Sensors and Actuators B 269 (2018) 8–14

Contents lists available at ScienceDirect

Sensors and Actuators B: Chemical

journal homepage: www.elsevier.com/locate/snb

Continuous protease assays using liquid crystal as a reporter

Mahbuba Jannat, Kun-Lin Yang ∗
Department of Chemical and Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117576, Singapore

a r t i c l e i n f o a b s t r a c t

Article history: Abnormal protease activities are associated with cancers, vascular diseases and Alzheimer diseases.
Received 6 October 2017 Therefore, detection of protease activities has become increasingly important in recent years. Herein,
Received in revised form 2 March 2018 we report a semi-quantitative, liquid crystal (LC)-based protease assay for naked-eye detection of pro-
Accepted 22 April 2018
tease activity. In this assay, casein molecules are cleaved by proteases into small peptide fragments, which
Available online 24 April 2018
can be quantified either by using Lowry’s method or using LC. In the latter, peptide fragments adsorb on a
solid surface and disrupt LC to produce a bright spot for naked-eye detection. The bright spot is observed
Keywords:
only when the surface-adsorbed peptide density exceeds a critical value. In the assay, a major challenge
Continuous assay
Protease
is how to separate undigested casein and remaining protease from peptide fragments to prevent their
Liquid crystal interferences with LC. To overcome this issue, trichloroacetic acid (TCA) is added to precipitate casein
Trichloro acetic acid and proteases. By using the assay, we are able to detect 10 ng/mL of protease (activity 7.23 U/mg pro-
Syringe pump tease, R2 = 0.991) or 6.5 ␮g/mL of peptide fragments. Finally, a continuous protease assay is developed to
minimize manual sampling and reduce errors in kinetic studies.
© 2018 Elsevier B.V. All rights reserved.

1. Introduction
Proteases are enzymes which are able to catalyze the hydroly-
sis of peptide bonds. They play important roles in food digestion,
healing of wounds and blood coagulation [1–3]. Proteases attract
much attention in recent years because they are associated with
cancers [4,5], vascular diseases [6] and Alzheimer diseases [7].
Therefore, development of facile yet sensitive protease assays for
the detection of protease activities is highly important. Tradi-
tionally, protease activities can be determined after incubation
of proteases with their substrates, such as proteins or peptides.
After the incubation period, peptide fragments released from
the substrates (due to protease activity) are detected by using
analytical instrumentation such as high performance liquid chro-
matography coupled with mass spectroscopy (HPLC–MS), liquid
nitrogen assisted spray ionization mass spectrometry (LNASI-MS),
matrix-assisted laser desorption/ionization time-of-flight mass
spectrometry (MALDI-TOF-MS) and surface assisted laser desorp-
tion/ionization time-of-flight mass spectrometry (SALDI-TOF-MS)
[8–11]. Based on peptide mass fingerprints and peak heights, one
can determine protease cleavage sites and protease activities. This
method is accurate, label-free and highly specific, but it relies on
expensive instrumentation and is time-consuming [12,13].
∗ Corresponding author.
E-mail address: cheyk@nus.edu.sg (K.-L. Yang).
Recently, many simple and portable protease assays have been
developed [14–20]. These assays can be classified as specific and
non-specific assays depending on the substrates used. For specific
protease assays, custom-synthesized peptides with a well-defined
sequence are used as a substrate for a particular protease. On
the other hands, non-specific protease assays use low-cost pro-
teins such as casein as a substrate, which can be digested by
proteases into peptide fragments [21]. After precipitation of undi-
gested casein and remaining proteases, the peptide fragments can
be quantified by using Lowry’s method with Folin–Ciocalteu (FC)
reagent, which reacts with tyrosine residue in the peptide frag-
ments and gives a blue color. Although this method is simple and
quantitative, the limit of detection (LOD) is only 0.1 U/mL [22].
To address this issue, many researchers used labeled caseins as
substrates to lower LODs. For example, Manicourt and Lefebvre
used [3 H] and [125 I]-labeled casein to detect matrix metallopro-
teinases (MMPs) and plasmin with LODs of 2.5 ng/mL and 5 ␮g/mL,
respectively [23,24]. However, labeling of casein with radioisotopes
is tedious, time-consuming, expensive and requires special han-
dling of radioactive materials. Alternatively, casein can be labeled
with fluorescent dyes for fluorescence detection. For example,
Schade et al. and Mancini et al. employed 4,4-difluoro-4-bora-
3a,4a-diaza-s-indacene (BODIPY) and fluorescein-labeled caseins
to detect protease activities. They were able to achieve LODs of
50 ng/mL and 170 nM, respectively [25–27]. Although fluorescence
technique is highly sensitive, fluorescein becomes non-fluorescent
in acidic solutions. Alternatively, some researchers derivatized

https://doi.org/10.1016/j.snb.2018.04.125
0925-4005/© 2018 Elsevier B.V. All rights reserved.
 M. Jannat, K.-L. Yang / Sensors and Actuators B 269 (2018) 8–14 9

casein with a chromogenic agent 4-(dimethylamino) azobenzene-

4 -sulfonyl chloride (DABS-C1) for detecting protease activities.
For trypsin and ␣-chymotrypsin, LODs can go down to as low as
64.16 ng and 89.86 ng, respectively [28]. However, the selectivity
of the reaction is poor as the chromophores can bind to any amino
acids, peptides and protein molecules. Finally, protease assays
based on electrophoretic methods and charge-changing substrates
were also reported in the literature. In these assays, LODs could
reach 0.3 ng/mL and 500 ng/mL for trypsin and ␣-chymotrypsin,
respectively [29,30]. However, they require custom-synthesized
substrates in the assays.
Recently, liquid crystals (LCs) were exploited by many
researchers to develop label-free bioassays with naked-eye
detection capability [31–35]. LC-based bioassays are portable, easy-
to-use, and can be used to detect a wide variety of biomolecules
with good sensitivity. For example, Yang and co-workers used LC
to detect peptides of different lengths [36–39]. In this assay, LC was
directly applied to a surface functionalized with peptides. When
long peptides were immobilized on the surface, they disrupted ori-
entations of the LC and gave a bright color. On the other hand, Fig. 1. (A) Working principle of a nonspecific protease assay. Casein is used as a
when short peptides were immobilized on the surface, they did substrate for protease and TCA is used to precipitate remaining casein and proteases.
not disrupt the orientations of the LC and the LC remained dark. Detection of peptide fragments is accomplished by using absorbance measurement
or LC. (B) A continuous protease assay in which reaction products are mixed with
The capability of using LC to differentiate long and short peptides TCA at a T-junction. Peptide fragments emerging from the millifluidic device were
was further exploited to develop a heterogeneous protease assay. further analyzed by using LC.
In the presence of proteases, surface-immobilized long peptides
were shortened by protease activity, and the LC gave a dark spot
2.2. Measurement of protease activity
[40]. However, in the heterogeneous protease assay, the substrates
are surface-immobilized peptides which are different from free
Initially, casein (6.5 mg/mL) was dissolved in potassium phos-
peptides in buffer solutions. Therefore, there is a need to develop
phate buffer (50 mM, pH 7.5) and used as a substrate solution
a LC-based, homogeneous protease assays in which free protein
[22]. After that, protease solutions of different concentrations were
molecules can be used as substrates.
prepared in sodium acetate (10 mM, pH 7.5) containing 5 mM of
In this study, we developed a LC-based protease assay by using
calcium acetate (protease diluent). To measure protease activity,
casein as a substrate and used it to determine protease activity
approximately 400 ␮L of the casein solution was incubated at 37 ◦ C
and perform kinetic studies. Unlike other LC-based bioassays, this
for 5 min under constant shaking (100 rpm). Then, 100 ␮L of a pro-
assay is a homogenous assay in which proteases and substrates
tease solution was added to the casein solution and incubated for
were mixed homogeneously in a buffer solution to release pep-
another 10 min. Subsequently, 600 ␮L of TCA solution (110 mM)
tide fragments. Thus, the protease activities can be determined
was added to the reaction mixture to stop the reaction. The mix-
more precisely than heterogeneous assays. However, undigested
ture was incubated for another 30 min to precipitate all casein and
substrates and remaining proteases must be separated from the
protease. At the end of the incubation, the reaction mixture was
solution before peptide fragments can be analyzed. In this study, we
centrifuged for 1 min. A clear supernatant (200 ␮L) was transferred
first studied how to precipitate casein and proteases by using TCA
into a vial with 500 ␮L of sodium carbonate (500 mM) and 100 ␮L of
and avoid their interferences with LC. Parameters such as casein and
FC reagent (500 mM). FC reagent was used to form blue color com-
protease concentrations, mixng ratios of TCA to casein and protease
plex with tyrosine of peptide fragments in the supernatant [22]
were optimized for complete precipitation of casein and protease.
.The solution was incubated for another 30 min for color develop-
Next, LC was used to detect peptide fragments produced by pro-
ment. For absorbance measurement, 200 ␮L of the solution was
tease activities. Finally, to minimize manual sampling, a continuous
transferred into a 96-well microplate. Absorbance was measured
protease assay with a syringe pump was developed.
by using microplate reader at 660 nm and 20 ◦ C. Fig. 1A shows the
mechanism of protease assay.
2. Experimental
2.3. Surface modification
2.1. Materials
Glass slides were washed with deionized (DI) water twice. Sub-
Polytetrafluoroethylene (PTFE) tubing (ID 0.8 mm × OD sequently, the glass slides were immersed in a 5% Decon 90 solution
2.4 mm) and 96-well microplates were purchased from Cole overnight for cleaning. After that, the glass slides were sonicated
Palmer (Singapore). A T-shape polypropylene connector was and rinsed rigorously with DI water. Afterwards, the glass slides
purchased from Kartell (Italy). N, N-dimethyl-n-octadecyl- were sonicated twice for 15 min in DI water. Then, the glass slides
3aminopropyltrimethoxy- silyl chloride (DMOAP), protease from were immersed in an aqueous solution containing 0.1% of DMOAP
bovine pancreas type I (≥5 U/mg solid), sodium acetate, potassium for 5 min and washed with DI water. DMOAP-coated glass slides
phosphate buffer, calcium acetate, casein from bovine pancreas, were heated in a vacuum oven at 100 ◦ C for 15 min to allow cross-
trichloroacetic acid (TCA), Folin & Ciocalteu’s phenol reagent linking of DMOAP [38].
(FC reagent), sodium carbonate, Tween 20 and L-tyrosine were
purchased from Sigma Aldrich (Singapore) and used as received. 2.4. LC-based protease assays
Liquid crystal 4-cyano-4-pentylbiphenyl (5CB) was purchased
from Merck (Singapore), Glass slides were purchased from Fisher To detect protease activity by using LC, we collected 0.5 ␮L of
Scientific (U.S.A). a sample solution after TCA was added and incubated for 30 min.
10 M. Jannat, K.-L. Yang / Sensors and Actuators B 269 (2018) 8–14
The solution was spotted onto a DMOAP-coated glass slide. The
slide was stored inside a humidified chamber at room temperature,
allowing the adsorption of peptide fragments on the DMOAP-
coated surface. After 30 min, the slide was dipped into an aqueous
solution containing 0.005% (v/v) of Tween 20 briefly to clean the
surface. It was followed by washing the surface with DI water and
blown-dried with compressed nitrogen gas. Afterwards, two strips
of 6 ␮m polyester spacer films and another DMOAP-coated glass
slide were put on the top of the first glass slide to form an opti-
cal cell for LC. Subsequently, two binder clips were used to secure
two edges of the cell. Approximately 10 ␮L of LC was added into
the space between the two glass slides, allowing it to flow by cap-
illary action. After 10–20 min, the optical cell was placed under a
polarized microscope (Nikon ECLIPSE LV100POL, Japan). A digital
camera attached on the microscope was used to capture images
with an exposure time of 20 ms.
2.5. Kinetic study
Casein solutions (6.5 mg/mL) were dispensed into vials and
incubated in a water bath at 37 ◦ C for about 5 min with constant
shaking (100 rpm). To prepare blank samples, 400 ␮L of solution
from each vial was transferred to another vial containing 600 ␮L
of TCA solution (110 mM) and 100 ␮L of protease solutions were
added [22]. For real samples, protease solutions were added to the
remaining casein solutions. The protease-to-casein volume ratio
was maintained at 1:4. Subsequently, at every 5 min time inter-
vals, 500 ␮L of reaction mixture was taken out from each vial and
transferred to a TCA containing vial to quench the reaction. After
collecting different time interval reaction mixtures and stopping
the reaction with TCA, absorbance-based measurement was carried
out.
2.6. Continuous protease assay
Two syringes, one containing 200 ␮L of protease solution and
1 mL of casein solution (referred to as a reaction mixture) and
the other containing TCA solution (165 mM) were prepared. Both
syringes were connected to a piece of PTFE tubing of same length
and driven by a syringe pump. The two pieces of tubings were
connected in a T-junction as shown in Figs. 1B and S1. After that,
both solutions were delivered at a flow rate of 500 ␮L/min with
a total volume of 100 ␮L. After the injection, the PTFE tubing con-
taining the reaction mixture was incubated in a water bath at 37◦ C.
Every 5 min, the reaction mixture was pushed out from the tubing
and mixed with the TCA solution to stop the reaction immedi-
ately. The final product was collected in a vial. After that, the vials
were incubated at 37 ◦ C bath with continuous shaking (300 rpm)
on a thermomixer for 30 min. Subsequently, the vial was then cen-
trifuged for 1 min to settle down all precipitants. For the LC-based
assay, clear supernatant from the top was collected and spotted on
a DMOAP coated glass slide for detection.
3. Results and discussion
3.1. Disruption of LC by surface-adsorbed casein and protease
LC is used in the protease assay as a signal reporter as it is able
to detect surface-adsorbed peptide fragments. However, because
peptide fragments are mixed together with undigested casein and
remaining proteases in a buffer solution, both casein and proteases
must be separated from the reaction mixture before LC-based pro-
tease assay to avoid their interferences with LC. To understand
effects of casein and proteases on LC, we spotted casein solutions
with different concentrations on glass slides and observed how
they affect LC. Fig. 2A shows that surface-adsorbed casein led to a
Fig. 2. Effect of surface adsorbed (A) casein and (B) protease on optical textures of
LC supported on the surface. The number above each spot indicates concentrations
(␮g/mL) of casein or protease.
bright LC spot when the initial casein concentration was 20 ␮g/mL
or higher. In contrast, in regions where no casein solution was
dispensed, or when the concentration of casein was lower than
20 ␮g/mL, the optical texture of LC was dark. This result suggests
that casein with a concentration of 20 ␮g/mL can adsorb on the
glass surface and disrupt LC. This phenomenon is similar to other
types of surface-adsorbed proteins such as bovine serum albumin
(BSA) reported in the literature [41]. Thus, in an LC-based protease
assay, the final casein concentration must be kept below 20 ␮g/mL
to avoid interference. Similarly, we also spotted protease solutions
with different concentrations (100–10 ␮g/mL) on the same glass
surface. This concentration range was chosen to be in line with
protease concentrations used in our protease assays. Fig. 2B shows
that even when 100 ␮g/mL of the protease was used, the surface-
adsorbed protease only slightly disrupted the orientations of LC and
caused a non-saturated LC spot with a grey color. When the con-
centration of protease was lower than 100 ␮g/mL, protease was
unable to disrupt the orientation of LC as the LC remained dark.
This result suggests that either the protease did not adsorb on the
surface or its density was too low to disrupt the orientations of LC.
These results mean that the final concentration of protease should
be below 100 ␮g/mL to avoid interferences from protease.
3.2. Precipitation of casein and protease by using TCA
In a standard protease assay, the casein concentration was
6.5 mg/mL. However, it has been shown that even at a concen-
tration as low as 20 ␮g/mL, casein can adsorb on the surface and
cause interferences in the LC-based assay. Therefore, casein must
be separated from the reaction mixture before LC can be applied
to detect peptide fragments. In the literature, an effective way to
precipitate proteins from solutions is to use trichloro acetic acid
(TCA). The acidic nature of TCA and its trichloro moiety are known to
cause conformational changes in proteins and protein precipitation
[42]. To determine how much TCA is needed to precipitate casein
from the solution, we prepared a stock TCA solution (110 mM) [22]
and mixed it with a casein solution (6.5 mg/mL) in different mixing
ratios (based on volumes). Fig. 3A shows different solutions right
after the addition of TCA to casein. When the TCA to casein ratio
was 0.25:1, the solution appeared clear, suggesting that the amount
of TCA was insufficient to precipitate casein. When the ratio was
0.43:1 the mixture became cloudy and some protein precipitated
at the bottom. When the TCA to casein ratio was increased to 0.67:1,
the protein precipitates at the bottom became more compact and
the supernatant became clearer, suggesting that most casein has
been precipitated by TCA at this mixing ratio. Finally, when the
 M. Jannat, K.-L. Yang / Sensors and Actuators B 269 (2018) 8–14
Fig. 3. Precipitation of casein (6.5 mg/mL) by using a TCA solution (110 mM) at dif-
ferent volume ratios. Numbers indicate TCA to casein ratios. (A) Visual appearance
of mixtures containing both TCA and casein after the mixing. (B–D) LC was used
to detect remaining casein in the supernatant after the precipitation of casein by
using TCA. The initial casein concentration was (B) 6.5 mg/mL, (C) 0.65 mg/mL and
(D) 65 ␮g/mL, respectively.
ratio reached at 4:1, the supernatant was completely clear. At this
point, probably all casein molecules have precipitated. However,
it is difficult to determine the exact amount of TCA needed to
precipitate all casein molecules from the solution based on this
experiment. To find the exact amount of TCA required, we repeated
the same experiment and used LC to detect residual casein in the
supernatant. LC was used because we have shown that 20 ␮g/mL
of casein in the solution leads to a bright LC spot (Fig. 2A). Thus,
if LC gives a dark spot, then it can be concluded that the casein
concentration is less than 20 ␮g/mL. As shown in Fig. 3B, when
the TCA to casein ratio was 0.25:1, the corresponding LC spot was
bright. The LC result suggests that the remaining casein concentra-
tion in the supernatant was higher than 20 ␮g/mL. Thus, it can be
concluded that the amount of TCA was not enough to precipitate
casein. Next, when the TCA to casein ratio was 1:1, a bright LC spot
was still observed. This result suggests that the amount of TCA was
still insufficient to precipitate all casein molecules. Finally, when
the TCA to casein ratio was 1.5:1, a dark LC spot appeared. At this
mixing ratio, it seemed that all casein molecules (6.5 mg/mL) were
precipitated by TCA. Subsequently, similar TCA precipitation exper-
iments were conducted by using two lower casein concentrations
at 0.65 mg/mL and 65 ␮g/mL, respectively. Fig. 3C shows that when
the casein concentration was 0.65 mg/mL, a TCA to casein ratio of
0.67:1 was sufficient to precipitate all casein molecules. Similarly,
Fig. 3D shows that when the casein concentration was 65 ␮g/mL, a
TCA to casein ratio at 0.43:1 was sufficient to precipitate all casein
molecules. These results, when combined, suggest that the TCA to
casein ratio for complete precipitation of 65 mg/mL, 0.65 mg/mL
and 65 ␮g/mL of casein were 1.5:1, 0.67:1 and 0.43:1, respectively.
A similar experiment was also repeated for protease to determine
how much TCA is required to precipitate protease (Fig. S2). Since the
casein concentration was much higher than that of protease in the
11
Fig. 4. Protease assays for determining protease activities. Released peptide frag-
ments were quantified (as tyrosine equivalent) either by using (A) Lowry’s method
or (B) LC method. Number above each spot indicates the protease concentration
(␮g/mL).
assay, the amount of TCA required to precipitate protease is much
less and did not need to be taken into account. In the following
experiments, TCA to casein ratio was fixed at 1.5:1 to make sure all
undigested casein and remaining protease have been precipitated
from the reaction mixture.
3.3. Measurement of protease activity
After the precipitation of casein and protease, the remaining
peptide fragments in the reaction mixture can be quantified to
obtain protease activity. Firstly, we used Lowry’s method to deter-
mine the amount of peptide fragments (as tyrosine equivalent)
after incubation of casein with different concentrations of protease
for 10 min [22]. Fig. 4A shows the amount of peptide fragments
in the supernatant as a function of the protease concentration. It
is apparent from the figure that the amount of peptide fragments
increased linearly with increasing protease concentration. From
the slope of the regression line, protease activity was found to be
7.23 U/mg protease (R2 = 0.991). This activity is comparable to man-
ufacturer’s data (≥5 U/mg protease). From Fig. 4A, it can also be
seen that 5 ␮g/mL (or 0.04 U/mL) of protease can be detected by
using Lowry’s method. At this protease concentration, 3.26 mg/mL
of peptide fragments were released after 10 min of protease reac-
tion. In addition, it is also found that this activity is not affected by
the pH of protease solution within 3–11.5 range (Fig. S3)
In addition to Lowry’s method, peptide fragments were also
detected by using the LC method. In this case, reaction mixtures
from the protease assays were first mixed with TCA and then
centrifuged to remove casein and protease. Subsequently, clear
supernatant was spotted on glass slides and LC cells were fabri-
12 M. Jannat, K.-L. Yang / Sensors and Actuators B 269 (2018) 8–14

Fig. 5. Protease kinetic study using a batch process. Protease concentrations were
100 ␮g/mL (䊏), 30 ␮g/mL (䊉), 5 ␮g/mL () and 1 ␮g/mL (), respectively. Original
data of this kinetic study is provided at supplementary document (Table S1).

cated to detect peptide fragments. Fig. 4B shows optical images

of the LC cells. When the protease concentration was 100 ␮g/mL,
a bright LC spot was observed. The result implies that during the
reaction, casein was digested by proteases and peptide fragments
were released. Unlike casein and proteases, these peptide frag-
ments were too small to be precipitated by TCA and remained in the
supernatant after centrifugation. When the supernatant was spot-
ted on the glass slide, peptide fragments adsorbed on the surface
and disrupted the orientations of LC as shown in Fig. 4B. When
protease concentrations were lowered to 80, 50 and 30 ␮g/mL,
bright LC spots were still observed in Fig. 4B. It shows that the
amount of peptide fragments in the supernatant was still enough to
disrupt the orientations of LC. However, when the protease concen-
tration was 5 ␮g/mL, the LC displayed a grey color and some area
remained dark. One possible explanation is that the amount of pep-
tide fragments produced by the protease activity was insufficient to
cover the entire surface area. When the protease concentration was
reduced further, both the brightness and the grey area decreased
with decreasing protease concentration. When the protease con-
centration was 0.001 ␮g/mL (or 1 ng/mL), the LC spot only had a
grey lining. Under this condition, the amount of peptide fragments
was too low to disrupt the orientation of LC. Thus, on the basis Fig. 6. Kinetic study of protease activity by using LC as a reporter. The assay was
of LC result shown in Fig. 4B, approximately 10 ng/mL of protease operated in a continuous mode with casein (6.5 mg/mL) as the substrate. LC was
used to quantify the amount of peptide fragment produced at different time inter-
can be detected by using the LC method. Since the protease activity
vals. Protease concentrations were (A) 100 ␮g/mL, (B) 30 ␮g/mL, (C) 5 ␮g/mL, (D)
was 7.23 U/mg protease, we can estimate that 6.5 ␮g/mL of peptide 1 ␮g/mL, (E) 100 ng/mL and (F)10 ng/mL. The number above each spot indicates
fragments can be detected by using the LC assay. This is far more reaction time (min).
sensitive than the Lowry’s method which has a LOD of 3.26 mg/mL
of peptide fragments. It is also noted that the LOD of the LC-based was limited by the substrate (casein) after 10 min. Secondly, when
assay is similar or even lower than other assays such as fluores- the protease concentration was 30 ␮g/mL, the amount of peptide
cence method (LOD ∼ 14 ng/mL for bovine pancreatic trypsin) [43], fragments increased linearly with time during the entire duration
UV–vis absorption method (LOD ∼ 390 ng/mL for bovine pancreatic of 30 min, suggesting that the reaction was not substrate-limited
trypsin) [44] and LC–MS (LOD ∼20 ␮g/mL of peptide fragments for when 30 ␮g/mL of protease was used. Thirdly, when the pro-
bovine pancreatic trypsin) [45]. tease concentration was 5 ␮g/mL, peptide fragments could only be
detected after 10 min. Finally, at 1 ␮g/mL of protease, no peptide
3.4. Kinetic study fragment was detected and the line became totally flat. There-
fore, it can be concluded that when protease concentration was
In the previous section, protease activity was measured at a fixed 100 ␮g/mL, the reaction rate was constant in the first 10 min and
time period of 10 min. To confirm that the reaction rate was not lim- then the reaction slowed down and became substrate-limited.
ited by the substrate concentration within 10 min, we performed For lower protease concentrations, the reaction did not reach the
a kinetic study by collecting samples at different time intervals. substrate-limiting region. From this kinetic study, it is clear that a
Lowry’s method was used to quantify peptide fragments in these reaction time of 10 min is reasonable even at the highest protease
samples. Fig. 5 shows the amount of peptide fragments as a func- concentration of 100 ␮g/mL.
tion of time. When the protease concentration was 100 ␮g/mL,
the amount of peptide fragments increased linearly with time 3.5. Continuous protease assays
during the first 10 min. Then, the reaction started to slow down
after 10 min and the peptide fragment concentration reached a In the kinetic study mentioned above, samples were taken at dif-
plateau after 25 min. This is probably because the reaction rate ferent time intervals and then TCA was added manually to stop the
 M. Jannat, K.-L. Yang / Sensors and Actuators B 269 (2018) 8–14
reaction. This process unavoidably leads to experimental errors. To
reduce the error, we developed a continuous protease assay with
a millifluidic device in which a TCA solution was injected into the
reaction mixture automatically. Mixing in the millifluidic device is
more efficient due to the small dimensions of millifluidic devices.
Moreover, each droplet emerged from the tubing has a constant vol-
ume, which can be controlled by the syringe pump precisely. Both
factors will contribute to better assaying results. Another advan-
tage of the continuous assay is that periodic sampling is not needed
as the reaction mixture is pushed out from the reactor at regu-
lar time intervals. Despite these advantages, each droplet in the
millifluidic device was too small to be used for absorbance mea-
surements by using Lowry’s method. Therefore, only LC was used
to detect peptide fragments in the droplet, as the LC-based method
only requires a very small sample volume (∼0.5 ␮L) to complete
the assay. As shown in Fig. 6A, when the protease concentration
was 100 ␮g/mL, all samples collected from the millifluidic reactor
contained peptide fragments (as judged from the bright LC spots).
However, Fig. 6B shows that when the protease concentration was
reduced to 30 ␮g/mL, the brightness of all LC spots reduced. This
phenomenon implies that less amount of peptide fragments was
produced due to the lower protease concentration. The trend con-
tinued for other protease concentrations (Fig. 6C–E) as expected, as
the concentration of peptide fragment is largely dependent on the
protease concentration. Finally, Fig. 6F shows that when the pro-
tease concentration was 10 ng/mL and the reaction time was 5 min,
LC gave a light grey color. When the reaction time was 10 min,
the LC spot displayed more grey color and the bright area also
expanded. This result shows that very few peptide fragments were
produced at 10 ng/mL of protease. However, when the incubation
time was extended to 20 min, a grey LC spot was clearly visible.
This result is comparable to protease assays performed in batches
where 10 ng/mL protease activity could be detected (Fig. S4). There-
fore, this observation suggests that the LC-based assay can be used
to detect protease activity down to 10 ng/mL.
4. Conclusion
In this study, LC-based continuous protease assays were devel-
oped by using a millifluidic device and a syringe pump. LC was
used to detect peptide fragments released from casein after pro-
tease activity. Peptide fragments were able to adsorb on a solid
surface and disrupt the orientations of LC which further produced
bright LC spots for visual detection. Using LC as a signal reporter
has several advantages. For example, the sample volume can be as
small as 0.5 ␮L, and casein molecules do not need to be labeled.
However, the test results could be affected by undigested casein or
remaining proteases, unless TCA was used to precipitate all casein
and protease molecules before the peptide fragments were ana-
lyzed by using LC. By using the assay, we can detect 10 ng/mL of
protease and performed kinetic studies in both batch and continu-
ous modes. This is the first report of a homogeneous protease assay
with LC as a reporter.
Acknowledgement
This project was funded by Agency for Science, Technology and
Research (A* STAR) in Singapore under project no. 1421200079.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.snb.2018.04.125.
13
References
[1] P.G. Noone, Z. Zhou, L.M. Silverman, P.S. Jowell, M.R. Knowles, J.A. Cohn, Cystic
fibrosis gene mutations and pancreatitis risk: relation to epithelial ion
transport and trypsin inhibitor gene mutations, Gastroenterology 121 (2001)
1310–1319.
[2] L.E. Littlepage, M.D. Sternlicht, N. Rougier, J. Phillips, E. Gallo, Y. Yu, et al.,
Matrix metalloproteinases contribute distinct roles in neuroendocrine
prostate carcinogenesis, metastasis, and angiogenesis progression, Cancer Res.
70 (2010) 2224–2234, http://dx.doi.org/10.1158/0008-5472.CAN-09-3515.
[3] A. Brik, C.-H. Wong, HIV-1 protease: mechanism and drug discovery, Org.
Biomol. Chem. 1 (2003) 5–14.
[4] A.V. Galkin, L. Mullen, W.D. Fox, J. Brown, D. Duncan, O. Moreno, et al.,
CVS-3983, a selective matriptase inhibitor, suppresses the growth of androgen
independent prostate tumor xenografts, Prostate 61 (2004) 228–235.
[5] R.P. Beckett, A.H. Davidson, A.H. Drummond, P. Huxley, M. Whittaker, Recent
advances in matrix metalloproteinase inhibitor research, Drug Discov. Today
1 (1996) 16–26.
[6] Y. Hua, S. Nair, Proteases in cardiometabolic diseases: pathophysiology,
molecular mechanisms and clinical applications, Biochim. Biophys. Acta Mol.
Basis Dis. 1852 (2015) 195–208.
[7] R. Vassar, B.D. Bennett, S. Babu-Khan, S. Kahn, E.A. Mendiaz, P. Denis, et al.,
Beta-secretase cleavage of Alzheimer’s amyloid precursor protein by the
transmembrane aspartic protease BACE, Science 286 (1999) 735–741.
[8] S. Kumar Kailasa, N. Hasan, H.-F. Wu, Identification of multiply charged
proteins and amino acid clusters by liquid nitrogen assisted spray ionization
mass spectrometry, Talanta 97 (2012) 539–549.
[9] S.K. Kailasa, H.-F. Wu, Recent developments in nanoparticle-based MALDI
mass spectrometric analysis of phosphoproteomes, Microchim. Acta 181
(2014) 853–864.
[10] S.K. Kailasa, H.-F. Wu, Functionalized quantum dots with dopamine
dithiocarbamate as the matrix for the quantification of efavirenz in human
plasma and as affinity probes for rapid identification of microwave tryptic
digested proteins in MALDI-TOF-MS, J. Proteom. 75 (2012) 2924–2933.
[11] S.K. Kailasa, H.-F. Wu, One-pot synthesis of dopamine dithiocarbamate
functionalized gold nanoparticles for quantitative analysis of small molecules
and phosphopeptides in SALDI- and MALDI-MS, Analyst 137 (2012)
1629–1638.
[12] H. Morisaka, K. Hata, J. Mima, T. Tanigawa, M. Furuno, N. Ishizuka, et al.,
Enhanced sequence coverage in tryptic fragment analysis by
two-dimensional HPLC/MS using a monolithic silica capillary column, Biosci.
Biotechnol. Biochem. 70 (2006) 2154–2159.
[13] Q.M. Wang, R.B. Johnson, G.A. Cox, E.C. Villarreal, R.J. Loncharich, A continuous
colorimetric assay for rhinovirus-14 3C protease using peptidep-nitroanilides
as substrates, Anal. Biochem. 252 (1997) 238–245.
[14] O. Schulz, H.F. Sewell, F. Shakib, A sensitive fluorescent assay for measuring
the cysteine protease activity of Der p 1, a major allergen from the dust mite
Dermatophagoides pteronyssinus, Mol. Pathol. 51 (1998) 222–224.
[15] X. Ding, D. Ge, K.-L. Yang, Colorimetric protease assay by using gold
nanoparticles and oligopeptides, Sens. Actuators B Chem. 201 (2014)
234–239.
[16] L. Zhang, J. Du, A sensitive and label-free trypsin colorimetric sensor with
cytochrome c as a substrate, Biosens. Bioelectron. 79 (2016) 347–352.
[17] G. Ertürk, M. Hedström, B. Mattiasson, A sensitive and real-time assay of
trypsin by using molecular imprinting-based capacitive biosensor, Biosens.
Bioelectron. 86 (2016) 557–565.
[18] K.E. Sapsford, D. Farrell, S. Sun, A. Rasooly, H. Mattoussi, I.L. Medintz,
Monitoring of enzymatic proteolysis on a electroluminescent-CCD microchip
platform using quantum dot-peptide substrates, Sens. Actuators B Chem. 139
(2009) 13–21.
[19] N. Marmé, J.-P. Knemeyer, J. Wolfrum, M. Sauer, Highly sensitive protease
assay using fluorescence quenching of peptide probes based on photoinduced
electron transfer, Angew. Chem. Int. Ed. 43 (2004) 3798–3801.
[20] K.A. Kilian, T. Böcking, K. Gaus, M. Gal, J.J. Gooding, Peptide-modified optical
filters for detecting protease activity, ACS Nano 1 (2007) 355–361.
[21] I.L.H. Ong, K.-L. Yang, Recent developments in protease activity assays and
sensors, Analyst 142 (2017) 1867–1881.
[22] C. Cupp-Enyard, Sigma’s non-specific protease activity assay-casein as a
substrate, J. Vis. Exp.: JoVE (2008).
[23] D.H. Manicourt, V. Lefebvre, An assay for matrix metalloproteinases and other
proteases acting on proteoglycans, casein, or gelatin, Anal. Biochem. 215
(1993) 171–179.
[24] R.F. Highsmith, R.D. Rosenberg, A rapid and sensitive proteolytic assay for
human plasminogen and plasmin using radioiodinated ␣-casein, Thromb. Res.
11 (1977) 131–140.
[25] S.Z. Schade, M.E. Jolley, B.J. Sarauer, L.G. Simonson, BODIPY–␣-casein, a
pH-independent protein substrate for protease assays using fluorescence
polarization, Anal. Biochem. 243 (1996) 1–7.
[26] F. Mancini, M. Naldi, V. Cavrini, V. Andrisano, Multiwell fluorometric and
colorimetric microassays for the evaluation of beta-secretase (BACE-1)
inhibitors, Anal. Bioanal. Chem. 388 (2007) 1175–1183.
[27] M.E. Jolley, Fluorescence polarization assays for the detection of proteases
and their inhibitors, J. Biomol. Screen. 1 (1996) 33–38.
14 M. Jannat, K.-L. Yang / Sensors and Actuators B 269 (2018) 8–14
[28] E. Anderson, K.W.C. Sze, S.K. Sathe, New colorimetric method for the
detection and quantitation of proteolytic enzyme activity, J. Agric. Food Chem.
43 (1995) 1530–1534.
[29] R.B. Lefkowitz, G.W. Schmid-Schönbein, M.J. Heller, Whole blood assay for
trypsin activity using polyanionic focusing gel electrophoresis,
Electrophoresis 31 (2010) 2442–2451.
[30] R.B. Lefkowitz, J.Y. Marciniak, C.-M. Hu, G.W. Schmid-Schönbein, M.J. Heller,
An electrophoretic method for the detection of chymotrypsin and trypsin
activity directly in whole blood, Electrophoresis 31 (2010) 403–410.
[31] H. Tan, S. Yang, G. Shen, R. Yu, Z. Wu, Signal-enhanced liquid-crystal DNA
biosensors based on enzymatic metal deposition, Angew. Chem. 122 (2010)
8790–8793.
[32] J. Rim, H.J. Kim, C.-H. Jang, Parallel detection of anti-tuberculosis antibodies
upon a liquid crystal-based optical sensor, Bull. Korean Chem. Soc. 37 (2016)
1644–1651.
[33] A.M. Lowe, B.H. Ozer, Y. Bai, P.J. Bertics, N.L. Abbott, Design of surfaces for
liquid crystal-based bioanalytical assays, ACS Appl. Mater. Interfaces 2 (2010)
722–731.
[34] C.-H. Jang, M.L. Tingey, N.L. Korpi, G.J. Wiepz, J.H. Schiller, P.J. Bertics, et al.,
Using liquid crystals to report membrane proteins captured by affinity
microcontact printing from cell lysates and membrane extracts, J. Am. Chem.
Soc. 127 (2005) 8912–8913.
[35] D. Hartono, S.L. Lai, K.-L. Yang, L.-Y.L. Yung, A liquid crystal-based sensor for
real-time and label-free identification of phospholipase-like toxins and their
inhibitors, Biosens. Bioelectron. 24 (2009) 2289–2293.
[36] Q. Zhu, K.-L. Yang, Microfluidic immunoassay with plug-in liquid crystal for
optical detection of antibody, Anal. Chim. Acta 853 (2015) 696–701.
[37] C.-H. Chen, K.-L. Yang, Functional protease assay using liquid crystals as a
signal reporter, Biosens. Bioelectron. 35 (2012) 174–179.
[38] L.H. Ong, X. Ding, K.-L. Yang, Mechanistic study for immobilization of
cysteine-labeled oligopeptides on UV-activated surfaces, Colloids Surf. B 122
(2014) 166–174.
[39] C.-Y. Xue, K.-L. Yang, Dark-to-bright optical responses of liquid crystals
supported on solid surfaces decorated with proteins, Langmuir 24 (2008)
563–567.
[40] X. Bi, S.L. Lai, K.L. Yang, Liquid crystal multiplexed protease assays reporting
enzymatic activities as optical bar charts, Anal. Chem. 81 (2009) 5503–5509.
[41] S.R. Kim, R.R. Shah, N.L. Abbott, Orientations of liquid crystals on mechanically
rubbed films of bovine serum albumin: a possible substrate for biomolecular
assays based on liquid crystals, Anal. Chem. 72 (2000) 4646–4653.
[42] D. Rajalingam, C. Loftis, J.J. Xu, T.K. Kumar, Trichloroacetic acid-induced
protein precipitation involves the reversible association of a stable partially
structured intermediate, Protein Sci. 18 (2009) 980–993.
[43] X. Wang, Y. Wang, H. Rao, Z. Shan, A Sensitive fluorescent assay for trypsin
activity in biological samples using BSA-Au nanoclusters, J. Braz. Chem. Soc.
23 (2012) 2011–2015.
[44] X. Ding, K.-L. Yang, Quantitative serine protease assays based on formation of
copper(ii)-oligopeptide complexes, Analyst 140 (2015) 340–345.
[45] E.C. Huang, J.D. Henion, LC/MS and LC/MS/MS Determination of protein
tryptic digests, J. Am. Soc. Mass Spectrom. 1 (1990) 158–165.
Biographies
Mahbuba Jannat is now pursuing her phD degree in the department of Chemical
and Biomolecular Engineering at the National University of Singapore (NUS). Before
joining NUS, she received her B.S degree in Chemical Engineering from Bangladesh
University of Engineering and Technology (BUET) in 2013. Her current research
interest is liquid crystal based biosensor development.
Dr. Yang Kun-Lin is an associate professor in the Department of Chemical and
Biomolecular Engineering at the National University of Singapore (NUS). Before
joining NUS, he was a post-doctoral researcher in the Chemical and Biomolecu-
lar Department at the University of Wisconsin −Madison working on liquid crystal
sensors. He received his BS and MS degrees in Chemical Engineering from National
Taiwan University and PhD degree from Georgia Tech. He is the recipient of the
Defense Innovation Research Program Award and the A*Star Research Grant Award
for his work on liquid crystal sensors. His present research interests include biosen-
sors, liquid crystals and microfluidics.