Article

pubs.acs.org/ac

Paper Machine for Molecular Diagnostics

John T. Connelly,*, Jason P. Rolland, and George M. Whitesides

Diagnostics For All, 840 Memorial Drive, Cambridge, Massachusetts 02139, United States
Department of Chemistry and Chemical Biology, Harvard University, Cambridge, Massachusetts 02138, United States

S Supporting Information
*

ABSTRACT: Clinical tests based on primer-initiated amplication of specic nucleic acid sequences achieve high levels of
sensitivity and specicity. Despite these desirable characteristics,
these tests have not reached their full potential because their
complexity and expense limit their usefulness to centralized
laboratories. This paper describes a device that integrates sample
preparation and loop-mediated isothermal amplication (LAMP)
with end point detection using a hand-held UV source and
camera phone. The prototype device integrates paper microuidics (to enable uid handling) and a multilayer structure, or a
paper machine, that allows a central patterned paper strip to
slide in and out of uidic path and thus allows introduction of
sample, wash buers, amplication master mix, and detection
reagents with minimal pipetting, in a hand-held, disposable device intended for point-of-care use in resource-limited
environments. This device creates a dynamic seal that prevents evaporation during incubation at 65 C for 1 h. This interval is
sucient to allow a LAMP reaction for the Escherichia coli malB gene to proceed with an analytical sensitivity of 1 doublestranded DNA target copy. Starting with human plasma spiked with whole, live E. coli cells, this paper demonstrates full
integration of sample preparation with LAMP amplication and end point detection with a limit of detection of 5 cells. Further, it
shows that the method used to prepare sample enables concentration of DNA from sample volumes commonly available from
ngerstick blood draw.

NAATs intended for use in resource-limited settings must

operate within the constraints of those environments. One
diagnostic technology that has been successfully implemented
at the point of care is lateral ow assays, particularly lateral ow
immunochromatographic strips.6 These systems require minimal sample preparation and no instrumentation. They typically
provide a visual result within 10 to 20 min.6 Tests for direct or
indirect detection of a range of pathogens (including
Trypanosoma cruzi,7 diphtheria toxin,8 Plasmodium falciparum,9
and Hemophilis ducreyi10) have been reported and tested and
are available in resource-limited settings. Lateral ow assays
have also been used for end point detection of PCR and
isothermal amplication products, either from reactions
conducted using standard equipment1114 or integrated into
microuidic systems.1517 Though lateral ow end point
detection can simplify the nal readout of NAATs, the
instrumentation required (often prohibitively) complicating
factors, from sample preparation through amplication, and
may include open-channel microuidic systems with external
pumps.
Despite the obstacles to implementation of NAATs at the
point of care, a number of systems are being developed for this

he unparalleled sensitivity and specicity of nucleic acid

amplication tests (NAATs), led by those based on
polymerase chain reaction (PCR), have become increasingly
available in the developed world for the diagnosis of infectious,
genetic, and other diseases. In developing countries, these tests
are much less used because of their cost, complexity, and
requirements for expensive equipment and reagents. PCR
requires three precisely controlled temperatures to enable
melting of target DNA, annealing, and elongation of primers
and is typically carried out using expensive thermocyclers.
Additionally, prior to amplication, NAATs usually require
sample preparation involving a series of steps to release the
DNA and to remove the many inhibitory compounds found in
clinical samples.1,2 These steps often require the use of complex
laboratory equipment and, therefore, highly trained personnel.
This combination of factors, expensive and delicate equipment,
trained personnel, and complex procedures, has kept NAATs
from being broadly adopted in resource-limited countries
(although they are, of course, used to limited extents in
centralized hospitals and government laboratories). Because
transporting patients and/or samples to these central facilities
can be dicult, slow, and expensive in resource-limited
environments with under-developed transportation systems,
local healthcare outposts and traveling healthcare workers are
often the primary, or only, providers of healthcare.36
2015 American Chemical Society

Received: January 30, 2015

Accepted: June 23, 2015
Published: June 23, 2015
7595

DOI: 10.1021/acs.analchem.5b00411
Anal. Chem. 2015, 87, 75957601

Article

Analytical Chemistry

isothermal amplication, and detection; the cost of materials for

the device is US $0.60, and the total cost per assay is US $1.83
(Table S1, Supporting Information). These costs, of course, do
not indicate the cost of a fully developed product based on this
type of system, but they do indicate that the cost of materials
will not limit aordability. Figure 1a shows the multilayer

type of application. The Cepheid GeneXpert (Sunnyvale, CA)

is a notable current example. The GeneXpert reduces user
handling of samples by integrating sample preparation steps
with PCR amplication and real-time uorescence detection in
the same cartridge.18 Though meeting the ideal of a sample-in,
answer-out device, the GeneXpert has complex optical and
electronic systems, required for real-time PCR, and pumps to
move sample and reagents through the cartridge and is still not
a general solution for resource-limited environments.19
We have chosen to use isothermal amplication methods in
the system described here. These methods simplify molecular
diagnostics by eliminating thermocycling and reducing equipment costs and power requirements, without sacricing
sensitivity or specicity. In particular, loop-mediated isothermal
amplication (LAMP)20,21 has a high tolerance for common
components present in clinical samples that inhibit PCR and
other reactions.2225 This robust tolerance toward inhibitors
further streamlines assays, because it simplies sample
preparation and complexity/expense of equipment.26 Although
devices integrating LAMP amplication with detection have
entered the market (for example, the IllumiGene system;
Meridian Bioscience, Cincinnati, OH), they have not directly
incorporated the steps necessary for sample preparation into
the instrument.27 Further, as products targeted toward
developed world settings, they employ complex optical systems
and user interfaces that increase cost and, especially, the
requirements for maintenance and repair.
By merging the useful characteristics of lateral-ow assays
with those of open-channel microuidic systems, paper
microuidic platforms using three-dimensional stacked patterned paper2831 or two-dimensional networks of cut
paper3234 allow complex uid handling without the need for
external pumps and power. Paper as a substrate for diagnostics
has other benecial properties as well: (i) It is biocompatible
(and has been used for decades for storage of clinical samples in
the form of dried blood spots and is also widely used as a
chromatography medium); (ii) Because DNA adsorbs to
cellulose,35 paper provides a simple matrix for its capture and
isolation; (iii) The cellulose ber network provides a high
surface area and facilitates storage of dried reagents.36 A paperbased test for liver function (Diagnostics For All) has been
successful in eld trials; this test provides semiquantitative
measurements of liver transaminases from whole blood.37,38
Despite these advances, there have, however, been no
demonstrations of NAATs in complete systems (blood to
read-out) in paper microuidic systems.
In important steps toward the objective of a complete
system, Rohrman and Richards-Kortum have demonstrated
recombinase polymerase amplication (RPA) in a paper-based
device (using sample preparation and detection as separate
steps).39 The Ellington group demonstrated detection of
LAMP products in paper microuidic devices using their
catalytic hairpin assembly (CHA) reaction, which uses a series
of specially designed oligonucleotide hairpins to specically
identify the target and provide an additional signal enhancement. Here, the amplication reaction was conducted ochip.40
Our goal is to develop a NAAT device that is compact and
inexpensive and fully integrates sample preparation, amplication, and detection in a low-cost, single-use format appropriate
for point-of-care diagnostics in resource-limited settings. To
this end, we describe a prototype paper microuidic device that
enables sample preparation from clinically relevant matrices,

Figure 1. Schematic of sliding-strip device. The exploded view (a)

shows the three major layers of the sliding-strip architecture and their
components. The assembled device (b), which was used throughout
this work, accommodates a single sample or control. An image of a
prototype concept device (c) contains three reactions discs, one
sample and a negative (NC) and positive (PC) control, with ports to
add the samples, wash buer, amplication Master Mix (MM), and
detection reagents and visual the signal (Read).

architecture of the device, which we hereafter refer to as a

sliding strip; this design allows for the middle layer, the strip,
containing the reaction vessel, a paper disc in which reaction
and washing steps occur, to slide into dierent uidic paths.
The system is, thus, a paper machine: that is, a device
fabricated in paper, with moving parts, which accomplishes a
task.
The linear motion of the sliding strip, relative to the
stationary surrounding structures, acts as a valve to control the
serial introduction of sample, wash buer, amplication
reagents, and detection reagents, while also dynamically sealing
to prevent evaporation during incubation steps. The sample
preparation steps are simplied by fabricating the reaction disc
out of Whatman FTA paper, which contains a proprietary blend
of lytic reagents dried into the cellulose matrix. FTA paper has
been shown to lyse a wide range of cells and viruses,4143 and
in its normal use, after extensive washing in tubes with
vortexing, it can be added directly to PCR reactions. Our device
replaces this instrumented, multistage, high-volume washing
with two, through-ow washes before the introduction of
LAMP master mix. Also, by exploiting the hydrophilicity of
paper, we simplify uid handling by replacing external
mechanical systems for pumping with uid ow driven by
capillary action. Final detection of the amplied product uses
SYBR Green I, a uorescent intercalating dye, and a hand-held
UV source and camera phone. For a proof-of-concept, we used
LAMP to amplify a portion of the malB gene common to all
Escherichia coli.44
7596

DOI: 10.1021/acs.analchem.5b00411
Anal. Chem. 2015, 87, 75957601

Article

Analytical Chemistry

MATERIALS AND METHODS

Oligonucleotide primers were ordered from Eurons Operon
MWG (Huntsville, AL) with HPLC purication and suspended
in 1 Tris-EDTA. Bst 2.0 DNA polymerase, 10 Isothermal
Amplication Buer, dNTPs, DNA/DNase-free stocks of
MgSO4, and restriction endonucleases were purchased from
New England Biolabs (Ipswich, MA), and betaine was from
Sigma-Aldrich (St. Louis, MO). Pooled human plasma in
sodium heparin derived from whole blood donations was
acquired from Valley Biomedical Products and Services, Inc.
(Winchester, VA). Precast 2% agarose gels and SYBR Green I
were purchased from Invitrogen. All other reagents were
supplied by Fisher Scientic.
Fabrication of Magnetic Sliding-Strip Devices and
Conrmation of EvaporationResistance. The magnetic
material used to fabricate the devices is 0.75 mm thick synthetic
rubber-bonded ferrite magnetic sheeting with 3 mm bands of
alternating poles on both faces obtained from McMaster-Carr
(Robbinsville, NJ). The magnetic sheets, as well as all laminate
layers, were cut using a design le from Adobe Illustrator and a
Graphtec CraftRoboPro (Graphtec Irvine, CA) knife plotter. A
thin lm of Krytox LVP lubricant (DuPont Wilmington, DE) is
applied with a roller to the inner faces of the top and lower
layers as well as to both sides of the sliding strip. The reaction
disc is 4.75 mm diameter and consists of Whatman FTA paper
while the wash vias are 6.35 mm diameter discs cut from
Ahlstromm 226 paper. The various layers were stacked,
laminated, and pressed as illustrated in Figure 1a.
The attractive force produced by the magnetic material forms
an evaporation-resistant seal in combination with the thin lm
of lubricant applied to the layer interfaces. This seal was tested
by applying 10 L of water to the paper reaction disc within a
sliding-strip device, which was predried to remove any ambient
moisture, and slid to seal. The mass of the device before and
after the addition of the water was recorded to serve as the
baseline. This device was then incubated at 65 C for 1 h,
weighed, and slid to unseal the disc. The device was returned to
the incubator to completely evaporate any water remaining, and
the nal mass was recorded. The masses of the device after
incubation and after complete evaporation were used to
determine the amount of water lost during incubation by
comparison to the baseline.
E. coli Culture. We used E. coli BL21(DE3)pLysS as a
model organism in this work. Cells were cultured overnight at
37 C, on a shaker platform, from frozen glycerol stocks in LB
broth containing 34 g/mL chloramphenicol. We determined
the concentration of cells by measuring the optical density at
600 nm and comparing this value to a growth curve. When
directly used to spike sample, the culture was centrifuged, the
supernatant aspirated, and the pellet resuspended to the desired
concentration with fresh media, twice, to remove any free
DNA. This stock of cells was then serially diluted in human
plasma for experiments.
LAMP Amplication. The primer set (Table S2,
Supporting Information) targeting the malB gene of E. coli,
including loop primers, as described by Hill et al. was employed
in this work.44 LAMP was performed using the Bst 2.0 DNA
Polymerase, an in silico designed homologue of Bst DNA
Polymerase I, Large Fragment, commonly used in this reaction,
and its supplied buer.45 In transitioning the reaction from
tubes to paper, the concentration of betaine, primers, and
enzyme were optimized, as was the incubation time, to

minimize nonspecic reactions and maximize sensitivity, as

shown in Figure S1 and discussed in the Supporting
Information. The nal reactions contained 1 isothermal
amplication buer (10 mM Tris-HCl, 10 mM (NH4)2SO4, 50
mM KCl, 2 mM MgSO4, 0.1% Tween-20) supplemented with
an additional 6 mM MgSO4, 0.9 M betaine, 1.4 mM of each
dNTP, 1.6 M of each inner primer (FIP, BIP), 0.8 M of each
loop primer (LF, LB), 0.2 M of each outer primer (F3, B3),
and 0.32 U of polymerase. The optimal incubation was 65 C
for 60 min. The reaction was always run with two negative
controls, a no-template control (NTC) and a no-amplication
control (NA+), containing a heat-inactivated enzyme and 1000
copies of target or 1000 cells, as appropriate. Reactions in tubes
were carried out at 25 L; in paper, the reaction volume was 10
L.
Product Analysis by Gel Electrophoresis. To conrm
that positive signals were the result of specic, template-driven
reactions, LAMP products were analyzed by agarose gel
electrophoresis (AGE). First, a positive reaction conducted in
a tube using 1000 copies of the target was digested using
restriction endonuclease PvuII, which recognizes a sequence in
region F1 of the target, and run on a 2% agarose gel with a
DNA ladder and the undigested product to determine if the
observed LAMP banding pattern was the result of templatespecic amplication and to mark said banding pattern as the
specic reaction product for future comparison.
General Sliding-Strip Operation. The sliding-strip
devices enable the serial operations of sample preparation,
including cell lysis, DNA isolation, and purication, as well as
LAMP amplication and detection. Unless otherwise noted, the
procedure for sliding-strip assays, illustrated in Figure 4a, began
with the application to the reaction disc of a 10 L sample
through the sample port. The strip was then slid to align the
reaction disc with the wash port, and 40 L of FTA purication
buer and 80 L of nuclease-free water were sequentially
applied and allowed to transit through the disc to the paper vias
and nally to the removable wash pad. The disc was then dried
completely by being placed in an oven at 65 C for 5 min; the
strip was slid to align the disc with the amplication reagent
port, and 10 L of LAMP Master Mix was applied. The strip
was then slid again, sealing the disc (by the magnetic force and
the lubricant lm) within the amplication zone. The device
was placed in an incubator at 65 C. After a 60 min incubation
time, the devices were removed from the incubator and the
strip was slid to the detection window; the disc was dried
completely, again at 65 C for 5 min. For detection, 10 L of
100 SYBR Green I in 1 TE was applied to the disc and it
was excited using a hand-held shortwave UV source and imaged
using a camera phone. Images were then processed, converted
to grayscale, and analyzed using ImageJ to quantify mean pixel
intensity as gray value. The mean of the NTC replicates plus 3
times their standard deviation was used as the LOD, and the
threshold value, for determining positive and negative reactions.
For ease of imaging and data presentation of multiple samples
and replicates, the reaction discs were removed from the device
and arrayed before imaging.
Determination of Analytical Sensitivity. The analytical
sensitivity of amplication and detection in the sliding-strip
device was determined using a double-stranded 200 bp analog
of the malB target sequence produced by Integrated DNA
Technologies, Inc. (Coralville, IA) as two single-stranded DNA
Ultramers, the forward strand and the reverse complement.
These strands were then combined to yield an equimolar
7597

DOI: 10.1021/acs.analchem.5b00411
Anal. Chem. 2015, 87, 75957601

Article

Analytical Chemistry
solution, heated to 95 C for 5 min, and slowly cooled to room
temperature to anneal. The concentration was conrmed by
spectrometry using the absorbance at 260 nm. Serial dilutions
were prepared, and 10 L was applied to dry reaction discs,
which had been pretreated with 10 L of human plasma, 40 L
of FTA purication buer, and 80 L of nuclease-free water,
through the amplication reagent port of the sliding-strip
devices. The discs were then dried completely, and the general
sliding-strip operations outlined above were followed starting
from the application of LAMP Master Mix.
Integrated Assay with Spiked Plasma Samples. To
examine the eectiveness of the sliding-strip device for sample
preparation from complex matrices, the assay was challenged
with samples of human plasma spiked with whole, live E. coli
prepared from overnight culture as described above. All steps in
the general sliding-strip operation outlined previously were
followed. Here, parameters such as the number of washes, wash
volumes, and wash buer composition were optimized as
shown Figure S2, Supporting Information.
Collecting and Concentrating DNA from Large
Sample Volumes. To examine the ability to isolate DNA
from larger volumes of dilute samples, we applied 10, 20, 30,
40, or 50 L of a 0.05 cell/L sample of E. coli in human plasma
to the reaction discs. After each sample wicked through the
reaction disc, the general sliding-strip operation procedure was
followed as described.

Figure 2. Demonstration of evaporationresistance. The magnetic

substrate used to fabricate the device, in conjunction with the thin lm
of lubricant, creates an evaporation-resistant seal, with an average loss
of less than 0.02 L of 10 L water applied (n = 4) after 1 h of heating
at 65 C.

the combination of incubation time and betaine concentration

eliminated nonspecic amplications.
To assess the performance of the LAMP reaction when
conducted in buer contained within a paper disc that is
housed in the sliding strip, we examined serial dilutions of the
200 bp dsDNA target sequence in nuclease-free water. By
drying these samples onto reaction discs, which had been
treated with human plasma and washed, we replicated the
condition of the discs that would exist in the full assay prior to
amplication and isolated potential amplication inhibition
from sample preparation loses, both of which could yield false
negative results. As shown in Figure 3a, the reaction proceeds
uninhibited and reactions starting with as few as 1 copy of the
double stranded target yield a statistically signicant signal over
the no template control (NTC) (n = 8, p < 1 108), with
positive results from all eight replicates. On the basis of a most
probable number (MPN) analysis (described in the Supporting
Information), this dilution would have an MPN of 1.93 copies/
disc, which as both strands of the target can serve as a template
for amplication, is an expected result for this double stranded
target. As is evident in Figure 3b, a camera phone image of the
reaction discs under UV excitation, these optimal conditions
and idealized sample produce a binary result, common with end
point detection of isothermal amplication reactions.
Though no sequence homology exists between primers in
the set, in suboptimal conditions, mismatches may be tolerated.
To conrm that uorescent signal is the result of the specic
template-driven reaction, a restriction endonuclease analysis
was performed as described in the Supporting Information and
shown in Figure S4.
Integrated Assay with Spiked Plasma Samples. The
assay integrated with the sliding-strip device was challenged
with whole, live E. coli serially diluted in human plasma. The
samples were subjected to all of the steps of sample
preparation, in the order as described in the Materials and
Methods section, by linear actuation of the sliding strip to the
appropriate ports for application of the wash buer, LAMP
Master Mix, and detection reagents, as detailed in Figure 4a.
The optimized washing conditions, detailed in Figure S2,

RESULTS AND DISCUSSION

Fabrication of Sliding-Strip Devices with EvaporationResistant Dynamic Seals. Nucleic acid-based diagnostic tests
require serial operations to be performed on a complex sample
in order to yield reliable results. DNA must be released from
cells, isolated from the cell debris, proteins, and other sample
components, puried of these potential polymerase inhibitors,
amplied, and detected. Our device is designed to perform
these operations, typically done using laboratory equipment,
with minimal user steps and at a fraction of the cost. The core
architecture of the sliding-strip device, illustrated in Figure 1a,b,
consists of three major layers: the top layer, which provides
ports to access a reaction disc; the middle layer, which consists
primarily of the reaction disc housed in the sliding strip, which
is linearly actuated to serially address the ports; and the bottom
layer, which contains paper vias to contact the reaction disc and
the wash pad to transit waste wash out of the device. Each of
these layers is fabricated out of a low-cost, exible magnetic
substrate, the attractive force of which, when combined with
the thin lm of lubricant, creates a dynamic seal that allows for
extended incubation at 65 C with negligible evaporation
(Figure 2). As shown in Figure 1c (not used in the experiments
described), these devices are small, only 0.225 cm thick, and the
prototype device, capable of running one sample and two
controls, is 4.7 cm wide and 10 cm from the top to the end of
the strip-pull when it is in the starting position and will
ultimately include an inexpensive electronic heater. This, in
combination with on-board storage of lyophilized reagents,
enables point-of-care use.
Analytical Sensitivity in Paper. Prior to assessing the
analytical sensitivity of the LAMP reaction in the paper
machine, we investigated two DNA polymerases (Figure 1a)
and optimized the incubation time (Figure S1b, Supporting
Information) and betaine concentration (Figure S1c, Supporting Information). The selected DNA polymerase provided
improved amplication at lower concentrations of target, and
7598

DOI: 10.1021/acs.analchem.5b00411
Anal. Chem. 2015, 87, 75957601

Article

Analytical Chemistry

Figure 3. Analytical sensitivity of LAMP in sliding-strip device. All

reactions containing as few as 1 double-stranded copy of the 200 bp
dsDNA target yielded detectable amplication (a). Positive replicates
are those producing signals greater than the limit of detection (LOD)
calculated as the average plus three times the standard deviation of the
no template control (NTC) (n = 8). Average signal data is presented
in Figure S3, Supporting Information. The data is generated by
photographing the discs under UV excitation (b) after application of
10 L of 100 SYBR Green I and quantifying the mean gray values
with ImageJ. To control for potential variations in exposure, all
reaction discs were removed from their devices, arrayed, and imaged,
yielding the image shown.

Figure 4. Assay integrated with sliding-strip device using spiked

plasma. The entire bioanalytical assay, from sample preparation
through amplication and detection, was integrated with the slidingstrip devices. Visualized in cross-section (a), the sample consisting of
whole, live E. coli spiked into human plasma was applied to the
reaction disc, and then washing, reagent addition, incubation, and
detection were controlled in series via linear actuation of the sliding
strip. As before, the replicates were binned as positive and negative
(b), on the basis of mean gray values quantied via ImageJ (Figure S5,
Supporting Information) (n = 6).

Supporting Information, were 40 L of FTA purication buer

followed by 80 L of nuclease-free water.
Upon imaging, positive results were obtained from four of six
samples containing as few as 1 cell (Figure 4b); this outcome is
consistent with stochastic sample error as predicted by the
MPN method (described in the Supporting Information). As
shown in Figure S5, Supporting Information, statistical
signicance (p < 0.001) is achieved for all concentrations
over 5 cells per sample (e.g., 0.5 cells/L or 500 cells/mL).
Collecting and Concentrating DNA from Larger
Sample Volumes. Our intended use of a product based on
this system includes direct application of 30 L of nger-stick
blood, the average volume obtained from a safety lancet. As the
limit of detection of 500 cells/mL was achieved using a 10 L
sample volume, we investigated the eects of larger sample
volumes. By varying the volume of a sample containing 0.05

cells/L (an order of magnitude lower than the above

established limit of detection) in increments of 10 L applied
to dierent reaction discs, as shown in Figure 5, we noted an
increase in the number of replicates resulting in positive
amplication with 10 and 20 L volumes (approximately 0.5
and 1 cell/disc, respectively) producing no positive signals, 40
L (approximately 2 cells/disc) producing 7 positive replicates,
and 60 L (approximately 3 cells/disc) yielding positive signals
7599

DOI: 10.1021/acs.analchem.5b00411
Anal. Chem. 2015, 87, 75957601

Article

Analytical Chemistry

Using a dsDNA analog of the target region within the malB

gene of E. coli, we demonstrated the LAMP reaction can be run
within the pores of the cellulose ber matrix of a paper disc
with an analytical sensitivity of a single copy of the doublestranded target sequence. When integrated with the entire
procedure for sample preparation, and beginning from whole,
live E. coli cells in human plasma, we were able to consistently
detect 5 cells, and the presence of the organism was detected
on 4 out of 6 samples containing 1 cell.
We also demonstrated the ability to concentrate target in the
reaction disc by applying larger sample volumes. This ability to
concentrate may permit less precise metering of sample in a
product based on this system (e.g., direct application of a
nger-stick blood drop), making it more user-friendly and
eliminating the need for pipettors. Further, the use of larger
sample volumes may improve detection at low concentrations
by increasing the probability of capturing the target in the
sample volume.
The current prototype uses an inexpensive detection scheme,
excitation of an intercalating uorophore with a UV source and
camera phone-based imaging, that can be adapted directly to a
marketable product. This system replaces complicated and
expensive optics, such as those found in a uorimeter. The
requirement of even this simple equipment, and the incubator
required to heat the device, may limit use in some
environments. To address this potential limitation, current
work is focused on replacing this detection scheme with one
that produces a colorimetric signal visible with the unaided eye
in ambient light. Similarly, we are designing and testing a lowcost, electronic heater built-in to the device to eliminate the
incubator.
The use of liquid amplication reagents in this proof-ofconcept will prevent application in resource-limited environments as they have a limited shelf life in the absence of
refrigeration. Ongoing work is focused on lyophilizing these
reagents on-board to address this problem.
A key metric in the development of this device is cost; in
order for a product based on this system to be widely adopted
in resource-limited countries, it must be aordable. This paper
machine, as presented, is low-cost through its use of paper
microuidics and an isothermal amplication reaction with end
point detection. Through continued development of this
system, we envision a total-disposable, noninstrumented device
that can dramatically reduce the cost of nucleic acid diagnostic
tests for infectious diseases, genetic conditions, and cancer and
bring it to the point-of-care in resource-limited settings.

Figure 5. Collection and concentration of target from larger sample

volumes. Samples of 0.05 cells/L E. coli in human plasma were
applied in 10 L increments to reactions discs in sliding-strip devices
and processed through the integrated sample preparation, amplication, and detection procedures of the device (n = 8). Signals from all
replicates of 10 and 20 L (0.5 and 1 cell applied to discs) samples
remained below the LOD, determined as the mean plus 3 times the
standard deviation of a NTC run in parallel, and were thus scored as
negative. With 40 and 60 L samples (2 and 3 cells), the number of
replicates scoring positive, above the LOD, increased to 7 and 8,
respectively. Sample volumes over 60 L began to reduce the number
of positive replicates likely as a result of washing eects. Average signal
and average positive signal data is shown in Figure S6, Supporting
Information.

in all 8 replicates. This demonstrates that the additional cells

present in the larger volume samples are being successfully
lysed and their DNA captured by the cellulose bers of the
paper reaction discs. This concentration eect is limited.
Volumes over 60 L degrade these gains, with 80 and 100 L
volumes (approximately 4 and 5 cells/disc, respectively)
dropping to 5 and 3 positive replicates, respectively. This
decline is likely due to washing eects; both washing away of
the dried lysis chemistry of the FTA paper disc, preventing the
release of additional target, and washing away of previously
captured target.

SUMMARY AND CONCLUSIONS

We designed, fabricated, and tested an inexpensive technology
for molecular diagnostics that enables sample preparation steps
of cell lysis, DNA isolation, concentration, and purication
integrated with LAMP amplication and detection. The basis
for fabrication uses a patterned strip that contains the reaction
mixture in a paper disc and slides this disc from station to
station to accomplish addition of sample, lysis of cells (if
relevant), adsorption of DNA, washing, and addition of
reagents. This simple sliding motion of the strip in the
paper machine, in conjunction with the ability of paper to
cause the transport of water by capillary wicking, replaces most
of the equipment needed to conduct NAATs on the benchtop:
centrifuges, vortex mixers and, in a consumer product,
pipettors. The assembly is held together using synthetic
rubber-bonded ferrite magnets and laminates. The compressive
force of the magnets on the strip and thin lm of lubricant
between the layers creates a dynamic seal that resists
evaporation and allows for incubation of the amplication
reaction at 65 C for 1 h.

ASSOCIATED CONTENT

S Supporting Information
*

Table S1: Bill of Materials for Device; Table S2: LAMP Primer
Set; Figure S1: Optimization of LAMP Reaction Conditions in
FTA Paper Discs; Figure S2:Optimization of Wash-Based
Purication; Figure S3: Average Signals for Analytical
Sensitivity of LAMP; Figure S4: Restriction Endonuclease
Digest for Specicity Conrmation; Figure S5: Average Signals
for Integrated Assay; Figure S6: Average Signals for Increased
Sample Volume. The Supporting Information is available free
of charge on the ACS Publications website at DOI: 10.1021/
acs.analchem.5b00411.
7600

DOI: 10.1021/acs.analchem.5b00411
Anal. Chem. 2015, 87, 75957601

Article

Analytical Chemistry

(24) Kaneko, H.; Kawana, T.; Fukushima, E.; Suzutani, T. J. Biochem.

Biophys. Methods 2007, 70, 499501.
(25) Wang, F.; Jiang, L.; Ge, B. J. Clin. Microbiol. 2012, 50, 9197.
(26) Boehme, C. C.; Nabeta, P.; Henostroza, G.; Raqib, R.; Rahim,
Z.; Gerhardt, M.; Sanga, E.; Hoelscher, M.; Notomi, T.; Hase, T.;
Perkins, M. D. J. Clin. Microbiol. 2007, 45, 19361940.
(27) Boyanton, B. L.; Sural, P.; Loomis, C. R.; Pesta, C.; GonzalezKrellwitz, L.; Robinson-Dunn, B.; Riska, P. J. Clin. Microbiol. 2011, 50,
640645.
(28) Carrilho, E.; Martinez, A. W.; Whitesides, G. M. Anal. Chem.
2009, 81, 70917095.
(29) Martinez, A. W.; Phillips, S. T.; Butte, M. J.; Whitesides, G. M.
Angew. Chem., Int. Ed. Engl. 2007, 46, 13181320.
(30) Martinez, A. W.; Phillips, S. T.; Whitesides, G. M. Proc. Natl.
Acad. Sci. U. S. A. 2008, 105, 1960619611.
(31) Martinez, A. W.; Phillips, S. T.; Whitesides, G. M.; Carrilho, E.
Anal. Chem. 2009, 82, 310.
(32) Fu, E.; Lutz, B.; Kauffman, P.; Yager, P. Lab Chip 2010, 10,
918920.
(33) Lutz, B. R.; Trinh, P.; Ball, C.; Fu, E.; Yager, P. Lab Chip 2011,
11, 42744278.
(34) Osborn, J. L.; Lutz, B.; Fu, E.; Kauffman, P.; Stevens, D. Y.;
Yager, P. Lab Chip 2010, 10, 26592665.
(35) Alberts, B. M.; Amodio, F. J.; Jenkins, M.; Gutmann, E. D.;
Ferris, F. L. Cold Spring Harbor Symp. Quant. Biol. 1968, 33, 289305.
(36) Yetisen, A. K.; Akram, M. S.; Lowe, C. R. Lab Chip 2013, 13,
22102251.
(37) Pollock, N. R.; Rolland, J. P.; Kumar, S.; Beattie, P. D.; Jain, S.;
Noubary, F.; Wong, V. L.; Pohlmann, R. A.; Ryan, U. S.; Whitesides,
G. M. Sci. Transl. Med. 2012, 4, 152ra129.
(38) Pollock, N. R.; McGray, S.; Colby, D. J.; Noubary, F.; Nguyen,
H.; Nguyen, T. A.; Khormaee, S.; Jain, S.; Hawkins, K.; Kumar, S.;
Rolland, J. P.; Beattie, P. D.; Chau, N. V.; Quang, V. M.; Barfield, C.;
Tietje, K.; Steele, M.; Weigl, B. H. PLoS One 2013, 8, No. e75616.
(39) Rohrman, B. A.; Richards-Kortum, R. R. Lab Chip 2012, 12,
30823088.
(40) Allen, P. B.; Arshad, S. A.; Li, B.; Chen, X.; Ellington, A. D. Lab
Chip 2012, 12, 29512958.
(41) Li, C.-C.; Beck, I. A.; Seidel, K. D.; Frenkel, L. M. J. Clin.
Microbiol. 2004, 42, 38473849.
(42) Picard-Meyer, E.; Barrat, J.; Cliquet, F. J. Virol. Methods 2007,
140, 174182.
(43) Turse, J. E.; Pei, J.; Ficht, T. A. Front. Microbiol. 2011, 2, 5466.
(44) Hill, J.; Beriwal, S.; Chandra, I.; Paul, V. K.; Kapil, A.; Singh, T.;
Wadowsky, R. M.; Singh, V.; Goyal, A.; Jahnukainen, T.; Johnson, J.
R.; Tarr, P. I.; Vats, A. J. Clin. Microbiol. 2008, 46, 28002804.
(45) Tanner, N. A.; Zhang, Y.; Evans, T. C., Jr. Biotechniques 2012,
53, 8189.

AUTHOR INFORMATION

Corresponding Author

*E-mail: jconnelly@dfa.org. Tel: +1 617 494 0700. Fax: +1 617

494 1162.
Notes

The authors declare no competing nancial interest.

ACKNOWLEDGMENTS
This work was supported by the Defense Advanced Research
Projects Agency (Grant Number HR0011-12-2-0010).

REFERENCES

(1) Al-Soud, W. A.; Radstrom, P. J. Clin. Microbiol. 2001, 39, 485

493.
(2) Wilson, I. G. Appl. Environ. Microbiol. 1997, 63, 37413751.
(3) Cordray, M. S.; Richards-Kortum, R. R. Am. J. Trop. Med. Hyg.
2012, 87, 223230.
(4) LaBarre, P.; Hawkins, K. R.; Gerlach, J.; Wilmoth, J.; Beddoe, A.;
Singleton, J.; Boyle, D.; Weigl, B. PLoS One 2011, 6, No. e19738.
(5) Mabey, D.; Peeling, R. W.; Ustianowski, A.; Perkins, M. D. Nat.
Rev. Microbiol. 2004, 2, 231240.
(6) Weigl, B. H.; Boyle, D. S.; de los Santos, T.; Peck, R. B.; Steele,
M. S. Expert Rev. Med. Devices 2009, 6, 461464.
(7) Barfield, C. A.; Barney, R. S.; Crudder, C. H.; Wilmoth, J. L.;
Stevens, D. S.; Mora-Garcia, S.; Yanovsky, M. J.; Weigl, B. H.;
Yanovsky, J. IEEE Trans. Biomed. Eng. 2011, 58, 814817.
(8) Engler, K. H.; Efstratiou, A.; Norn, D.; Kozlov, R. S.; Selga, I.;
Glushkevich, T. G.; Tam, M.; Melnikov, V. G.; Mazurova, I. K.; Kim,
V. E.; Tseneva, G. Y.; Titov, L. P.; George, R. C. J. Clin. Microbiol.
2002, 40, 8083.
(9) Labbe, A. C.; Pillai, D. R.; Hongvangthong, B.; Vanisaveth, V.;
Pomphida, S.; Inkathone, S.; Hay Burgess, D. C.; Kain, K. C. Ann.
Trop. Med. Parasitol. 2001, 95, 671677.
(10) Patterson, K.; Olsen, B.; Thomas, C.; Norn, D.; Tam, M.;
Elkins, C. J. Clin. Microbiol. 2002, 40, 36943702.
(11) Aveyard, J.; Mehrabi, M.; Cossins, A.; Braven, H.; Wilson, R.
Chem. Commun. 2007, 42514253.
(12) Connelly, J. T.; Nugen, S. R.; Borejsza-Wysocki, W.; Durst, R.
A.; Montagna, R. A.; Baeumner, A. J. Anal. Bioanal. Chem. 2008, 391,
487495.
(13) Soliman, H.; El-Matbouli, M. Mol. Cell. Probes 2010, 24, 3843.
(14) Mens, P. F.; van Amerongen, A.; Sawa, P.; Kager, P. A.; Schallig,
H. D. F. H. Diagn. Microbiol. Infect. Dis. 2008, 61, 421427.
(15) Chen, D.; Mauk, M.; Qiu, X.; Liu, C.; Kim, J.; Ramprasad, S.;
Ongagna, S.; Abrams, W. R.; Malamud, D.; Corstjens, P. L.; Bau, H. H.
Biomed. Microdevices 2010, 12, 705719.
(16) Chen, Z.; Mauk, M. G.; Wang, J.; Abrams, W. R.; Corstjens, P.
L. A. M.; Niedbala, R. S.; Malamud, D.; Bau, H. H. Ann. N.Y. Acad. Sci.
2007, 1098, 429436.
(17) Roskos, K.; Hickerson, A. I.; Lu, H.-W.; Ferguson, T. M.;
Shinde, D. N.; Klaue, Y.; Niemz, A. PLoS One 2013, 8, No. e69355.
(18) Boehme, C. C.; Nabeta, P.; Hillemann, D.; Nicol, M. P.; Shenai,
S.; Krapp, F.; Allen, J.; Tahirli, R.; Blakemore, R.; Rustomjee, R.;
Milovic, A.; Jones, M.; OBrien, S. M.; Persing, D. H.; Ruesch-Gerdes,
S.; Gotuzzo, E.; Rodrigues, C.; Alland, D.; Perkins, M. D. N. Engl. J.
Med. 2010, 363, 10051015.
(19) Evans, C. A. PLoS Med. 2011, 8, No. e1001064.
(20) Nagamine, K.; Hase, T.; Notomi, T. Mol. Cell. Probes 2002, 16,
223229.
(21) Notomi, T.; Okayama, H.; Masubuchi, H.; Yonekawa, T.;
Watanabe, K.; Amino, N.; Hase, T. Nucleic Acids Res. 2000, 28, e63.
(22) Ihira, M.; Akimoto, S.; Miyake, F.; Fujita, A.; Sugata, K.; Suga,
S.; Ohashi, M.; Nishimura, N.; Ozaki, T.; Asano, Y.; Yoshikawa, T. J.
Clin. Virol. 2007, 39, 2226.
(23) Iseki, H.; Kawai, S.; Takahashi, N.; Hirai, M.; Tanabe, K.;
Yokoyama, N.; Igarashi, I. J. Clin. Microbiol. 2011, 48, 25092514.
7601

DOI: 10.1021/acs.analchem.5b00411
Anal. Chem. 2015, 87, 75957601