LETTER

pubs.acs.org/ac

Self-Powered Sensor for Naked-Eye Detection of Serum Trypsin

Brian A. Zaccheo and Richard M. Crooks*
Department of Chemistry and Biochemistry, Center for Electrochemistry, and Center for Nano and Molecular Science and Technology,
The University of Texas at Austin, 1 University Station, A5300, Austin, Texas 78712-0165, United States
bS Supporting Information
ABSTRACT: Here, we report a device for the detection of the proteolytic enzyme trypsin,
which is a biomarker for pancreatitis. The sensor is self-powered, easy to use, and signals the
presence of trypsin via a light-emitting diode (LED) that is visible to the unaided eye. Assay
time is ∼3 h, and the limit of detection is 0.5 μg/mL, which is within the range required for
detection of trypsin at levels signaling acute pancreatitis. The sensing mechanism relies on
trypsin digestion of a gelled protein layer. Partial digestion of the protein layer permits
hydroxide penetration and subsequent etching of an underlying Al membrane. Degradation of both the protein and Al layers exposes an
underlying Mg anode and closes an electrochemical circuit that produces ∼2.2 V. This is sufficient voltage to illuminate the LED. A
logarithmic relationship is observed between the time required for LED illumination and trypsin concentration. The device is equally
effective for trypsin dissolved in buffer or serum media.

H ere, we report a self-powered device for detection of trypsin

in serum at pathological concentrations that uses a light-
emitting diode (LED) to deliver a response visible to the unaided
serum trypsin concentration is at a maximum 24 h after the onset of
symptoms, but it is decreased at 48 h and again at 72 h.4 For
diagnostic detection, a system must be sensitive to ∼1 μg/mL
eye. The assay is complete within ∼3 h, and it provides enzyme, capable of distinguishing between small variations in
quantitative results without requiring an external power source. concentration, and sufficiently accessible so as to be implemented
The device, illustrated in Scheme 1, is an Mg//Fe3þ galvanic cell immediately after the appearance of pancreatic symptoms.
with protein and Al passivating layers blocking Mg oxidation at Ionescu et al. have published an indirect electrochemical
the anode. In the presence of trypsin and hydroxide, however, method for trypsin detection using a catalyst-modified Pt disk
these passivating layers are etched, and this action connects the electrode that generates an electroactive species that can be
anode and cathode, thereby illuminating the LED. However, in detected by cyclic voltammetry.5 In this study, electrodes were
the absence of trypsin, the protein passivating layer blocks LED modified with electropolymerized glucose oxidase (GOx) and
illumination. In this article, we show dose-response assays using coated with a layer of gelatin. When immersed in a solution
trypsin in both buffer and serum, and we report a relationship containing trypsin (9380 U/mg) and glucose, the enzyme
between the concentration of trypsin and the LED illumination digested the gelatin layer, and GOx converted glucose to
time. The limit of detection is 0.5 μg/mL, even in the presence of electroactive peroxide.
the blood-borne proteins, sugars, and salts found in serum. The The sensing device described here relies on an LED to signal
novelty of this single-use sensor arises from three factors. First, it the onset of an electrochemical reaction. Previous electrochemi-
is self-powered by the Mg//Fe3þ galvanic cell discussed earlier. cal studies have used LEDs to report the passage of current in a
Second, the presence of trypsin is reported by an LED, and battery6 and as a component in more complex circuits to
therefore, only the unaided eye is required for detection. Third, interrogate and quantify electroactive species under external
the device is simple to operate. potential control.7,8 The LED used here has a minimum turn-
Trypsin is a digestive enzyme produced in the pancreas that on voltage of ∼1.7 V and is self-powered by the two half cells
cleaves proteins on the c-terminal side of arginine and lysine comprising the sensor. Specifically, Mg oxidation and Fe3þ
residues. It is most active under slightly alkaline conditions and reduction yield an observed cell voltage of ∼2.2 V. Self-powered
in the presence of Ca2þ, Mg2þ, and Mn2þ.1 Trypsin is formed sensors have been reported previously. For example, biofuel cells
initially as the proenzyme trypsinogen, which self-cleaves to yield developed by Willner,9 Minteer,10 and their co-workers use
the more active form as needed.2 This self-regulating process can be anodes modified with enzymes or mitochondria to detect sugars
adversely affected by pathologies, such as pancreatitis, which result and nitroaromatics, respectively.
in organ damage and release of enzyme into the blood. An In this article, we present a method for the indirect detection
immunoassay-based, quantitative study found healthy individuals of trypsin in serum. The limit of detection is ∼0.5 μg/mL,
to have a mean serum trypsin concentration of 0.25 ( 0.1 μg/mL, which is sufficiently low for detection of pathological conditions.
whereas acute pancreatitis patients exhibited a higher concentration
of 1.4 ( 0.6 μg/mL.3 Speed is a factor to consider when trypsin was Received: November 26, 2010
used as an indicator of pancreatitis. For example, a time-resolved Accepted: January 11, 2011
enzyme linked immunosorbent assay (ELISA) study found that the Published: January 19, 2011

r 2011 American Chemical Society 1185 dx.doi.org/10.1021/ac103115z | Anal. Chem. 2011, 83, 1185–1188
Analytical Chemistry LETTER

Scheme 1 solutions, it was important to limit bubble formation to avoid

pinholes that could lead to premature etching of the Al foil. After
sonication, the molten protein was aliquoted and stored at 4 °C
until needed. Just prior to running an assay, these aliquots were
placed in a sonicating bath at 60 °C for 2 min, and then, 250 μL of
the solution was added to the anodic well of each device. Cathodic
wells were filled with 50 μL of 0.2 M FeCl3. The devices were then
chilled at 4 °C for 30 min to gel the protein. Devices were
inspected before use, and those having a smooth, bubble-free
protein layer and a nonilluminated LED were retained for assays.
Trypsin Assays. The assays were carried out using 100.0 μL
volumes of buffer or serum with or without added trypsin. The
evaluation of the activity of trypsin is described in the Supporting
Information. Stock trypsin solutions were prepared daily at a
concentration of 10.0 mg/mL in either buffer or serum and
diluted to concentrations ranging from 0.50 to 100 μg/mL.
Buffer trials used trypsin in 46 mM, pH 8.0 TRIS containing
11.5 mM CaAc. For serum trials, trypsin was diluted into 100.0 μL
of equine serum having a native pH of 7.9, and then, an additional
13.0 μL of 0.115 M CaAc was added. Negative control assays used
100.0 μL volumes of buffer or serum, with an additional 13.0 μL of
0.115 M CaAc added to the serum assays. The test solution was
added to the anodic well of each device, which was then placed
in a closed Petri dish inside an air incubator at 26 ( 1 °C and left for
The device provides a logarithmic response that spans more than 2 2 h. During incubation, the incubator maintained a constant
orders of magnitude in trypsin concentration, regardless of whether temperature compatible with the stability of the protein layer and
the analyte is present in buffer or serum. Detection is achieved with the activity of trypsin while ensuring uniform conditions for all
just the unaided eye within 3 h of analyte injection. Finally, the assays. After incubation, an additional 50.0 μL of 0.20 M Fe3þ
device is fully self-powered and, therefore, appropriate for many solution was added to the cathodic well to compensate for solution
point-of-care applications, including those in remote locations. sorbed into the gel during incubation. Devices having an illumi-
nated LED at this stage were discarded. Finally, the Al dissolution
’ EXPERIMENTAL SECTION process was initiated by adding 300 μL of 1.0 M NaOH to the
anodic well. The time between hydroxide addition and LED
Chemicals and Materials. Trypsin from porcine pancreas illumination was recorded.
(Type IX), gelatin, Type A, 300 bloom electrophoresis grade
from porcine skin, heat-inactivated and sterile filtered equine ’ RESULTS AND DISCUSSION
serum, and calcium acetate monohydrate (>99%), henceforth
referred to as CaAc, were purchased from Sigma (St. Louis, MO). Device Overview. This trypsin sensor is designed around an
Genetic-analysis grade agarose and NR-p-tosyl-L-arginine methyl electrochemical cell composed of two half cells in which Fe3þ is
ester (TAME) hydrochloride were obtained from Thermo reduced and Mg is oxidized. Prior to adding the analyte, the two
Fisher Scientific (Waltham, MA). BiPRO whey protein isolate, half cells are prevented from discharging by the presence of Al
henceforth referred to as WPI, was purchased from Davisco and protein barrier layers positioned above the Mg anode
Foods International (LaSueur, MN). Magnesium alloy AZ31B (Scheme 1). However, when trypsin is added to the anodic well,
was obtained in 1 mm-thick, flat sheets from A1 Alloys (San it degrades the protein layer. Subsequent addition of hydroxide
Diego, CA). This alloy is ∼95% Mg and has electrochemical etches the Al passivating layer,11 and this exposes the Mg anode
properties similar to those of pure Mg. Additional materials used to the KCl-impregnated agarose salt bridge.12 Now, the two half
in this study are described in the Supporting Information. cells are in electrochemical contact, and sufficient voltage is
Cell Assembly. Components of the sensing platform are shown produced to illuminate an LED. Hence, the LED signals the
in Scheme 1. The device components are labeled and displaced presence of trypsin. If the added analyte does not contain trypsin,
vertically with respect to order of assembly. Specific details on then the protein layer remains intact and retards transport of
component assembly are provided in the Supporting Information. hydroxide to the Al layer. Therefore, the electrochemical circuit
Gelatin and WPI were selected as the components of the remains open, and LED illumination is prevented. More detailed
protein layer. Protein solutions were prepared in 0.7 M HEPES information about device operation is provided later.
buffer at pH 7.55. The concentrations of the proteins were Dose-Response Assays. Dose-response assays were car-
individually optimized for buffer and serum. For buffer-based ried out for trypsin in buffer and serum. The assay was initially
assays, the protein solution contained 5% gelatin and 10% WPI optimized using trypsin in buffer, and then, the protein gel was
(both w/v). For detection of trypsin in serum, the protein solution reoptimized for detection of trypsin in serum. Specifically, the
contained 6% gelatin and 9% WPI (both w/v). The protein amount of gelatin was increased from 5% to 6%, and WPI was
powders were added to solution in conical tubes and dispersed concomitantly decreased to keep the total amount of protein
using a vortex mixer. The tubes were placed in a sonicating water constant. The additional gelatin used for the serum assays
bath at 60 °C for 15 min to dissolve the proteins and defoam the prevented delamination of the protein layer from the Al, which
solution. During the preparation and handling of the protein was observed only when serum media was used.
1186 dx.doi.org/10.1021/ac103115z |Anal. Chem. 2011, 83, 1185–1188
Analytical Chemistry LETTER

The difference in illumination times arises from the properties

of the two proteins used in the barrier layer. Gelatin gels by heat-
induced (>50 °C)13 denaturation followed by hydrogen-bond-
driven coalescence.14 However, gelatin is unstable in the pre-
sence of hydroxide, so WPI is added to improve its tolerance to
basic conditions. WPI stabilizes the gel under basic conditions,
because it denatures at alkaline pH, thereby exposing buried
cysteine residues that can polymerize via disulfide bonding15-17
to form a rigid, proteinaceous solid.18,19 Once polymerized, a
WPI-containing gel is hydroxide resistant unless it has previously
been treated with trypsin to degrade the gelatin component.20,21
In this latter case, hydroxide moves through the protein layer and
then encounters and degrades the Al barrier. In the absence of
trypsin, the added base polymerizes WPI and rigidifies the gel,
thereby increasing the time required for base attack on the
Figure 1. Histogram correlating the average time between addition of protective Al membrane. The key point is that the rate of
hydroxide and illumination of the LED to the concentration of trypsin hydroxide transport through the gel depends on the concentra-
present in buffer or serum. The negative control was obtained for a tion of trypsin (Figure 1).
trypsin-free sample. The inset is a semilog, least-squares plot correlating
the average illumination time to the concentration of trypsin. Error bars ’ SUMMARY AND CONCLUSIONS
represent one standard deviation from the average of data obtained for
3-5 independently prepared sensors. We have described a self-powered trypsin sensor that is easy to
use, detects trypsin at pathological concentrations in serum, and
signals detection with an LED in ∼3 h. The range of trypsin
Scheme 2
concentrations that can be quantitated includes those that are
characteristic of acute pancreatitis. The discharge of the galvanic
cell is gated by the two-component protein gel. That is, in the
presence of trypsin, the gel is degraded and this results in faster
transport of hydroxide and subsequent corrosion of the under-
lying Al barrier. In the absence of trypsin, however, base
polymerizes the WPI component of the protein layer and
increases the time required for hydroxide to etch the Al barrier.
Another important aspect of this sensor is that it is self-powered.
In essence, the device is a battery having a trypsin-selective switch
that closes the circuit between the anode and cathode. The
sensor works equally well in buffer or serum, demonstrating the
Data from the trypsin assay in buffer and serum are shown in resiliency of this method to the endogenous proteins, sugars, and
Figure 1. Here, the average time between addition of hydroxide salts present in real biological fluids.
and LED illumination time for 3-5 different devices is plotted
against the different trypsin concentrations and a trypsin-free ’ ASSOCIATED CONTENT
negative control. The error bars represent one standard devia-
tion from the mean. It is important to note that even for the
lowest trypsin concentration (0.50 μg/mL) the error bars do
bS Supporting Information. Additional information about
the chemicals and materials, instrumentation, cell assembly, enzyme
not overlap with the negative control assay. The inset in activity assay, and controls. This material is available free of charge via
Figure 1 shows that there is a linear relationship between the the Internet at http://pubs.acs.org.
mean time for LED illumination and the logarithm of the
trypsin concentration. The R2 values for the buffer and serum ’ AUTHOR INFORMATION
fits are 0.96 and 0.99, respectively. Because the inset plot is
concerned with the response of the system to changes in trypsin Corresponding Author
concentration, the trypsin-free, negative control data are not *E-mail: crooks@cm.utexas.edu. Phone: 512-475-8674.
included.
Role of the Protein Gel and Al Barrier Layers. A diagram
illustrating the roles of the protein and Al barrier layers is shown ’ ACKNOWLEDGMENT
in Scheme 2. In panel a, the protein gel layer is exposed to a We gratefully acknowledge the U.S. Department of Energy,
solution containing trypsin. Trypsin partially digests the protein Office of Basic Energy Sciences (Contract No. DE-FG02-
layer, which renders it more permeable to hydroxide than the 06ER15758) for support of this work. We also thank the Robert
native gel. After trypsin digestion for 2 h, hydroxide is introduced A. Welch Foundation (Grant F-0032) for sustained support.
into the anodic well. Panel b shows that hydroxide penetrates the
digested protein layer and partially dissolves the Al layer to ’ REFERENCES
expose the Mg strip. This closes the electrical circuit and results (1) Jeohn, G.-H.; Serizawa, S.; Iwamatsu, A.; Takahashi, K. J. Biol.
in LED illumination. However, in the absence of trypsin diges- Chem. 1995, 270, 14748–14755.
tion (panel c), migration of hydroxide through the gel layer is (2) Schomberg, D., Schomberg, I., Eds. In Springer Handbook of
slowed and, therefore, LED illumination is delayed. Enzymes, 2nd ed.; Springer-Verlag: New York, 2002; pp 11-24.

1187 dx.doi.org/10.1021/ac103115z |Anal. Chem. 2011, 83, 1185–1188

Analytical Chemistry LETTER

(3) Artigas, J. M.; Garcia, M. E.; Faure, M. R.; Gimeno, A. M.

Postgrad. Med. J. 1981, 57, 219–222.
(4) Regner, S.; Manjer, J.; Appelros, S.; Hjalmarsson, C.; Sadic, J.;
Borgstrom, A. Pancreatology 2008, 8, 600–607.
(5) Ionescu, R. E.; Cosnier, S.; Marks, R. S. Anal. Chem. 2006, 78,
6327–6331.
(6) Eggen, P. O.; Gronneberg, T.; Kvittingen, L. J. Chem. Educ. 2006,
83, 1201–1203.
(7) Faulkner, L. R.; Walsh, M. R.; Xu, C. J. . Contemporary Electro-
analytical Chemistry; Ivaska, A., Lewenstam, A., Sara, R. , Eds.; Plenum
Press Div Plenum Publishing Corp: New York, 1990.
(8) Sun, L.; Crooks, R. M. J. Electrochem. Soc. 2005, 152, E371–E377.
(9) Katz, E.; B€uckmann, A. F.; Willner, I. J. Am. Chem. Soc. 2001, 123,
10752–10753.
(10) Germain, M. N.; Arechederra, R. L.; Minteer, S. D. J. Am. Chem.
Soc. 2008, 130, 15272–15273.
(11) Streicher, M. A. Ind. Eng. Chem. 1949, 41, 818–819.
(12) Howell, B. A.; Cobb, V. S.; Haaksma, R. A. J. Chem. Educ. 1983,
60, 273.
(13) Johnston-Banks, F. A. . In Food Gels; Harris, P., Ed.; Elsevier
Applied Science: New York, 1990; p 233-289.
(14) Finer, E. G.; Franks, F.; Phillips, M. C.; Suggett, A. Biopolymers
1975, 14, 1995–2005.
(15) Foegeding, E. A.; Davis, J. P.; Doucet, D.; McGuffey, M. K.
Trends Food Sci. Tech. 2002, 13, 151–159.
(16) Ruben, M.-P.; William, R. P.; Wilson, D. I. Int. J. Food Sci.
Technol. 2008, 43, 1379–1386.
(17) Mercade-Prieto, R.; Paterson, W. R.; Dong Chen, X.; Ian
Wilson, D. Chem. Eng. Sci. 2008, 63, 2763–2772.
(18) Bottomley, R. C.; Evans, M. T. A.; Parkinson, C. J. . In Food
Gels; Harris, P., Ed.; Elsevier Applied Science: New York, 1990; p
435-466.
(19) Veith, P. D.; Reynolds, E. C. J. Dairy Sci. 2004, 87, 831–840.
(20) Mercade-Prieto, R.; Chen, X. D. AIChE J. 2006, 52, 792–803.
(21) Mercade-Prieto, R.; Falconer, R. J.; Paterson, W. R.; Wilson,
D. I. Biomacromolecules 2007, 8, 469–476.

1188 dx.doi.org/10.1021/ac103115z |Anal. Chem. 2011, 83, 1185–1188