Sequence-selective detection of double-stranded DNA sequences using pyrrole-imidazole polyamide microarrays

Article abstract of DOI:10.1021/ja309677h

We describe a microarray format that can detect double-stranded DNA sequences with a high degree of sequence selectivity. Cyclooctyne-derivatized pyrrole-imidazole polyamides were immobilized on azide-modified glass substrates using microcontact printing and a strain-promoted azide-alkyne cycloaddition (SPAAC) reaction. These polyamide-immobilized substrates selectively detected a seven-base-pair binding site incorporated within a double-stranded oligodeoxyribonucleotide sequence even in the presence of an excess of a sequence with a single-base-pair mismatch.

Full text of DOI:10.1021/ja309677h

Article
pubs.acs.org/JACS
Sequence-Selective Detection of Double-Stranded DNA Sequences
Using Pyrrole−Imidazole Polyamide Microarrays
Ishwar Singh,† Christian Wendeln,‡ Alasdair W. Clark,§ Jonathan M. Cooper,§ Bart Jan Ravoo,*,‡
and Glenn A. Burley*,†
†Department of Pure & Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow G1 1XL, United Kingdom
‡Organic Chemistry Institute and Centre for Nanotechnology (CeNTech), Westfalische Wilhelms-Universitat Munster, Correnstrasse
̈
̈
̈
40, 48149 Munster, Germany.
̈
§Division of Biomedical Engineering, School of Engineering, University of Glasgow, Oakfield Avenue, Glasgow G12 8LT, United
Kingdom
S
* Supporting Information
ABSTRACT: We describe a microarray format that can
detect double-stranded DNA sequences with a high degree of
sequence selectivity. Cyclooctyne-derivatized pyrrole−imida-
zole polyamides were immobilized on azide-modified glass
substrates using microcontact printing and a strain-promoted
azide−alkyne cycloaddition (SPAAC) reaction. These poly-
amide-immobilized substrates selectively detected a seven-
base-pair binding site incorporated within a double-stranded
oligodeoxyribonucleotide sequence even in the presence of an
excess of a sequence with a single-base-pair mismatch.
a microarray setup that examines the binding affinity and
selectivity of dsDNA-binding molecules where dsDNA is the
analyte present in solution has not been reported.
INTRODUCTION
■
Functionalized microarrays that are capable of recognizing
specific biomolecules hold great promise as tools for nano-
technological as well as medical diagnostic applications.1−4
Compared with traditional low-throughput methods for the
analysis of biomolecular binding events such as cell culture
assays, microarray technology enables the detection and
analysis of the relative abundance and variation of specific
biomolecules in a high-throughput screening format.5,6 DNA
microarrays in particular have been widely exploited for the
analysis of genotypic variations by detecting the binding of a
single-stranded DNA (ssDNA) analyte using complementary
ssDNA probes immobilized on a solid substrate.7,8 The
transduction of this binding event into a quantifiable signal
such as a change in fluorescence emission can be used to detect
as well as quantify the abundance of the corresponding ssDNA
analyte.8,9
Recent innovations in microarray technology have extended
their present capability to include the assessment of recognition
profiles of transcription factors9,10 and small molecules11−13 in
a parallel and high-throughput fashion using double-stranded
DNA (dsDNA) probe sequences immobilized on a solid
substrate.11−15 Detection of a dsDNA binding event is achieved
either by a reporter directly attached to the DNA-binding
molecule12,13 or via a fluorescently tagged antibody specific for
the transcription factor of interest.9 Both of these microarray
formats utilize immobilized dsDNA oligodeoxyribonucleotides
(ODNs) to probe the dsDNA binding characteristics of the
DNA binder present in solution. To the best of our knowledge,
To develop an array setup of this type and assess the scope
and limitations of the corresponding format, three develop-
ments are required: (i) efficient access to a suite of probe
molecules that recognize dsDNA sequences selectively; (ii) the
ability of these probe molecules to distinguish a matched
dsDNA sequence from mismatched sequences when immobi-
lized on a solid substrate; and (iii) the availability of
chemoselective and high-yielding bioconjugation chemistry to
immobilize the probe molecules on a substrate in a microarray
format. Of the various families of molecules available that bind
to dsDNA sequences, we chose pyrrole−imidazole polyamides
(PAs) as a paradigm for this study. PAs bind to dsDNA with
nanomolar binding affinity in a sequence-selective manner by
reading the hydrogen-bonding landscape of the edges of
Watson−Crick base pairs (bp) in the minor groove of B-type
dsDNA (Figure 1a).16 They are highly programmable, enabling
the identification of dsDNA sequences from 5 to 16 bp in
length, thus addressing criteria (i) and (ii) above.12,13,17
In this article, we describe a PA microarray format and
demonstrate its use for direct and selective detection of dsDNA
sequences. The microarray design utilizes microcontact printing
(μCP)18,19 and strain-promoted azide−alkyne cycloaddition
(SPAAC)20,21 to immobilize PA (1) onto a glass substrate
Received: September 30, 2012
Published: February 4, 2013
© 2013 American Chemical Society
3449
dx.doi.org/10.1021/ja309677h | J. Am. Chem. Soc. 2013, 135, 3449−3457

Journal of the American Chemical Society
Article
Figure 1. (a) Structure of PA (1) used in this study. (b) Schematic depiction of the process of fabricating PA-derivatized surfaces using μCP and
SPAAC. A PDMS stamp inked with PA (1) is brought into direct contact with an azide-modified surface to produce a PA-patterned surface. The PA-
derivatized surface selectively detects fluorophore-tagged ODN duplexes consisting of matched binding sequences.
Figure 2. Structure of PA (2).
a
Scheme 1
a
Reagents and conditions: (i) DSC, DIPEA, MeCN, room temperature, 1 h; 76%.
(Figure 1b). Single-base-pair sequence discrimination is
observed by the selective detection of fluorescently tagged
ODN duplexes containing a 7 bp target binding site matched to
the core PA (2) (Figure 2), even in the presence of a
competing sequence with a single-base-pair mismatch.
EXPERIMENTAL SECTION
■
Synthesis of 4-(Cyclooct-2-yn-1-yloxy)butyl (2,5-Dioxopyr-
rolidin-1-yl) Carbonate (4). To a solution of 4-(cyclooct-2-yn-1-
yloxy)butan-1-ol (3)22 (70 mg, 0.357 mmol) in acetonitrile (2 mL)
was added N,N′-disuccinimidyl carbonate (DSC) (110 mg, 0.428
mmol) followed by N,N-diisopropylethylamine (DIPEA) (75 μL,
3450
dx.doi.org/10.1021/ja309677h | J. Am. Chem. Soc. 2013, 135, 3449−3457

Journal of the American Chemical Society
Article
a
Scheme 2
a
Reagents and conditions: (i) TEA, DCM, room temperature, 20 min; (ii) 2, HATU, DIPEA, DMF, 1 h; 44% over two steps.
0.428 mmol). The reaction mixture was stirred at room temperature in
air for 1 h and then concentrated in vacuo. Column chromatography
(SiO2, gradient from 9:1 to 3:1 hexane/EtOAc) afforded (4) (91 mg,
76% yield) as a colorless oil (Scheme 1). 1H NMR (400 MHz,
CDCl3): δ 4.38−4.34 (m, 2H), 4.15−4.12 (m, 1H), 3.61−3.55 (m,
1H), 3.37−3.33 (m, 1H), 2.83 (s, 4H), 2.25−2.06 (m, 3H), 1.92−1.81
(m, 6H), 1.70−1.61 (m, 5H). 13C NMR (100 MHz, CDCl3): δ 168.1,
151.0, 99.3, 92.5, 71.9, 70.3, 67.8, 41.8, 33.7, 29.2, 25.8, 25.1, 25.0,
24.9, 20.1. HRMS (ESI+) m/z: calcd for C17H24O6N [M + H]+,
338.1604; found, 338.1598 (Figure S35 in the Supporting
Information).
Synthesis of Polyamide−Cyclooctyne Conjugate (1). To a
solution of 15-amino-4,7,10,13-tetraoxapentadecanoic acid (5) (3 mg,
0.011 mmol) in 1 mL of dry dichloromethane (DCM) was added a
solution of (4) (3.8 mg, 0.0113 mmol) in dry DCM (1 mL) followed
by triethylamine (TEA) (3 μL, 0.0026 mmol). The reaction mixture
was stirred at room temperature for 20 min and then concentrated in
vacuo. The crude mixture was redissolved in 1 mL of dry N,N-
dimethylformamide (DMF), and 2-(1H-7-azabenzotriazol-1-yl)-
1,1,3,3-tetramethyluronium hexafluorophosphate (HATU) (1.8 mg,
0.0047 mmol) and DIPEA (2.8 μL, 0.015 mmol) were added. After
the reaction mixture was stirred for 5 min at room temperature, PA
(2) (2 mg, 0.0015 mmol) was added, and the resulting mixture was
stirred for 1 h at room temperature. Purification by semipreparative
reversed-phase HPLC (t = 15.1 min) afforded PA (1) as a yellow
powder after freeze-drying (1.2 mg, 44% yield). For information
concerning the buffer system and gradient used in the analysis and
purification, see the Supporting Information. HRMS (ESI+) m/z:
calcd for C83H115N24O18 [M + H]+, 1735.8821; found, 1735.8824
(Figure S37 in the Supporting Information).
Synthesis of n-Hexyl Polyamide Cyclooctyne Click Con-
jugates (7). To a solution of PA (1) in dimethyl sulfoxide (DMSO)
(75 μL, 0.039 μmol, 533 μM, 1 equiv) was added a solution of n-hexyl
azide in DMSO (780 μL, 7.8 μmol, 10 mM, 200 equiv). The reaction
mixture was shaken at room temperature in air for 24 h. Reversed-
phase HPLC revealed the formation of two new peaks (t = 15.7 and 16
min), corresponding to the triazole products (7a) and (7b) according
positive-ion-mode electrospray ionization (ESI+) mass spectrometry
(MS). For information concerning the buffer system and gradient used
in the HPLC analysis and purification, see the Supporting Information.
HRMS (ESI+) m/z: calcd for C89H128N27O18 [M + H]+, 1862.9931;
found, 1863.0051 (Figure S38 in the Supporting Information).
Preparation of Undecyl Azide-Modified Silicon and Glass
Substrates. Undecyl azide-modified substrates were prepared using a
modified protocol of a literature procedure.23 Glass and silicon
substrates were cut into small pieces and cleaned by sonication in
pentane, acetone, and finally Milli-Q water. The substrates were
treated with freshly prepared piranha solution (3:1 H2SO4/H2O2;
Caution!) for 30 min. The surfaces were thoroughly washed with
Milli-Q water to remove all traces of sulfuric acid and dried under a
stream of Ar. The substrates were immersed in a solution of 11-
bromoundecyltrichlorosilane (0.1 vol %) in toluene for 40 min and
then cleaned by extensive washing with DCM, ethanol, and Milli-Q
water. The dry alkyl bromide-functionalized surfaces were converted to
azide-functionalized substrates by exposure to a saturated solution of
NaN3 in DMF for 48 h at 70 °C. Finally, the substrates were washed
with Milli-Q water and ethanol and dried in a stream of Ar.
3451
dx.doi.org/10.1021/ja309677h | J. Am. Chem. Soc. 2013, 135, 3449−3457

Journal of the American Chemical Society
Article
a
Scheme 3
a
Conditions: (i) DMSO, room temperature, 24 h.
Microcontact Printing. Polydimethylsiloxane (PDMS) stamps
(flat, patterned) were oxidized in an UV ozonizer (PSD-UV, Novascan
Technologies Inc.) for 55 min directly before use and stored under
distilled water. The PDMS stamps were dried in a stream of Ar and
subsequently covered with 10−20 μL of a solution of cyclooctyne-
modified PA (1) in 1:1 methanol/ethanol (concentration ranging from
100 nM to 100 μM). After 1 min, the stamps were blow-dried in a
stream of Ar and placed on the undecyl azide-modified substrate
surfaces. A weight of 30 g was applied on the top of the PDMS stamps
at ambient temperature. After 20−60 min, the PDMS stamps were
removed from the surfaces, and the modified substrates were washed
with Milli-Q water and ethanol, sonicated in Milli-Q water (1 min),
and dried in a stream of Ar (see Figure 3).
dsDNA Binding Protocol Using PA-Derivatized Microarrays.
A concentration series of 5′-Cy3-labeled dsDNA (ODN1−4) was
prepared [1 μM/500 nM/10 nM/1 nM in 50 mM phosphate buffer
(pH 7.5) containing 0.01% Tween 20] and applied to PA-modified
glass surfaces. The surfaces were transferred to a moist chamber and
stored at room temperature overnight. The unbound dsDNA was
removed by washing with 50 mM phosphate buffer (pH 7.5)
containing 0.01% Tween 20 for 3 min at room temperature and
then for 3 min at 50 °C, after which the washed glass slides were dried
under N2. DNA binding was assessed using confocal fluorescence
microscopy.
studies showed that DNA microarrays can be prepared easily
and cost-effectively by combining μCP with copper-free click
chemistry to immobilize DNA strands on glass substrates.26 In
the current work, a SPAAC protocol20,27−29 was adopted for
the immobilization of PA (1) ink onto an azide substrate, as it
dispels the need for copper(I) catalysis as required in
conventional copper-catalyzed Huisgen [3 + 2] azide−alkyne
cycloaddition (CuAAC).30,31 We avoided utilizing CuAAC
chemistry because PAs incorporate imidazole units, which have
known affinity for copper ions.32 The core sequence of PA (1)
was chosen as a paradigm for this study by virtue of its high
binding affinity for the target DNA sequence 5′-WWGGWCW-
3′ (where W = A/T).33 A tetraethylene glycol tether was
installed between the cyclooctyne unit and the PA core to
impart water solubility and endow PA (1) with sufficient
flexibility to access optimal binding conformations with its
corresponding dsDNA target binding site.
Synthesis of PA (1). Amine-terminated PA (2) was
prepared in 24% yield by solid-phase synthesis of the
heterocyclic core using our triphosgene-mediated coupling
strategy (Figure 2 and Scheme S1 in the Supporting
Information)34 followed by resin cleavage using 3,3′-diamino-
N-methyldipropylamine (DMPA). The cyclooctyne linker (4)
was prepared using N,N′-disuccinimidyl carbonate (Scheme 1).
Coupling of the NHS ester of (4) with amino acid (5) afforded
linker (6), which was used directly in the next step without any
purification. Finally, compound (6) was coupled with PA (2)
using HATU as the activating agent to obtain PA (1) in 44%
yield (Scheme 2).
Competitive dsDNA Binding Assay on the Microarray.
Competitive binding assays were performed using a series of mixtures
of ODN1 and ODN2 (see Table 3). The binding and washing
protocols were performed as described above.
RESULTS AND DISCUSSION
■
Experimental Design. This work had two principal aims:
(i) to develop an efficient and robust synthetic methodology for
immobilizing PAs onto solid substrates and (ii) to assess the
ability of these PA-derivatized surfaces to bind matched dsDNA
sequences selectively. Our design focused on the utility of
μCP24,25 in immobilizing cyclooctyne-modified PA (1) onto
azide-functionalized glass surfaces (Figure 1).23 Previous
Assessment of the Performance of Strain-Promoted
Click Chemistry with PA (1). In the presence of 200 equiv of
n-hexyl azide, the two triazole regioisomers (7a) and (7b) were
formed from PA (1) in quantitative yield in 24 h (Scheme 3),
according to HPLC analysis and high-resolution mass
3452
dx.doi.org/10.1021/ja309677h | J. Am. Chem. Soc. 2013, 135, 3449−3457

Journal of the American Chemical Society
Article
Figure 3. Light microscopy images of water condensation experiments on PA (1)-modified glass surfaces: (a) 100 μM PA (1), 60 min printing time;
(b) 10 μM, 60 min; (c) 1 μM, 60 min; (d) 500 nM, 60 min; (e) 100 nM, 60 min. (f) No pattern was observed when μCP was conducted using an
octadecyl-modified glass substrate [100 μM PA (1), 10 μm line, and 5 μm space].
a
Table 1. dsDNA Sequences Used in This Study
a
The location of the mismatched sequence is italicized and underlined. The PA target sequence for this study was 5′-WWGGWCW-3′, where W =
A/T.
spectrometry (HRMS) (see Figures S33 and S38 in the
Supporting Information). This confirmed that the cyclooctyne
moiety of PA (1) was available to react with a corresponding
azide and did not form an unfavorable conformation to
sequester its reactivity with the azide surface.
Immobilization of PA (1) on Solid Surfaces by μCP.
Oxidized PDMS stamps with a relief structure were inked with
a solution of PA (1) and brought into contact with undecyl
azide-modified substrates. The SPAAC reaction was induced in
the area of contact to produce stable, covalent attachment of
PA (1) on the glass substrate. Because of the more hydrophilic
nature of PAs relative to the azide surfaces, immobilization led
to a significant decrease in the water contact angle. After μCP
of PA (1) with a flat stamp (100 μM, 20 min), the advancing
contact angle decreased from 82
monolayer to 62 3° for the PA-modified surface. The
receding contact angle decreased even more significantly from
70 3° to 15 5°, providing further evidence of covalent
attachment of PAs to the glass surface. Water condensation
experiments further confirmed the attachment of PA (1), as
visualized by the hydrophilic line patterns shown in Figure 3a−
e. We observed that the surface density of the immobilized PA
residues could be controlled by the concentration of PA (1)
used for the ink solutions. In the case of concentrated solutions
(100 μM), well-defined condensation patterns were obtained,
whereas dilute solutions (100 nM) did not cause any visible
arrangement of droplets. As a negative control, PA (1) was
3° for the undecyl azide
3453
dx.doi.org/10.1021/ja309677h | J. Am. Chem. Soc. 2013, 135, 3449−3457

Journal of the American Chemical Society
Article
Table 2. UV−Vis Melting Temperature Analysis of PA (2) in Complex with ODN1−4
printed (100 μM) on an octadecyl-modified glass substrate. No
pattern was observed (Figure 3f).
Further characterization of PA-derivatized surfaces by atomic
force microscopy (AFM) and X-ray photoelectron spectrosco-
py (XPS) was also undertaken. PA-patterned surfaces (5 μm
lines spaced by 3 μm) were clearly visible by AFM in height
(Figure S1 in the Supporting Information) as well as phase
contrast (Figure S2 in the Supporting Information), with the
observed height difference of 1.5−2 nm between the azide- and
PA-covered surfaces. XPS confirmed the attachment of PA (1)
to the azide-modified surface. In the case of an undecyl azide
monolayer, only saturated C 1s signals (285 eV) and signals for
carbons attached to nitrogen (approximately 287 eV) were
observed (Figure S3 in the Supporting Information). After μCP
of PA (1) using a flat stamp, additional shoulders for carbon
atoms in a C−O environment (286−287 eV) and in an amide
environment (288 eV) were visible, verifying the attachment of
PAs to the substrate (Figure S3). Moreover, the absence of the
two nitrogen signals in the N 1s region of the undecyl azide
monolayer (corresponding to internal and neighboring azide
nitrogens) and the presence of a single, broad nitrogen peak for
all surface nitrogens (triazole, amide, amine) further corrobo-
rated that immobilization of PA (1) on the substrate had been
Figure 4. Fluorescence confocal scanning microscopy image of PA-
derivatized surfaces (10 μm lines spaced by 5 μm) after incubation
with ODN1 (1 μM) and washing (three washing steps in total) at 23
°C.
achieved (Figure S4 in the Supporting Information).
PA-Derivatized Surfaces Are Selective for Their Target
dsDNA Binding Sequence. Table 1 summarizes the dsDNA
sequences used to assess the binding affinity and selectivity of
PA (2) [i.e., the precursor of (1)] in solution. The ODN1
duplex contained a matching 7 bp sequence 5′-ATGGTCA
[the core binding sites of the heterocycles of PA (2) are
highlighted in bold]. Fluorophore-tagged DNA duplexes
containing a 1 bp (ODN2), 2 bp (ODN3), or 3 bp (ODN4)
mismatch were also used in this study. UV−vis melting studies
of PA (2) in complex with ODN1 [i.e., PA(2)@ODN1]
revealed an 8.5 °C stabilization of the duplex melting
temperature (ΔTm) (Figure S5 in the Supporting Information)
relative to free ODN1. In contrast, a stabilization of only 4.2 °C
(Figure S6 in the Supporting Information) for PA(2)@ODN2
was observed for the 1 bp mismatched ODN2. The
stabilizations for the 2 bp (ODN3, ΔTm = 2.0 °C; Figure S7
in the Supporting Information) and 3 bp (ODN4, ΔTm = 0.2
°C; Figure S8 in the Supporting Information) mismatched
versions were even less pronounced. This study confirmed the
selectivity of PA(2) to bind to its matched sequence, ODN1
(Table 2).
Assessment of dsDNA Binding to Surfaces Derivat-
ized with PA (1). The ability of the PA-derivatized surfaces to
bind target dsDNA sequences selectively was then assessed.
Fluorescently tagged Cy3−DNA strands were used as a
reporter of a PA−dsDNA binding event. The PA-derivatized
surface was incubated in a buffered solution containing
ODN1−4 (1 μM and 500 nM) at 23 °C for 12 h. An
important aspect in assessing selective binding was washing
away excess dsDNA. Initial washing steps of the PA-derivatized
surface (three in total) were performed at 23 °C (Figure 5a) in
the presence of either the matched ODN1 or the 1 bp
mismatched ODN2. An intense fluorescence emission was
observed after either excess ODN1 (Figure 4) or ODN2 was
washed off, indicating poor sequence discrimination by the
immobilized PA surface. In contrast, a 2-fold reduction in the
fluorescence intensity of the patterned areas was observed for
the 2 bp mismatched duplex ODN3. Finally, no fluorescence
emission in the patterned areas was observed in the presence of
the 3 bp mismatched ODN4 (Figure 5a). We therefore
conclude that our PA-derivatized surface selectively binds to
ODN1 (matched) and ODN2 (1 bp mismatch), but
discrimination between ODN1 and ODN2 was not possible
using washing steps conducted at 23 °C.
To improve the sequence selectivity of the process, washing
steps (again three in total) were conducted at an elevated
temperature of 50 °C. After washing at 50 °C, a 3-fold decrease
in the fluorescence emission was observed for the 1 bp
mismatched ODN2 relative to the matched ODN1. No
fluorescent pattern was visible (i.e., no distinction between
the fluorescence emission in the patterned area and the
background signal) for either the 2 bp mismatched ODN3 or
the 3 bp mismatched ODN4 (Figure 5b). We surmise that
enhanced destabilization of the binding of the mismatched
sequences (ODN2−4) relative to the matched system (ODN1)
3454
dx.doi.org/10.1021/ja309677h | J. Am. Chem. Soc. 2013, 135, 3449−3457

Journal of the American Chemical Society
Article
Table 3. Competitive Binding Experiments Using PA-
a
Derivatized Substrates
ODN conc. (nM)
entry
fluorescence (arb. units)
ODN1
ODN2
ratio
Cy3−ODN1/ODN2
1
2
3
4
5
P: 56.97 6.74
500
500
500
100
50
50
100
500
500
500
10:1
5:1
N: 42.8 5.4
P: 51.13 5.8
N: 37.54 5.94
P: 55.55 8.6
N: 39.25 6.09
P: 45.19 5.16
N: 34.26 5.33
P: 49.3 5.46
N: 34.59 5.37
1:1
1:5
1:10
ODN1/Cy3−ODN2
6
no pattern
500
500
1:1
a
P = patterned area; N = nonpatterned area.
Figure 5. Comparative analysis of the fluorescence intensities of
patterned and nonpatterned areas of glass surfaces derivatized with PA
(1) and exposed to 1 μM concentration of ODN1 (black), ODN2
(gray), ODN3 (red), and ODN4 (green). (a) Washings were
conducted in 50 mM phosphate buffer (pH 7.5) containing 0.01%
Tween 20 at 23 °C for 3 min. Room temperature intensity comparison
(where patterned and nonpatterned areas were indistinct, the same
value was used for both; further details of the image analysis can be
found in Figures S5 and S9−16 in the Supporting Information). (b)
Three washing steps were conducted in 50 mM phosphate buffer (pH
7.5) containing 0.01% Tween 20 at 50 °C for 3 min.
Figure 6. Competitive binding analysis of the fluorescence intensities
of patterned and nonpatterned areas of glass surfaces derivatized with
PA (1) and exposed to Cy3−ODN1/ODN2 mixtures: (1) 500 nM
Cy3−ODN1/50 nM ODN2; (2) 500 nM/100 nM; (3) 500 nM/500
nM; (4) 100 nM/500 nM; (5) 50 nM/500 nM. Further details of the
image analysis can be found in Figures S21−25 in the Supporting
Information.
emission in the presence of increasing amounts of ODN2 was
observed in the patterned areas (entries 1−4 and Figure 6),
even in the presence of a 10-fold excess of ODN2 (entry 5 and
Figure 6). This suggests that the binding preference of PA-
derivatized surfaces for ODN1 is not significantly perturbed in
the presence of the competing 1 bp mismatched ODN2. The
data in entries 1−5 all were obtained with the fluorescence
reporter covalently linked to the matched strand. As an
additional verification of the results, the fluorescent reporter
was transferred to the mismatched ODN2 (Cy3−ODN2; entry
6). A crucial result was the lack of a defined fluorescence
pattern observed using an equimolar amount of Cy3−ODN2
and unlabeled ODN1 when this mixture was applied to a PA-
printed surface (Figure S26 in the Supporting Information),
indicating that the unlabeled matched ODN1 outcompetes
occurs when a washing step is conducted at an elevated
temperature.
Competition Assay Using PA-Derivatized Surfaces.
The capacity of our PA-derivatized surfaces to detect ODN1
selectively in the presence of 1 bp mismatched ODN2 was then
assessed by exposing the surfaces to a series of mixtures of
Cy3−ODN1 and ODN2. Within this series, the concentration
of Cy3−ODN1 was varied from 500 nM (Table 3, entry 1)
down to 50 nM (entry 5) with a commensurate increase in the
concentration of ODN2 (i.e., 50 nM in entry 1 to 500 nM in
entry 5). Strikingly, very little variation in the fluorescence
3455
dx.doi.org/10.1021/ja309677h | J. Am. Chem. Soc. 2013, 135, 3449−3457

Journal of the American Chemical Society
Article
Cy3−ODN2 for the PA-binding sites on the surface. We
therefore conclude that our PA-derivatized surface is selective
for ODN1 even in the presence of an excess of the competing
mismatched sequence ODN2.
sequence ODN1 can outcompete the challenging 1 bp
mismatched competitor ODN2 in this 50−500 nM regime
using our PA-derivatized surfaces.
PAs Maintain Their dsDNA Sequence Selectivity
When Immobilized on a Solid Substrate. A significant
outcome arising from this study is the ability of PAs to retain
their sequence selectivity when bound to a flat surface. PAs
tethered to gold nanoparticles, for example, exhibit excellent
mismatch discrimination of dsDNA duplexes using a gold
aggregation assay.39 This study has shown that this
phenomenon can be extended to a microarray format using
μCP as the fabrication method. We predict that further
optimization of the selectivity and sensitivity of these arrays
could be achieved by (i) optimizing the temperature and
washing steps and (ii) changing the reporter module from
fluorescence emission to a more sensitive technique such as
surface-enhanced resonance Raman spectroscopy.39,40
Nanomolar Detection of Matched dsDNA Sequences
Using PA-Derivatized Surfaces. Finally, the limit of
detection of our PA-derivatized microarray exemplar was then
assessed as a function of the concentration of ODN1−4
(concentration series ranging from 1 μM to 1 nM) using our
optimized binding and washing protocol. A distinct difference
in the fluorescence emissions of the patterned and non-
patterned areas was observed using ODN1 down to 10 nM
(Figure S27 in the Supporting Information). No fluorescence
emission pattern was observed at this concentration in case of
the 1 bp mismatched ODN2. Furthermore, no pattern was
visible (i.e., no distinction between the fluorescence emission
and the background signal could be observed) when PA-
derivatized surfaces were exposed to 2 bp mismatched ODN3
or 3 bp mismatched ODN4 at 10 nM, thus confirming that PAs
can detect dsDNA sequences selectively when immobilized on
solid substrates.
CONCLUSIONS
■
We have demonstrated for the first time a new microarray
format that can detect dsDNA sequences with nanomolar
sensitivity and very good mismatch discrimination. Since PAs
have demonstrated binding affinity and selectivity for sequences
up to 16 bp in length,17 the development of PA-based
microarrays targeting unique dsDNA genomic sequences (i.e.,
>12 bp) is indeed a possibility. Since microarray technology
offers superior throughput capabilities and is generally less
expensive than other methods such as next-generation
sequencing,41,42 the development of a binding platform of
this type could offer an initial high-throughput screening tool to
interrogate the dsDNA binding profile of small-molecule
ligands that could be used upstream of a more detailed but
lower-throughput technique such as deep sequencing.43−45
A second and potentially powerful application of this
technology could be its utility in structural DNA nano-
technology. PAs are effective ligands for programming the
assembly of functional materials along DNA duplexes46 and
nanostructures.47,48 Our proposed technology could therefore
offer the DNA nanotechnology field new opportunities to
facilitate the immobilization of DNA nanostructures in
predefined locations, for example.47−50 Future developments
will now focus on optimizing the sensitivity, sequence
selectivity, and throughput as well as developing these methods
into a more generalized label-free platform for the analysis of
PA binding for both biological and nanotechnological
applications.
DISCUSSION
■
This study was designed to test whether PAs maintained their
ability to bind to their target dsDNA binding sequence
selectively when immobilized on a flat surface. We discuss
here several conclusions that emerged from our results.
The Temperature of the Washing Step Enhances the
Sequence Selectivity of the PA-Derivatized Microarray.
Discrimination between the matched sequence (ODN1) and
the 1 bp mismatched sequence (ODN2) was observed when a
washing procedure was applied at an elevated temperature, in
this case 50 °C. The application of a washing step conducted at
an elevated temperature is a well-established method for
suppressing mismatched binding in DNA microarrays.35,36 The
ionic strength of the buffer is another possible parameter that
can be tuned to enhance the discrimination between matched
and mismatched binding regimes in DNA microarrays.35−38 In
this study, only the temperature variable was investigated, but
there is significant scope to optimize the selectivity of dsDNA
binding using PA-derivatized arrays by investigating variation of
both the temperature and the ionic strength in future testing
cascades. Recent innovations in engineered microfluidic
devices, for example, have been developed for DNA micro-
arrays and could be adapted to investigate their utility in
optimizing selective dsDNA binding with PA-derivatized
microarrays.35,36
An unexpected result was the high degree of sequence
selectivity for the matched sequence (Cy3−ODN1) even in the
presence of an excess of the mismatched one (ODN2). The
fluorescence emission observed from the PA microarray was
effectively the same whether the concentration of Cy3−ODN1
was 500 nM (Table 3, entry 1) or 50 nM (entry 5). We
rationalize this result as follows: At 50 nM Cy3−ODN1, all of
the PA-binding sites on the substrate were saturated. Thus, the
use of more concentrated samples of Cy3−ODN1 (e.g., entries
1−4) did not result in a further enhancement of the
fluorescence emission in the patterned areas. When the
fluorophore was transferred from ODN1 to the mismatched
sequence (Cy3−ODN2; entry 6), the PAs present on the
surface were saturated with nonfluorescent ODN1, and as a
result, no defined fluorescence pattern was observed in the PA-
patterned areas. Therefore, we can conclude that the matched
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthesis and characterization of PAs, characterization of PA-
derivatized substrates, and fluorescence analysis of DNA-
binding to PA-derivatized surfaces. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
b.j.ravoo@uni-muenster.de; glenn.burley@strath.ac.uk
Notes
The authors declare no competing financial interest.
3456
dx.doi.org/10.1021/ja309677h | J. Am. Chem. Soc. 2013, 135, 3449−3457

Journal of the American Chemical Society
Article
(38) Binder, H.; Krohn, K.; Burden, C. J. BMC Bioinf. 2010, 11, 291.
ACKNOWLEDGMENTS
■
̌
́
(39) Krpetic, Z.; Singh, I.; Su, W.; Guerrini, L.; Faulds, K.; Burley, G.
G.A.B. acknowledges the EPSRC for financial support via an
Advanced Research Fellowship (EP/E055095/1) and a Follow-
on-Fund Grant (EP/H02915X/1). B.J.R. thanks the DFG
(Ra1732/2) for financial support.
A.; Graham, D. J. Am. Chem. Soc. 2012, 134, 8356.
(40) Guerrini, L.; Graham, D. Chem. Soc. Rev. 2012, 41, 7085.
(41) Roh, S. W.; Abell, G. C. J.; Kim, K.-H.; Nam, Y.-D.; Bae, J.-W.
Trends Biotechnol. 2010, 28, 291.
(42) Hurd, P. J.; Nelson, C. J. Briefings Funct. Genomics Proteomics
2009, 8, 174.
(43) Bock, C. Nat. Rev. Genet. 2012, 13, 705.
(44) Minoshima, M.; Bando, T.; Sasaki, S.; Fujimoto, J.; Sugiyama, H.
Nucleic Acids Res. 2008, 36, 2889.
(45) Meier, J. L.; Yu, A. S.; Korf, I.; Segal, D. J.; Dervan, P. B. J. Am.
Chem. Soc. 2012, 134, 17814.
(46) Su, W.; Schuster, M.; Bagshaw, C. R.; Rant, U.; Burley, G. A.
Angew. Chem., Int. Ed. 2011, 50, 2712.
(47) Cohen, J. D.; Sadowski, J. P.; Dervan, P. B. J. Am. Chem. Soc.
2008, 130, 402.
(48) Cohen, J. D.; Sadowski, J. P.; Dervan, P. B. Angew. Chem., Int.
Ed. 2007, 46, 7956.
REFERENCES
■
(1) Yoo, S. M.; Choi, J. H.; Lee, S. Y.; Yoo, N. C. J. Microbiol.
Biotechnol. 2009, 19, 635.
(2) Sassolas, A.; Leca-Bouvier, B. D.; Blum, L. J. Chem. Rev. 2008,
108, 109.
(3) Carter, N. P. Nat. Genet. 2007, 39, S16.
(4) Gilad, Y.; Borevitz, J. Curr. Opin. Genet. Dev. 2006, 16, 553.
(5) Syvanen, A. C. Nat. Rev. Genet. 2001, 2, 930.
(6) Ghosh, I.; Stains, C. I.; Ooi, A. T.; Segal, D. J. Mol. BioSyst. 2006,
2, 551.
(7) Howell, A.; Sims, A. H.; Ong, K. R.; Harvie, M. N.; Evans, D. G.
R.; Clarke, R. B. Nat. Clin. Pract. Oncol. 2005, 2, 635.
(8) Hoheisel, J. D. Nat. Rev. Genet. 2006, 7, 200.
(9) Berger, M. F.; Philippakis, A. A.; Qureshi, A. M.; He, F. S.; Estep,
P. W., III; Bulyk, M. L. Nat. Biotechnol. 2006, 24, 1429.
(10) Stormo, G. D.; Zhao, Y. Nat. Rev. Genet. 2010, 11, 751.
(11) Carlson, C. D.; Warren, C. L.; Hauschild, K. E.; Ozers, M. S.;
Qadir, N.; Bhimsaria, D.; Lee, Y.; Cerrina, F.; Ansari, A. Z. Proc. Natl.
Acad. Sci. U.S.A. 2010, 107, 4544.
(49) Pinheiro, A. V.; Han, D.; Shih, W. M.; Yan, H. Nat. Nanotechnol.
2011, 6, 763.
(50) Kershner, R. J.; Bozano, L. D.; Micheel, C. M.; Hung, A. M.;
Fornof, A. R.; Cha, J. N.; Rettner, C. T.; Bersani, M.; Frommer, J.;
Rothemund, P. W. K.; Wallraff, G. M. Nat. Nanotechnol. 2009, 4, 557.
NOTE ADDED AFTER ASAP PUBLICATION
■
The version of this article posted ASAP on February 21, 2013,
contained errors in Schemes 2 and 3. The corrected version was
reposted on February 22, 2013.
(12) Warren, C. L.; Kratochvil, N. C. S.; Hauschild, K. E.; Foister, S.;
Brezinski, M. L.; Dervan, P. B.; Phillips, G. N.; Ansari, A. Z. Proc. Natl.
Acad. Sci. U.S.A. 2006, 103, 867.
(13) Puckett, J. W.; Muzikar, K. A.; Tietjen, J.; Warren, C. L.; Ansari,
A. Z.; Dervan, P. B. J. Am. Chem. Soc. 2007, 129, 12310.
(14) Tse, W. C.; Boger, D. L. Acc. Chem. Res. 2004, 37, 61.
(15) Tse, W. C.; Boger, D. L. Chem. Biol. 2004, 11, 1607.
(16) Dervan, P. B.; Edelson, B. S. Curr. Opin. Struct. Biol. 2003, 13,
284.
(17) Trauger, J. W.; Baird, E. E.; Dervan, P. B. J. Am. Chem. Soc.
1998, 120, 3534.
(18) Wendeln, C.; Singh, I.; Rinnen, S.; Schulz, C.; Arlinghaus, H. F.;
Burley, G. A.; Ravoo, B. J. Chem. Sci. 2012, 3, 2479.
(19) Wendeln, C.; Ravoo, B. J. Langmuir 2012, 28, 5527.
(20) Jewett, J. C.; Bertozzi, C. R. Chem. Soc. Rev. 2010, 39, 1272.
(21) Sletten, E. M.; Bertozzi, C. R. Angew. Chem., Int. Ed. 2009, 48,
6974.
(22) Singh, I.; Heaney, F. Chem. Commun. 2011, 47, 2706.
(23) Balachander, N.; Sukenik, C. N. Langmuir 1990, 6, 1621.
(24) Mehlich, J.; Ravoo, B. J. Org. Biomol. Chem. 2011, 9, 4108.
(25) Ravoo, B. J. J. Mater. Chem. 2009, 19, 8902.
(26) Rozkiewicz, D. I.; Gierlich, J.; Burley, G. A.; Gutsmiedl, K.;
Carell, T.; Ravoo, B. J.; Reinhoudt, D. N. ChemBioChem 2007, 8, 1997.
(27) Singh, I.; Freeman, C.; Madder, A.; Vyle, J. S.; Heaney, F. Org.
Biomol. Chem. 2012, 10, 6633.
(28) Singh, I.; Freeman, C.; Heaney, F. Eur. J. Org. Chem. 2011, 6739.
(29) Singh, I.; Vyle, J. S.; Heaney, F. Chem. Commun. 2009, 3276.
(30) Tornoe, C. W.; Christensen, C.; Meldal, M. J. Org. Chem. 2002,
67, 3057.
(31) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem., Int. Ed.
2001, 40, 2005.
(32) Birker, P. J. M. W. L.; Hendriks, H. M. J.; Reedijk, J.; Verschoor,
G. C. Inorg. Chem. 1981, 20, 2408.
(33) White, S.; Szewczyk, J. W.; Turner, J. M.; Baird, E. E.; Dervan, P.
B. Nature 1998, 391, 468.
(34) Su, W.; Gray, S. J.; Dondi, R.; Burley, G. A. Org. Lett. 2009, 11,
3910.
(35) Petersen, J.; Poulsen, L.; Birgens, H.; Dufva, M. PLoS One 2009,
4, No. e4808.
(36) Petersen, J.; Poulsen, L.; Petronis, S.; Birgens, H.; Dufva, M.
Nucleic Acids Res. 2008, 36, No. e10.
(37) Pozhitkov, A. E.; Stedtfeld, R. D.; Hashsham, S. A.; Noble, P. A.
Nucleic Acids Res. 2007, 35, No. e70.
3457
dx.doi.org/10.1021/ja309677h | J. Am. Chem. Soc. 2013, 135, 3449−3457