Phase-switching depolymerizable poly(carbamate) oligomers for signal amplification in quantitative time-based assays

Article abstract of DOI:10.1021/ma4007413

This article describes the use of poly(carbamate) oligomers that depolymerize from head-to-tail as phase-switching reagents for increasing the sensitivity of quantitative point-of-care assays that are based on measurements of time. The poly(carbamate) oligomers selectively react with hydrogen peroxide (a model analyte) and provide sensitivity by depolymerizing in the presence of the analyte to convert from water-insoluble oligomers to water-soluble products. This switching reaction enables a sample to wick through a three-dimensional paper-based microfluidic device, where the flow-through time reflects the quantity of the analyte in the sample. Oligomers as short as octamers enable quantitative detection to low nanomolar concentrations of the analyte.

Full text of DOI:10.1021/ma4007413

Article
pubs.acs.org/Macromolecules
Phase-Switching Depolymerizable Poly(carbamate) Oligomers for
Signal Amplification in Quantitative Time-Based Assays
Gregory G. Lewis,‡ Jessica S. Robbins,‡ and Scott T. Phillips*
Department of Chemistry, The Pennsylvania State University, University Park, Pennsylvania 16802, United States
S
* Supporting Information
ABSTRACT: This article describes the use of poly-
(carbamate) oligomers that depolymerize from head-to-tail as
phase-switching reagents for increasing the sensitivity of
quantitative point-of-care assays that are based on measure-
ments of time. The poly(carbamate) oligomers selectively react
with hydrogen peroxide (a model analyte) and provide
sensitivity by depolymerizing in the presence of the analyte
to convert from water-insoluble oligomers to water-soluble
products. This switching reaction enables a sample to wick
through a three-dimensional paper-based microfluidic device,
where the flow-through time reflects the quantity of the analyte
in the sample. Oligomers as short as octamers enable
quantitative detection to low nanomolar concentrations of the analyte.
∼500 000-fold improvement over our previous system. These
improvements in sensitivity provide an important step toward
realizing a generalizable quantitative POC assay platform for
use in resource-limited environments since many analytes of
interest are present in samples at (or below) micro- and
nanomolar levels.
INTRODUCTION
■
A long-standing challenge in the area of point-of-care (POC)
diagnostics has been the development of operationally simple
and inexpensive platforms for conducting reproducible and
rapid quantitative assays.1−3 The ideal quantitative POC assay,
particularly for use in resource-limited environments such as
the developing world, should be inexpensive, be straightforward
to operate, and provide rapid and reproducible quantitative
results without the need for specialized electronic devices to
measure the output of the assay.4,5 As a step toward this goal,
we recently described an approach for quantitative POC assays
that requires only measurements of time as the readout (Figure
1).6 The selectivity in the assay is based on a hydrophobic-to-
hydrophilic switch of a hydrogen-peroxide-responsive small
molecule (compound 1, Figure 1d). This novel assay strategy,
however, was limited in sensitivity (e.g., hydrogen peroxide was
measured only to ∼3 mM levels) and thus was restricted in the
types of applications to which it could be applied.
Herein we describe an approach for substantially improving
the sensitivity of the model assay by using analyte-triggered
oligomers, rather than small molecules, to achieve signal
amplification in the assay. Specifically, we rationally designed
water-insoluble poly(carbamate) oligomers that depolymerize
from head-to-tail7−10 to reveal water-soluble products in
response to hydrogen peroxide, thus inducing a large
hydrophobic-to-hydrophilic switch. This approach improves
the sensitivity 4 orders of magnitude compared with our
previous system (limits-of-detection of hydrogen peroxide are
now 146 nM). Moreover, by optimizing the number of layers of
paper containing the oligomer in the assay platform, we further
improved the limit of detection by approximately another order
of magnitude, providing a limit of detection of 6 nM and a
RESULTS AND DISCUSSION
■
Design of the Assay Platform. Figure 1a provides an
illustration of the assay platform. The assay is conducted in a
three-dimensional (3D) paper-based microfluidic device that
consists of stacked, alternating layers of (i) paper that has been
patterned with wax into hydrophobic and hydrophilic regions,
and (ii) double-sided adhesive tape, which has holes patterned
into it using a CO2 laser cutter.11−13 The holes in the tape are
filled with disks of hydrophilic paper such that hydrophilic
regions in paper connect with the hydrophilic disks in the tape.
The result is a disposable and inexpensive multilayered 3D
device that wicks aqueous fluids into defined regions within the
device. Each layer of the device (except layers 1, 4, and 7) is
preloaded with a reagent and dried before assembly, such that
addition of sample to the top of this device allows the assay to
occur automatically without user intervention.
As the sample passes from layer 1 into layer 2, it redissolves
HEPES buffer salts to control the pH of the sample. In layers 3
and 5, the sample encounters compound 1 (Figure 1d), which
is hydrophobic and water-insoluble, and thus alters the wetting
properties of the paper. In the absence of the analyte, the
Received: April 9, 2013
Revised: June 5, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ma4007413 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
required for the sample to pass from the top of the device to
the bottom, which is established by the appearance of green
color in layer 7.14 The selectivity for the assay is provided by
the selective oxidative cleavage of the aryl boronate in 1 via
hydrogen peroxide,6,15,16 although presumably other activity-
based detection events could be employed if the aryl boronate
is replaced with a substrate for another target analyte.16,17
Effect of the Quantity of 1 on the Limit of Detection
(LOD). To render this type of assay general, we needed to
increase the sensitivity of the system so that it could be applied
to a variety of analytes that are present in samples at
concentrations below low millimolar levels. In initial studies,
we observed that there is a direct relationship between the
concentration of hydrogen peroxide in the sample and the time
required for the sample to flow through the paper-based device
(Figure 2a). This behavior provides the basis for the
Figure 1. Strategy for obtaining quantitative assay results by measuring
the time required for a sample to pass from top to bottom of a paper-
based microfluidic device.6 (a) Graphical representation of the device,
including the dry reagents that are included in each layer. (b)
Photographs of the top of the device (where the sample is added) and
(c) the bottom (after the sample has redissolved the dried food
coloring and filled the hydrophilic circle on layer 7 of the device). The
dimensions of the devices are 1 cm ×1 cm ×0.9 mm. (d) Compound
1, which is used in layers 3 and 5. Compound 1 converts from a
hydrophobic molecule to hydrophilic molecules selectively in response
to hydrogen peroxide (the model analyte) through a proposed
quinone methide elimination mechanism.
Figure 2. Relationship between 1 and the time required for a sample
containing hydrogen peroxide to flow through the paper-based
microfluidic device depicted in Figure 1. (a) Effect of the quantity
of 1 on the flow-through time for the assay. The green data was
acquired using 6.1 nmol of 1 per mm3 of paper, the blue data using 9.2
nmol of 1 per mm3 of paper, and the black data using 18.4 nmol of 1
per mm3 of paper. The data points are the averages of 12
measurements, and the error bars reflect the standard deviations
from these averages. The assay was stopped at 45 min, regardless of
whether the sample had wicked through the device. (b) Relationship
between the quantity of 1 per mm3 of paper and the limit of detection
for the assay. The limits of detection were calculated using graphs such
as those in part a, where the linear region of the exponential response
was used along with the equation LOD = (3 × σo)/slope.
sample stops wicking (or slowly wicks, depending on the
quantity of 1 in the paper) through hydrophobic layers 3 and 5.
In the presence of hydrogen peroxide, 1 selectively degrades
into hydrophilic products (Figure 1d), thus switching phases
from insoluble 1 into soluble products, and, consequently,
changing the wetting properties of the paper back to
hydrophilic. After this switching reaction occurs, the sample
wicks to layer 6 where it redissolves dried food coloring to
convert the sample into a brightly colored solution, which
becomes visible when the sample fills the hydrophilic circular
region in layer 7 (Figure 1c). The quantity of hydrogen
peroxide is measured in this device by tracking the time
B
dx.doi.org/10.1021/ma4007413 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
quantitative assay, but it also offers an opportunity to increase
the sensitivity of the method. For example, as revealed in Figure
2a, different quantities of 1 predeposited into layers 3 and 5
substantially affect the sensitivity of the assay: i.e., the green
data were obtained from devices that contained 6.1 nmol of 1
per mm3 of paper in layers 3 and 5, while the black data
originates from devices that contained 18.4 nmol of 1 per mm3
of paper. The limit of detection18 for the former is 4.1× better
than the latter (i.e., 3.1 mM H2O2 vs 12.8 mM), indicating that
the quantity of 1 is critical for achieving the optimal limit of
detection for the assay. In fact, this relationship between the
quantity of 1 and the limit of detection for the model assay for
hydrogen peroxide is parabolic, as indicated in Figure 2b, thus
suggesting that a well-defined quantity of 1 must be present in
the paper to convert into hydrophilic products upon reaction
with hydrogen peroxide to allow the sample to pass through
layers 3 and 5 in an appropriate time frame. With too much of 1
in the device, samples containing low concentrations of
hydrogen peroxide flow through with rates equal to samples
lacking hydrogen peroxide since the relative change from
hydrophobic 1 into hydrophilic products is small. With too
little of 1, samples containing different low concentrations of
hydrogen peroxide flow through equally quickly. Clearly the
absolute quantity of 1 per volume of paper is important for
maximizing the sensitivity of the assay, but so too is the relative
quantity of 1 that converts to hydrophilic products upon
reaction with the analyte. On the basis of this observation, we
predicted that the sensitivity of the assay could be improved if
we altered the magnitude of phase-switching provided by
derivatives of 1 in response to the target analyte.
studies on controlled release reagents,19 we found that addition
of a methyl ether (labeled blue in Figure 3a) ortho to the
benzylic leaving group on the benzene ring substantially
accelerates the rate of quinone methide elimination reactions,
which is the proposed degradation pathway for 1 and 2 to
convert to hydrophilic products. We predicted that this
quinone methide elimination reaction was the rate-limiting
step in converting 1 into hydrophilic products during the assay.
Therefore, by accelerating this quinone methide elimination
reaction, we reasoned that more of hydrophobic compound 2
than 1 would convert to hydrophilic products in the time frame
that it takes for (i) hydrogen peroxide to react with 2 and (ii)
for the aqueous solution to pass through layers 3 and 5 in the
device. If more of 2 converts to hydrophilic products than 1
within the time frame of flow-through, then less hydrogen
peroxide would be needed to enable flow-through using 2 than
1, and thus the sensitivity for the assay should increase.
In the second approach (Figure 3b), we predicted that
oligomers modeled after 2 would provide a greater change in
hydrophobicity than 2 by converting from a large hydrophobic
molecule (an oligomer) to small hydrophilic products upon
reaction with hydrogen peroxide, thus further increasing the
sensitivity for the assay since the change in wetting properties
will be amplified relative to 2.20 The proposed mechanism of
response for the oligomers is as follows: oxidative cleavage of
the aryl boronate in 3−6 would generate a phenol, which would
initiate a cascade head-to-tail depolymerization reaction21−23
through quinone- and azaquinone−methide-mediated path-
ways,7 similar to the mechanism depicted in Figure 1d for
compound 1. We modeled the oligomers after 2 because we
anticipated that the rate of depolymerization would substan-
tially impact the limit of detection for the assays. Therefore, we
included methyl ethers on each repeating unit to accelerate the
rate of formation of azaquinone methide during the
depolymerization reaction,7 much like we included a methyl
ether on 2 to increase the rate of formation of quinone
methide.
Chemical Strategies for Improving the Limit of
Detection. We hypothesized that two different methods of
altering the phase-switching molecule would increase the
sensitivity of the assay. In the first approach (Figure 3a), we
Synthesis of Poly(carbamate) Oligomers. Compound 2
was prepared from 4-(hydroxymethyl)-3-methoxyphenylbor-
onic acid pinacol ester and p-nitrophenyl isocyanate. Com-
pounds 3 and 4 were prepared in a stepwise fashion, as
described previously,7 while oligomers 5 and 6 were prepared
according to the route outlined in Scheme 1. This route
involved a tin-catalyzed polymerization8 of monomer 8, where
the length of the oligomer was controlled by polymerization
time. The polymerization reaction was quenched by addition of
the aryl boronate end-cap, and a post polymerization
modification was used to append the p-nitrophenyl carbamate.
Effect of Reaction Rate on the Limit of Detection. After
optimizing the quantities of 1 and 2 needed for this quantitative
flow-through assay (Tables S9−S13, Supporting Information),
we found that 2 provided a 17-fold improvement in the
sensitivity for the assay compared to 1 (i.e., LOD = 1707 μM
for 1 vs 103 μM for 2). Thus, this result suggests that formation
of quinone methide indeed is the rate-limiting step under the
conditions of the assay.
Figure 3. Hypotheses for increasing the sensitivity of the quantitative
assay by replacing 1 with (a) a derivative that degrades faster than 1
when exposed to hydrogen peroxide, or (b) oligomers that provide a
greater hydrophobic-to-hydrophilic change than 1 when exposed to
hydrogen peroxide. In part b, the oligomers are designed to
depolymerize once the end-cap (aryl boronate) is cleaved from the
polymer and converted into a terminal phenol. The proposed
mechanism of depolymerization is similar to the arrow pushing
mechanism depicted in Figure 1d. The structural changes for 2−6 in
relation to 1 are highlighted in blue.
Effect of Depolymerization on the Limit of Detection.
Oligomers 3−6 also impart significant improvements in
sensitivity to the assay compared to 1 and 2 (Figure 4). For
example, after optimizing the quantity of each reagent for the
flow-through assay (Tables S1−S38, Supporting Information),
we found that addition of one repeating unit (i.e., oligomer 3)
improved the limit of detection 6× compared to 2 and 106×
designed a derivative of 1 (i.e., compound 2) that should
degrade into hydrophilic products faster than 1. In previous
C
dx.doi.org/10.1021/ma4007413 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
a
since the depolymerization intermediates (i.e., truncated
oligomers) are expected to retain substantial hydrophobic
character relative to the products of complete depolymeriza-
tion. There appears to be a match in depolymerization time and
flow-through time when approximately 2−5 repeating units are
added to 2, at which point less dramatic improvement in
sensitivity is obtained when using longer oligomers that
depolymerize more slowly than 4 (e.g., oligomer 6, n = 8,
has a LOD = 0.15 μM, whereas 5, n = 5, has a LOD = 0.96
μM).
Scheme 1. Synthesis of Oligomers 5 and 6.
While we were unable to verify the depolymerization kinetics
of 2−6 in the context of the paper-based microfluidic device,25
solution-phase kinetics reveal the expected linear increase in
half-life for depolymerization as the number of repeating units
in the oligomers increases (Figure 4, blue data). This
relationship between sensitivity of the assay, length of the
oligomer, and depolymerization kinetics suggests that further
improvements in sensitivity can be realized if longer polymers
are used that have substantially faster depolymerization kinetics
than the poly(carbamate) oligomers in Figure 3b.26
Modifications to the Device for Improving the Limit
of Detection. Having established that both the magnitude of
the hydrophobic to hydrophilic switch and the rate of
depolymerization of the phase-switching reagent improve the
sensitivity of the assay, we next explored whether the number of
layers containing the oligomeric reagent within the device
would impact the sensitivity and dynamic range for measuring
hydrogen peroxide. On one hand, we reasoned that several
layers containing small quantities of the phase-switching
reagent might provide improved sensitivity for the assay by
establishing repetitive hydrophobic to hydrophilic switching
interactions between the sample and the oligomeric reagent.
On the other hand, a single layer containing a large quantity of
the phase-switching reagent (relative to a device containing
multiple layers of the compound) might provide the maximum
impact on controlling the flow of the sample in relation to the
concentration of hydrogen peroxide in the sample.
a
Reagents and conditions: (a)TsOH, THF−H2O (87%); (b) (i)
DBTL, (ii) 4-(hydroxymethyl)-3-methoxyphenylboronic acid pinacol
ester, DMSO 110 °C (33% for 9, 64% for 10); (c) 4-nitrophenyl
isocyanate, TEA, DMF (71% for 5, 92% for 6).
To test these scenarios, we prepared devices that were similar
to the device in Figure 1a, but with either one, two, or three
layers of paper that were modified with oligomer 3 (Figures
S1−S3, Supporting Information). We then characterized the
relationship between the limit of detection and the total
quantity of oligomer 3 in the devices to determine the
minimum quantity of 3 needed to provide the lowest limit of
detection for a particular design (Tables S14−S18 and S39−
S49, Supporting Information). These experiments revealed that
a device containing one layer of 3 provided a limit of detection
for hydrogen peroxide that is nearly 5× better than a
comparable device containing three layers of 3 (i.e., the LOD
for one layer of 3 = 9.4 μM, while three layers of 3 = 46 μM).
In fact, the limit of detection worsens by ∼2× every additional
layer of 3 incorporated into the device. Likewise, the dynamic
range for the assay worsens as the number of layers containing
3 increases. For example, the device containing one layer of 3
(Figure S1, Supporting Information) has a dynamic range of 9.4
μM to 1000 μM, whereas the device containing three layers of
3 (Figure S3, Supporting Information) has a smaller dynamic
range of 46−250 μM. Clearly the best device design for
improving the sensitivity of the assay involves providing a single
layer where the magnitude of phase-switching is substantial,
rather than providing small stages of phase-switching over
several layers in a device.
Figure 4. Limit of detection (LOD) (black data) and depolymeriza-
tion half-lives (blue data) versus the number of repeating units in
oligomers 2−6. The half-lives were measured in a 5:4:1 solution of
dioxanes−DMSO−H2O using 20 equiv H2O2. Although the half-lives
of carbamate oligomers depend on the polarity of the environment,7 a
similar linear relationship between depolymerization half-life and the
number of repeating units in an oligomer likely exists in the
environment associated with wet paper as well. The half-life data
points are the averages of three measurements and the error bars
reflect the standard deviations from these averages. The typical assay
time for the limit of detection measurements was 25 min.
compared to 1. In fact, a nearly linear relationship exists
between limit of detection and the number of repeating units in
the oligomer8,24 until, presumably, the rate of hydrogen
peroxide-induced depolymerization becomes competitive with
the residence time of the sample within layers 3 and 5 of the
device (Figure 4). Incomplete depolymerization would affect
the magnitude of hydrophobic-to-hydrophilic phase-switching
D
dx.doi.org/10.1021/ma4007413 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
cannula. Organic solutions were concentrated by rotary evaporation
(25−40 mmHg) at 30 °C. All reagents were purchased commercially
and were used as received unless otherwise noted. 4-Nitrophenyl
isocyanate was recrystallized from petroleum ether prior to use. N,N-
Dimethylformamide, dimethyl sulfoxide, tetrahydrofuran, and triethyl-
amine were purified by the method of Pangborn et al.36 Flash-column
chromatography was performed as described by Still et al.,37
employing silica gel (60 Å pore size, 32−63 μm, standard grade).
Thin-layer chromatography was carried out on silica gel TLC plates
(20 × 20 cm w/h, F-254, 250 μm). Deionized water was purified by
filtration and irradiation with UV light. The papers used were
Whatman Chromatography Paper grade I and Boise Aspen 30 Printer
Paper (92 brilliant, 30% postconsumer content), and the tape was Ace
Hardware Plastic carpet tape (part no. 50106).
Combining Features to Improve the Limit of
Detection. We now have the opportunity to combine the
two approaches for improving sensitivity (i.e., improved
reagents and device design) to ascertain the overall sensitivity
limits for the current assay. Toward this end, we created a
single layer device (Figure S1, Supporting Information) and
determined the optimum quantity of 6 (the oligomer that
provided the best LOD) needed to provide the lowest limit of
detection for quantifying hydrogen peroxide in a sample. This
revised assay requires only 1.9 μg of 6, yet now provides
measurements of hydrogen peroxide down to 6 nM, which is a
LOD that is 500 000× better than our original assay reported in
ref 6.
Methods. Proton nuclear magnetic resonance (1H NMR) spectra
and carbon nuclear magnetic resonance spectra (13C NMR) were
recorded using either a 300, 360, or 400 MHz NMR spectrometer at
25 °C, as indicated in the Experimental Section. Proton chemical shifts
are expressed in parts per million (ppm) and are referenced to residual
protium in the NMR solvent (CHCl3 δ 7.26 ppm, CO(CH3)2 δ 2.05
ppm, or SO(CH3)2 δ 2.50 ppm).38 Data are represented as follows:
chemical shift, multiplicity (s = singlet, bs = broad singlet, d = doublet,
t = triplet, m = multiplet and/or multiple resonances), integration, and
coupling constant (J) in Hertz. Carbon chemical shifts are expressed in
parts per million and are referenced to the carbon resonances of the
NMR solvent (CDCl3 δ 77.0 ppm or CO(CH3)2 δ 29.8 and 206.3
ppm). UV/vis spectroscopic data were obtained using a six-cell
spectrometer. Low resolution and high resolution mass spectra were
acquired using mobile phases containing 5 mM ammonium formate.
GPC data were acquired on a 300 × 7.5 mm, 3−100 μm particle size
styrene divinylbenzene copolymer column using 1 mL/min N,N-
dimethylformamide as the mobile phase. Molecular weights were
calculated from low-angle and right-angle light scattering data. The
system was calibrated using polystyrene standards.
Preparation of Compound 2. Triethylamine (52 μL, 0.38 mmol,
2.0 equiv) was added dropwise to a solution of 4-(hydroxymethyl)-3-
methoxyphenylboronic acid pinacol ester (50 mg, 0.19 mmol, 1.2
equiv)7 and 4-nitrophenyl isocyanate (26 mg, 0.16 mmol, 1.0 equiv) in
tetrahydrofuran (2.0 mL). The reaction mixture was stirred at 23 °C
for 4 h. The solvent was removed by rotary evaporation and the
residue was purified by silica gel flash column chromatography (10%
ethyl acetate in hexanes, increasing to 20% ethyl acetate in hexanes) to
afford compound 2 as a white, amorphous solid (46 mg, 0.11 mmol,
67%): IR (cm−1) 3313, 2977, 2360, 1738, 1600, 1549, 1508; 1H NMR
δ (360 MHz, CO(CH3)2) 9.46 (bs, 1H), 8.22 (d, 2H, J = 9.3 Hz), 7.82
(d, 2H, J = 9.3 Hz), 7.41 (d, 1H, J = 7.3 Hz), 7.36 (d, 1H, J = 7.4 Hz),
7.31 (s, 1H), 5.26 (s, 2H), 3.89 (s, 3H), 1.34 (s, 12H); 13C NMR δ
(360 MHz, CO(CH3)2) 157.6, 154.0, 146.4, 143.4, 129.3, 128.4, 127.8,
125.7, 118.6, 116.4, 84.6, 62.9, 55.8, 25.2 (there appear to be
overlapping peaks in the aromatic region of the 13C spectrum); MS
(TOF MS AP−) 427.2 (M − H+); HRMS (TOF MS AP−) calcd for
C21H24N2O7B (M − H+) 427.1677, found 427.1657.
Preparation of Compound 8. p-Toluenesulfonic acid mono-
hydrate (0.35 g, 1.9 mmol, 0.30 equiv) was added in one portion to a
solution of compound 7 (2.4 g, 6.2 mmol, 1.0 equiv)7 in 4:1
tetrahydrofuran−water (62 mL) under an atmosphere of air. The
reaction mixture was stirred at 23 °C for 4 h. Ethyl acetate (50 mL)
and saturated aqueous sodium bicarbonate (10 mL) were added, each
in one portion, and the layers were separated. The organic layer was
washed with saturated aqueous sodium bicarbonate solution (1 × 50
mL) and was dried over sodium sulfate. The sodium sulfate was
removed by filtration, the solvent was removed by rotary evaporation,
and the residue was purified by silica gel flash column chromatography
(20% ethyl acetate in hexanes, increasing to 60% ethyl acetate in
hexanes) to afford compound 8 as a white, amorphous solid (1.5 g, 5.4
mmol, 87%): IR (cm−1) 3540, 3470, 3269, 2963, 1727, 1615, 1547; 1H
NMR δ (400 MHz, CDCl3) 7.41−7.16 (m, 8H), 6.74 (d, 1H, J = Hz),
4.64 (s, 2H), 3.81 (s, 3H), 2.43 (bs, 1H); 13C NMR δ (300 MHz,
CDCl3) 158.0, 151.8, 150.4, 138.4, 129.4, 129.1, 125.8, 124.3, 121.6,
110.2, 101.6, 61.5, 55.3; MS (Q MS APCI+) 256.1 (M − OH−);
CONCLUSION
■
In conclusion, this article describes the use of depolymerizable
poly(carbamate) oligomers8−10,27 for improving the sensitivity
of an assay that enables quantitative measurements of hydrogen
peroxide by simply tracking the time required for a sample to
pass through a 3D paper-based microfluidic device. The assay is
selective for hydrogen peroxide based on the aryl boronate
functionality in compounds 1−6, quantitative (it has a LOD of
6 nM), does not require substantial input from the user, and is
simple (it uses only paper, tape, and a poly(carbamate)
oligomer that depolymerizes from an insoluble oligomer into
water-soluble products in response to hydrogen peroxide).
Oligomers that depolymerize from head-to-tail in response to
specific analytes offer a unique role in this assay, both as a
hydrophobic-to-hydrophilic switch, and as a means of achieving
signal amplification. This method of signal amplification20 is
unique among POC assays, both in the mechanism of
amplification and in the realization that we cannot adjust the
assay time to increase the sensitivity of the assay, as can be done
with most signal amplification strategies. Instead, we increase
the sensitivity by using oligomers that depolymerize in response
to the target analyte, thus providing a greater hydrophobic-to-
hydrophilic switch than is possible for a small molecule (e.g.,
1). Further improvements in sensitivity are expected if longer
polymers are used (i.e., compared to 6, which has 8 repeating
units), but with the caveat that the polymers must
depolymerize faster than the residence time of the sample in
the device. Efforts to achieve this goal are currently in progress,
as are studies to expand the scope of the method to measure
analytes other than hydrogen peroxide.16 Solutions to these
aspects of sensitivity and selectivity will require thoughtful
designs of new polymers to balance (i) the solubility of the
polymer vs the monomer and (ii) the rate of depolymerization
of the polymer,28 particularly ensuring complete depolymeriza-
tion in seconds to minutes of solid-state polymers.29−32
While an assay platform that offers further improvements in
sensitivity will be useful, it is worth noting that the current limit
of detection for hydrogen peroxide of 6 nM is sufficiently
sensitive to measure hydrogen peroxide in rain and other
sources of water where the presence of hydrogen peroxide is
indicative of pollution.33−35 Such an application may be ideally
suited to a simple, inexpensive, rapid, “reader-less” quantitative
assay that enables users to cheaply and easily track the quality
of various sources of water for environmental analyses.
EXPERIMENTAL SECTION
Materials. All reactions were performed in flame-dried glassware
under a positive pressure of argon unless otherwise noted. Air- and
moisture-sensitive liquids were transferred via syringe or stainless steel
■
E
dx.doi.org/10.1021/ma4007413 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
HRMS (TOF MS AP+) calcd for C15H14NO3 (M − OH−) 256.0974,
found 256.0967.
Preparation of Oligomer 5. Compound 8 (0.60 g, 2.2 mmol, 1.0
equiv) was added in one portion to stirring dimethyl sulfoxide (2.2
mL) at 110 °C. Dibutyltin dilaurate (0.26 mL, 0.44 mmol, 0.2 equiv)
was added in one portion and the reaction mixture was stirred for 2.75
min at 110 °C. 4-(Hydroxymethyl)-3-methoxyphenylboronic acid
pinacol ester (2.0 g, 7.6 mmol, 3.5 equiv) was added in one portion
and the reaction mixture was stirred for 2 h at 110 °C. The reaction
mixture was cooled to 23 °C and poured into 0 °C methanol (20 mL).
A yellow precipitate formed that was washed using a solid phase
washing vessel by adding methanol, bubbling N2 through the solution
at a vigorous rate (see the Supporting Information of ref 29 for a video
of the bubbling rate) for 15 min, then draining the solvent. This
process was repeated three times. The solids were dried under vacuum
for 12 h to give oligomer 9 as an off-white powder (0.17 g, 0.15 mmol,
1
33%): H NMR δ (360 MHz, SO(CH3)2) 9.90 (bs, 1H), 9.80 (bs,
3H), 9.65 (bs, 1H), 7.29−7.19 (m, 12H), 7.00−6.98 (m, 6H), 5.15 (s,
2H), 5.04 (s, 8H), 4.84 (bs, 1H), 4.39 (s, 2H), 3.84−3.70 (m, 18H),
1.29 (s, 12H). GPC: Mn = 1.2 kDa, Mw = 1.7 kDa, PDI = 1.44.
Triethylamine (0.12 mL, 0.86 mmol, 10 equiv) was added dropwise
to a solution of oligomer 9 (0.10 g, 86 μmol, 1.0 equiv) and 4-
nitrophenyl isocyanate (71 mg, 0.43 mmol, 5.0 equiv) in
dimethylformamide (1.7 mL). The reaction mixture was stirred for
16 h at 23 °C, after which the solvent was removed by rotary
evaporation. The residue was washed using a solid phase washing
vessel with methanol (3×) followed by acetonitrile (2×). The solids
were dried under vacuum for 12 h to give oligomer 5 as a peach-
colored powder (0.81 g, 61 μmol, 71%): 1H NMR δ (360 MHz,
SO(CH3)2) 10.45 (bs, 1H), 9.90 (bs, 1H), 9.80 (bs, 4H), 8.19 (d, 2H,
J = 9.3 Hz), 7.68 (d, 2H, J = 9.1 Hz), 7.40−7.21 (m, 12H), 6.99 (d,
6H, J = 8.3 Hz), 5.15 (s, 2H), 5.09−5.03 (m, 10H), 3.83−3.75 (m,
18H), 1.29 (s, 12H). GPC: Mn = 1.4 kDa, Mw = 1.7 kDa, PDI = 1.3.
Preparation of Oligomer 6. Compound 8 (0.60 g, 2.2 mmol, 1.0
equiv) was added in one portion to stirring dimethyl sulfoxide (2.2
mL) at 110 °C. Dibutyltin dilaurate (0.26 mL, 0.44 mmol, 0.2 equiv)
was added in one portion and the reaction mixture was stirred for 5.0
min at 110 °C. 4-(Hydroxymethyl)-3-methoxyphenylboronic acid
pinacol ester (2.0 g, 7.6 mmol, 3.5 equiv) was added in one portion
and the reaction mixture was stirred for 2 h at 110 °C. The reaction
mixture was cooled to 23 °C and poured into 0 °C methanol (20 mL).
A yellow precipitate formed that was washed using a solid phase
washing vessel with methanol (3 × ). The solids were dried under
vacuum for 12 h to give oligomer 10 as a light yellow powder (0.30 g,
Figure 5. Expanded view of the device shown in Figure 1a. The device
is 10 mm wide × 10 mm long × 0.9 mm thick.
paper disk (2.5 mm-diameter ×180 μm thick) soaked in HEPES buffer
(45 mM, pH 8.0) and dried. Disks were loaded with HEPES by adding
600 μL of HEPES buffer to 320 mg of disks and dried under vacuum.
Disks were Whatman Chromatography Paper grade I paper fabricated
using an Epilog Mini 24 Laser (CO2 laser). Layers 3 and 5 were loaded
with 0.25 μL of one of the hydrophobic detection reagents (i.e., 1−4)
dissolved in EtOAc (compounds 5 and 6 were dissolved in THF). A
solution of a hydrophobic detection reagent was spotted using a
Drummond 0.25 μL disposable micropipet. Layer 6 contains a paper
disk soaked in green dye. The disks were loaded with the green dye by
adding 600 μL of green food coloring (1:5 food coloring−deionized-
water) to 300 mg of disks followed by drying the disks under vacuum.
After assembly, the devices were pressed using a rolling pin, applying
medium pressure.
Procedure for Measuring Flow-Through. We measured the
time required for a sample to flow through the device in Figure 5 as
follows: to layer 1 was added 8 μL of an aqueous solution of H2O2. A
timer was started immediately upon addition of the sample to the
device. The device was turned over so that layer 7 was visible. The
flow-through time was recorded when the hydrophilic region of layer 7
had completely changed color. Fourteen replicate tests were
performed for each concentration of H2O2 and the two highest and
two lowest flow-through times were removed from the data set to
account for errors arising from failures during the device fabrication
procedure.
1
0.18 mmol, 64%): H NMR δ (400 MHz, SO(CH3)2) 9.89 (bs, 1H)
9.78 (bs, 6H), 9.63 (bs, 1H), 7.41−7.19 (m, 18H), 7.03−6.95 (m,
9H), 5.16 (s, 2H), 5.04 (s, 14H), 4.82 (t, 1H, J = 5.6 Hz), 4.40 (d, 2H,
J = 5.7 Hz), 3.94−3.67 (m, 27H), 1.30 (s, 12H). GPC: Mn = 1.3 kDa,
Mw = 2.3 kDa, PDI = 1.73.
Triethylamine (0.16 mL, 1.2 mmol, 10 equiv) was added dropwise
to a solution of oligomer 10 (0.20 g, 0.12 mmol, 1.0 equiv) and 4-
nitrophenyl isocyanate (97 mg, 0.59 mmol, 5.0 equiv) in
dimethylformamide (2.4 mL). The reaction mixture was stirred for
16 h at 23 °C, after which the solvent was removed by rotary
evaporation. The residue was washed using a solid phase washing
vessel with methanol (3×) followed by acetonitrile (2×). The solids
were dried under vacuum for 12 h to give oligomer 6 as a peach-
1
colored powder (0.19 g, 0.11 mmol, 92%): H NMR δ (360 MHz,
Procedure for Measuring Depolymerization Kinetics. p-
Dioxanes (250 μL), dimethyl sulfoxide (190 μL), and phosphate
buffered water (40 μL, 0.01 M, pH 7.1) were added to a 2 mL vial and
mixed by swirling the solution. A solution containing the oligomer (10
μL from a 0.01 M solution in DMSO) was added to the vial and
vortexed for 5 s. Hydrogen peroxide (10 μL from a 0.2 M solution in
phosphate buffered water, 0.01 M, pH 7.1) was added and the
combined solution was aspirated using a pipet. The solution was
transferred to a quartz cuvette (500 μL, 0.1 cm path length) and the
absorbance value at 385 nm was monitored continuously. Half-lives
were calculated based on the relative quantity of released p-
nitroaniline, using the method described in reference7.
SO(CH3)2) 10.46 (bs, 1H), 9.91 (bs, 1H) 9.80 (bs, 6H), 8.20 (d, 2H, J
= 9.1 Hz), 7.69 (d, 2H, J = 9.2 Hz), 7.38−7.22 (m, 18H), 7.01−6.99
(m, 9H), 5.16 (s, 2H), 5.09−5.04 (m, 16H), 3.85−3.76 (m, 27H), 1.30
(s, 12H). GPC: Mn = 1.7 kDa, Mw = 1.9 kDa, PDI = 1.2.
Procedure for Fabricating the Paper-Based Microfluidic
Device. The paper was patterned using a wax printer according to the
procedures described in refs 6 and12. The wax was melted into the
paper by placing the patterned paper in an oven at 150 °C for 105 s.
The devices were assembled according to the procedures in ref 11; the
layout of the device is shown in Figure 5. The paper used for all layers
of wax-patterned paper in the devices was Boise Aspen 30 Printer
Paper (92 brilliant, 30% postconsumer content). Layer 2 contains a
F
dx.doi.org/10.1021/ma4007413 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
(23) Wong, A. D.; DeWit, M. A.; Gillies, E. R. Adv. Drug Delivery Rev.
2012, 64, 1031−1045.
ASSOCIATED CONTENT
* Supporting Information
Tables of primary data, effect of number of layers on sensitivity,
calculation of limits-of-detection, depolymerization kinetics,
and figures of NMR spectra and GPC traces. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
S
(24) Weinstain, R.; Sagi, A.; Karton, N.; Shabat, D. Chem.Eur. J.
2008, 14, 6857−6861.
(25) Because only ∼0.5 μg quantities of 2−6 are used in the paper-
based microfluidic device, Raman spectroscopy was unable to
distinguish background signal from paper from signal arising from
the oligomers. Likewise, reflectance measurements targeting the color
produced by the released p-nitroaniline upon depolymerization were
unsuccessful due to the limited quantity of color produced during
flow-through.
AUTHOR INFORMATION
Corresponding Author
*E-mail: (S.T.P.) sphillips@psu.edu.
■
(26) A logical extension of this work would involve adding two
methyl ether groups ortho to the benzylic leaving group on each
repeating unit in the oligomer. In previous work focused on the
controlled release of phenols, we demonstrated that incorporation of
two methyl ethers enhanced the rate of release of phenol 35−fold
compared to an analogous controlled release reagent that incorporated
only one methyl ether.19 In the context of carbamate oligomers,
however, the benzylic leaving group is a carbamic acid rather than a
phenol, and thus is a better leaving group. The consequence is that
incorporation of two methyl ethers in carbamate oligomers caused
substantial background decomposition that not only made their
synthesis difficult, but also prevented their use in the assay.
(27) de Gracia Lux, C.; McFearin, C. L.; Joshi-Barr, S.;
Sankaranarayanan, J.; Fomina, N.; Almutairi, A. ACS Macro Lett.
2012, 1, 922−926.
Author Contributions
‡These authors contributed equally. The manuscript was
written through contributions of all authors. All authors have
given approval to the final version of the manuscript.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported in part by the NSF (CHE-1150969),
the Arnold and Mabel Beckman Foundation, the Camille and
Henry Dreyfus Foundation, and Louis Martarano. S.T.P.
acknowledges support from the Alfred P. Sloan Research
Fellows Program.
(28) Chen, E. K. Y.; McBride, R. A.; Gillies, E. R. Macromolecules
2012, 45, 7364−7374.
REFERENCES
■
(29) Seo, W.; Phillips, S. T. J. Am. Chem. Soc. 2010, 132, 9234−9235.
(30) Zhang, H.; Yeung, K.; Robbins, J. S.; Pavlick, R. A.; Wu, M.; Liu,
R.; Sen, A.; Phillips, S. T. Angew. Chem., Int. Ed. 2012, 51, 2400−2404.
(31) DiLauro, A. M.; Abbaspourrad, A.; Weitz, D. A.; Phillips, S. T.
Macromolecules 2013, 46, 3309−3313.
(1) Yager, P.; Domingo, G. J.; Gerdes, J. Annu. Rev. Biomed. Eng.
2008, 10, 107−144.
(2) Yager, P.; Edwards, T.; Fu, E.; Helton, K.; Nelson, K.; Tam, M.
R.; Weigl, B. H. Nature 2006, 442, 412−418.
(3) Giljohann, D. A.; Mirkin, C. A. Nature 2009, 462, 461−464.
(4) Urdea, M.; Penny, L. A.; Olmsted, S. S.; Giovanni, M. Y.; Kaspar,
P.; Shepherd, A.; Wilson, P.; Dahl, C. A.; Buchsbaum, S.; Moeller, G.;
Hay Burgess, D. C. Nature 2006, 444, 73−79.
(32) DiLauro, A. M.; Robbins, J. S.; Phillips, S. T. Macromolecules
2013, 46, 2963−2968.
(33) Ortiz, V.; Rubio, M. A.; Lissi, E. A. Atmos. Environ. 2000, 34,
1139−1146.
(5) Peeling, R. W.; Holmes, K. K.; Mabey, D.; Ronald, A. Sex.
Transm. Infect. 2006, 82, v1−6.
́
(34) Claeys, M.; Wang, W.; Ion, A. C.; Kourtchev, I.; Gelencser, A.;
Maenhaut, W. Atmos. Environ. 2004, 38, 4093−4098.
(35) Taniai, T.; Sakuragawa, A.; Okutani, T. Anal. Sci. 2000, 16, 275−
281.
(6) Lewis, G. G.; DiTucci, M. J.; Phillips, S. T. Angew. Chem., Int. Ed.
2012, 51, 12707−12710.
(7) Robbins, J. S.; Schmid, K. M.; Phillips, S. T. J. Org. Chem. 2013,
78, 3159−3169.
(36) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518−1520.
(37) Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923−
2925.
(38) O’Neil, M. J.; Heckelman, P. E.; Koch, C. B.; Roman, K. J. The
Merck Index, 14th ed.; Merck Co., Inc.: Whitehouse Station, NJ, 2006.
(8) Sagi, A.; Weinstain, R.; Karton, N.; Shabat, D. J. Am. Chem. Soc.
2008, 130, 5434−5435.
(9) Esser-Kahn, A. P.; Sottos, N. R.; White, S. R.; Moore, J. S. J. Am.
Chem. Soc. 2010, 132, 10266−10268.
(10) DeWit, M. A.; Gillies, E. R. J. Am. Chem. Soc. 2009, 131, 18327−
18334.
(11) Martinez, A. W.; Phillips, S. T.; Whitesides, G. M. Proc. Natl.
Acad. Sci. U.S.A. 2008, 105, 19606−19611.
(12) Noh, H.; Phillips, S. T. Anal. Chem. 2010, 82, 4181−4187.
(13) Noh, H.; Phillips, S. T. Anal. Chem. 2010, 82, 8071−8078.
(14) Unlike many standard colorimetric assays, the intensity of the
green color is not relevant to the quantitative readout. Instead, it is
used as a visual indicator that the assay has reached completion.
(15) Lippert, A. R.; Van de Bittner, G. C.; Chang, C. J. Acc. Chem.
Res. 2011, 44, 793−804.
(16) Cho, D.-G.; Sessler, J. L. Chem. Soc. Rev. 2009, 38, 1647−1662.
(17) Efforts to vary the reagent in order to detect and quantify
analytes other than hydrogen peroxide are currently in progress.
(18) Defined as (3 × σ0)/slope of the linear exponential region of the
response.
(19) Schmid, K. M.; Jensen, L.; Phillips, S. T. J. Org. Chem. 2012, 77,
4363−4374.
(20) Scrimin, P.; Prins, L. J. Chem. Soc. Rev. 2011, 40, 4488−4505.
(21) Esser-Kahn, A. P.; Odom, S. A.; Sottos, N. R.; White, S. R.;
Moore, J. S. Macromolecules 2011, 44, 5539−5553.
(22) Peterson, G. I.; Larsen, M. B.; Boydston, A. J. Macromolecules
2012, 45, 7317−7328.
G
dx.doi.org/10.1021/ma4007413 | Macromolecules XXXX, XXX, XXX−XXX