Download Low Cost Nanoribbon Sensors for Protein Analysis in Human Serum

Low Cost Nanoribbon Sensors for Protein Analysis in Human Serum
Using a Miniature Bead-Based Enzyme-Linked Immunosorbent
Assay
Chunxiao Hu,† Ioannis Zeimpekis,† Kai Sun†, Sally Anderson,‡ Peter Ashburn,† and Hywel Morgan*,†
†
Department of Electronics and Computer Science, and Institute for Life Sciences, University of Southampton, United
Kingdom
‡
Sharp Laboratories of Europe, Oxford, United Kingdom
ABSTRACT: We describe a low cost thin-film transistor (TFT) nanoribbon sensor for detection of the inflammatory biomarker Creactive protein (CRP) in human serum via a miniature bead-based Enzyme-Linked Immunosorbent Assay (ELISA). The TFT sensor
measures the reaction products from the ELISA via pH changes. The bead-based ELISA decouples the protein functionalization steps
from the sensor surface, increasing the signal and simplifying the assay. The ability to directly sense proteins in human serum in this
way overcomes the Debye length limitation associated with nanowire and nanoribbon biosensors. Compared to classically fabricated
nanowires, the TFT nanoribbons are simple, extremely easy to fabricate and should therefore be much cheaper to manufacture. TFT
nanoribbons, configured as pH sensors were used for quantitative detection of CRP spiked into human serum at concentrations as
low as 1 ng/mL, which is ten thousand times lower than needed for diagnostic purposes, providing the potential for applications that
require very high sensitivity.
Quantification and analysis of biological processes,
especially determination of low concentrations of proteins, is
important in healthcare.1 The most widely used assay is the
Enzyme Linked Immunosorbent Assay or ELISA2 which uses
an enzyme to amplify the biochemical binding event. Surface
Plasmon Resonance (SPR)3 sensors on the other hand are label
free but usually require expensive optical equipment for
detection. Recently, Field Effect Transistor (FET) devices such
as silicon nanowire and nanoribbon biosensors have been
developed for direct, high sensitivity, label-free sensing of
biomolecules. These devices are small and could be integrated
into point of care (PoC) systems.4-8 Nanowire/nanoribbon
devices configured as ion-sensitive FETs (ISFETs) have been
used for a large variety of applications, including pH sensing,4,912
DNA sensing,13-19 protein sensing,4,7,8,16,20-23 and enzyme
detection.24 However, the technology has not yet evolved into a
robust platform for analysis of samples in human body fluids
such as blood due to several limitations. One is the limitation
on suspending medium composition imposed by the screening
effect of the electrical double layer. Label free FET mediated
biosensing of biomolecules is based on the fact that the charge
of a molecule bound to the surface of the transistor gate changes
the surface potential, thus varying the carrier density inside the
channel. In physiological media, the Debye length is around 0.7
nm; any charge from a bound molecule beyond this distance
will be effectively screened by the medium. Several methods
have been reported to overcome this limitation. Generally the
approach is to perform measurements in a low salt
concentration (e.g. 1. mM) buffer, thereby increasing the Debye
screening length.8,25,26 Other methods include the use of short
molecular receptors to reduce the distance between the surface
and the analyte,27,28 or modification of the sensor surface with
PEG polymers to increase the effective screening length.29
Recently, these devices have been used to measure proteins
through a modified ELISA where the electronic readout is via
pH change. .30,31 However, as for the label-free assays, , the
receptor molecules were linked directly to the sensor surface,
for example immobilized to the gold leads near the nanowire
gate or directly onto the nanoribbon itself. These approaches
limit the flexibility of the sensor and also the sensitivity and
modify the kinetics. It also complicates manufacturing since
each sensor has to be specifically functionalized with a different
capture moiety.
Here we report a more flexible approach for the detection of
C-reactive protein (CRP) in human serum using a miniature
bead-based ELISA with pH readout coupled with a low cost
TFT nanoribbon sensor for readout. Recently a magnetic beads
based assay was demonstrated in a microfluidic chip with insitu washing. Negatively charged single stranded DNA was
used as a label to modify the charge density and the signal read
out when the beads were attracted to the surface of a metal oxide
using a magnet. However, the use of miniature beads coupled
with an enzyme amplified pH readout provides a highly flexible
assay format that is simple and uses disposable electronics to
directly read out the signal. The use of beads as carriers also
makes the system scalable. For example large area very low cost
TFT arrays containing hundreds or thousands of sensors could
be used to measure multiple analytes simultaneously. CRP is
one of the most important acute-phase proteins that exists in
blood and is synthesized by the liver.32 It is widely used as a
general biomarker for inflammation and infection, and has
become a good indicator for evaluating risks of cardiovascular
diseases.33,34 Generally, the CRP concentration in blood is less
than 10 mg/L for healthy individuals, but can increase 1000 fold
during an acute phase of inflammation.35 Fast detection of CRP
has been demonstrated using FET biosensors with pH readout,36
but the proteins were immobilized onto a gold sensor surface
and the lowest concentration of CRP that could be detected was
29 ng/ml. In this paper we have modified a conventional beadbased ELISA to produce a pH change that is read using the
nanoribbon. Since molecular detection is measured indirectly
through an enzyme-substrate reaction, the Debye screening
effect is eliminated and both specificity and sensitivity are
increased. Furthermore, the assays are implemented using
magnetic beads on the substrate rather than the sensor surface,
which increases the capture cross section and improves the
reaction rate. It also provides significant advantages in terms of
manufacturing biosensors since the sensors do not need to be
chemically functionalized – the chemistry is entirely confined
to the carrier beads which can be introduced into the sensor
prior to the assay. The urea-urease reaction was first used as a
model system to evaluate the sensitivity of the TFT nanoribbon
sensors for enzyme-substrate detection. The same enzyme was
then used to detect CRP in human serum using a miniature
bead-based ELISA. The urease-induced pH change was
monitored to determine the quantity of protein in the serum. In
fact, any bead-based ELISA could be read out in this way.
Compared to conventional nanowire fabrication techniques,
e.g. top-down15,37,38 or bottom–up4,39,40 approaches, the
polysilicon thin-film transistor biosensor was fabricated with a
simple three-mask approach,41 which produces a technology
that is simple, cheap and suitable for the mass manufacture of
disposable biosensors for PoC applications.
EXPERIMENTAL
Device Fabrication and measurement configuration. A
three-mask fabrication method41 was used to manufacture the
TFT nanoribbon sensors shown in Fig. 1. Briefly the fabrication
process involved defining the polysilicon transistor using
photolithography, , contacting Source and Drain using metal
tracks followed by deposition and patterning of the thick resist
SU8 to create a sensing window. Fabrication details have been
described previously.41 The nanoribbon transistor has two gates.
The top gate is a liquid gate controlled by applying a potential
to an integrated on-chip Ag/AgCl electrode (Fig. 1a). The
bottom gate is the back of the substrate, which is grounded for
all the measurements described in this paper. Silicon dioxide is
used as the gate dielectric, which is in contact with the liquid.
An SU8 layer insulates all the contacts and defines the sensing
windows (Fig. 1b) where the nanoribbon transistors are in
contact with test solution which is confined with a small well
made from PDMS. A dual-channel picoammeter/voltage source
(Model 6482, Keithley) and a connection jig (Sharp
Laboratories of Europe) were used to apply voltage and acquire
data.
(a)
Vlg
Analyte
Solution
Ag/
AgCl
Ids
Vds
(b)
PDMS
SU-8
SiO2
TiN
Poly-silicon
Buried Oxide
Substrate
On-chip Ag/AgCl
electrode
SU-8 window
100 µm
Figure 1. (a) Schematic diagram of the nanoribbon sensor showing
the small droplet of sample (not to scale); (b) Microscope image of
one sensor, containing 30 nanoribbons (40 µm long) connected in
parallel with the integrated reference electrode.
Enzyme reaction measurement. A simple enzymesubstrate reaction was demonstrated using urea (Sigma) and
urease (Sigma). Five different concentrations of urea were
prepared in a measurement buffer (0.01×PBS + 150 mM
NaCl).30 Prior to each measurement, the entire sensing area
(including the silver/silver chloride reference electrode) was
incubated in 3% bovine serum albumin (BSA) in PBS to
minimize non-specific binding. Urea solution (100 μL at
various concentrations) was pipetted into the sensing window
and the source-drain current monitored against time (sourcedrain voltage Vds = 0.1 V; Liquid gate voltage Vlg = 0.05 V) to
provide a baseline signal. Subsequently urease (50 μL at 0.45
mg/mL) was added and the solution was quickly mixed by
pipetting up and down several times. Urease catalyzes the
hydrolysis of urea to ammonia and carbon dioxide leading to an
increase in the pH as the reaction proceeds. This is read as a
decrease in the source-drain current (for n-type nanoribbon
sensors).
Bead functionalization and measurement. A bead based
electronic ELISA was implement using magnetic beads that
were modified sequentially as shown in Fig. 2a. Two different
assays were performed. In one, the entire chemistry was
performed in a test tube “off-chip” and the products of th
reaction measured with the TFT transistor. In the other, the
entire reaction was performed in a 3L volume “on-chip”
Commercial superparamagnetic beads (3 μm in diameter) with
CRP capture antibody were purchased from R&D systems.
These beads were re-suspended in PBS to a concentration of
5000 beads/μL. For the off-chip assay, an aliquot (10 μL) of
these beads was mixed with a solution of CRP (R&D systems)
(10 μL) in the well of a microplate. The CRP stock was prepared
by spiking protein into human serum (Sigma) to give a
concentration of 10 mg/L CRP. This is the normal concentration
in the blood of a healthy person. This was diluted with PBS to
produce the desired concentrations for the assay. The bead
suspension was incubated at room temperature for 30 minutes
on a horizontal orbital microplate shaker set at 800±50rpm. The
beads were then washed three times in PBS by concentrating
the beads with a magnet and aspirating the supernatant.
Biotinylated CRP detection antibody solution (10 μL, 0.1
mg/ml in PBS; R&D systems) was added to the well and
incubated at room temperature for 10 minutes with shaking,
followed by three washes. Streptavidin (10 μL, 0.1 mg/ml in
PBS; R&D systems) was then added to the well and incubated
at room temperature for another 10 minutes on the shaker,
followed by three washes. Finally, biotinylated urease (10 μL,
0.2 mg/mL; Insight Biotechnology) was added and the solution
incubated at room temperature for 10 minutes followed by
further bead washing. The beads were then re-suspended in the
working buffer (0.01×PBS + 150 mM NaCl) to a volume of 3
μL and transferred to the sensor chip.
Nanoribbon
2
(a)
Biotinylated urease
Streptavidin
Biotinylated Anti-CRP
Detection
C-reactive protein
incubated for 30minutes, followed by a stop solution (50 μL;
Thermo Fisher Scientific). Since the magnetic beads could
affect the optical density of the fluid, the solution was
transferred to another well leaving the beads in place. The
optical density (O.D.) of each well was measured by a
microtiter plate reader with a filter set to 450 nm.
Anti-CRP Capture
Urea
H+
Figure 2. Schematic diagram of the magnetic bead sandwich assay.
(a) Magnetic beads with CRP capture antibody were sequentially
incubated with CRP at different concentrations, followed by
biotinylated CRP detection antibody, streptavidin, and biotinylated
urease. (b) Urea was added to initiate the reaction. The change of
pH is related to the concentration of CRP and was detected by the
nanoribbon sensor.
After adding the magnetic bead suspension, the source-drain
current signal (source-drain voltage Vds = 0.1 V; liquid gate
voltage Vlg = 0.05 V) was measured as a function of time. Then
urea solution (3 µL, 10 µM diluted with working buffer) was
added and rapidly mixed, and the pH change measured from the
Source-Drain current.
The same assay was also carried out with the entire reaction
implemented in the sensor well “on-chip”. In this case, beads
were suspended in PBS to a concentration of 5000 beads/μL and
a 3-μL aliquot added to the sensor well. A solution of CRP in
human serum (3 μL) was added by pipetting and the mixture
mixed by successive pipetting and incubated at room
temperature for 30 minutes. The beads were then washed three
times in PBS by concentrating the beads with magnet and
aspirating the supernatant. Biotinylated CRP detection antibody
solution (3 μL, 0.3 mg/ml in PBS) was added to the well, mixed
and incubated at room temperature for 10 minutes, followed by
three washes. Streptavidin (3 μL, 0.3 mg/ml in PBS) was then
added to the well, mixed and incubated at room temperature for
another 10 minutes, followed by three washes. Finally,
biotinylated urease (3 μL, 0.6 mg/mL) was added, mixed and
incubated at room temperature for 10 minutes followed by
washing. The bead suspension was then re-suspended by adding
the working buffer to a volume of 3 μL.
A conventional colorimetric bead based ELISA was also
carried out. In this case beads were suspended in PBS to a
concentration of 5000 beads/μL. An aliquot (100 μL) of these
beads was mixed with a 100 μL solution of CRP in human
serum in the well of a microplate. The mixture was incubated at
room temperature for 2 hours on a horizontal orbital microplate
shaker. Beads were washed with PBS before adding
biotinylated CRP detection antibody solution (100 μL, 0.1
mg/ml in PBS), followed by incubation and washing. Protein
concentration was assayed by adding Streptavidin-HRP (100
μL, 0.1 mg/ml in PBS; Thermo Fisher Scientific), followed by
incubation followed by three washes. TMB substrate solution
(100 μL; Thermo Fisher Scientific) was then added and
Electrical characteristics. The pH sensitivity of
nanoribbons with a silicon dioxide dielectric surface was first
characterized from IdsVlg sweeps with calibrated pH buffers. As
shown in Figure 3a, four different pH buffers (pH = 9.0, 7.0, 5.0
and 3.0 respectively) were used to characterize the electrical
response of the devices by sweeping the liquid gate voltage and
recording the source-drain current (Vds = 0.1 V). The
nanoribbon is n-type and therefore the source-drain current Ids
increases with decreasing pH. The transconductance g m =
∂Ids/∂Vlg is used to characterize individual nanoribbons, and this
value was extracted from the linear region of the IdsVlg curve (0.1 to 0.1 V in this case) and found to be nearly constant at
~2×10-7 A/V for all four pH values. The threshold voltage shift
ΔVth = Ids/gm calculated from the IdsVlg curves is shown in Fig.
3b. The figure shows that for liquid gate voltages ranging from
-0.1 to 0.1 V, ΔVth is a linear function of pH with a slope of ~
33 mV/pH in line with literature values for SiO2 surfaces.42
(a) 10
Ids (A)
(b)
RESULTS AND DISCUSSION
-6
10-7
pH9
pH7
pH5
pH3
Vds = 0.1 V
gm = 2.01×10-7 A/V
10-8
0
(b)
1
2
Vlg (V)
1.6
-100 mV
-60 mV
-20 mV
0 mV
20 mV
60 mV
100 mV
1.5
Vlg =
1.4
∆Vth (V)
Magnetic bead
1.3
1.2
1.1
Vds = 0.1 V
Slope = 33.5±1.9 mV/pH
3
4
5
6
7
8
9
pH
Figure 3. (a) Nanoribbon source-drain current (Ids) vs liquid gate
voltage measured for different pH buffer solutions. The
transconductance gm=∂Ids/∂Vlg is extracted from the linear region.
(b) Threshold voltage change ΔVth for 7 different liquid gate
voltages. The voltage is a linear function of pH with a subNernstian slope of 33 mV/pH.
3
1
6.8
7.1
0 mM Urea
2 mM Urea
5 mM Urea
10 mM Urea
20 mM Urea
30 mM Urea
-20
-30
7.4
7.7
-40
pH
-10
8.0
Substrate
8.3
-50
(1)
Enzyme
-60
Urease (at 0.45 mg/mL) was pipetted into a well around the
nanoribbons, and the change in pH was measured upon the
addition of urea at six different concentrations (0 to 30 mM).
Figure 4a shows the recorded current when a small volume of
urea solution (100 μL) was added (and mixed) to the sensing
window (point 1 on graph). As shown, the current remains
stable with time. After 400 s urease (50μL) was added and the
electrical current monitored (for typically 10 minutes). The
transconductance gm obtained during the electrical
characterization of a single device was used to normalize the
response from different devices. The data shows that the sourcedrain current decreases as the enzyme reaction proceeds and the
solution becomes more basic. The reaction rate was determined
from the initial slope of the reaction (dashed square in Fig. 4a)
and fitted to the Michaelis–Menten equation (Fig. 4b):
6.5
2
0
H+
8.6
8.9
-70
0
200
400
600
800
1000
Time (s)
(b)
Reaction Rate (M/min)
(𝑁𝐻2 )2 𝐶𝑂 + 𝐻2 𝑂 → 𝐶𝑂2 + 2𝑁𝐻3
(a) 10
(Id-Id0)/gm (mV)
Enzyme reactions. A simple enzyme-substrate reaction was
demonstrated using urea-urease. Urea serves an important role
in metabolic processes and is the main nitrogen-containing
substance in the urine of mammals. Its concentration is an
important indicator of some diseases for example heart and
renal failure.43,44 Urease catalyzes the hydrolysis of urea into
one carbon dioxide and two ammonia molecules increasing the
pH:
5
4
3
2
pH meter
Nanoribbon
1
0
0
5
10
15
20
25
30
Concentration (mM)
𝑣=
𝑉max [𝑆]
𝐾𝑀 +[𝑆]
(2)
𝑣 is the reaction rate (current or threshold voltage change
with time), Vmax is the maximum rate and KM, is the Michaelis
constant.
Figure 4. Current vs time for the urea-urease reaction measured
with the n-type nanoribbon field effect transistor. (a) Source-drain
current normalized against transconductance (left y-axis) for
different urea concentration with device operating in sub-threshold.
Number 1 indicates the point at which urea was added to the sensor,
and 2 when urease was added. Calculated equivalent pH changes
are shown on the right y-axis. (b) Michaelis–Menten curves
obtained from change in current during the first few minutes
(dashed square in Figure 4a). The same reaction was also carried
out in a test tube and the rate measured using a pH meter. Data are
the mean+SEM (three measurements with two devices).
For comparison, the urea/urease reaction was also monitored
in a larger volume using a standard pH meter. Data were
recorded every 5 seconds for the first 5 minutes and every 30
seconds for the following 10 minutes. The Michaelis constants
(KM) were calculated and found to be 15.9 ± 2.1 mM
(nanoribbon) and 9.2 ± 0.7 mM (pH meter) respectively. This
can be compared with the literature value of 25 mM45 (Table 1).
The KM determined using nanoribbon is slightly higher than
measured conventionally, this may be due to differences in
mixing or temperature between the two systems. The turnover
number (kcat) was also determined from
𝑘𝑐𝑎𝑡 =
𝑉𝑚𝑎𝑥
[𝐸]0
(3)
where [E] 0 is the enzyme concentration. Turnover numbers
were 3.7×104 s-1 and 3.4×104 s-1 for the nanoribbon and pH
meter respectively, which compares with literature values of
around 104 s-1.46,47 Table 1 summarizes the data. These
4
measurements indicate that the TFT nanoribbons can be used to
reliably measure enzyme-substrate reaction rates.
Bead-based ELISA. In a bead-based ELISA, the capture
antibodies are immobilized on the surface of small beads, rather
than the plastic surfaces of microtiter plates. Using beads rather
than a planar surface significantly increases the surface area
available for capture, and also makes mixing easier thereby
speeding up the reaction. These assays require relatively small
sample volumes and demonstrate increased sensitivity.48 For
this assay, magnetic beads functionalized with CRP capture
antibodies were used as capture surfaces. The assay was read
out by adding urea to the bead suspension, leading to an
increase in the pH of the solution that is proportional to the
amount of captured CRP. Seven concentrations of CRP spiked
into human serum were measured, ranging from 0 to 500 ng/ml.
The same reaction was performed in two different ways. Either
“off-chip” in a small tube or entirely “on chip”. In this case the
beads were washed on chip by concentrating them with a small
hand held magnet (see experimental section). The enzyme
reaction was measured through the change in current as a
function of time and the data is shown in Figure 5. The sourcedrain current was continuously measured and the bead
suspension added at point 1. Immediately after adding urea to
the bead suspension at point 2, a sharp decrease in current was
observed (Fig. 5a) caused by agitation of the liquid when the
suspension was mixed. The signal stabilized and the current
decreased after around 30 seconds. This time point was chosen
as the beginning of the reaction. The change in threshold
voltage is plotted in the figure and the slope of the curve over
the first minute (marked by a red dashed square) was used to
calculate the reaction rate (using Matlab). A plot of initial
reaction rate against CRP concentration is shown in Figure 5b.
The end point current after 500 seconds was also measured.
In total, eight sets of measurements were performed on the
nanoribbon sensors, four with the entire assay performed onchip (red curve) and four with the end product of the reaction
measured on chip (black curve). The data was fitted to the Hill
equation (Fig. 5b) to obtain the binding constant Kd. This was
determined for each separate experiment and the mean binding
constant determined as 43 ± 10 ng/ml for off-chip and 38 ± 13
ng/ml for on-chip. Figure 5c shows a similar standard curve
extracted from the end point voltage change instead of the initial
rates. This gave slightly lower binding constants of 30 ± 11
ng/ml for off-chip and 29 ± 10 ng/ml for on-chip assay. To
verify these binding constants, a conventional colorimetric endpoint bead based ELISA was performed (Fig 5d), (n = 3). This
gave a binding constant Kd = 43± 6 ng/ml, which is similar to
that measured on chip.
Table 1. Comparison of KM and kcat values for the urea-urease reaction determined with the nanoribbon sensor (three
measurements with two devices) and also a pH meter in bulk.
Nanoribbon
pH Meter
Literature
KM (mM)
15.88 ± 2.09
9.18 ± 0.69
25 45
R2
0.9937
0.9984
-
3.724×104
3.356×104
3×104 46
kcat
(s-1)
5
Figure 5. Detection of C-reactive protein using bead-based ELISA on nanoribbon sensors. (a) Change in threshold voltage for different
concentrations of CRP (with on-chip chemistry). The change in source-drain current is measured and converted to potential shift by dividing
by the transconductance gm. The equivalent change in pH is shown on the right y-axis. Number 1 indicates the time point when the beads are
pipetted onto the sensor surface, and point 2 when urea was added. (b) Plot of the initial reaction rate (determined within the dashed square
in Figure 5a) vs concentration of protein. Red curve is data for on-chip assay; black curve for off-chip assay (c) Plot of the final current (after
500 sec) against protein concentration. The plots were fitted to the Hill equation. Data are the mean ± SD (four measurements with two
devices). (d) Standard curve measured from the endpoint of a conventional colorimetric bead based ELISA. Optical density measured with
a plate reader.
The results from on-chip or off-chip chemistry are very
similar and indicate that the concentration of CRP in human
serum can be measured at a concentration as low as 0.2 ng/ml
in a volume of only 3 µL. The assay time for the on-chip
chemistry protocol is approximately 1 hour. This could be
reduced with further optimization of the fluidics and
development of magnetic bead agitation protocols for in-situ
continuous mixing. The method presented here is applicable to
the detection of many other proteins in high salinity buffers
such as human serum (where optimized bead based ELISAs are
available). Scaling up the system could provide a simple but
cost-effective approach for analyzing large numbers
(thousands) of different proteins using arrays of thin film
transistors.
CONCLUSION AND OUTLOOK
Low cost thin-film transistor nanoribbon sensors with
integrated reference electrodes have been fabricated and used to
analyze enzyme-substrate reactions via pH changes. The device
was tested using the urease-urea reaction to characterise
enzyme properties. An assay for the quantitative detection of
the inflammatory biomarker C-reactive protein in human serum
has been developed using a miniature bead-based ELISA with
pH readout. The protein assay can be performed using the TFT
nanoribbons in high ionic strength buffer, unlike conventional
label-free approaches that require low-ionic strength and a large
Debye length. The urea-urease system demonstrates that
enzyme kinetics can be reliably analyzed using the sensor. A
magnetic bead-based ELISA was used to detect CRP in human
serum at concentrations down to 0.2 ng/mL in a volume of 3
µL. The entire assay, including functionalization, mixing and
detection was performed on the sensor in under an hour.
Compared with current analytical methods for CRP
quantification, no expensive detection equipment is required
and the volume of the test sample is very small. An important
advantage of this assay is that all the functionalization steps are
6
performed on the magnetic beads, not on the sensor surface.
This simplifies manufacturing and assay development and also
allows the device to be reused. Many different bead-based
ELISA are commercially available, therefore minimal surface
chemistry development or assay optimization is required in
transferring these to the device Further optimization should lead
to smaller assay volumes, speeding up the reaction and enabling
the detection of very low amounts of proteins in small volumes
of serum. This electronic ELISA brings low-cost nanoribbon
sensor technology a step closer to a point of care diagnostic
system.
AUTHOR INFORMATION
Corresponding Author
* Email: [email protected]
ACKNOWLEDGMENT
The authors would like to acknowledge the Technology Strategy
Board (TSB) and the Engineering and Physical Sciences Research
Council (EPSRC: EP/K502327/1) for funding this work. We would
also like to thank Gregory Gay, Ben Hadwen,Chris J. Brown and
Jonathan Buse of Sharp Laboratories Europe for many useful
discussions and development of the jig etc to go with the Keithley.
REFERENCES
(1) Chen, K. I.; Li, B. R.; Chen, Y. T. Nano Today 2011, 6, 131-154.
(2) Engvall, E.; Perlmann, P. Immunochemistry 1971, 8, 871-874.
(3) Zeng, S.; Baillargeat, D.; Ho, H. P.; Yong, K. T. Chem Soc Rev
2014, 43, 3426-3452.
(4) Cui, Y.; Wei, Q.; Park, H.; Lieber, C. M. Science 2001, 293,
1289-1292.
(5) Tarasov, A.; Wipf, M.; Stoop, R. L.; Bedner, K.; Fu, W. Y.;
Guzenko, V. A.; Knopfmacher, O.; Calame, M.; Schonenberger, C.
Acs Nano 2012, 6, 9291-9298.
(6) Rajan, N. K.; Duan, X. X.; Reed, M. A. Wires Nanomed Nanobi
2013, 5, 629-645.
(7) Elfstrom, N.; Karlstrom, A. E.; Linnros, J. Nano Lett 2008, 8,
945-949.
(8) Hakim, M. M. A.; Lombardini, M.; Sun, K.; Giustiniano, F.;
Roach, P. L.; Davies, D. E.; Howarth, P. H.; de Planque, M. R. R.;
Morgan, H.; Ashburn, P. Nano Lett 2012, 12, 1868-1872.
(9) Knopfmacher, O.; Tarasov, A.; Fu, W. Y.; Wipf, M.; Niesen, B.;
Calame, M.; Schonenberger, C. Nano Lett 2010, 10, 2268-2274.
(10) Chen, S. Y.; Bomer, J. G.; Carlen, E. T.; van den Berg, A. Nano
Lett 2011, 11, 2334-2341.
(11) Peretz-Soroka, H.; Pevzner, A.; Davidi, G.; Naddaka, V.;
Tirosh, R.; Flaxer, E.; Patolsky, F. Nano Lett 2013, 13, 3157-3168.
(12) Zeimpekis, I.; Sun, K.; Hu, C. X.; Thomas, O.; de Planque, M.
R. R.; Chong, H. M. H.; Morgan, H.; Ashburn, P. Nanoscale Res Lett
2015, 10.
(13) Gao, A. R.; Lu, N.; Wang, Y. C.; Dai, P. F.; Li, T.; Gao, X. L.;
Wang, Y. L.; Fan, C. H. Nano Lett 2012, 12, 5262-5268.
(14) Hahm, J.; Lieber, C. M. Nano Lett 2004, 4, 51-54.
(15) Li, Z.; Chen, Y.; Li, X.; Kamins, T. I.; Nauka, K.; Williams, R.
S. Nano Lett 2004, 4, 245-247.
(16) Bunimovich, Y. L.; Shin, Y. S.; Yeo, W. S.; Amori, M.; Kwong,
G.; Heath, J. R. J Am Chem Soc 2006, 128, 16323-16331.
(17) Zhang, G. J.; Zhang, G.; Chua, J. H.; Chee, R. E.; Wong, E. H.;
Agarwal, A.; Buddharaju, K. D.; Singh, N.; Gao, Z. Q.;
Balasubramanian, N. Nano Lett 2008, 8, 1066-1070.
(18) Xie, P.; Xiong, Q.; Fang, Y.; Qing, Q.; Lieber, C. M. Nat
Nanotechnol 2012, 7, 119-125.
(19) Hao, Y. Q.; Zhou, B. B.; Wang, F. B.; Li, J.; Deng, L.; Liu, Y.
N. Biosensors & Bioelectronics 2014, 52, 422-426.
(20) Zheng, G. F.; Lieber, C. M. Methods Mol Biol 2011, 790, 223237.
(21) Stern, E.; Klemic, J. F.; Routenberg, D. A.; Wyrembak, P. N.;
Turner-Evans, D. B.; Hamilton, A. D.; LaVan, D. A.; Fahmy, T. M.;
Reed, M. A. Nature 2007, 445, 519-522.
(22) Kumar, M. A.; Jung, S.; Ji, T. Sensors-Basel 2011, 11, 50875111.
(23) Yu, R. M.; Pan, C. F.; Wang, Z. L. Energ Environ Sci 2013, 6,
494-499.
(24) Mu, L. Y.; Droujinine, I. A.; Rajan, N. K.; Sawtelle, S. D.;
Reed, M. A. Nano Lett 2014, 14, 5315-5322.
(25) Stern, E.; Wagner, R.; Sigworth, F. J.; Breaker, R.; Fahmy, T.
M.; Reed, M. A. Nano Lett 2007, 7, 3405-3409.
(26) Duan, X. X.; Li, Y.; Rajan, N. K.; Routenberg, D. A.; Modis,
Y.; Reed, M. A. Nat Nanotechnol 2012, 7, 401-407.
(27) Elnathan, R.; Kwiat, M.; Pevzner, A.; Engel, Y.; Burstein, L.;
Khatchtourints, A.; Lichtenstein, A.; Kantaev, R.; Patolsky, F. Nano
Lett 2012, 12, 5245-5254.
(28) Maehashi, K.; Katsura, T.; Kerman, K.; Takamura, Y.;
Matsumoto, K.; Tamiya, E. Anal Chem 2007, 79, 782-787.
(29) Gao, N.; Zhou, W.; Jiang, X. C.; Hong, G. S.; Fu, T. M.; Lieber,
C. M. Nano Lett 2015, 15, 2143-2148.
(30) Stern, E.; Vacic, A.; Li, C.; Ishikawa, F. N.; Zhou, C. W.; Reed,
M. A.; Fahmy, T. M. Small 2010, 6, 232-238.
(31) Aroonyadet, N.; Wang, X. L.; Song, Y.; Chen, H. T.; Cote, R.
J.; Thompson, M. E.; Datar, R. H.; Zhou, C. W. Nano Lett 2015, 15,
1943-1951.
(32) Thompson, D.; Pepys, M. B.; Wood, S. P. Struct Fold Des 1999,
7, 169-177.
(33) Miller, V. M.; Redfield, M. M.; McConnell, J. P. Curr Vasc
Pharmacol 2007, 5, 15-25.
(34) Mygind, N. D.; Harutyunyan, M. J.; Mathiasen, A. B.; Ripa, R.
S.; Thune, J. J.; Gotze, J. P.; Johansen, J. S.; Kastrup, J.; Grp, C. T.
Inflamm Res 2011, 60, 281-287.
(35) Clyne, B.; Olshaker, J. S. J Emerg Med 1999, 17, 1019-1025.
(36) Lin, C. L.; Kang, Y. W.; Chang, K. W.; Chang, W. H.; Wang,
Y. L.; Lee, G. B. IEEE 27th International Conference on Micro Electro
Mechanical Systems (Mems) 2014, 250-253.
(37) Lin, H. C.; Lee, M. H.; Su, C. J.; Huang, T. Y.; Lee, C. C.;
Yang, Y. S. IEEE Electr Device L 2005, 26, 643-645.
(38) Wenga, G.; Jacques, E.; Salaun, A. C.; Rogel, R.; Pichon, L.;
Geneste, F. Biosensors & Bioelectronics 2013, 40, 141-146.
(39) Whang, D.; Jin, S.; Wu, Y.; Lieber, C. M. Nano Lett 2003, 3,
1255-1259.
(40) Zheng, G. F.; Patolsky, F.; Cui, Y.; Wang, W. U.; Lieber, C. M.
Nature biotechnology 2005, 23, 1294-1301.
(41) Sun, K.; Zeimpekis, I.; Lombardini, M.; Ditshego, N. M. J.;
Pearce, S. J.; Kiang, K. S.; Thomas, O.; de Planque, M. R. R.; Chong,
H. M. H.; Morgan, H.; Ashburn, P. IEEE T Electron Dev 2014, 61,
2170-2176.
(42) Prodromakis, T.; Liu, Y.; Constandinou, T.; Georgiou, P.;
Toumazou, C. IEEE Electr Device L 2010, 31, 1053-1055.
(43) Aronson, D.; Mittlernan, M. A.; Burger, A. J. Am J Med 2004,
116, 466-473.
(44) Carvounis, C. P.; Nisar, S.; Guro-Razuman, S. Kidney Int 2002,
62, 2223-2229.
(45) In University Science Books; University Science Books, 2014,
p 372.
(46) Baker, R. R.; Murray, R. K. In PDQ series; BC Decker:
Hamilton, Ont., 2001, pp ix, 310 p.
(47) Burton, J. J.; Garten, R. L. Advanced materials in catalysis;
Academic Press: New York, 1977, p xii, 329 p.
(48) Scholler, N.; Crawford, M.; Sato, A.; Drescher, C. W.; O'Briant,
K. C.; Kiviat, N.; Anderson, G. L.; Urban, N. Clinical cancer research
2006, 12, 2117-2124.
7