of 6
DNA sensing by electrocatalysis with hemoglobin
Catrina G. Pheeney, Luis F. Guerra, and Jacqueline K. Barton
1
Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA 91125
Edited by Royce W. Murray, University of North Carolina at Chapel Hill, Chapel Hill, NC, and approved May 22, 2012 (received for review March 23, 2012)
Electrocatalysis offers a means of electrochemical signal amplifica-
tion, yet in DNA-based sensors, electrocatalysis has required high-
density DNA films and strict assembly and passivation conditions.
Here, we describe the use of hemoglobin as a robust and effective
electron sink for electrocatalysis in DNA sensing on low-density
DNA films. Protein shielding of the heme redox center minimizes
direct reduction at the electrode surface and permits assays on
low-density DNA films. Electrocatalysis with methylene blue that
is covalently tethered to the DNA by a flexible alkyl chain linkage
allows for efficient interactions with both the base stack and he-
moglobin. Consistent suppression of the redox signal upon incor-
poration of a single cytosine-adenine (CA) mismatch in the DNA
oligomer demonstrates that both the unamplified and the electro-
catalytically amplified redox signals are generated through DNA-
mediated charge transport. Electrocatalysis with hemoglobin is
robust: It is stable to pH and temperature variations. The utility
and applicability of electrocatalysis with hemoglobin is demon-
strated through restriction enzyme detection, and an enhancement
in sensitivity permits femtomole DNA sampling.
DNA charge transport
DNA sensors
mismatch detection
D
NA-modified electrodes have been extensively studied in the
development of sensitive DNA-based sensors using DNA-
mediated charge transport (CT) chemistry (1
18). Sensors based
on DNA-mediated CT are exquisitely sensitive to structural
perturbations caused by single-base mismatches, lesions, and
DNA-binding proteins (2
4). Due to the high specificity, robust
nature, and low cost of these DNA-based electrochemical sen-
sors, they are being developed as new cancer diagnostics through
the detection of transcription factors and low abundance micro-
RNAs (4, 5, 19, 20). Despite the intrinsic specificity of these
devices, these DNA-CT sensors currently lack the required sen-
sitivity to detect biologically relevant levels of cancer markers.
Therefore, a means to reliably and robustly amplify the DNA-
mediated signal is essential for the application of these technol-
ogies to the rapid detection of cancer markers directly from cell
lysates.
Different electrochemical strategies have been used to achieve
higher sensitivities, including the use of gold nanoparticles, con-
ducting polymers, and catalysis (21
27). Electrocatalysis, how-
ever, represents a preferred means of signal amplification in
electrochemical sensors (25
27). The degree of electrocatalytic
signal amplification depends upon the coupling of the catalysis
with the electrode surface and the electrode sink (28). Addition-
ally, robust reporting by electrocatalysis is dependent on the rig-
orous shielding of the electron sink from the electrode surface,
which prevents direct reduction by the electrode and false posi-
tive outputs (27). In electrocatalysis, the redox reporter is
coupled with a freely diffusing electron sink, typically ferricya-
nide. Methylene blue (MB) is typically used as the redox reporter
for DNA-mediated electrocatalysis as it is well-coupled to the
base stack through intercalation (26). Reduction of the reporter
leads to the reduction of the electron sink, returning the reporter
back to its oxidized form, and allowing for the repeated interro-
gation of the base stack. To complete this catalytic cycle reporting
on DNA-mediated changes, the reporter must interact with both
the DNA film and the freely diffusing electron sink (26). Finally,
in order for DNA-modified electrodes to be viable for the detec-
tion of DNA-binding proteins and the hybridization of oligonu-
cleotides, the surface-bound DNA duplex must be sufficiently
accessible (29
32). All three of these constraints must be met for
electrocatalysis to be applied as a detection platform in DNA CT-
based systems: dual reporter interactions, robust shielding of the
electron sink from surface reduction, and DNA accessibility.
Initial electrocatalysis experiments were performed using non-
covalently bound DNA reporters and high-density DNA films
(2, 3, 25, 26). The freely diffusing noncovalent reporter was able
to readily interact with both the electron sink and the base stack
while the high-density DNA films limited localization of the re-
porter to the distal end of the duplex. The high-density films also
imposed a strong kinetic barrier to the direct reduction of ferri-
cyanide at the electrode surface (2). This platform demonstrated
enhanced sensitivity due to electrocatalytic signal amplification
and allowed for efficient detection of single-base mismatches
(2, 3, 25, 26). The major limitation of this platform originated
from the need for well-packed films to passivate the surface and
ensure that the redox probe bound to the top of the DNA film.
These high-packing densities can be difficult to achieve reprodu-
cibly and lead to poor accessibility of the duplex DNA for analyte
binding (29
32).
Low-density DNA films are now the standard means for pro-
moting efficient DNA binding and hybridization events, as they
provide higher sensitivity and specificity. Low-density DNA films,
however, require covalent tethering of the reporter to the duplex
to ensure placement of the reporter near the top of the surface.
This covalent tethering of the reporter is furthermore essential
for label-free assays of DNA or protein targets. However, low-
density films expose the electrode surface, as the negatively
charged phosphate backbone of the immobilized DNA is insuffi-
cient to impose a kinetic barrier against the direct reduction of
positively charged species, such as ferricyanide. A previous at-
tempt at electrocatalysis with low-density films utilized negatively
charged functionalized mercapto-alkanes to shield ferricyanide,
the electron sink, from the electrode (27). This platform has not
been widely employed due to the limited solution accessibility
of the reporter and the difficulty imposing sufficient kinetic
barriers to the reduction of ferricyanide to drive electrocatalysis.
Additionally, this platform did not address the poor interaction
between the reporter and the electron sink imposed by the short,
rigid covalent tether linking the reporter to the duplex (27).
Therefore, a new means to promote efficient interactions be-
tween the reporter with the base stack and electron sink, as well
as to robustly shield the electron sink from the electrode surface,
must be developed for the application of electrocatalytic signal
amplification with these low-density DNA films.
Redox-active proteins may be useful as electron sinks for elec-
trocatalysis. Enzymes such as horseradish peroxide and glucose
oxidase have been shown to readily interact electronically with
phenothiazine dyes, such as MB, and have had wide application
Author contributions: C.G.P. and J.K.B. designed research; C.G.P. and L.F.G. performed
research; C.G.P., L.F.G., and J.K.B. analyzed data; and C.G.P. and J.K.B. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1
To whom correspondence should be addressed. E-mail: jkbarton@caltech.edu.
This article contains supporting information online at
www.pnas.org/lookup/suppl/
doi:10.1073/pnas.1201551109/-/DCSupplemental
.
11528
11533
PNAS
July 17, 2012
vol. 109
no. 29
www.pnas.org/cgi/doi/10.1073/pnas.1201551109
as recognition elements for molecular sensing (33
42). Detection
platforms based on these enzymes are desirable due to their
inherent catalytic activity. The application of a protein as an elec-
tron sink is also appealing because the protein shell itself can
shield the redox core from the electrode surface (34
41). Hemo-
globin exhibits this property and has been shown to couple with
Nile blue (42). Additionally, unlike many of the proteins typically
employed for signal amplification, the reduction of hemoglobin
does not generate any DNA-damaging byproducts like peroxide.
Until now, no single DNA-CTsensor has been compatible both
with use of low-density DNA films, required for hybridized DNA
or protein targets, and robust electrocatalytic signal amplifica-
tion, required for sufficient sensitivity. In this work, we report
a DNA-CTsensor that robustly and efficiently amplifies the signal
from low-density DNA films using electrocatalysis. This advance-
ment is the result of a new flexible alkyl linkage between the re-
porter and the duplex coupled with a novel electron sink. The
new flexible linkage allows for the efficient shuttling of electrons
between the duplex and the electron sink (28). The use of hemo-
globin allows electronic interaction with MB while inherently
being shielded from the electrode surface by the protein shell.
No specialized backfilling agents or assembly conditions were re-
quired, demonstrating the general applicability of this system.
This general technique, electrocatalysis of MB with hemoglobin,
can therefore be utilized with current DNA-detection strategies
without imposing any new constraints to the platform. Most im-
portantly, the enhanced sensitivity permits femtomole DNA
sampling.
Results and Discussion
Electrocatalytic Signal Amplification of MB-Modified DNA by Hemo-
globin.
In this study, MB is covalently tethered to DNA with a
flexible
ð
CH
2
Þ
8
linkage, such that, upon reduction, the tethered
MB mimics the behavior of a noncovalent reporter (Fig. 1).
Hemoglobin is known to interact electrochemically with organic
dyes, such as MB, and can function as an electronic sink for cat-
alytically reduced MB (42). The general electrocatalytic cycle
developed for this study utilizing MB-DNA and hemoglobin is
presented in Fig. 2. The two-electron reduction of MB to leuco-
methylene blue (LB) significantly decreases the binding affinity
of the reporter, resulting in its dissociation from the duplex.
LB proceeds to electrochemically interact with, and reduce, two
equivalents of freely diffusing hemoglobin, returning the reporter
to the oxidized form and completing the electrocatalytic cycle.
The low-density MB-DNA films employed in this study were
assembled without MgCl
2
and had typical surface coverages of
1
.
3

0
.
2
pmol
cm
2
. Prior to the addition of hemoglobin, the re-
sulting cyclic voltammograms (CVs) demonstrate a reversible re-
dox couple consistent with a signal generated from MB reduction
via DNA-mediated CT: a midpoint potential of
340
mV, peak
splitting of
50
mV, and a ratio of near unity for the cathodic and
anodic peak areas (Fig. 2). MB-DNA modified electrodes were
then examined in the presence of freely diffusing hemoglobin
(25
μ
M). The electrocatalytic signal amplification is apparent in
the CV by the characteristic nonreversible signal, where the re-
ductive peak is amplified and the oxidative peak is not observed.
Hemoglobin was also added to DNA in the absence of MB, show-
ing that hemoglobin alone has no effect on the observed electro-
chemistry (Fig. 2). Additionally, the electronic spectrum of hemo-
globin showed no significant variation before and after electro-
catalysis, indicating that protein integrity was retained (
Fig. S1
).
In the absence of hemoglobin, chronocoulometry at
450
mV
vs. Ag/AgCl, a potential appropriate for reduction MB-DNA,
showed a rapid initial accumulation of charge (0.04
μ
C over
0.3 s), with minimal charge accumulation (0.1
μ
C) over the
subsequent 10 s (Fig. 2). This modest accumulation of charge
is consistent with the reduction of surface-bound MB without
an electron sink for catalytic cycling. In the presence of hemoglo-
bin, there is the same rapid initial accumulation of charge as in
the noncatalytic case, but over the subsequent 10 s, a total of
0
.
99

0
.
04
μ
C of charge is accumulated. A comparison of the
accumulated charge as a function of the presence of hemoglobin
yields insight into the catalytic behavior of MB. An accumulated
charge of
0
.
99

0
.
04
μ
C in the presence of freely diffusing
hemoglobin correlates to roughly 10 turnovers per MB-DNA.
This electrocatalytic amplification of the accumulated charge
due to MB-DNA reduction demonstrates that covalently tether-
ing MB to DNA through a
ð
CH
2
Þ
8
linkage does not disrupt the
ability of hemoglobin to electrochemically couple with MB.
Electrocatalytic Amplification of MB-DNA with Intervening Mis-
matches.
In order to demonstrate that the electrocatalytically
amplified signal is generated via DNA-mediated CT, the signal
suppression due to the incorporation of a single-base mismatch
was examined. Previously, the degree of signal attenuation upon
incorporation of a single-base mismatch indicated that MB cova-
lently tethered to the DNA duplex through a flexible
ð
CH
2
Þ
12
linkage can be reduced by both DNA CT and direct surface re-
duction (43). Here, we utilized a shorter
ð
CH
2
Þ
8
linkage in order
to reduce contributions from the direct surface reduction of MB.
Using this assembly, Fig. 3 illustrates electrocatalysis with and
without an intervening mismatch. Cytosine-guanine (CG)-rich
thiol-modified DNA was annealed with complementary well-
matched DNA as well as a mismatched sequence containing a
single CA mismatched base pair in the middle of the 17-mer
duplex. Well-matched and mismatched CG-rich DNAs were each
assembled as low-density DNA films on eight electrodes of a
multiplexed chip. The use of equivalent thiol strands and multi-
plexed chips removes variability due to surface quality and film
formation.
As is evident in Fig. 3, signal attenuation is observed in this
system with an intervening mismatch. The average reductive peak
area from the CVs for well-matched MB-DNA is
5
.
0

0
.
4
nC
and that of mismatched MB-DNA is
3
.
2

0
.
4
nC. This signal
attenuation in the presence of a single-base mismatch confirms
that the signal observed is generated via DNA-mediated CT.
Upon the addition of hemoglobin (25
μ
M), the MB-DNA signals
are electrocatalytically amplified, and the accumulated charge
for well-matched and mismatched DNA were compared. The
accumulated charge over 10 s from the chronocoulometry data
are
0
.
99

0
.
04
μ
C and
0
.
62

0
.
02
μ
C for well-matched and
mismatched MB-DNA, respectively (Fig. 3). The accumulated
charge obtained for the mismatched MB-DNA was again only
62% of the well-matched MB-DNA signal upon electrocatalytic
amplification. The consistency in the signal suppression due to
mismatch incorporation indicates that the catalytic reduction of
MB-DNA occurs via DNA CT. Therefore, the sensitivity of the
observed signal to subtle structural perturbations to the
π
-stack is
still apparent with electrocatalytic amplification.
Variations in pH and Temperature on Amplification.
DNA-modified
electrodes are generally stable over the pH range of 5.5
8.5,
where minimal damage occurs to DNA. The reduction peak area
of MB-DNA, without amplification, is unaffected by changes in
pH, while the peak potential is shifted positive with increasing
acidity at a rate of 25 mV per pH unit (
Fig. S2
). This is expected,
HN
N
O
O
O
O
HO
H
N
N
H
O
N
S
+
N
N
Fig. 1.
Structure of MB-modified DNA with a flexible
ð
CH
2
Þ
8
linkage cova-
lently tethering MB to the DNA.
Pheeney et al.
PNAS
July 17, 2012
vol. 109
no. 29
11529
BIOCHEMISTRY
CHEMISTRY
SPECIAL FEATURE
as MB is protonated upon reduction. Previous work with enzyme-
mediated electrocatalytic cascades has demonstrated a nonlinear
response to pH changes due to the pH dependence of the enzy-
matic activity (34
42). In these cases, where catalysis is limited by
enzyme activity, constraints are imposed not only on the pH but
also on other running conditions relevant to enzyme activity
(34
42). In the system presented here, the electrocatalysis is
not dependent on hemoglobin reactivity with oxygen, only the
capability of hemoglobin to function as an electron sink. The elec-
trocatalytic amplification of MB should therefore be optimal at a
pH lower than the isoelectric point of hemoglobin (p
I
¼
6
.
8
)
where hemoglobin will be less negatively charged (42). We de-
monstrated this to be true by examining the accumulated charge
in the presence of hemoglobin at a range of pH values (5.5
8.5).
All solutions were buffered and had the same salt and hemoglo-
bin concentrations. Between pH 5.5 and 8.5, the accumulated
charge increases linearly with increasing acidity at a rate of
0.15
μ
C per pH unit (
Fig. S2
). This indicates that the electroca-
talytic amplification obtained is not limited by hemoglobin activ-
ity and behaves as would be expected for a proton-dependent
electrocatalytic cycle. Therefore, hemoglobin activity does not
impose new restrictions on the running conditions of DNA-mod-
ified electrodes compared to free iron, demonstrating the utility
of hemoglobin as an electron sink in electrocatalysis.
The dependence of the total accumulated charge was also
examined as a function of the temperature to demonstrate both
the stability and necessity for protein activity. The temperature
was systematically increased from 24 °C to 38 °C at a rate of
0
.
5
°C
min. The accumulation of charge was found to increase
at a rate of
20
μ
C
°C in the range of 24
32 °C (Fig. 4). Within
this temperature range, mismatch discrimination was still ob-
served in the catalytic signal. At increased temperatures ranging
from 32 °C to 40 °C, the accumulated charge increased at a rate of
60
μ
C
°C, roughly three times faster than at lower temperatures,
and the discrimination due to mismatch incorporation was lost
(Fig. 4). While hemoglobin is stable under physiological condi-
tions, the thermal denaturation of hemoglobin is highly pH
and concentration-dependent. In the buffer used for this study,
a transition was observed by circular dichroism at 32
35 °C
(
Fig. S3
). The accelerated charge accumulation and the loss in
Fig. 2.
Electrocatalytic reduction of MB-modified DNA by hemoglobin.
Left
: Schematic representation of the electrocatalytic cycle of MB-DNA. In this model,
upon reduction of MB to LB, the DNA-binding affinity is significantly decreased, resulting in the dissociation of reporter from the duplex and its sub
sequent
oxidation by freely diffusing hemoglobin.
Center
: Cyclic voltammetry (scan rate
¼
30
mV
s) of MB-DNA in the absence (black), DNA without MB in presence
(gray) and MB-DNA in the presence (red) of 25
μ
M hemoglobin in phosphate buffer (5 mM phosphate, 5 mM NaCl, 40 mM MgCl
2
, 5 mM spermidine, and
pH 7.0).
Right
: Corresponding chronocoulometry signal (V
¼
450
mV) acquired over 10 s intervals.
Fig. 3.
Signal suppression of MB-DNA due to incorporation of a single CA mismatch.
Left
: Low-density MB-DNA films were assembled on multiplexed chips.
Sequence for both well-matched MB-DNA (blue) and mismatched MB-DNA (red) are indicated.
Center
: In the absence of hemoglobin, CV scans
(scan rate
¼
100
mV
s) were used to demonstrate the characteristic signal attenuation due to incorporation of a single CA mismatch.
Right
: Chronocoulometry
scans (V
¼
450
mV) were used for electrocatalytically amplified MB-DNA in the presence of 25
μ
M hemoglobin. All scans were acquired in phosphate buffer
(5 mM phosphate, 5 mM NaCl, 40 mM MgCl
2
, 5 mM spermidine, and pH 7.0).
11530
www.pnas.org/cgi/doi/10.1073/pnas.1201551109
Pheeney et al.
mismatch discrimination indicate that upon structural changes to
hemoglobin the heme-bound iron is no longer effectively shielded
by the protein shell. These low-density DNA films are not de-
signed for passivation against the direct reduction of iron. This
temperature profile demonstrates the necessity for intact
hemoglobin to shield the iron from the electrode surface in order
to observe the electrocatalytically amplified MB-DNA signal and
avoid false positives.
Detection of Restriction Enzyme Activity.
Having demonstrated that
hemoglobin is an effective electron sink in DNA-mediated elec-
trocatalysis, we can apply this electrocatalysis to the detection of
DNA-binding proteins. The restriction enzyme
RsaI
sequence-
specifically cuts duplex DNA that contains the binding site
5
-GTAC-3
(8).
RsaI
activity was compared between DNA
sequences with and without the binding site (Fig. 5). Half the
electrodes of a multiplex chip were treated with
RsaI
(20
μ
L
of 50 nM) while the other half were left untreated. A comparison
between the treated and untreated electrodes was made for each
sequence in order to demonstrate the sequence-specificity of
RsaI
. Additionally,
RsaI
activity and electrocatalytic signal ampli-
fication were compared using both high-density and low-density
DNA films to demonstrate the effect of duplex accessibility on
the detection of DNA-binding proteins.
Low-density films of the
RsaI
cutting sequence showed accu-
mulated charges of
0
.
55

0
.
02
μ
C for untreated samples and
0
.
32

0
.
01
μ
C for samples treated with
RsaI
(Fig. 5). With these
low-density DNA films, this corresponds to a 40% signal attenua-
tion due to treatment with
RsaI
. This signal attenuation was se-
quence-specific, as electrodes modified with low-density DNA
films of the control sequence showed no differential from those
treated with
RsaI
(Fig. 5).
Using the traditionally implemented high-density DNA films,
no statistically relevant decrease could be discerned from either
the
RsaI
sequence or the control sequence upon
RsaI
treatment.
The
RsaI
sequence had accumulated charges of
0
.
36

0
.
05
μ
C
for untreated and
0
.
34

0
.
08
μ
C for treatment with
RsaI
, while
the control sequence had accumulated charges of
0
.
38

0
.
02
μ
C
for untreated and
0
.
41

0
.
03
μ
C for treatment with
RsaI
(Fig. 5).
In addition to the lack of detectable
RsaI
activity with high-den-
sity DNA films, there was significantly higher variability in the
electrocatalytically amplified signal with high-density films.
Overall, the low-density DNA films had enhanced electroca-
talytic activity and decreased variability compared to the high-
density films which can be attributed to the enhanced accessibility
of both
RsaI
and hemoglobin. Thus, electrocatalysis with hemo-
globin clearly requires duplex accessibility.
Sensitivity of DNA Detection.
The sensitivity to the reduction of
MB-DNA is compared with and without hemoglobin to demon-
strate the enhanced sensitivity that is achieved with the overall
signal gain of electrocatalysis. The reduction signal obtained from
electrodes assembled with decreasing amounts of MB-DNA was
compared with and without electrocatalysis. In this experiment,
Fig. 4.
Temperature dependence of the electrocatalysis of MB-DNA. The
temperature was increased at a rate of
0
.
5
°C
min by a Peltier device placed
beneath the multiplexed chip. The electrocatalytic amplification was moni-
tored by chronocoulometry at two-second intervals.
Left
: The accumulated
charge is plotted as a function of the measured solution temperature.
Two different linear regimes can be discerned with rates of
20
μ
C
°C and
60
μ
C
°C at moderate (24
32 °C) and high (32
40 °C) temperatures, respec-
tively.
Right
: In each regime, the effect on mismatch discrimination was
examined by chronocoulometry; well-matched (blue) and mismatched
(red) MB-DNA are shown at both 25 °C (
Bottom
) and 38 °C (
Top
). All scans
were acquired in phosphate buffer (5 mM phosphate, 5 mM NaCl, 40 mM
MgCl
2
, 5 mM spermidine, and pH 7).
Fig. 5.
Electrocatalytic detection of sequence-specific restriction enzyme
RsaI
activity.
Left
: Sequences for the control and
RsaI
binding sequence are indicated.
The binding site for
RsaI
is indicated in purple. All electrodes were incubated with BSA (1 mM for 1 h to minimize any signal attenuation observed from
nonspecific protein binding). Electrodes were either
Left
untreated (red) or treated (green) with
RsaI
(20
μ
L of 50 nM) for 2 h.
Right
: The accumulated charge
from the chronocoulometry data (V
¼
450
mV for 10 s) of MB-DNA in the presence of 25
μ
M hemoglobin for the control sequence is plotted for both high-
density (dark) and low-density (light) DNA films. High-density films were formed by the addition of 100 mM MgCl
2
during DNA film self-assembly. The ana-
logous data for the
RsaI
sequence is also presented.
Pheeney et al.
PNAS
July 17, 2012
vol. 109
no. 29
11531
BIOCHEMISTRY
CHEMISTRY
SPECIAL FEATURE
each set of four electrodes on a 16-electrode multiplexed chip was
assembled with 20
μ
L of MB-DNA of varying concentrations.
The electrodes were exposed to MB-DNA ranging from 25
μ
M
(500 pmol) down to 250 pM (5 fmol) and allowed to assemble
overnight. Once thoroughly washed and backfilled with mercap-
tohexanol, the electrodes were examined by CV in the absence
and chronocoulometry in the presence of electrocatalysis with he-
moglobin (25
μ
M).
The area of the reduction signal from MB-DNA in CV mea-
surements was found to decrease with decreasing amounts of
MB-DNA from 500 pmol down to 500 fmol (Fig. 6). This repre-
sents a dynamic range of only three orders of magnitude; assem-
bly with less than 500 fmol of MB-DNA results in no discernible
peaks from MB-DNA reduction in the CV. Upon addition of
hemoglobin, the accumulated charge after 10 s measured by
chronocoulometry was examined as a function of the amount
of MB-DNA. Electrocatalytic signal amplification with hemoglo-
bin resulted in a significantly enhanced dynamic range, where the
decrease in the accumulated charge due to decreasing moles of
MB-DNA during monolayer formation was discernible from
500 pmol down to 5 fmol, corresponding to an enhanced dynamic
range of five orders of magnitude (Fig. 6). This enhancement
corresponds to a 100-fold improvement in the sensitivity of the
device and a dynamic range down to femtomoles.
Implications.
In this study, a new means to improve the sensitivity
of DNA-modified electrodes by electrocatalysis is presented
using the novel electron sink hemoglobin. The use of low-density
films is integral for current DNA-binding protein and oligonu-
cleotide detection strategies, and electrocatalysis with hemoglo-
bin is shown to be the first compatible means of enhancing the
overall signal gain of these systems. Using a new flexible linkage
between MB and the DNA, and hemoglobin as an electron sink,
DNA-mediated CT in low-density DNA films was electrocataly-
tically amplified. As an electron sink, hemoglobin is effective
through physiologically relevant ranges of pH and temperature
and, as a compact protein, does not carry out direct reduction
at the surface, minimizing false positives. Restriction enzyme
reaction can be readily detected with high sensitivity, given the
duplex accessibility of the low-density films and high efficiency
of MB reaction with hemoglobin. Most importantly, the sensitiv-
ity for MB-DNA reduction was demonstrated to be improved
100
-fold upon electrocatalytic signal amplification. This en-
hanced sensitivity, which allows a high signal-to-noise ratio from
an electrode assembled with femtomoles of MB-DNA, should
provide the response necessary to detect the unlabeled cellular
markers, both DNA and protein targets, at low concentra-
tions (19, 20). As electrocatalytic signal amplification with hemo-
globin is a general means of enhancing the sensitivity of current
DNA-modified electrode technologies, it should be applicable
to many DNA electrochemical strategies based on redox report-
ing of MB.
Materials and Methods
All chemicals were purchased from Sigma Aldrich and used without further
purification. All proteins were purchased from New England Biolabs. Oligonu-
cleotides without unnatural modifications were purchased from Integrated
DNA Technologies while modified oligonucleotides were synthesized on a
3400 Applied Biosystems DNA synthesizer. Modified phosphoramidites were
purchased from Glen Research. Hemoglobin was obtained from Sigma Aldrich
as a lyophilized powder from Bovine Blood.
Oligonucleotides Preparation.
MB-modified DNA was prepared by the synth-
esis of amino-modified DNA utilizing an Amino-
ð
CH
2
Þ
2
-dT phosphoramidite
(Fig. 1). Amino-modified DNA, as well as thiol-modified DNA, was purified
through standard procedures, as previously reported (6). MB modified with
a terminal carboxylic acid was synthesized and coupled to the amino mod-
ified DNA (43). All single-stranded DNA was purified by reverse-phase HPLC
and characterized with matrix-assisted laser desorption/ionization-time of
flight (MALDI-TOF) mass spectrometry.
Single-stranded DNA stock solutions were prepared in phosphate buffer
(5.0 mM phosphate, 50 mM NaCl, and pH 7) and quantified by UV
Vis
spectroscopy based on their absorbance at 260 nm. The extinction coeffi-
cients for the single-strand DNA were estimated using SciTools from IDT.
The extinction coefficient for MB-modified DNA at 260 nm was corrected
for the contribution of MB to the absorbance by adding
10
;
300
M
1
cm
1
to the IDT-calculated extinction coefficient. All DNA solutions were thor-
oughly deoxygenated with argon prior to annealing. Equimolar amounts
of single-stranded stocks were combined and annealed by heating to
90 °C and cooling to ambient temperature over 90 min.
DNA-Modified Electrodes.
Multiplex chips were employed for the electroche-
mical experiments and were fabricated as previously reported (6). Chips con-
sisted of 16 gold electrodes (
2
mm
2
area) that were divided into four
quadrants of four electrodes. Duplex DNA (25
μ
Lof25
μ
M, except for con-
centration dependence experiment) was assembled overnight (20
24 h) in a
humid environment, allowing for monolayer formation. Low-density films
were assembled in the absence of MgCl
2
while high-density films were
assembled with the addition of 100 mM MgCl
2
. Once DNA films were
assembled and thoroughly washed with phosphate buffer, the electrodes
were backfilled with 1 mM 6-mercaptohexanol (MCH) for 45 min in phos-
phate buffer with 5% glycerol. The electrodes were scanned in a common
running buffer (5.0 mM phosphate, 5 mM NaCl, 40 mM MgCl
2
, 5 mM sper-
midine, and pH 7). Electrocatalysis data was acquired following the addition
of 25
μ
M hemoglobin (5 mM phosphate, 5 mM NaCl, 40 mM MgCl
2
,5mM
spermidine, and pH 7) to the central well.
Electrochemical Measurements.
Electrochemical measurements were per-
formed with a CHI620D Electrochemical Analyzer and a 16-channel multi-
plexer from CH Instruments. A three-electrode setup was used with a
common Ag/AgCl reference and a Pt wire auxiliary electrodes placed in the
central buffer solution. Cyclic voltammetry data were collected at
100
mV
s
unless otherwise indicated, and chronocoulometry data were acquired at an
applied potential of
450
mV for an interval of 10 s unless otherwise indi-
cated. Nonelectrocatalytic CV data were quantified by the area of the reduc-
tion peak for MB while electrocatalytic chronocoulometry data were quan-
tified by the accumulated charge after 10 s. The surface coverage of the
MB-DNA was based on the area of the reduction peak observed for the dis-
tally bound MB and calculated as previously reported (6). The catalytic turn-
over was calculated based on the ratio of the accumulated charge in the
presence and absence of hemoglobin.
Temperature Dependence.
Temperature dependence experiments were per-
formed by assembling the chip in a holder with a Peltier (thermoelectric)
device (Melcor Corp.) placed underneath the chip. The temperature was mea-
sured using a digital probe placed in the central well containing the running
Fig. 6.
Enhanced sensitivity for MB-DNA reduction upon electrocatalytic
signal amplification. Electrodes were incubated with 20
μ
L of MB-DNA of
varying concentrations, ranging from 25
μ
M down to 250 pM, resulting in
the electrodes being exposed to 500 pmol down to 5 fmol of MB-DNA during
monolayer formation. The surfaces were thoroughly washed and backfilled
with mercaptohexanol prior to measurement. Scans were performed in phos-
phate buffer (5 mM phosphate, 5 mM NaCl, 40 mM MgCl
2
, 5 mM spermidine,
and pH 7) with and without hemoglobin (25
μ
M). The reduction of MB-DNA
was quantified from the reductive peak area in the CV (black), without
hemoglobin, and the accumulated charge after 10 s in the chronocoulometry
(red), with hemoglobin.
11532
www.pnas.org/cgi/doi/10.1073/pnas.1201551109
Pheeney et al.
buffer. Chronocoulometry data were repeatedly acquired at pulse widths of
two seconds, while heating the chip at a rate of
0
.
5
°C
min over a tempera-
ture range of 24
38 °C.
Restriction Enzyme Detection.
DNA-modified electrodes used for
RsaI
detec-
tion were incubated with BSA (1 mM in 5.0 mM phosphate, 50 mM NaCl,
4.0 mM MgCl
2
, 4.0 mM spermidine, 50
μ
M EDTA, 10% glycerol, and pH 7)
for 1 h to minimize any signal attenuation observed from nonspecific protein
binding. The electrodes were then washed and scanned in the running buffer
(8). Half of the electrodes were then incubated for 2 h with
RsaI
restriction
enzyme (20
μ
L of 50 nM) in
RsaI
reaction buffer (10 mM Tris, 50 mM NaCl,
4 mM spermidine, 10 mM MgCl
2
, and pH 7.9) and then washed and scanned
in phosphate buffer.
ACKNOWLEDGMENTS.
This research was supported by the National Institute of
Health (GM61077) and ONR (N00014-09-1-1117). We would like to thank
Donald S. Clark for an undergraduate fellowship to L.F.G. The authors thank
N. Muren for discussions and contributions in fabricating the multiplexed
chips. This work was completed in part in the Caltech Micro/Nano Fabrication
Laboratory.
1. Kelley SO, Barton JK (1997) Electrochemistry of methylene blue bound to a DNA-mod-
ified electrode.
Bioconjug Chem
8:31
37.
2. Kelley SO, Boon E, Barton JK, Jackson N, Hill M (1999) Single-base mismatch detection
based on charge transduction through DNA.
Nucleic Acids Res
27:4830
4837.
3. Boon E, Ceres D, Drummond T, Hill M, Barton JK (2000) Mutation detection by elec-
trocatalysis at DNA-modified electrodes.
Nat Biotechnol
18:1096
1100.
4. Boon E, Salas J, Barton JK (2002) An electrical probe of protein
DNA interactions on
DNA-modified surfaces.
Nat Biotechnol
20:282
286.
5. Gorodetsky A, Buzzeo MC, Barton JK (2008) DNA-mediated Electrochemistry.
Biocon-
jug Chem
19:2285
2296.
6. Slinker J, Muren N, Gorodetsky A, Barton JK (2010) Multiplexed DNA-modified elec-
trodes.
J Am Chem Soc
132:2769
2774.
7. Kelley SO, Jackson NM, Hill MG, Barton JK (1999) Long-range electron transfer
through DNA films.
Angew Chem Int Ed
38:941
945.
8. Slinker J, Muren N, Renfrew S, Barton JK (2011) DNA charge transport over 34 nm.
Nat
Chem
3:230
235.
9. Murphy CJ, et al. (1993) Long-range photoinduced electron transfer through a DNA
helix.
Science
262:1025
1029.
10. Cai Z, Sevilla MD (2004) Long range charge transfer in DNA II.
Topics in Current Chem-
istry
, ed GB Schuster (Springer, Berlin, Heidelberg), Vol 237 p 103.
11. Wagenknecht HA, ed. (2005)
Charge Transfer in DNA: From Mechanism to Application
(Wiley-VCH Verlag GmbH & Co KGaA, Weinheim).
12. Holmlin ER, Dandliker PJ, Barton JK (1997) Charge transfer through the DNA base
stack.
Angew Chem Int Ed
36:2714
2730.
13. Genereux JG, Barton JK (2010) Mechanisms for DNA charge transport.
Chem Rev
110:1642
1662.
14. Hall DB, Holmlin RE, Barton JK (1996) Oxidative DNA damage through long-range
electron transfer.
Nature
382:731
735.
15. Kelley SO, Barton JK (1999) Electron transfer between bases in double helical DNA.
Science
283:375
381.
16. Kelley SO, Holmlin RE, Stemp EDA, Barton JK (1997) Photoinduced electron transfer in
ethidium-modified DNA duplexes: Dependence on distance and base stacking.
JAm
Chem Soc
119:9861
9870.
17. Delaney S, Barton JK (2003) Long-range DNA charge transport.
J Org Chem
68:6475
6483.
18. Genereux JC, Boal AK, Barton JK (2010) DNA-mediated charge transport in redox sen-
sing and signaling.
J Am Chem Soc
132:891
905.
19. Husale S, Persson H, Sahin O (2009) DNA nanomechanics allows direct digital detection
of complementary DNA and microRNA targets.
Nature
462:1075
1079.
20. Fang Z, et al. (2009) Direct profiling of cancer biomarkers in tumor tissue using a multi-
plexed nanostructured microelectrode integrated circuit.
ACS Nano
3:3201
3213.
21. Shao Z, Liu Y, Xiao H, Li G (2008) PCR-free electrochemical assay of telomerase activity.
Electrochem Commun
10:1502
1504.
22. Xia F, et al. (2010) An electrochemical supersandwich assay for sensitive and selective
DNA detection in complex matrices.
J Am Chem Soc
132:14346
14348.
23. Rowe A, et al. (2011) Electrochemical biosensors employing an internal electrode
attachment site and achieving reversible, high gain detection of specific nucleic acid
sequences.
Anal Chem
83:9462
9466.
24. Das J, Yang H (2009) Enhancement of electrocatalytic activity of DNA-conjugated gold
nanoparticles and its application to DNA detection.
J Phys Chem C
113:6093
6099.
25. Boon E, et al. (2002) An electrochemical probe of DNA stacking in an antisense oligo-
nucleotide containing a C3
-endo-locked sugar.
Angew Chem Int Ed
41:3402
3405.
26. Boon E, et al. (2003) Reduction of ferricyanide by methylene blue at a DNA-modified
rotating-disk electrode.
Langmuir
19:9255
9259.
27. Gorodetsky A, Hammond W, Hill M, Slowinski K, Barton JK (2008) Scanning electro-
chemical microscopy of DNA monolayers modified with Nile blue.
Langmuir
24:14282
14288.
28. Anson F (1980) Kinetic behavior to be expected from outer-sphere redox catalysts con-
fined within polymeric films on electrode surfaces.
J Phys Chem
84:3336
3338.
29. Zhang J, Wang L, Pan D, Song S, Fan C (2007) DNA hybridization
turns on
electro-
catalysis at gold electrodes.
Chem Commun
43:1154
1156 Cambridge.
30. Lapierre M, O
Keefe M, Taft B, Kelley S (2003) Electrocatalytic detection of pathogenic
DNA sequences and antibiotic resistance markers.
Anal Chem
75:6327
6333.
31. Lapierre-Devlin M, et al. (2005) Amplified electrocatalysis at DNA-modified nanowires.
Nano Lett
5:1051
1055.
32. Biagiotti V, et al. (2012) Probe accessibility effects on the performance of electroche-
mical biosensors employing DNA monolayers.
Anal Bioanal Chem
402:413
421.
33. Pelossof G, Tel-Vered R, Elbaz J, Willner I (2010) Amplified biosensing using the
horseradish peroxidase-mimicking DNAzyme as an electrocatalyst.
Anal Chem
82:4396
4402.
34. Du P, Zhou B, Cai C (2008) Development of an amperometric biosensor for glucose
based on electrocatalytic reduction of hydrogen peroxide at the single walled carbon
nanotube/Nile blue A nanocomposite modified electrode.
Electroanal Chem
614:149
156.
35. Qian J, Liu Y, Liu H, Yu T, Deng J (1996) An amperometric new methylene blue N-med-
iating sensor for hydrogen peroxide based on regenerated silk fibroin as an immobi-
lization matrix for peroxidase.
Anal Biochem
236:208
214.
36. Ni F, Feng H, Gorton L, Cotton T (1990) Electrochemical and SERS studies of chemically
modified electrodes: Nile blue A, a mediator for NADH oxidation.
Langmuir
6:66
73.
37. Ju H, Shen C (2001) Electrocatalytic reduction and determination of dissolved oxygen
at a poly(Nile blue) modified electrode.
Electroanalysis
13:789
793.
38. Xu J, Zhu J, Wu Q, Hu Z, Chen H (2003) An amperometric biosensor based on the coim-
mobilization of horseradish peroxidase and methylene blue on a carbon nanotubes
modified electrode.
Electroanalysis
15:219
224.
39. Liu H, Lu J, Zhang M, Pang D (2002) Electrochemical properties of Nile blue covalently
immobilized on a self-assembled thiol-monolayer modified gold electrodes.
Anal Sci
18:1339
1344.
40. Du P, Wu P, Cai C (2008) A glucose biosensor based on electrocatalytic oxidation
of NADPH at single-walled carbon nanotubes functionalized with poly(nile blue A).
J Electroanal Chem
624:21
26.
41. Evtugyn G, Porfierva A, Hianik T, Cheburova M, Budnikov H (2008) Potentiometric
DNA sensor based on electropolymerized phenothiazines for protein detection.
Electroanalysis
20:1300
1308.
42. Kuramitz H, et al. (1999) Electrocatalytic reduction of hemoglobin at a self-assembled
monolayer electrode containing redox dye, Nile blue as an electron-transfer mediator.
Anal Sci
15:589
592.
43. Pheeney CG, Barton JK (2012) DNA electrochemistry with tethered methylene blue.
Langmuir
28:7063
7070.
Pheeney et al.
PNAS
July 17, 2012
vol. 109
no. 29
11533
BIOCHEMISTRY
CHEMISTRY
SPECIAL FEATURE