during the
process of workup, the resulting mixture without further separation and
purification was directly stirred for another 2 h at pH 5, adjusted with
10% aqueous hydrochloric acid. The reaction solution was then washed
with 5 mL of 5% aqueous potassium hydrogen sulfate and 10% aqueous
potassium sulfate, and 5 mL of brine, was dried over sodium sulfate,
filtered and dried after the removal of solvents to give 239.3 mg of a
white solid with a yield of 92.6% and 98% minimum purity by HPLC.
The final product was confirmed by ESI-Mass to be the desired 17a.
Data for 16a: Rf 0.46 (acetate/hexane/acetic acid, 2/1/0.01, v/v).
Analytical HPLC: tR ) 25.83 min. ESI-Mass: m/z calculated for
C26H35NO5Si: 469.23; found: 470.21 [M + H]+, 492.23 [M + Na]+.
Data for 17a: Rf 0.28 (acetate/hexane/acetic acid, 2/1/0.01, v/v).
3
(m, 2H, OCH2), 4.22 (m, 1H, Fluoren. H), 3.97 (dd, JH,H ) 8.8 Hz,
3JH,H ) 4.0 Hz, 1H, CH2CH), 3.10 (m, 2H, NHCH2), 1.78 and 1.59
(m, 2H, CH2CH). ESI-Mass: calculated m/z for C19H19NO5: 341.13;
found: 342.12 [M + H]+.
14b was synthesized according to the above procedure with a yield
of 95.8% and 98% minimum purity by HPLC. Data for 14b: Rf 0.71
(DCM/methanol/acetic acid 4/1/0.2, v/v). Analytical HPLC: tR 19.75
3
min. 1H NMR (400 MHz, DMSO-d6): δ 7.87 (d, JH,H ) 7.6 Hz, 2H,
3
Fluoren. H), 7.67 (d, JH,H ) 7.2 Hz, 2H, Fluoren. H), 7.40 (m, 2H,
Fluoren. H), 7.32 (m, 2H, Fluoren. H), 4.26 (m, 2H, OCH2), 4.20 (m,
1H, Fluoren. H), 3.96 (dd, 3JH,H ) 8.7 Hz, 3JH,H ) 4.0 Hz, 1H, CH2CH),
3.08 (m, 2H, NHCH2), 1.79 and 1.59 (m, 2H, CH2CH). ESI-Mass:
calculated m/z for C19H19NO5: 341.13; found: 342.12 [M + H]+,
364.12 [M + Na]+, 380.10 [M + K]+.
(2S)-2-tert-Butyloxy-4-(9-fluorenylmethoxy)carbonylaminobu-
tyrate tert-Butyl Ester (15a) and (2R)-2-tert-Butyloxy-4-(9-fluore-
nylmethoxy)carbonylaminobutyrate tert-Butyl Ester (15b). Isobu-
tylene (20 mL) was introduced to a solution of 14a (1 g, 2.93 mmol)
in 20 mL of dry DCM in the presence of 40 µL of concentrated sulfuric
acid (98%) in a pressure flask with stirring at low temperature. The
mixture was then very slowly warmed to room temperature and
continuously stirred for another 40 h. The reaction was monitored by
TLC with a mixture of ethyl acetate/hexane (1/2, v/v). After completion
of the reaction, the solution was exposed to air at room temperature to
remove excess isobutylene. Then, the solution was transferred to a
separatory funnel, washed with 3 mL of water, and adjusted to pH 7.0
with 2 M aqueous sodium hydroxide. The dichloromethane layer was
separated and the organic solvent carefully removed by rotary evapora-
tion. Ethyl acetate (5 mL) and 3 mL of water were added to the residue,
and the organic layer was washed with 3 mL of brine, dried with
anhydrous sodium sulfate for 5 h, filtered, and dried after the removal
of solvents to obtain 1.41 g of crude product as a light-yellow oil. The
product was subjected to flash column chromatography on 63-200
µm silica gel using gradient elution of a mixture of ethyl acetate/hexane
from 1/3 to 1/2 as eluents to afford finally 1.01 g of light-yellow oil
with a yield of 76.5% at a 98% minimum purity by HPLC. The final
product was confirmed by ESI-Mass to be the desired compound 15a.
Data for 15a: Rf 0.52 (ethyl acetate/hexane, 1/2, v/v). Analytical
1
Analytical HPLC: tR ) 24.01 min. H NMR (400 MHz, DMSO-d6):
δ 12.3 (br, 1H, COOH), 7.87 (d, 3JH,H ) 7.6 Hz, 2H, Fluoren. H), 7.67
(d, 3JH,H ) 7.2 Hz, 2H, Fluoren. H), 7.40 (t,3JH,H ) 7.6 Hz, 2H, Fluoren.
3
3
H), 7.30 (t, JH,H ) 7.6 Hz, 2H, Fluoren. H), 7.25 (t, JH,H ) 4.8 Hz,-
1H, NH), 4.27 (d, 3JH,H ) 6.4 Hz, 2H, OCH2), 4.19 (t, 3JH,H ) 6.8 Hz,
1H, Fluoren. H), 3.94 (dd, 3JH,H ) 8.4 Hz, 3JH,H ) 3.6 Hz, 1H, CH2CH),
3.04 (m, 2H, NHCH2), 1.71 & 1.60 (m, 2H, CH2CH), 1.13 (s, 9H,
OCMe3). ESI-Mass: m/z calculated for C23H27NO5: 397.19; found:
398.26 [M + H]+, 420.17 [M + Na]+; negative ionization, 396.18 [M
- H]-.
17b was synthesized according to the above procedure at a yield of
88% and 98% minimum purity by HPLC. Data for 16b: Rf 0.48
(acetate/hexane/acetic acid, 2/1/0.01, v/v). Analytical HPLC: tR ) 25.9
min. ESI-Mass: m/z calculated for C26H35NO5Si: 469.23; found:
470.22 [M + H]+, 492.25 [M + Na]+. Data for 17b: Rf 0.28 (acetate/
hexane/acetic acid, 2/1/0.01, v/v). Analytical HPLC: tR ) 24.0 min.
1H NMR (400 MHz, DMSO-d6): δ 12.3 (br, 1H, COOH), 7.88 (d,
3JH,H ) 7.6 Hz, 2H, Fluoren. H), 7.67 (d, 3JH,H ) 7.2 Hz, 2H, Fluoren.
3
H), 7.40 (t,3JH,H ) 7.4 Hz, 2H, Fluoren. H), 7.31 (t, JH,H ) 7.6 Hz,
2H, Fluoren. H), 7.26 (t, 3JH,H ) 4.6 Hz,1H, NH), 4.28 (m, 2H, OCH2),
1
3
3
HPLC: tR ) 29.70 min. H NMR (400 MHz, DMSO-d6): δ 7.88 (d,
4.19 (t, JH,H ) 6.0 Hz, 1H, Fluoren. H), 3.95 (dd, JH,H ) 8.2 Hz,
9
8774 J. AM. CHEM. SOC. VOL. 128, NO. 27, 2006

Discrimination of Hairpin Polyamides
A R T I C L E S
3JH,H ) 3.6 Hz, 1H, CH2CH), 3.05 (m, 2H, NHCH2), 1.71 & 1.60 (m,
2H, CH2CH), 1.11 (s, 9H, OCMe3). ESI-Mass: m/z calculated for
C23H27NO5: 397.19; found: 398.23 [M + H]+, 420.16 [M + Na]+;
negative ionization, 396.17 [M - H]-.
Manual Synthesis Protocol. To obtain polyamides 5-12 with
amino- or hydroxyl-modified monomers, the following additional step
was involved. After polyamides were cleaved from resins 5a-12a in
the presence of (dimethylamino)propylamine, purified by HPLC and
dried, Boc- or tert-butyl-protected polyamides were obtained and
confirmed by 1H NMR and ESI-Mass or TOF-Mass. Then, the products
in the presence of a mixture of TFA/TIS/TMS/H2O (91/3/3/3, v/v) (0.5
µL) were stirred at room temperature for 30 min under nitrogen, and
solvents were removed by coevaporation with DCM in vacuo.
Hydrochloric acid (0.5 mL, 5 M) was added to the residue, and solvent
was removed in vacuo with a cold trap containing potassium hydroxide.
The residue was subjected to analysis, then separation by HPLC,
confirmation by 1H NMR and ESI-Mass or TOF-Mass, and finally
lyophilization to yield the final desired polyamides 5-12 with amino-
or hydroxyl-modified monomers as white or light-yellow powders.
Determination of the Enantiomeric Purity of 17a and 17b. To
17a or 17b (7.95 mg, 0.02 mmol) in DMF (0.45 mL) were added HATU
(22.8 mg, 0.06 mmol) and DIEA (4.5 µL). The resultant mixture was
stirred at room temperature for approximately 15 min, and then (S)-
(-)-1-(1-naphthyl)ethylamine (7.78 mg, 0.0454 mmol) was added. The
mixture was continuously stirred at room temperature for 2 h. The
reaction was monitored by TLC with a mixture of ethyl acetate/hexane
(1/1, v/v). After the completion of the reaction, the solution was diluted
with 1 mL of ethyl acetate, washed with 10% aqueous citric acid (2 ×
1 mL) and brine (1 × 1 mL), respectively. The resulting organic layer
was dried (anhydrous Na2SO4), filtered, concentrated, and chromato-
graphed on a column packed with 63-200 µm silica gel using a mixture
of ethyl acetate/hexane (2/1, v/v) to obtain 10.9 mg of the amide from
17a (S) at a yield of 99% and 10.8 mg of the amide from 17b (R) at
a yield of 99% as white solids. The final products were analyzed by
1H NMR (400 MHz, DMSO-d6) and confirmed to be targets by ESI-
Mass. The amide from 17a (S) showed a triplet (J ) 5.4) at 7.17 ppm
and a doublet (J ) 8.0) at 6.83 ppm, and the amide from 17b (R)
showed a triplet (J ) 5.6) at 7.23 ppm and a doublet (J ) 7.6) at 6.96
ppm. In both spectra, almost no evidence of diastereomeric contamina-
tion (<3%) was observed.
Polyamide Synthesis by Fmoc Chemistry. Machine-Assisted
Solid-Phase Protocols. All machine-assisted polyamide syntheses were
performed on a Pioneer Peptide Synthesizer (Applied Biosystems) with
a computer-assisted operation system at a 0.10 mmol scale (200 mg of
clear resin, 0.52 mequiv/g) by using Fmoc chemistry. Reagent positions
were as follows: wash position, DMF; deblock position, 20% piperidine
in DMF; activator position 1, 0.5 M HATU in DMF; activator position
2, 1.0 M DIEA in DMF; auxiliary reagent position 1, methanol; and
auxiliary reagent position 3, 5% acetic anhydride and 5% pyridine in
DMF. The following conditions were used in all polyamide solid-phase
syntheses for each cycle: deblocking for 5 min with 20% piperidine/
DMF, activating for 2 min with 0.5 M HATU/DMF and 1.0 M DIEA/
DMF, coupling for 60 min, capping for 5 min with 5% acetic anhydride/
5% pyridine/DMF, and final washing with methanol. A double-couple
cycle was employed when coupling any monomer to imidazole and
coupling of imidazole itself; all other couplings were carried out with
single-couple cycles.
Surface Plasmon Resonance (SPR) Assay. All SPR experiments
were performed on a Biacore X instrument at 25 °C. The biotinylated
hairpin DNAs: 5′-biotin-GGCCGATGN′GCATGCTATAGCATGC-
NCATCGGCC-3′ (N‚N′: T‚A, A‚T, G‚C, and C‚G) were obtained from
Proligo. Streptavidin-functionalized SA sensor chips were purchased
from Biacore. Immobilization of these hairpin DNAs onto the sensor-
chips, measurements of binding curves of polyamides to the hairpin
DNAs, and data processing were performed according to following
procedure.25
Immobilization of Biotinylated DNAs. After docking the strepta-
vidin-coated sensor chip SA into the instrument, priming the system
three times with HBS-EP buffer, normalizing the system with BIA-
CORE normalizing solution (40% (w/w) glycerol in H2O), and allowing
the buffer to run for about 10 min at a rate of 10 µL/min until the
baseline was stable, the sensor chip SA was preconditioned to remove
any unbound streptavidin from the surface of the sensor chip with a
solution of 1 M NaCl/50 mM NaOH at a rate of 10 µL/min with
injection of 10 µL three to five times, and then allowing the buffer to
run for 10 min to remove any trace of NaOH until the baseline was
stable. Simultaneously, 20 µL of buffer was injected three times to
wash the loop at the same flow rate. Before immobilizing DNA on the
sensor chip, 100 µL of 5′-biotinylated DNA solution in running buffer
without DMSO was renatured by heating to 85 °C for 2 min and slowly
cooled to room temperature. DNA was then immobilized onto the
chosen flow cell by injecting appropriate amounts of DNA solution at
a flow rate of 2 µL/min to obtain the wanted immobilization level.
After the DNA injection was complete, the buffer was kept running
through the cell until the response units on the sensorgram became
stable. An unmodified flow cell was used as a control for matrix effects.
Typically, Fmoc-â-Ala-clear-acid resin (200 mg, 0.10 mmol) was
swollen in 3 mL of DMF in a 15-mL plastic centrifuge tube for 30
min and loaded onto the reaction column. Glass scintillation tubes with
loading monomers in DMF were placed in the racks, in order, from
the front position to the back position. The synthesis was then started,
controlled by the computer using the established program. All lines
were emptied with argon before and after solution transfers.
General Procedures for SPR Binding Studies. Different concen-
trations of polyamide solutions in HBS-EP buffer with 0.1% DMSO
were prepared by dilutions from 10 mM of stock solution in DMSO,
and concentrations were checked by UV spectroscopy at 294 nm. Unless
otherwise noted, all binding experiments were carried out using HBS-
EP buffer with 0.1% DMSO as running buffer, and the running buffer
or 10 mM glycine-HCl (pH ) 1.5) was used as regeneration solution.
Typically, the buffer was injected at a flow rate of 5 µL/min as a blank
control after a stable baseline was obtained, and then samples were
injected under identical conditions to those of buffer injection at
concentrations ranging from 5 nM to 5 µM. To minimize carryover,
samples were injected in the order of increasing concentration. Kinetic
information was obtained by global fitting of the response units vs time
using a model with mass transport effects with the BIA evaluation 4.1
program.
After the completion of the synthesis on the peptide synthesizer,
the resin was washed with a mixture of methanol/DCM (1/1, v/v) (3
× 1 mL) and dried in a desiccator at room temperature in vacuo. The
resin was then placed in a 20-mL glass scintillation vial, 5 mL of
(dimethylamino)propylamine was added, and the solution stirred at 55
°C overnight. Resin was removed by filtration through a pad of Celite
and washed thoroughly with methanol/DCM (1/1, v/v). The resultant
filtrate was acidified to pH 4-5 with 10% aqueous hydrochloric acid,
and solvents were removed in vacuo. The residue was dissolved in
approximately 0.5 mL of DMF, and insoluble substances were removed
carefully using a 0.45 µm disposable syringe filter unit. The resulting
polyamide solution was analyzed and purified by analytical HPLC at
254 nm and preparatory HPLC at 254 or 350 nm, respectively. After
confirmation by 1H NMR and ESI-Mass or TOF-Mass, appropriate
fractions were lyophilized to give the final desired polyamides 1, 2, 3,
and 4 as white or light-yellow powders.
Molecular Modeling Studies. Minimizations were performed with
the Discover (MSI, San Diego, CA) program using CVFF force-field
(25) (a) Supekova, L.; Pezacki, J. P.; Su, A. I.; Loweth, C. J.; Riedl, R.;
Geierstanger, B.; Schultz, P. G.; Wemmer, D. E. Chem. Biol. 2002, 9, 821-
827. (b) Wong, C. H.; Hendrix, M.; Manning, D. D.; Rosenbohm, C.;
Greenberg, W. A. J. Am. Chem. Soc. 1998, 120, 8319-8327. (c) Rich, R.
L.; Myszka, D. G. Curr. Opin. Biotechnol. 2000, 11, 54-61.
9
J. AM. CHEM. SOC. VOL. 128, NO. 27, 2006 8775

A R T I C L E S
Zhang et al.
parameters. The starting structure was built based on the NMR structure
of the ImPyPy-γ-PyPyPy-5′-d(CGCTAACAGGC)-3′/5′-d(GCCTGT-
TAGCG)-3′ complex.22a Aliphatic monomers were inserted into the
polyamide by using standard bond lengths and angles. Twenty-two Na
cations were placed at the bifurcating position of the O-P-O angle at
a distance of 2.21 Å from the phosphorus atom. The resulting complex
was soaked in a 10 Å layer of water. The whole system was minimized
without constraints to the stage where the RMS was less than 0.001
kcal/mol Å.
Acknowledgment. We gratefully acknowledge financial
support from COE Project of Japan. We also thank Mr. H.
Ayame for synthesizing three polyamides.
Supporting Information Available: Data for synthesis and
characterization of polyamides 1-12, sensorgrams of poly-
amides 6 and 10 for recognition of 5′-TGCNCA-3′ (N ) T,
A), and the free energy changes before and after binding of
polyamides 1-12. This material is available free of charge via
the Internet at http://pubs.acs.org.
JA0580587
9
8776 J. AM. CHEM. SOC. VOL. 128, NO. 27, 2006