A DNA-mediated homogeneous binding assay for proteins and small molecules

Article abstract of DOI:10.1021/ja505519b

Optical detection of molecular targets typically requires immobilization, separation, or chemical or enzymatic processing. An important exception is aptamers that allow optical detection in solution based on conformational changes. This method, however, requires the laborious selection of aptamers with high target specificity and affinity, and the ability to undergo the required conformational changes. Here we report on an alternative generic scheme for detecting small molecules and proteins in solution based on a shift in the equilibrium of DNA-based strand displacement competition reaction. The shift occurs upon binding of a protein, for example, an antibody to its target. We demonstrate nanomolar detection of small molecules such as biotin, digoxigenin, vitamin D, and folate, in buffer and in plasma. The method is flexible, and we also show nanomolar detection of the respective antibodies or protein targets of these molecules. The detection scheme provides a generic alternative to aptamers for detection of analytes.

Full text of DOI:10.1021/ja505519b

Article
pubs.acs.org/JACS
A DNA-Mediated Homogeneous Binding Assay for Proteins and
Small Molecules
Zhao Zhang,† Christian Hejesen, Michael B. Kjelstrup, Victoria Birkedal, and Kurt V. Gothelf*
Center for DNA Nanotechnology (CDNA) at the Interdisciplinary Nanoscience Center (iNANO), Aarhus University, DK-8000
Aarhus C, Denmark
S
* Supporting Information
ABSTRACT: Optical detection of molecular targets typically
requires immobilization, separation, or chemical or enzymatic
processing. An important exception is aptamers that allow
optical detection in solution based on conformational changes.
This method, however, requires the laborious selection of
aptamers with high target specificity and affinity, and the ability to undergo the required conformational changes. Here we report
on an alternative generic scheme for detecting small molecules and proteins in solution based on a shift in the equilibrium of
DNA-based strand displacement competition reaction. The shift occurs upon binding of a protein, for example, an antibody to its
target. We demonstrate nanomolar detection of small molecules such as biotin, digoxigenin, vitamin D, and folate, in buffer and
in plasma. The method is flexible, and we also show nanomolar detection of the respective antibodies or protein targets of these
molecules. The detection scheme provides a generic alternative to aptamers for detection of analytes.
INTRODUCTION
■
Detection of analytes in binding assays such as enzyme-linked
immunosorbent assay (ELISA) relies on immobilization of the
binding moiety, e.g., an antibody at a surface.1 Homogeneous
binding assays on the other hand can be performed in a single
tube containing the specimen and all reagents, obviating any
immobilization, separation, or washing steps, thus minimizing
the chances of contamination.2 Most of these assays are either
based on aptamers,3,4 or are limited to DNA-binding
proteins.5,6 Other assays such as quenchbodies,7,8 binding-
induced annealing9−13 or proximity ligation,14−16 often involve
the events of multivalent binding or the procedure of DNA−
protein conjugation. The method of fluorescence polarization is
less general since it requires a fluorescent ligand.17
Here we report on a method for detection of small molecule
or protein targets in homogeneous solution, which is based on
Figure 1. Tm change induced by ligand−protein binding on DNA
duplex. (Left) Melting curves of a 18 bp DNA duplex with one biotin
the binding of a protein to a small molecule ligand conjugated
to one of the strands in a DNA duplex. We have found that
upon binding of streptavidin (STV) the melting temperature
(Tm) of the duplex decreases. As shown in Figure 1, the
melting temperature of a duplex containing biotin attached to
one of the bases (ds-1b) decreases around 10 °C after binding
STV, which is even more than what is caused by a mismatch
(Supporting Information Figure S1). Moreover, the Tm will
decrease further when STV binds to two biotins on one strand,
in which case part of the strand might tend to wrap around the
protein via bivalent binding (Figure 1 and Figure S1).
Recently, Zhang et al. reported on a method to optimize the
specificity in nucleic acid hybridization and to detect
mismatches.18 In this method, two DNA strands compete
about the hybridization to a third strand which has toeholds19
at each terminus unique for each of the competing strand. The
introduction of a mismatch in one of the competing strands in
(black), one biotin and one streptavidin (STV) (red), or two biotins
and one STV (blue). (Upper right) Illustration of each configuration.
(Lower right) Tm derived from the curves.
the toehold exchange reaction results in a significant change of
the equilibrium compared to the fully matched case.20
By implementing the ligand−protein induced changes of Tm
in toehold exchange probes, we developed the strand
displacement competition (SDC) assay for detection of
proteins shown in Figure 2. Without target protein, the three
strands, A, B modified with a ligand, and S, will reach a dynamic
equilibrium which mainly depends on the Tm difference
between AS and BS18 (see Supporting Information Figure S2).
If a protein target (T) binds to the small molecule conjugated
Received: June 2, 2014
Published: July 22, 2014
© 2014 American Chemical Society
11115
dx.doi.org/10.1021/ja505519b | J. Am. Chem. Soc. 2014, 136, 11115−11120

Journal of the American Chemical Society
Article
and S or their equivalents in equal stoichiometric ratio (70 μL, 20
nM), in 1× TAE/Mg2+ buffer pH 8. The target protein or control
(usually water) was also added at the same time. Then the whole
mixture was incubated at room temperature (RT) overnight in the
dark. Since this method is based on thermodynamics instead of
kinetics, the order of addition of different components is irrelevant to
the final readout.
Construction of the SDC Assay for Small Molecule
Detection. As illustrated in Supporting Information Figure S12,
two strategies can be adopted for this purpose. In the inhibition
strategy, the specimen with or without analyte molecules was premixed
with the corresponding protein first. After incubation at room
temperature (RT) for 3 h, the mixture was added into the standard
SDC assay and incubated at RT overnight in the dark. In the
replacement strategy, the protein and assay are mixed first and left at
RT overnight. The specimen is then added and the mixture incubated
at RT for another 5 h.
FRET Experiments. Fluorescence measurements were performed
using a scanning spectrofluorometer (Fluoro-Max-3, HORIBA Jobin
Yvon Inc.). The assay-specimen mixture was pipetted into a quartz
cuvette (65 μL), which was washed by 1× TAE/Mg2+ three times
between different samples. All spectral measurements were recorded
with 0.5 s integration time and 1 nm wavelength interval, at 25 °C.
For the FRET pair of Alexa555 and Alexa647 as in most assays,
excitation was performed at 550 nm for Alexa555. Relative FRET
efficiencies were calculated as Er = IDA/(IDA + IDD), where IDA is the
acceptor peak fluorescence intensity at 667 nm and IDD is the donor
peak fluorescence intensity at 567 nm.
For the three-dye system, excitation was performed at 450 nm for
Alexa488, 550 nm for Alexa555, and 600 nm for Alexa647. Relative
FRET efficiency of the pair of Alexa488 and Alexa555 was calculated as
Er = ID1A1/(ID1A1 + ID1D1), where ID1A1 is the acceptor peak
fluorescence intensity at 567 nm and ID1D1 is the donor peak
fluorescence intensity at 519 nm. Relative FRET efficiency of the pair
of Alexa555 and Alexa647 was calculated as Er = IA2A2/ID2A2, where
IA2A2 is the acceptor peak fluorescence intensity at 667 nm when
Alexa555 is excited, and ID2A2 is the acceptor peak fluorescence
intensity at 667 nm when Alexa647 is excited.
Figure 2. Detection of biomolecules based on the strand displacement
competition assay. (a) Schematic illustration of the strand displace-
ment competition reaction. Both strand A and strand B are
complementary to a part of strand S, and they are mixed in a 1:1:1
ratio. In a typical pathway, A (or B) will displace B (or A) from S,
through the process of toehold binding, branch migration, and toehold
dissociation. This reaction keeps proceeding at both directions; thus, a
dynamic equilibrium is reached after certain amount of time. (b) In the
presence of target protein (T), which can bind to the ligand modified
on strand B to form B(T), the final equilibrium will be shifted toward
AS since the Tm of B(T)S is lower after binding.
to the B strand, the equilibrium shifts due to decrease of the
Tm of B(T)S, thus more AS will be formed. The population
change can be detected by PAGE or by an optical signal such as
Forster resonance energy transfer (FRET) if dyes are
̈
conjugated to the DNA strands. This binding-induced blocking
mechanism also relates to Plaxco et al.’s electrochemical
sensor,21 however, in that assay immobilization is a require-
ment.
EXPERIMENTAL SECTION
■
Materials. All the modified and unmodified oligonucleotides
except two were purchased from DNA Technology A/S (Denmark)
or Sigma-Aldrich (St. Louis, MO). HPLC purification was performed
by the supplier directly after synthesis. The DNA strand with a shorter
linker for the amino modifier (B4a2−2) was synthesized in-house on a
Mermade-12 DNA synthesizer (BioAutomation, TA), using the 5-
aminoallyl-dU phosphoramidite (Berry & Associates, Dexter, MI)
according to instructions. The DNA strand with a 2′-amino
modification was bought from IBA (Germany) and purified by
HPLC. The concentration of each oligonucleotide was measured by
Nanodrop 1000 spectrophotometer (Thermo Fisher Scientific,
Cambridge, MA).
For the real-time FRET measurement, the pair of Alexa488 and
Alexa555 was used. Fluorescence was recorded at 519 nm (Alexa488)
and 567 nm (Alexa555) with an excitation at 450 nm, at a time interval
of 60 s. Relative FRET efficiencies were calculated as Er = IDA/(IDA
+
IDD), where IDA is the acceptor peak fluorescence intensity at 567 nm
and IDD is the donor peak fluorescence intensity at 519 nm.
Nondenaturing Polyacrylamide Gel Electrophoresis (PAGE).
Mixtures of sample (0.16 pmol for each strand) and 6× nondenaturing
loading buffer (0.25% bromophenol blue +30% glycerol) were loaded
on native PAGE gel (acrylamide/bisacrylamide; 19:1, 6%), which was
run in a fridge using a PowerPac 1000 power supply (Bio-Rad) (70 V,
1.5 h). TBE/Mg2+ (1×; 12.5 mM) was used for both gel and running
buffer.
Typhoon Scanning. After electrophoresis, the gel was scanned by
Typhoon scanner (Amersham Biosciences) without staining. Red laser
(633 nm) for excitation and 670 nm band-pass filter for emission were
chosen. The scanned gels were analyzed by ImageQuant TL 7.0 (GE
Healthcare). Lane creation and band detection were done manually.
Since Alexa647 was mostly modified on strand A, the intensity of the
bands only represent the amount of A. Normally there are two
detectable bands: the upper band is A in the form of AS duplex, while
the lower band is single-stranded A (ssA). We use the value of ssA/
(ssA + AS) as a robust indicator for evaluation of the population
change in solution.
Alexa Fluor Succinimidyl Esters (647, 555 and 488) were purchased
from Invitrogen (Carlsbad, CA). Polyclonal anti-digoxigenin from
sheep was purchased from Roche Applied Science (Germany).
Adenosine 5′-triphosphate (ATP) was purchased from New England
Biolabs (Ipswich, MA). Monoclonal anti-folic acid antibody from
mouse was purchased from Acris Antibodies (Germany). Folic Acid-
N3 was purchased from Baseclick (Germany). All the other reagents
were purchased from Sigma-Aldrich.
UV Absorbance Melting Studies. Absorbance vs melting
temperature curves were recorded at 260 nm on a Varian Cary 100
spectrophotometer (Palo Alto, CA), in thermal mode. All melting
curves were recorded in 1× TAE buffer (40 mM Tris, 1 mM EDTA,
pH 8) containing 12.5 mM MgCl2, at a standard concentration of 1
μM. A ramp rate of 0.5 °C/min was used and the block underwent two
repeated thermal cycles from 25 to 70 °C then back to 25 °C. Data of
the second cycle from 25 to 70 °C was collected and analyzed by using
derivative calculations. All melting temperatures were recorded in
duplicate and they varied by less than 1 °C in all cases.
RESULTS
■
Detection of Proteins: Streptavidin. We first tested this
strategy for detection using the strong biotin-STV interaction as
a model. As shown in Figure 3a, strands A, B, and S were mixed
in equal stoichiometry, among which, A and S each are labeled
Construction of the SDC Assay for Protein Detection. 0.5 mL
BILATEC tubes (Sigma-Aldrich) were used for all the reactions. A
typical SDC assay was prepared by simply mixing DNA strand A, B
11116
dx.doi.org/10.1021/ja505519b | J. Am. Chem. Soc. 2014, 136, 11115−11120

Journal of the American Chemical Society
Article
The position of the ligand on B is important for the effect of
target binding and it was observed that the optimal position is
in the branch migration region (Supporting Information Figure
S4). By making use of the tetravalency of STV, we also
conjugated two biotins on one strand (Supporting Information
Figure S5), which results in even lower Tm of BS after bivalent
binding to STV (Figure 1). Both FRET and gel results showed
that in this two-biotin system, a higher amount of AS was
formed than in the one-biotin system in the presence of STV,
and only negligible single-stranded A was left (Figure 3c and
Supporting Information Figure S4).
The effect of target binding depends on the stoichiometry of
A, B and S. However, we found that other ratios than 1:1:1 can
also ensure significant effect of STV, although the absolute
FRET value is different (Supporting Information Figure S6).
Furthermore, it is worth noting that after normalization with
reference samples containing only AS or BS, the FRET signal
change became even more significant (Supporting Information
Figure S7).
The sensor specificity was tested by adding a variety of other
proteins, which should not interact with biotin to the biotin
assay, and none of them displayed any significant effect on the
equilibrium (Supporting Information Figure S8). When adding
various proteins into a control system without ligand on the
DNA strands, no FRET signal change was observed, validating
the robustness of the system (Supporting Information Figure
S9).
To make sure the SDC reaction reaches equilibrium, unless
otherwise specified, all the reactions in this study were left to
proceed at RT overnight, although the time-dependent FRET
measurement revealed that the FRET signal has already halfway
changed in the first half an hour (Figure 3d). Elongating the
toehold length on both A and B is expected to increase the rate
of the equilibrium,22 but our study also revealed that longer
toeholds lead to decreased FRET signal (Supporting
Information Figure S10).
The FRET signals of a 7 nM assay in response to STV of
increasing concentrations within the range of 0 to 2 nM are
shown in Figure 3e. The detection limit of this assay relies on
the sensitivity of the fluorometer, as well as the rate constant
which is dependent on concentration.22 We have demonstrated
that by PAGE it was possible to detect as low as 125 pM
streptavidin (Supporting Information Figure S11).
Small Molecule Detection: Biotin. The SDC assay can be
extended to detection of small molecules similar to the ligand
conjugated to the B strand, using an inhibition or a replacement
method (Supporting Information Figure S12). In the former
detection mode, a known concentration of the protein is
incubated with a sample containing an unknown concentration
of the ligand and this mixture is applied to the SDC assay. This
method was applied to the STV-biotin assay for the detection
of biotin. In the absence of biotin as target molecule, STV will
exert the same effect on the equilibrium as before. If the sample
contained biotin, it will block the corresponding binding-sites
on STV, thereby decreasing the number of sites that can bind
to the biotin-DNA conjugate (strand B) in the assay
(Supporting Information Figure S13). Figure 3f shows the
measured FRET efficiency using 14 nM of DNA, 20 nM of
STV, and a series of concentrations of biotin. The curve reflects
the tetravalency of streptavidin as the dynamic range is in the
interval of 60−80 nM biotin, where the last binding site of STV
is occupied. In the alternative Replacement method (Supporting
Information Figure S12), the sample containing the small
Figure 3. Detection of streptavidin (STV) and biotin as a model
system. (a) Schematic illustration of the SDC assay based on biotin-
STV interaction. (b) FRET results of STV detection. In the presence
of excessive STV, FRET signal increased more than 1-fold, indicating
more AS are formed. (c) Gel scanned on a Typhoon scanner. Left two
lanes are the system shown in (a), while right two lanes are a two-
biotin system shown in Supporting Information Figure S5. The band
intensity reflects the amount of A, since only the dye of Alexa647 on A
is excited and emitted. Upper and lower band represent AS and ssA,
respectively. The effect of STV was evaluated and compared using the
percentage of ssA in total A. (d) Time-dependent FRET for STV
detection. (e) Titration curve of STV detection ([A] = [B] = [S] = 7
nM). All the data are fit by a hyperbolic function. The same below. (f)
Titration curve of biotin detection by the inhibition method ([A] =
[B] = [S] = 14 nM, [STV] = 20 nM).
with fluorophores forming a FRET pair, while B is conjugated
to biotin which serves as the binding moiety for STV detection.
All the modifications are located on the branch migration
region. A and B endow 3 nt and 4 nt toeholds, respectively;
therefore, without STV, BS has slightly lower standard free
energy than AS. Upon binding of STV, the melting temperature
of the BS duplex decreases, resulting in a shift of the
equilibrium toward the AS duplex. This event can be observed
by both FRET (Figure 3b) and PAGE (Figure 3c). Here the
FRET signal is linked to the population of AS, while the gel
shows the ratio of A in single-stranded (ss) and double-
stranded (ds) form in the solution. It is observed that STV
causes the emergence of more AS and less ssA, which is
consistent with the scheme. The amount of BS is confirmed to
decrease in a three-dye system where strand B also has a
fluorophore (Supporting Information Figure S3).
11117
dx.doi.org/10.1021/ja505519b | J. Am. Chem. Soc. 2014, 136, 11115−11120

Journal of the American Chemical Society
Article
Figure 4. Detection of digoxigenin, vitamin D, folate and their respective protein/antibody binders. (a) Schematic illustration of the SDC assay. The
functional group that is used for conjugation to DNA is indicated with a green arrow. (b) Gel characterization of the effect of aD, DBP and aFA. The
intensity of bands represents the amount of strand A. More AS duplex and less ssA emerged after protein binding. (c−e) FRET results for titration of
aD and DIG, VD and DBP, and aFA and FA. Small molecules were detected using the inhibition method. (f and g) Bar plots comparing the
detection of aD and DIG, and aFA and FA in buffered samples and samples containing 14% human plasma.
molecule can be added directly to the SDC assay containing the
protein to replace the small molecules conjugated to DNA
bound to the protein with the small molecule target in solution.
The replacement method requires a reversible exchange of
ligands bound to the protein and cannot be applied to the
biotin-STV assay since the off rate is to low. Therefore, only the
inhibition method was tested for detection and quantification of
biotin.
Detection of 3 Other Proteins/Antibodies and Their
Small Molecule Targets. After having obtained proof of
concept for the SDC assay for detection of the very strong
biotin-STV interaction, we have extended it to detect other
small molecule-protein interactions. Three pairs of interactions
were examined: digoxigenin (DIG)/antidigoxigenin (aD),
vitamin D3 (VD)/vitamin D binding protein (DBP), and folate
(FA)/antifolate (aFA) (Figure 4). The former pair was chosen
as model for antigen−antibody binding, while both vitamin D
and folate (Vitamin B9) levels in the body are crucial to human
health and corresponding deficiency can cause serious
diseases.23−25
The three molecules DIG, VD and FA were conjugated to
amino-modified strand B (Figure 4a). However, when using the
standard C6 linker between the modified DNA base and the
small molecules that was used for biotin no changes in
equilibriums were detected when the corresponding binding
proteins were added (Supporting Information Figure S14). To
increase the impact of the interaction on the DNA strand, a dT
amino modifier with a shorter C3 linker was employed for this
purpose,26 and the protein binding events were confirmed by
PAGE (Supporting Information Figure S15). The feasibility of
another linker where the amine group is directly modified on
the C2′ of the deoxyribose was also confirmed (Supporting
Information Figure S16).
The changes in equilibrium caused by the addition of the
corresponding proteins/antibodies were analyzed by PAGE as
shown (Figure 4b). The changes in equilibrium are less
pronounced than the biotin streptavidin interaction, however
still significant, in particularly for the binding of DBP to VD.
We believe the smaller change in FRET is caused by a smaller
change in Tm upon protein binding. It was not possible to
obtain Tm curves for DIG and DBP due to denaturation of the
proteins; however, for the aFA/FA couple, we observed a
change in Tm of 3 °C in the presence of excess of the protein.
This corresponds well with the smaller change in FRET for the
aFA/FA couple. The FRET response to titration of increasing
concentrations of aD, DBP and aFA are shown in Figure 4c−e.
For these assays the concentration of the ABS strands were 20
nM and it was observed that all three titration curves have the
dynamic ranges from approximately 1 to 20 nM. The specificity
of each assay was confirmed by the absence of equilibrium
shifts upon adding other proteins to the assays (Supporting
Information Figure S17). With the inhibition method, the small
molecules DIG, VD and FA can also be detected by premixing
the small molecules and the corresponding protein followed by
addition to the SDC assay (Figure 4c-e). As it appears from the
titration curves, the small molecules can be detected by FRET,
and the assays have dynamic ranges of approximately 10−100
nM using the current setup. The replacement method by which
the specimen is added directly into the assay mixture including
proteins can also be adopted for detection of the small
molecules (Supporting Information Figure S18). The free
11118
dx.doi.org/10.1021/ja505519b | J. Am. Chem. Soc. 2014, 136, 11115−11120

Journal of the American Chemical Society
Article
ligands will compete with the ligands on strand B to bind with
proteins, thus have the opposite effect on the equilibrium.
Human Plasma and Saliva. Finally, the SDC assay was
tested for detection of the 6 analytes in 14% human plasma.
First it was attempted to examine the VD/DBP assay. However,
since the albumins in plasma bind to Vitamin D the equilibrium
shifted significantly in plasma and as a result neither VD nor
DBP could be detected. On the other hand, the detection of aD
and DIG and aFA and FA was feasible in plasma. For these
experiments, 14% plasma spiked with the target compounds
was used. Due to the numerous different constituents of human
plasma, PAGE was not suitable for monitoring the equilibrium
in plasma. The equilibrium changes were, however, clearly
detectable by FRET. As shown in Figure 4f,g, the addition of
plasma to the ABS system has only minor impact on the
equilibrium. The shift in FRET observed after addition of 2
equiv of aD is almost similar for the systems in plasma and in
buffer. For the detection of digoxigenin, addition of 10 equiv of
DIG to the ABS system and subsequent addition of aD shows
almost similar results in buffer and in plasma. Detection of aD
and DIG was also demonstrated in 84% plasma and in 14%
saliva (Supporting Information Figure S19).
binding interactions between proteins (antibody) and small
molecule can be applied, although it does require that the small
molecule can be linked to the oligonucleotide without
significantly decreasing the binding affinity to the protein. In
most cases, the conformational change of aptamers required for
providing an optical readout from a homogeneous solution s
another challenge and often requires substantial structural
analysis and structural engineering of selected aptamers. The
SDC system is, on the other hand, predesigned to make the
shift of equilibrium upon binding of the protein to the small
molecule ligand and the FRET couple is thereby completely
separated. Furthermore, the same sequences can be used for all
targets in the SDC assay. Adjustment of the initial equilibrium
can easily be achieved by adding or deleting bases in the
toehold region of sequence A or B.
The sensitivity of the SDC method is dependent on (i) the
affinity of the protein to the small molecule−DNA conjugate,
(ii) the concentration of the DNA sequences in the assay, (iii)
the emission from the FRET couple at the given concentration,
and (iv) the equilibrium shift. At the current state of
development of the assay. the main limiting factor is detection
of the emission from the FRET couple at low concentration of
the DNA assay. In the data presented here, the detection limit
for proteins/antibodies is in the low nanomolar range by FRET
and down to 125 pM by PAGE. In several cases, DNA and
RNA aptamers that bind to proteins with affinities in the
picomolar range have been selected.31 The SDC may, however,
have potential to reach similar low affinities for protein binding
by optimizing the FRET dyes. For small molecule , aptamers
with nanomolar sensitivity have been selected, however, in
most cases the affinities are micromolar.32 The examples shown
here by the SDC method demonstrate sensitivity in the
nanomolar regime in all cases, and as for the protein detection,
we also believe that the sensitivity of small molecule binding
can be further optimized.
DISCUSSION
■
The method appears to be general for detecting biomolecular
interactions involving small molecules and proteins as shown
for the couples of biotin/STV, DIG/aD, VD/DBP and FA/
aFA. However, the optimal length of the tether between the
DNA B sequence and the small molecule target seems to be
crucial and differs from protein to protein. This is probably
caused by different local configurations of the ligand−protein
interactions. We believe that the decrease of the Tm of the
binding duplex is mainly caused by steric repulsion between the
protein and the binding site in the major groove of the DNA
helix. Meanwhile, electrostatic interactions may also be
important for the interaction. In both cases, the more proximal
the protein is to the DNA components upon binding to the
ligand, the larger effect it will potentially have on the
equilibrium. The biotin-binding pocket in streptavidin is buried
relatively deep inside the protein,27 and therefore, the effect is
obtained with the relatively long C6 linker between the DNA
base and biotin. For the antibody/small molecule interactions,
the binding site is generally located at the surface of the
antibody,28 and thus, a shorter linker between the DNA B
strand is required to obtain a sufficient repulsion to shift the
equilibrium in the SDC assay. We believe this is also the case
for the VD/DBP couple.29
CONCLUSION
■
We introduced a method to detect small molecule analytes and
proteins, e.g., antibodies in homogeneous solutions with high
specificity. By the direct binding of a protein target to a small
molecule in the SDC assay, quantitative detection of nM
concentrations of proteins can be obtained by fluorescence or
PAGE. Small molecules can also be detected by incubation with
a known concentration of the respective protein, thus
outcompeting the binding to the identical DNA linked small
molecule. The competitive nature of the assay makes the
equilibrium highly invariable to changes in the environment
such as concentration, salts, temperature, etc.18 As an example,
we showed that the SDC assay could be applied for detection of
antibody and its small molecule hapten in plasma or saliva. In
perspective, we believe that the method can be extended to
other small molecules and protein couples and the detection
scheme has potential for application in clinics for quantification
of disease biomarkers, in human plasma.
The position of the small molecule ligand in the DNA B
sequence also appeared to be crucial for the equilibrium shift
induced upon binding of the protein as shown in Supporting
Information Figure S4. A ligand placed in the branch migration
region will shift the sensor equilibrium more significantly than
when positioned in the toehold region, while a ligand placed at
the terminal of the strand has almost no effect. This is in
accordance with the relation reported for the position of single
point mismatches in short DNA duplexes and their Tm.30 The
maximum decrease of Tm was observed when the mismatch
was at the center of the sequence.
The SDC method allows detection of small molecule -
protein interaction in homogeneous solution without the need
of immobilization or separation procedures and in that regard
the method compares to aptamers. The major advantage of the
SDC method is that no selection is required since existing
ASSOCIATED CONTENT
■
S
* Supporting Information
DNA sequences, supplementary schemes (S1−S3) and addi-
tional results (Figures S1−S19). This material is available free
of charge via the Internet at http://pubs.acs.org.
11119
dx.doi.org/10.1021/ja505519b | J. Am. Chem. Soc. 2014, 136, 11115−11120

Journal of the American Chemical Society
Article
Pantothenic Acid, Biotin, and Choline; National Academy Press.:
Washington, D.C., 1998; p 196.
(26) Lermer, L.; Roupioz, Y.; Ting, R.; Perrin, D. M. J. Am. Chem.
AUTHOR INFORMATION
Corresponding Author
kvg@chem.au.dk
■
Soc. 2002, 124, 9960.
Present Address
(27) Hendrickson, W. A.; Pahler, A.; Smith, J. L.; Satow, Y.; Merritt,
E. A.; Phizackerley, R. P. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 2190.
(28) Al-Lazikani, B.; Lesk, A. M.; Chothia, C. J. Mol. Biol. 1997, 273,
927.
(29) Verboven, C.; Rabijns, A.; De Maeyer, M.; Van Baelen, H.;
Bouillon, R.; De Ranter, C. Nat. Struct. Biol. 2002, 9, 131.
(30) Guo, Z.; Liu, Q.; Smith, L. M. Nat. Biotechnol. 1997, 15, 331.
(31) Hermann, T.; Patel, D. J. Science 2000, 287, 820.
(32) McKeague, M.; Derosa, M. C. J. Nucleic Acids 2012, 2012, 748.
†Nanobiology Institute, Yale University, West Haven, CT
06516, USA.
Notes
The authors declare the following competing financial
interest(s): The authors of the manuscript have filed a patent
application covering the main concepts of the work presented
here. The PCT application WO2014041024 was published on
May 20, 2014 under the title: Detection of non-nucleic acid
analytes using strand displacement exchange reactions.
ACKNOWLEDGMENTS
■
This research was supported by grants from the Danish
National Research Foundation to Center for DNA Nano-
technology (DNRF81) and the Danish Council for Independ-
ent Research for a Sapere Aude Grant (V.B.) and an Eliteforsk
Award (K.V.G.).
REFERENCES
(1) Lequin, R. M. Clin Chem. 2005, 51, 2415.
(2) Zhang, H.; Li, F.; Dever, B.; Li, X. F.; Le, X. C. Chem. Rev. 2013,
■
113, 2812.
(3) Nutiu, R.; Li, Y. J. Am. Chem. Soc. 2003, 125, 4771.
(4) Sassolas, A.; Blum, L. J.; Leca-Bouvier, B. D. Analyst 2011, 136,
257.
(5) Wang, J.; Li, T.; Guo, X.; Lu, Z. Nucleic Acids Res. 2005, 33, e23.
(6) Leung, C. H.; Chan, D. S.; He, H. Z.; Cheng, Z.; Yang, H.; Ma, D.
L. Nucleic Acids Res. 2012, 40, 941.
(7) Abe, R.; Ohashi, H.; Iijima, I.; Ihara, M.; Takagi, H.; Hohsaka, T.;
Ueda, H. J. Am. Chem. Soc. 2011, 133, 17386.
(8) Abe, R.; Jeong, H. J.; Arakawa, D.; Dong, J.; Ohashi, H.; Kaigome,
R.; Saiki, F.; Yamane, K.; Takagi, H.; Ueda, H. Sci. Rep. 2014, 4, 4640.
(9) Heyduk, E.; Dummit, B.; Chang, Y. H.; Heyduk, T. Anal. Chem.
2008, 80, 5152.
(10) Lundberg, M.; Eriksson, A.; Tran, B.; Assarsson, E.; Fredriksson,
S. Nucleic Acids Res. 2011, 39, e102.
(11) Li, F.; Zhang, H.; Wang, Z.; Li, X.; Li, X. F.; Le, X. C. J. Am.
Chem. Soc. 2013, 135, 2443.
(12) Li, F.; Zhang, H.; Lai, C.; Li, X. F.; Le, X. C. Angew. Chem., Int.
Ed. 2012, 51, 9317.
(13) Zhang, H.; Li, F.; Dever, B.; Wang, C.; Li, X. F.; Le, X. C. Angew.
Chem., Int. Ed. Engl. 2013, 52, 10698.
(14) Fredriksson, S.; Gullberg, M.; Jarvius, J.; Olsson, C.; Pietras, K.;
Gustafsdottir, S. M.; Ostman, A.; Landegren, U. Nat. Biotechnol. 2002,
20, 473.
(15) Fredriksson, S.; Dixon, W.; Ji, H.; Koong, A. C.; Mindrinos, M.;
Davis, R. W. Nat. Methods 2007, 4, 327.
(16) Gustafsdottir, S. M.; Schlingemann, J.; Rada-Iglesias, A.;
Schallmeiner, E.; Kamali-Moghaddam, M.; Wadelius, C.; Landegren,
U. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 3067.
(17) Rossi, A. M.; Taylor, C. W. Nat. Protoc. 2011, 6, 365.
(18) Zhang, D. Y.; Chen, S. X.; Yin, P. Nat. Chem. 2012, 4, 208.
(19) Yurke, B.; Turberfield, A. J.; Mills, A. P., Jr.; Simmel, F. C.;
Neumann, J. L. Nature 2000, 406, 605.
(20) Werntges, H.; Steger, G.; Riesner, D.; Fritz, H. J. Nucleic Acids
Res. 1986, 14, 3773.
(21) Cash, K. J.; Ricci, F.; Plaxco, K. W. J. Am. Chem. Soc. 2009, 131,
6955.
(22) Zhang, D. Y.; Winfree, E. J. Am. Chem. Soc. 2009, 131, 17303.
(23) DeLuca, H. F. Am. J. Clin. Nutr. 2004, 80, 1689S.
(24) Holick, M. F. N. Engl. J. Med. 2007, 357, 266.
(25) Board., N. A. o. S. I. o. M. F. a. N. In Dietary Reference Intakes for
Thiamin, Riboflavin, Niacin, Vitamin B6, Folate, Vitamin B12,
11120
dx.doi.org/10.1021/ja505519b | J. Am. Chem. Soc. 2014, 136, 11115−11120