Biosensors and Bioelectronics ] (]]]]) ]]]]]]

Contents lists available at SciVerse ScienceDirect

Biosensors and Bioelectronics

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

Macro to microuidics system for biological environmental monitoring

Cyril Delattre a, Cedric P. Allier a, Yves Fouillet a,n, Dorothee Jary a, Frederic Bottausci a, Denis Bouvier a,
Guillaume Delapierre a, Manuelle Quinaud a, Arnaud Rival a, Laurent Davoust b, Christine Peponnet a
a

CEA, LETI, MINATEC Campus, Microtechnologies for Biology and Healthcare Department, 17 rue des Martyrs, F-38054 Grenoble Cedex 9, France
Electromagnetic Processing of Materials (EPM) Group, Materials and Processes Science and Engineering Laboratory (SIMaP), Grenoble Institute of Technology (Grenoble INP),
38402 Saint Martin dHe res cedex, France
b

a r t i c l e i n f o

abstract

Article history:
Received 18 January 2012
Received in revised form
30 March 2012
Accepted 13 April 2012

Biological environmental monitoring (BEM) is a growing eld of research which challenges both
microuidics and system automation. The aim is to develop a transportable system with analysis
throughput which satises the requirements: (i) fully autonomous, (ii) complete protocol integration
from sample collection to nal analysis, (iii) detection of diluted molecules or biological species in a
large real life environmental sample volume, (iv) robustness and (v) exibility and versatility. This
paper discusses all these specications in order to dene an original uidic architecture based on three
connected modules, a sampling module, a sample preparation module and a detection module. The
sample preparation module highly concentrates on the pathogens present in a few mL samples of
complex and unknown solutions and puries the pathogens nucleic acids into a few mL of a controlled
buffer. To do so, a two-step concentration protocol based on magnetic beads is automated in a reusable
macro-to-micro uidic system. The detection module is a PCR based miniaturized platform using digital
microuidics, where reactions are performed in 64 nL droplets handled by electrowetting on dielectric
(EWOD) actuation. The design and manufacture of the two modules are reported as well as their
respective performances. To demonstrate the integration of the complete protocol in the same system,
rst results of pathogen detection are shown.
& 2012 Elsevier B.V. All rights reserved.

Keywords:
Biological environmental monitoring
Electrowetting
Sample preparation
Real-time PCR

1. Introduction
Sensitive and specic biological environmental monitoring
(BEM) are still far from eld applications, even though daily news
updates remind us that the need is real. Indeed, in industrialized
countries, nosocomial infection occurs in 2%12% of hospitalized
patients (Jarvis et al., 1991) of which a great part is probably due
to airborne transmission of pathogens (Cotterill et al., 1996;
Gehanno et al., 2009; Schultsz et al., 2003). In hospital rooms
for immuno-depressed patients, detection thresholds in the air
are very low, with values down to 100 cfu/m3 for Streptococcus
pneumoniae and 1 cfu/m3 (cfu: colony-forming unit) for Aspergillus.
Other environmental controls could benet from more frequent
monitoring, such as sanitary analysis for swimming conditions
in beaches, lakes and rivers (notably fecal coliforms, Enterococcus spp.,
cyanobacteria producing toxins like microcystin, Cryptosporidium
oocysts, Giardia). European legislation 1 has specied detection

Corresponding author. Tel.: 33 438 789 274; fax: 33 438 782 343.
E-mail address: yves.fouillet@cea.fr (Y. Fouillet).
1
Directive 2006/7/EC of the European Parliament and of the Council of
15 February 2006 concerning the management of bathing water quality and
repealing directive 76/160/EEC OJ L 64, p. 46.

thresholds to dene the water quality unsafe for swimming, i.e.

185 cfu/L for Enterococcus and 250 cfu/L for Escherichia coli.
Todays normative methods dened by the safety regulations
are slow and costly. Aqueous samples, typically 110 L, are
manually collected and sent to a central lab where a microbiological test is performed by cultivating and counting bacteria on a
Petri dish. It typically takes 24 days to have the result with a
regulatory approved control. Even if quicker analysis methods,
e.g. PCR based detection, are used to reduce the analysis time, a
huge part of the delay and cost is due to the manual sampling and
sample preparation. As a result BEM are performed at limited
frequency, usually once a month, at the most once per week.
Therefore, BEM could tremendously benet from fully automated
systems (from sampling to data) performing in a semi-continuous
mode several analyses per day and having a complete autonomy
of 1 or 2 weeks. In order to be statistically representative, typical
samples are 1100 m3 for air quality analysis and several liters for
water analysis. Thus, elevated concentration factor and purication yield of biological targets coupled to high detection sensitivity and specicity are a requisite.
In order to address these requirements, complex systems have
been developed in conjunction with devoted biological protocols.
In most cases they combine a sample preparation step and a

0956-5663/$ - see front matter & 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.bios.2012.04.024

Please cite this article as: Delattre, C., et al., Macro to microuidics system for biological environmental monitoring. Biosensors and
Bioelectronics (2012), http://dx.doi.org/10.1016/j.bios.2012.04.024

C. Delattre et al. / Biosensors and Bioelectronics ] (]]]]) ]]]]]]

detection step. The latter are commonly based on specic

molecular recognition of biological molecules, either proteins
(immunodetection) or nucleic acids (molecular biology). Immunoassays are particularly adapted for analysis when speed and
simplicity are preferred over sensitivity, as they require very
limited sample preparation before analysis (Rider et al., 2003;
Schultz et al., 2008). When high sensitivity is needed, real time
quantitative Polymerase Chain Reaction (qPCR) is the preferred
detection method. Extreme specicity and sensitivity, as low as
1 cfu or nucleic acid molecule, can be obtained using PCR based
detection. The Autonomous Pathogen Detection System (APDS)
developed by the Lawrence Livermore National Laboratory is the
rst fully automated tool integrated in the US Biowatch program.
It integrates both immunodetection and PCR capability, coupled
with an adapted sample preparation module (Regan et al., 2008).
The BioHazard Water Analyser is based on extraction of RNA from
pathogens, followed by identication using electrochemical
detection on a DNA chip. It can detect down to 110 pathogens
in 10 L of water within 23 h. Both systems integrate every
analytical step from collection to detection. However, they weigh
around 200 kg, which is a drawback for portable and in-the-eld
applications. As shown by the GeneXperts (Cepheid), a promising approach is to reduce both weights and footprints by
integrating the protocol at a microuidic scale. Sample preparation and qPCR are performed using a unique cartridge. Despite
excellent analytical results, it is limited to one-shot measurement
without any capability for semi-continuous monitoring.
To best meet the specic needs of BEM we have been
developing an autonomous system integrating generic modules
for onsite sampling, sample preparation and PCR based detection.
This fully automated and portable system will feature short timeto-result (0.51 h), multi-pathogens (113) identication and
quantication. In this paper we describe a rst system that
couples sample preparation and PCR detection. A PCR detection
module working in small volumes of few tens nL is demonstrated.
It represents several important advantages for the BEM applications, such as capability of multiplex analysis, minimized reagents
and waste, increased autonomy and portability. Yet, this places
a strong constraint on the sample preparation module which
has to handle volumes differing by three orders of magnitude,
from few mL input sample down to few mL. We present also an
original biological protocol based on magnetic beads capture
maximizing the concentration factor before PCR. This protocol
is integrated into an automated sample preparation module,
which is successfully coupled to the PCR detection module.
Their common automation demonstrates the feasibility of
macro to microuidic system for BEM. It will be further coupled
to a sample collection module depending on the application, i.e.
air or water sampling.

2. Material and methods

2.1. Biological models and analytical protocols
We have developed a DNA sample preparation protocol, based
on magnetic beads, to address the different needs for integration e.g.
very large concentration factor, generic reagent for pathogens
capture, no enzyme, no temperature variation and small volumes
of reagents. A generic capture was chosen because it allows multipathogens detection, e.g. virus, bacteria and spores. However, the
counterpart is the potential carryover of PCR inhibitors present in
the large sample volume which are also concentrated. To overcome
this difculty, we have worked on an original double capture
protocol based on two types of magnetic beads implying different

chemical interactions. The probability of having carryover inhibitors

binding strongly to both types of beads is thus very low.
The rst step of the pathogen capture is done on polycationic
beads #1 (SIMAG-ionex  1 mm beads, Chemicell). Typically 1
10 mL sample volume is mixed with 110 mL of beads #1
corresponding to 108109 beads. Up to 108 microorganisms can
be captured on the beads #1 in just a few minutes. Beads #1 are
then pelleted by a magnet and the supernatant is removed. 50 mL
of lysis buffer (5 M Guanidine thiocyanate, 1% N-Lauryl Sarcosine
sodium salt, 20 mM TrisHCl pH 8) is added on beads #1, which
are then dispersed. The lysis step occurs in 2 min. Then the beads
#1 are pelleted allowing the supernatant to be removed. The
latter is mixed with 200 mL of DNA capture buffer (Qiagen PB
bufferRef. 19066) and 2.5 mL of silanol beads #2 (SIMAG-basic
 1 mm beads, Chemicell) for the nucleic acids purication. Beads
#2 are pelleted and washed twice with the washing buffer
(Qiagen AW2 bufferRef. 19072) for 1 min. The beads are then
resuspended in 110 mL of elution buffer (10 mM TrisHCl, pH 8).
Elution step occurs in 3 min. Beads #2 are captured and removed
from the solution containing the extract of puried nucleic acids.
Otherwise noted, for biological validation the sample volume is
1 mL and the elution volume is 10 mL. This sample preparation
protocol is performed at room temperature within 20 min, with
no centrifugation steps and only reagents addition and removal.
The puried DNA from the sample preparation module is then
mixed with the PCR common reagents.
E. coli, Bacillus subtilis and S. pneumoniae strains have been
purchased from ATCC (ATCC 9637 (AmpR), ATCC-33608, ATCC
BAA-255). Human adenovirus2 (Ad2) and baculovirus were generously provided by Dr Pascal Fender (Institut de Biologie Structurale
de Grenoble, France). Cultures are made in LB buffer at 37 1C
overnight for E. coli and B. subtilis, and at 30 1C for S. pneumoniae.
Bacteria concentration was established by standard optical absorbance measurement at 600 nm in PBS buffer and correlated with
colony counting on agar plates. Mock samples were prepared by
spiking a known quantity of bacteria in 110 mL of 10 mM TrisHCl
pH 8, when not specied. The Qiagen QIAamp DNA Mini kit
(Ref. 51306) was chosen as the reference sample preparation
method for DNA extraction and purication yield comparison.
Reference DNA was purchased from Sigma Aldrich for E. coli and
from ATCC for the other bacterial strains. Quantication of Ad2 and
Baculovirus was obtained by directly running qPCR with the viruses.
Primers and probe (Taqman and MGB) molecules were purchased
from Applied Biosystems for S. Pneumoniae and from MWG for the
other strains. Specicity for each biological strain was validated
against all other strains. Final PCR mixture contains 0.25 U/mL of
AmpliTaq Gold and 1X PCR Gold buffer (Applied Biosystem), 3 mM
MgCl2, 200 mM of dNTP, 0.8 mM BSA, 600 nM of primers and probes.
A single amplication condition has been used for all the biological
models, consisting of an initial step of 10 min at 95 1C, followed by
40 cycles of 30 s. denaturation at 92 1C and 60 s annealing at 60 1C.
For all sets of primers and probes, PCR sensitivity of at least  10
DNA copies was measured with the standard PCR reference method,
performed on a MxPro 3005P Stratagene apparatus (Agilent) in
10 mL PCR mixture. Standard qPCR curves were made by 10 fold
serial dilutions of Tris buffer DNA (or viruses) solutions starting with
1051 copies/mL concentrations in PCR mixture. PCR efciency is
deduced with the following formula (10(1/slope)  1)100, using the
slope of the linear relationship between Ct (qPCR cycle threshold)
and Log (DNA concentration).
2.2. Sample preparation module
The automated sample preparation module performs all the
biological protocol steps in a unique polypropylene chamber of
20 mL (1.4 mm internal) that can handle the required 10 mL input

Please cite this article as: Delattre, C., et al., Macro to microuidics system for biological environmental monitoring. Biosensors and
Bioelectronics (2012), http://dx.doi.org/10.1016/j.bios.2012.04.024

C. Delattre et al. / Biosensors and Bioelectronics ] (]]]]) ]]]]]]

volume. As depicted in Figs. 1 and 2a, the module further includes

a one axis moveable Phynox needle ( 0.3 mm internal, Minitubes), a moveable magnet (1.4 T), and an ultrasound device.
Fluidic handling is ensured by standard tubing ( 1.6 mm external, PFA) for large volume manipulation, while the needle
addresses volumes down to mL range (e.g. reagents, elution).
The uidic network, described in Fig. 1a, features several pumps
(P1P4, Instech P625/66.133) and isolation valves (NRsearch
225T). P1 introduces the sample (110 ml) from the collection
module into the sample vessel where it is mixed with 110 mL of
beads #1. After 1 min, the sample is pumped by P2 into the 20 mL
preparation chamber and pumped out by P3 through the needle
set at the bottom of the chamber. The liquid travels in front of the
magnet at a ow rate of 0.5 mL/min, slow enough to capture

 100% of the magnetic beads. This leads to a quick decrease in

reaction volume, a concentration factor of 250 can be obtained
(5 mL/20 mL). In order to disperse the compact magnetic beads
pellet (Fig. 2b) we have developed a specic ultrasound device
(30 kHz). It is able to disperse up to 109 beads, within a volume of
550 mL (Fig. 2c). The lysis buffer is then added in the chamber.
After 30 s of beads mixing, the beads #1 are collected and the
supernatant is transferred in an external tube, sitting in the
reagents storing part of the system, containing beads #2 with
binding buffer. In the mean time the chamber is washed with
water to remove the beads #1. This new mixing is aspirated in the
chamber with beads collection during liquid ow. For the remaining steps of the protocol, washings and elution, uidic operations
consisting in liquid addition and removal are performed with P3
connected to the Phynox needle addressing multiple reagent
reservoirs. At the end of the sample preparation protocol, a
510 mL volume of puried DNA extracts is obtained. It is then
mixed with the same volume of PCR reagents and ushed through
the needle into the transfer chamber lled with oil. The PCR
mixture is moved into the rst EWOD chip reservoir using P4
connected to the chip outlet. Its arrival into the EWOD PCR chip is
automatically detected by the optical set-up of the detection
module.

2.3. PCR detection module

Fig. 1. Integrated platform: schematic of liquid volume handling from input

sample to detection in droplets (a). The overall system has a footprint of 30 cm
 50 cm  40 cm (b). It integrates a preparation and a detection module. The
preparation module is described in Section 2.2. It reduces the analytical volume
from 1 mL10 mL and puries the DNA starting from bacteria or viruses. Sample
and reagents are actuated by conventional pumps and valves, and all the
biological protocol steps are performed in a unique reaction chamber of 20 mL
(c). The detection module (described in Section 2.3) is a PCR based miniaturized
platform where droplets of few tens nL droplets are dispensed, moved and mixed
using EWOD actuation (d).

Fig. 2. Concentration step in the preparation chamber: schematic of the chamber,

magnet, ultrasound device, needle and compact magnetic beads pellet (a). Pictures
of the real sample preparation chamber before (b) and after (c) the beads pellet
dispersion.

EWOD (ElectroWetting On Dielectric) based digital microuidics was chosen for the PCR detection module. Recent reviews
(Fair, 2007; Malic et al., 2010) report on numerous biochemical
protocols performed on EWOD chips, including PCR, enzymatic
reaction (Jary et al., 2006), immunoassay (Sista et al., 2008),
protein analysis (Wheeler et al., 2004). By electrically controlling
the effective wettability of a droplet (Berge, 1993) EWOD is able
to handle submicroliter droplet with operations such as displacement, mixing or dispensing (Cho et al., 2003; Pollack et al., 2002).
Complete protocols including sample preparation were already
successfully integrated in different EWOD platform, but are
limited to sample volume of few microliters and mainly for POC
application. In comparison with existing EWOD systems, we
focused our work on developing an EWOD PCR chip dedicated
to BEM, In particular, this chip is ready-to-use, prelled with oil,
features an airtight packaging, embedded specic primers and
probes (at this time in 13 single-use PCR reaction chambers).
Furthermore the chip needs to be compatible with the liquid
loading from the sample preparation module.
The PCR detection module consists of a EWOD microuidic
chip (Fig. 1d) coupled with a dedicated instrument. The latter was
developed in order to simultaneously mechanically clamp the
EWOD PCR chip and make both electrical and uidic connections
in a robust way. The EWOD PCR chip (22  22 mm2) is based on a
standard and already published conguration (Fouillet et al.,
2008; Delattre et al., 2008; Malk et al., 2011). It is made of four
parts: an active silicon chip, a 100 mm spacer, a cover plate with
through holes and a uidic interface. It can perform every basics
operation such as reservoir loading, droplet dispensing, displacement and mixing. These operations are achieved with a network
of 130 electrodes (800  800 mm2), controlled with only 40
independent electrical contacts. The chip is designed with 13
reaction chambers, located around a central bus (Fig. 1a). The
specic qPCR reagents (probes and primers) are deposited in each
reservoir using a nanoliter dispensing system (Packard BioChipArray) on substrate size up to 200 mm. A 32.4 nL volume
containing reagents at twice the targeted concentration is dispensed with a piezoelectric pipette generating 300 pL droplets.
Reagents are dried under atmospheric conditions within a few

Please cite this article as: Delattre, C., et al., Macro to microuidics system for biological environmental monitoring. Biosensors and
Bioelectronics (2012), http://dx.doi.org/10.1016/j.bios.2012.04.024

C. Delattre et al. / Biosensors and Bioelectronics ] (]]]]) ]]]]]]

minutes. More detail on the EWOD chip fabrication can be found

in the online Supplementary information.

3. Results and discussion

3.1. Sample preparation
The sample preparation protocol developed for BEM applications was proven versatile, generic, sensitive and robust. It has
been evaluated manually on 50 data points corresponding to
7 different experiments with 104 bacteria E. coli or B. subtilis per
1 ml sample. The purication yield of the manual sample preparation protocol was calculated by comparing the Ct obtained on
the puried DNA sample with the expected Ct given by the
standard qPCR curve. All PCR measurements have been performed
with Stratagene apparatus. The same method was used to assess
the presence of PCR inhibitors by spiking the nal PCR mixture
with DNA from another strain. Here we deal only with inhibitors
resulting from the protocol itself, e.g. beads fragments and
chemical agents, and not from inhibitors that may be present in
real sample matrix. The purication yield of our dedicated sample
preparation reached typical value around 50%, comparable to the
manual reference Qiagen sample preparation kit that we used as
reference. The detection limit for each strain was estimated
between 10 and 100 pathogens per sample. It was dened by
the lowest DNA quantity for which a PCR amplication occurred.
Robustness of the purication protocol was evaluated by processing 1 mL mixed samples containing a large excess of one strain
(106 bacteria E. coli) and 3 other strains (adenovirus2, S. pneumoniae
and B. subtilis) with varying quantity (0, 1, 10, 102, 103 and 104
pathogens). Purication yields reaches about 50%, 740% corresponding to a Ct of about 23.571, with mixed sample, similar to
what obtained from single strain sample.
Using the automated sample preparation module, this protocol
allows processing a sample every hour. Between each analysis, all
uidic components are lled with a 2.6% sodium hypochlorite
aqueous solution for 20 min and then thoroughly rinsed with
water to prevent cross-contamination. Thorough controls have
been performed showing that neither cross-contamination nor
inhibition was found after this decontamination procedure. The
sample preparation reproducibility was established by performing the sample preparation on 1 mL of buffer containing 104 E. coli
bacteria. The yield reached values of about 20%, 715% corresponding to a Ct of about 25.3 71 over 84 experiments (see
Fig. 3).
3.2. Detection module
The EWOD elementary uidic steps to be tested are: (i) droplet
dispensing and displacement along the bus towards given locations,
(ii) dried reagents dissolution in specic droplets (Fig. 4ad). Using
an actuating voltage of 60 Vrms, droplet dispensing reproducibility
with CVs (Coefcient of Variation: standard deviation/mean
value100) of 0.5% for a given EWOD chip and 3% for independent
EWOD chips were measured, indicating that the dispensing step
is highly repeatable. The presence of a permanent thin oil lm
between the droplet and the hydrophobic surface during electrowetting actuation has already been described (Srinivasan et al.,
2004). Although it limits adsorption of molecules, thus minimizing contamination risk, it could be an obstacle to the dry reagent
dissolution. Fortunately this is not the case, the dried reagent spot
immediately dissolves inside the droplet of PCR mixture (Fig. 4e).
Indeed, the dried reagent spot produce a wetting defect; as a
consequence, the thin oil lm instantaneously breaks upon arrival
of the droplet. It was previously reported (Paik et al., 2003) that the

Fig. 3. Stratagene analysis of sample preparation experiments performed with the

automated system and in tube for samples of 15 mL spiked with 4  104 E. coli
bacteria. The elution is performed in 10 mL for all experiments and then diluted
twice with PCR reagents leading to an expected concentration of 2  103 copies/mL
in PCR tubes (Ct of 22.5).

Fig. 4. 64 nL uorescent droplets are dispensed from the inlet reservoir (a) and
(b). Droplets are moving back and forth (c) and (d) for enhanced mixing. Dried
reagent re-suspension, complete mixing is achieved after ten back and forth
EWOD displacements (e).

mixing of two droplets can be accelerated with a short back and

forth displacement along consecutive electrodes. Similarly, the dried
reagent dissolution was characterized during back and forth displacements by measuring the uorescence homogeneity of the
droplet. Complete dissolution and mixing was achieved with less
than ten back and forth displacements and within 10 s. All qPCR
were run on the detection module with the following conditions:
7 min activation at 92 1C, followed by 45 cycles with denaturation
step at 92 1C for 6 s. and annealing step at 60 1C for 30 s. Overall
cycling time is of roughly 40 min but has not been optimized.
During qPCR cycles uorescence of each droplet is simultaneously
measured at every cycle by capturing images of the overall chip with
a CCD camera. An example of an image is given in Fig. 5 at the end of
a qPCR assay. The stack of images is analyzed with homemade
software, and the plotted uorescence PCR curves (Figs. 5 and 6)
were corrected for baseline but not normalized.
First validations were achieved by running qPCR with the same
PCR mixtures simultaneously on the detection module (64 nL)
and on the Stratagene apparatus (10 mL). In both cases, the cycle
threshold (Ct) value was calculated for each PCR curve using the
second derivative method from the raw uorescence data (LuuThe et al., 2005). Both the efciency and the Ct values are identical
in 64 nL on EWOD PCR chip and with the reference method

Please cite this article as: Delattre, C., et al., Macro to microuidics system for biological environmental monitoring. Biosensors and
Bioelectronics (2012), http://dx.doi.org/10.1016/j.bios.2012.04.024

C. Delattre et al. / Biosensors and Bioelectronics ] (]]]]) ]]]]]]

as uidic handling, temperature control or uorescence homogeneity of the excitation signal.

In order to check the stability of on-chip embedded reagents,
an adenovirus DNA sample at 105 copies/ml was amplied on
EWOD PCR chip using embedded primers and probe stored at 4 1C
for 40 days. In parallel, a regular PCR was performed on the same
chip using solubilised primers and probe with the same DNA
sample (Fig. S.3). Very similar results were obtained with Ct
values of 18 and 17 respectively for dried and solubilised primers
and probe. Same level of uorescence signal was found for both
conditions showing no degradation of the sensitive Taqman probe
during the packaging process or during storage conditions in oil
for 40 days. Finally, same results were obtained on three different
primers and probe sets demonstrating the versatility towards
different oligonucleotide sequences.
Fig. 5. qPCR curves obtained on the EWOD PCR detection module for the
Baculovirus sample amplication at 5000, 500, 50 and 0 copies/mL corresponding
to 320, 32 and 4 copies in the 64 nL droplet. Same experiments in 10 mL reactions
were performed on Stratagene apparatus, leading to Ct values of respectively 24.9,
28.3, 31.5 for 5000, 500 and 50 copies/mL. Ct values obtain with the Stratagene
apparatus and the EWOD PCR chip are very similar. The photography shows the
uorescent droplets at the end of this qPCR assay.

Fig. 6. Detection results obtained with the integrated system: qPCR curves on a
positive control with 104 DNA molecules of E. coli and on puried DNA after
complete processing of 1 mL of 104 bacteria of E. coli (uorescence signals are
normalized with the end-point value).

3.3. System validation and further integration

Finally, the integrated system was tested on the complete
protocol. A 1 mL sample of 10 mM TrisHCL pH 8 buffer containing 104 E. coli bacteria was processed using the sample preparation module coupled with the detection module. Identical Ct of
24.8 was obtained with both embedded and solubilized reagents.
A PCR mixture with reference E. coli DNA was entered into the
detection module as a positive control and leads to a Ct of 23.6. A
purication yield around 50% was estimated from this Ct shift
(Fig. 6).
We also performed measurements with samples obtained with
commercial air collection module (SKC BioSamplers). We have to
further integrate this air collection module in the platform and
couple it with the inlet of the sample preparation module. These
results prove that all the biological protocols needed for pathogens quantication were successfully automated. Moreover, an
identical DNA purication yield was obtained with both classical
methods and integrated system. This work is original as the use of
a unique and automated system handling 1 mL buffer with
pathogens and taking advantages of a microuidic EWOD chip
for PCR detection has not yet been reported.

4. Conclusion
(Fig. 5). These results indicate that the PCR runs with the same
performances on both systems.
However, when comparing DNA copies numbers per PCR
reaction, 50 copies/mL corresponds to only 4 copies in a 64 nL
droplet but to 500 copies on the Stratagene reference. Nevertheless, on-chip detection is clear and non ambiguous with a Ct of
32.5 despite the very low copy number. This clearly shows that,
even with large and dilute samples, robust PCR detection can be
obtained with the combination of an elevated concentration
factor as well as the contribution of microuidic scale. The at
negative control curves perfectly show the absence of cross
contamination between droplets, even though the different positive and negative control droplets are successively displaced
along a common bus prior to PCR, when loading the chip.
Then, qPCR reproducibility of several droplets on the same
EWOD PCR chip was compared to multiple well reactions on the
Stratagene apparatus. 19 PCR droplets were amplied in parallel.
Mean Ct values for EWOD PCR chip and Stratagene apparatus
were respectively of 19.9 and 19. Both had an equivalent 2% CV on
their mean Ct value and feature equivalent PCR reproducibility.
Moreover, when the same experiment was repeated on 3 EWOD
PCR chips, a 2.5% CV was obtained on the Ct value demonstrating
a good overall chip to chip reproducibility. These results reect
the great optimization that was achieved on the detection
instrument regarding every aspect inuencing PCR results, such

In order to meet the requirements for BEM (automation,

complete integration from sample collection to nal analysis,
handling of highly diluted environmental sample), we have
developed an integrated system coupling sample preparation
and PCR-based detection modules. The overall system has a small
footprint (30 cm  50 cm  40 cm) and weighs approximately
25 kg. It was developed in a modular approach as to be very
versatile, with the possibility of implementing a wide variety of
BEM protocols. For both modules, the biological performances
(sensitivity, purication yield, reproducibility) were at the level of
those obtained with manual reference commercial systems.
Moreover, these modules were connected for the complete protocol integration from sample preparation to DNA detection, in a
semi-continuous mode of one analysis per hour. We conrmed
that EWOD based digital microuidics give a state-of-the-art level
of sensitivity with high potential for integration, miniaturization
and automation of PCR analysis. This is particularly demonstrated
in this study, where specic PCR reagents were embedded inside
an air-tight packaged chip, allowing us to drastically simplify
system automation, by reducing the number of uidic handling
steps in the detection module. To our knowledge, EWOD chip
with embedded dry reagents and air-tight packaging has not been
reported. These results constitute an original work paving the
way toward ready-to-use EWOD chip for dedicated PCR analysis.

Please cite this article as: Delattre, C., et al., Macro to microuidics system for biological environmental monitoring. Biosensors and
Bioelectronics (2012), http://dx.doi.org/10.1016/j.bios.2012.04.024

C. Delattre et al. / Biosensors and Bioelectronics ] (]]]]) ]]]]]]

Acknowledgments
We would like to thank G. Castellan, O. Constantin, C. Chabrol
M. Alessio and A. Bellemin Comte for EWOD chip fabrication and
packaging; C. Danjean and P. Claustre for protocol development
and biological validations; and C. Echampard, R. Sauze, R. Charles,
H. Grateau and Y. Bettassa for instrument development. We also
thank J.-P. Nikolovski for the development of the ultrasound
resonator.
This work was funded by the French joint ministerial CBRNE
project (driven by CEAAtomic Energy and Alternative Energies
Commission) and by the OSEO (French state innovation agency).
Cylergie (research center of COFELY) is also gratefully acknowledged for technical advices and nancial support.

Appendix A. Supporting information

Supplementary data associated with this article can be found
in the online version at http://dx.doi.org/10.1016/j.bios.2012.04.
024.

References
Berge, B., 1993. Comptes Rendus De LAcademie Des Sciences Serie II 317 (2),
157163.
Cho, S.K., Moon, H.J., Kim, C.J., 2003. Journal of Microelectromechanical Systems 12
(1), 7080.
Cotterill, S., Evans, R., Fraise, A.P., 1996. Journal of Hospital Infection 32 (3),
207216.

Delattre, C., Blanc, R., Castellan, G., Chabrol, C., David, N., Dubard, Constantin O.,
Fouillet, Y., Jary, D., Rival, A., Caillat, P., 2008. Micro Total Analysis Systems.
San Diego, USA.
Fair, R.B., 2007. Microuidics and Nanouidics 3 (3), 245281.
Fouillet, Y., Jary, D., Chabrol, C., Claustre, P., Peponnet, C., 2008. Microuidics and
Nanouidics 4 (3), 159165.
Gehanno, J.F., Louvel, A., Nouvellon, M., Caillard, J.F., Pestel-Caron, M., 2009.
Journal of Hospital Infection 71 (3), 256262.
Jarvis, W.R., Edwards, J.R., Culver, D.H., Hughes, J.M., Horan, T., Emori, T.G.,
Banerjee, S., Tolson, J., Henderson, T., Gaynes, R.P., Martone, W.J., 1991.
American Journal of Medicine 91, S185S191.
Jary, D., Chollat-Namy, A., Boutet, J., Chabrol, C., Peponnet, C., 2006. Nanotech.
Boston, USA.
Luu-The, V., Paquet, N., Calvo, E., Cumps, J., 2005. Biotechniques 38 (2), 287293.
Malic, L., Brassard, D., Veres, T., Tabrizian, M., 2010. Lab Chip 10 (4), 418431.
Malk, R., Fouillet, Y., Davoust, L., 2011. Sensors and Actuators B: Chemical 154 (2),
191198.
Paik, P., Pamula, V.K., Fair, R.B., 2003. Lab Chip 3 (4), 253259.
Pollack, M.G., Shenderov, A.D., Fair, R.B., 2002. Lab Chip 2 (2), 96101.
Regan, J.F., Makarewicz, A.J., Hindson, B.J., Metz, T.R., Gutierrez, D.M., Corzett, T.H.,
Hadley, D.R., Mahnke, R.C., Henderer, B.D., Breneman, J.W., Weisgraber, T.H.,
Dzenitis, J.M., 2008. Analytical Chemistry 80 (19), 74227429.
Rider, T.H., Petrovick, M.S., Nargi, F.E., Harper, J.D., Schwoebel, E.D., Mathews, R.H.,
Blanchard, D.J., Bortolin, L.T., Young, A.M., Chen, J.Z., Hollis, M.A., 2003. Science
301 (5630), 213215.
Schultsz, C., Meester, H.H.M., Kranenburg, A.M.H., Savelkoul, P.H.M., BoeijenDonkers, L.E.A., Kaisers, A.M., de Bree, R., Snow, G.B., Vandenbroucke-Grauls,
C.J.M.E, 2003. Journal of Hospital Infection 55 (4), 269275.
Schultz, E., Galland, R., Du Bouetiez, D., Flahaut, T., Planat-Chretien, A., Lesbre, F.,
Hoang, A., Volland, H., Perraut, F., 2008. Biosensors and Bioelectronics 23 (7),
987994.
Sista, R., Hua, Z.S., Thwar, P., Sudarsan, A., Srinivasan, V., Eckhardt, A., Pollack, M.,
Pamula, V., 2008. Lab Chip 8 (12), 20912104.
Srinivasan, V., Pamula, V.K., Fair, R.B., 2004. Lab Chip 4 (4), 310315.
Wheeler, A.R., Moon, H., Kim, C.J., Loo, J.A., Garrell, R.L., 2004. Analytical Chemistry
76 (16), 48334838.

Please cite this article as: Delattre, C., et al., Macro to microuidics system for biological environmental monitoring. Biosensors and
Bioelectronics (2012), http://dx.doi.org/10.1016/j.bios.2012.04.024