Biosensors and Bioelectronics 24 (2009) 3365–3371

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

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

Gold screen-printed-based impedimetric immunobiosensors for direct and

sensitive Escherichia coli quantisation
Vanessa Escamilla-Gómez, Susana Campuzano, María Pedrero, José M. Pingarrón ∗
Dpto. Química Analítica, Facultad de CC. Químicas, Universidad Complutense de Madrid, E-28040 Madrid, Spain

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

Article history: Label-free electrochemical impedance immunosensors for the detection and quantification of Escherichia
Received 6 March 2009 coli (E. coli) using self-assembled monolayers (SAMs)-modified gold screen-printed electrodes (AuSPEs)
Received in revised form 20 April 2009 were developed. Two different immunosensor configurations were tested and compared. In the first one,
Accepted 30 April 2009
the immunosensing design was based on the covalent immobilization of anti-E. coli at AuSPEs using the
Available online 7 May 2009
homobifunctional cross-linker 3,3 -dithiobis[sulfosuccinimidylpropionate] (DTSSP). The other one was
based on the immobilization of the thiolated antibody onto the electrode surface. In both cases, the eval-
Keywords:
uation of the developed immunosensors performance was accomplished through the monitoring of the
E. coli
Immunosensors
electron-transfer resistance detected by electrochemical impedance spectroscopy (EIS) in the presence
Electrochemical impedance spectroscopy of [Fe(CN)6 3− ]/[Fe(CN)6 4− ] as redox probe. The configuration using the thiolated antibodies gave rise to
Label-free detection a better analytical performance, exhibiting a linear relationship between the increment in the electron-
transfer resistance (Ret ) and the logarithmic value of the E. coli concentration in the 5–1.0 × 108 cfu mL−1
range. The limit of detection achieved, with no need for preconcentration or pre-enrichment steps was
3.3 cfu mL−1 . The developed immunosensors showed a high selectivity against Staphylococcus aureus (S.
aureus) and Salmonella choleraesuis (S. choleraesuis). The usefulness of the thiolated antibodies-based
design for the rapid analysis (1 h) of 10 cfu mL−1 E. coli inoculated river and tap water samples was
demonstrated.
© 2009 Elsevier B.V. All rights reserved.

1. Introduction for the detection of pathogens including miniaturized biochem-

ical tests, immunological assays, and methods based on nucleic
Nowadays, one of the challenges in food industry, environmental acid probes using polymerase chain reaction (PCR) (Yang et
monitoring and clinical diagnosis is the development of meth- al., 2004). Nowadays, biosensors play a significant role in the
ods for the rapid detection of pathogenic bacteria to prevent risks determination of pathogens. Regarding immunosensors, most of
of infection, bioterrorism, enteric diseases and economic losses them use labelled antibodies to monitor the formation of the
(Subramanian et al., 2006; Maalouf et al., 2007). The presence of antigen–antibody complex. Label-free immunosensors, in which
coliforms is usually employed as an indicator of faecal contamina- the immune interaction between antibody and antigen is directly
tion, being Escherichia coli generally considered as the most reliable monitored, exhibit some important advantages in terms of speed
indicator because it is directly related to this kind of contamination and simplicity of operation. Several approaches concerning the
(Dudak and Boyacı, 2007). development of label-free immunosensors for the detection of
Conventional methods for the detection of E. coli include bacteria have been reported. These include quartz crystal microbal-
multiple-tube fermentation, membrane filter, or plate count. These ance (QCM) immunosensors for the detection of Salmonella (Fung
methods are laborious and time-consuming, and, in spite of and Wong, 2001; Park and Kim, 1998; Park et al., 2000) and E.
their correctness, they cannot easily meet inspection demands coli (Kim and Parl, 2003), and surface plasmon resonance (SPR)
(Kim and Parl, 2003), considering that European Regulation No. immunosensors to detect Salmonella and Listeria (Koubova et al.,
1441/2007 establishes that non-contaminated drinkable water 2001).
should exhibit absence of E. coli in 100 mL of sample, thus mak- Electrochemical impedance spectroscopy (EIS) is an efficient
ing necessary the detection of one single bacteria in 100 mL. electrochemical technique for the transduction of biosensing events
Hence, a number of more rapid methods have been proposed on electrodes (Maalouf et al., 2007), including the monitoring of
the antigen–antibody interaction on electrode surfaces (Kharitonov
et al., 2000). Impedimetric immunosensors for the detection of
∗ Corresponding author. Tel.: +34 913 944315; fax: +34 913 944329. Salmonella typhimurium (Pournaras et al., 2008) and E. coli (Geng
E-mail address: [email protected] (J.M. Pingarrón). et al., 2008) have been described.

0956-5663/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.bios.2009.04.047
3366 V. Escamilla-Gómez et al. / Biosensors and Bioelectronics 24 (2009) 3365–3371

The immobilization of antibodies on the electrode surface Phosphate buffer solutions consisted of sodium dihydrogen
readily affects the biosensor analytical performance. Strategies phosphate dehydrate (Scharlau) and 0.05 M anhydrous disodium
for antibodies immobilization on electrodes include physical and hydrogen phosphate (Scharlau) (pH 7.0). 4 mM DTSSP and 40 mM
chemical adsorption for preparing oriented antibody molecular sulfo-LC-SPDP aqueous solutions were employed for the covalent
layers on solid surfaces (Sigal et al., 1996; Pei et al., 1998; Lo et antibody immobilization on the AuSPEs. A 0.1 M ethanolamine
al., 1999; Kanno et al., 2000), direct physical adsorption (Dan et (Sigma) solution in 0.05 M phosphate buffer (pH 7.0) was also used.
al., 2007), covalent attachment (Subramanian et al., 2006), and A 150 mM DTT solution, prepared in 0.05 M phosphate buffer (pH
polymer entrapment (Gupta and Chaudhury, 2007). Self-assembled 7.0), was used for cleaving of the disulfide bond resulting from the
monolayers (SAMs) provide a powerful tool to obtain monomolec- thiolation procedure. Anti-E. coli solutions (200 ␮g mL−1 and more
ular films of biological molecules on various substrates, and they diluted standards) were also prepared in 0.05 M phosphate buffer
have been used to prepare antibody molecular layers on solid sur- (pH 7.0). K3 Fe(CN)6 and K4 Fe(CN)6 (Sigma) solutions, 5 mM in each
faces by means of thiolated antibodies (Park and Kim, 1998), or component and prepared in 0.1 M KCl were used for CV and EIS
exploiting carboxyl–amine couplings (Subramanian et al., 2006; measurements.
Dong and Shanon, 2000; Park et al., 2004; Taylor et al., 2005; All chemicals used were of analytical-reagent grade, and
Escamilla-Gómez et al., 2007, 2008). deionised water was obtained from a Millipore Milli-Q purification
The development and analytical performance of label-free system.
electrochemical impedance immunosensors for the direct detec-
tion of bacteria are reported in this work. E. coli was used as
2.3. Bacteria cultivation
a model bacterium, and anti-E. coli antibodies were immobi-
lized on gold screen-printed electrodes (AuSPEs). Two different
Wild-type E. coli and S. aureus belong to the microorganism
immunosensor designs were compared. The first one implied
biosafety level 2 group, while S. choleraesuis belongs to the level
the covalent immobilization of the specific antibodies on AuS-
1 group. Consequently, all safety considerations concerning these
PEs modified with a 3,3 -dithiobis[sulfosuccinimidylpropionate]
groups were accomplished in the manipulation of these bacteria
(DTSSP) SAM. The other configuration involved the immobilization
(Anon, 2000). E. coli and S. aureus were grown overnight in Luria-
of thiolated antibodies onto the AuSPE surface. Antibodies were
Bertani Broth (LB) (Scharlau) while S. choleraesuis was grown in
thiolated by using sulfosuccinimidyl 6-[3 -(2-pyridyldithio) propi-
Tryptic Soy Broth (TSB) (Fluka), at 37 ◦ C with aeration and shak-
onamide] hexanoate (sulfo-LC-SPDP) as thiolation reagent (Tanaka
ing, which allowed the growing stationary phase to be reached.
et al., 1999). The performance of both designs as impedimetric
Then, several dilutions of these mother growth media were made
immunosensors upon binding of E. coli cells to the electrode was
(10-fold steps), and 10 ␮L of the diluted solutions were transferred
evaluated by measuring the electron-transfer resistance by EIS in
to LB agar plates (E. coli and S. aureus) or to TS agar (TSA) plates
the presence of [Fe(CN)6 3− ]/[Fe(CN)6 4− ] as redox probe. The selec-
(S. choleraesuis) and incubated for 24 h at 37 ◦ C, for enumeration
tivity against other bacteria was evaluated for both immunosensor
of colonies. At the same time, the stationary-phase cultures were
designs. Finally, the thiolated-anti-E. coli-AuSPE configuration was
diluted to 10–108 cfu mL−1 in deionised water for the impedimetric
applied to the analysis of inoculated river and tap water samples.
measurements.

2. Experimental
2.4. Immunosensors construction

2.1. Apparatus and electrodes

First, AuSPEs were pre-treated. A 50-␮L drop of a 0.1 M H2 SO4
solution was placed on the AuSPE surface. Then, 10 cyclic voltam-
Electrochemical impedance measurements were performed
mograms were scanned from 0.00 to 1.25 V at 100 mV s−1 , and the
using a ␮-Autolab type III with FRA2 software (ECO Chemie). Cyclic
electrodes were washed with deionised water.
voltammetry (CV) was carried out with an ECO Chemie Autolab
The immunosensor construction using DTSSP monolayers
P/GSTAT 10 (ECO Chemie, Autolab) controlled by GPES4.9 software
involved the deposition of a 10-␮L drop of a 4 mM DTSSP solu-
(General Purpose Electrochemical System).
tion on the AuSPEs, allowing the reaction to proceed overnight at
An optic Ivymen System constant temperature incubator shaker
4 ◦ C. Then, the modified electrodes were rinsed with 0.05 M phos-
(Comecta S.A.), a P-Selecta ultrasonic bath, a P-Selecta Agimatic
phate buffer (pH 7.0). DTSSP-AuSPEs were coated with 40 ␮L of a
magnetic stirrer and a Sky Line Shaker S-4 (Elmi) were also used.
10 ␮g mL−1 anti-E. coli solution and incubated during 1 h at room
Gold screen-printed electrodes (AuSPEs) were purchased from
temperature. The unbound antibodies were removed from the elec-
DropSens. They include a gold disk-shaped (12.6 mm2 ) working
trode surface by rinsing with 0.05 M phosphate buffer (pH 7.0).
electrode, a silver pseudo-reference electrode and a gold counter
Then the antibody-modified electrodes were treated with a 0.1 M
electrode, all of them screen-printed on a ceramic substrate
ethanolamine solution for 1 h to block the non-reacted and non-
(3.4 cm × 1.0 cm) and subjected to low-temperature curing. All
specific DTSSP SAM sites. Finally, the electrode was thoroughly
potential values were referred to the screen-printed silver pseudo-
rinsed with 0.05 M phosphate buffer (pH 7.0).
reference electrode.
The other configuration tested implied, in a first step, the anti-
body thiolation. The procedure was 20 ␮L of a 200 ␮g mL−1 anti-E.
2.2. Reagents and solutions coli solution was incubated with 10 ␮L of a 40 mM sulfo-LC-SPDP
solution at room temperature for 1 h (in a shaker at 36 r.p.m.). Then,
Wild-type E. coli (CECT 515), Staphylococcus aureus (CECT 35 ␮L of a 150 mM DTT solution was added to the mixture and left
59) and Salmonella choleraesuis (CECT 700) were obtained to react for 45 min in order to reduce the disulfide bond of the thio-
from the Spanish Collection of Type Cultures. 3,3 -Dithiobis- lated antibody. Next, 40 ␮L of the resulting solution were deposited
(sulfosuccinimidylpropionate) (DTSSP) and sulfosuccinimidyl 6- on the AuSPE and left overnight at 4 ◦ C. Finally, the electrode was
[3 -(2-pyridyldithio) propionamido] hexanoate (sulfo-LC-SPDP) rinsed with 0.05 M phosphate buffer (pH 7.0).
were supplied by Pierce Inc. (IL, USA), while dithiotreitol (DTT) In this case no blocking procedure was needed, because no bac-
and ethanolamine were supplied by Sigma. Anti-E. coli polyclonal teria immobilization was observed on the AuSPEs prepared in the
antibodies were obtained from Affinity BioReagents. absence of thiolated-anti-E. coli.
 V. Escamilla-Gómez et al. / Biosensors and Bioelectronics 24 (2009) 3365–3371 3367

2.5. E. coli detection ple, storing it for 1 h at room temperature, and rinsing with 0.05 M
phosphate buffer (pH 7.0).
A 40-␮L aliquot of E. coli cell cultures with different cell numbers
were deposited on the anti-E. coli-modified electrodes surface, and 3. Results and discussion
incubation was let to proceed for 1 h at ambient temperature. To
remove non-specifically bound cells, the immunosensor was rinsed Fig. 1 schematizes the two impedimetric immunosensor config-
with 0.05 M phosphate buffer (pH 7.0). urations developed in this work. Relative sizes of the components
Faradaic impedimetric and CV measurements were carried out are not drawn on real scale in order to visualize all of them.
after deposition of a 40-␮L drop of a 5 mM [Fe(CN)6 ]3−/4− redox The first design involved modification of the AuSPEs with a
probe solution on top of the modified AuSPE. Impedance spectra DTSSP SAM. This water-soluble, homobifunctional, cross-linker
were recorded in the frequency range from 0.01 to 10,000 Hz at the contains an amine-reactive N-hydroxysulfosuccinimide (sulfo-
formal potential of the [Fe(CN)6 ]3−/4− redox couple. The amplitude NHS) ester at each end of an 8-carbon spacer arm capable of
of the alternating voltage was 0.01 V. Experimental spectra, pre- reacting with primary amines at pH 7–9 to form stable amide bonds,
sented in the form of complex plane diagrams (i.e., Nyquist plots), along with release of the N-hydroxysulfosuccinimide leaving group.
were fitted with proper equivalent circuits using the facilities of When the antibodies were deposited on top of the DTSSP-modified
the FRA2 software 4.9.006 (EcoChemie). Electron-transfer resis- electrodes, the active NHS esters of the SAM were replaced by
tance (Ret ), as well as the increase in electron-transfer resistance the primary amines in the side chain of lysine (K) residues of the
(Ret ) values were taken as analytical signals, according to the antibodies, which were thus immobilized on the electrode surface
equation: through the amide bond.
Ret = Ret(antibody–bacteria) − Ret(antibody) The other design implied immobilization of 2-pyridyl-disulfide
activated antibodies. In this case, sulfo-LC-SPDP, a heterobifunc-
where Ret(antibody) is the electron-transfer resistance value mea- tional, thiol-cleavable cross-linker, with one N-hydroxysuccinimide
sured after antibody immobilization on the electrode, and ester and one pyridyl-disulfide residue, reacted with primary
Ret(antibody–bacteria) is the electron-transfer resistance value upon the amines and sulfhydryls in the antibodies, thus inserting 2-pyridyl-
antibody–E. coli conjugate formation. In the case of using anti-E. disulfide groups. The resulting disulfide bonds are cleavable by a
coli-DTSSP-AuSPE immunosensors, the Ret values were measured reducing agent such as DTT. Thereafter, the thiolated antibody–gold
as: complex was covalently formed on the screen-printed gold elec-
trodes (Park and Kim, 1998; Tanaka et al., 1999).
Ret = Ret(antibody–bacteria) − Ret(ethanolamine)
In both cases, a specific binding event occurred between the
where Ret(ethanolamine) is the resistance value after the blocking treat- covalent immobilized antibodies and the antigens on the E. coli cell
ment. surface.
Cyclic voltammetry (CV) in the −0.25 to 0.50 V range at a scan
rate of 100 mV s−1 , was used to monitor the construction of the 3.1. Formation of the antibody–bacteria complex at the AuSPE
immunosensors designs. surface
All the experiments were carried out at room temperature
(25 ± 1) ◦ C. The stepwise assembly with both immunosensor designs was
monitored by using EIS and CV.
2.6. Water analysis Fig. 2a shows the obtained Nyquist plots for a bare AuSPE (curve
1), a DTSSP-AuSPE (curve 2), an anti-E. coli-DTSSP-AuSPE (curve
E. coli was determined in river and tap water samples inoculated 3), this latter modified electrode after blocking with ethanolamine
at a 10 cfu mL−1 level, using the thiolated-anti-E. coli immunosensor (curve 4), and after formation of the antibody–E. coli complex
design. Impedimetric measurements were carried out after coating (curve 5). The bare AuSPE showed the expected fast electron-
the immunosensor surface with a 40-␮L drop of the water sam- transfer process with a diffusion limiting step, and the formation

Fig. 1. Schemes of the immunosensor configurations used in this work: (a) covalent immobilization of anti-E. coli antibodies on the DTSSP-modified AuSPE; (b) immobilization
of thiolated antibodies onto the AuSPE.
3368 V. Escamilla-Gómez et al. / Biosensors and Bioelectronics 24 (2009) 3365–3371

Fig. 2. Complex impedance plots (a and b) and cyclic voltammograms (c and d) obtained with a bare AuSPE (curves 1), a DTSSP-modified AuSPE (curve 2a), an anti-E. coli-
DTSSP-AuSPE (curve 3a), an anti-E. coli-DTSSP-AuSPE after blocking with ethanolamine (curve 4a), an E. coli-anti-E. coli-DTSSP-AuSPE (curve 5a), a thiolated-anti-E. coli-AuSPE
(curve 2b), and an E. coli-thiolated-anti-E. coli-AuSPE (curve 3b). Operating conditions: EIS, 5 mM [Fe(CN)6 ]3−/4− (1:1) in 0.1 M KCl, 0.01–10,000 Hz frequency range with a
0.01 V r.m.s. signal; CV, v = 100 mV s−1 . Experimental conditions—(a and c): [anti-E. coli] = 10 ␮g mL−1 , tanti-E. coli = 60 min, tE. coli = 60 min, [E. coli] = 1.0 × 105 cfu mL−1 ; (b and d):
[sulfo-LC-SPDP] = 40 mM, tsulfo-LC-SPDP = 60 min, [anti-E. coli] = 200 ␮g mL−1 , [DTT] = 150 mM, tDTT = 45 min, tthiolated-anti-E. coli = 15 h, [E. coli] = 1.0 × 103 cfu mL−1 .

of the DTSSP SAM resulted in the expected increase in the Ret bodies produced a further decrease in the peak current (curve 3),
value. When antibodies were immobilized on the DTSSP-AuSPE, and only a slight current increase was observed after deactivation
the penetration of the redox probe was reduced and, accordingly, of the excess NHS esters in the SAM using ethanolamine (curve
Ret increased notably. Blocking of the residual succinimidyl groups 4). Binding of bacteria to the immobilized antibodies (curve 5)
in the SAM with ethanolamine promoted the redox probe trans- produced a ∼15% increase in the separation of peak potentials
fer to the electrode surface, and a decrease in Ret was observed. in comparison to the separation observed in curve 4 as well as
This fact could be due to either electrostatic attractions or to the a remarkable decrease in the peak currents. These results are
ability of ethanolamine to replace weakly immobilized antibodies. consistent with the changes observed in the electron-transfer
The further attachment of bacterial cells to the modified electrode resistance by EIS.
surface gave rise to an additional barrier towards the access of A similar set of EIS/CV experiments to those commented above
the redox probe to the electrode, resulting in a further increase was carried out using 2-pyridyl-disulfide activated antibodies
in the electron-transfer resistance. As can be seen, Ret changed (Fig. 2b and d, respectively). This immunosensor configuration
from 761 to 998 , i.e. increased a 31% after the formation of the exhibited a much higher increase in the Ret value (162%) and,
antibody–E. coli complex. These results confirmed the possibility accordingly, the possibility of achieving a much higher sensitiv-
of using EIS to monitor the change in electron-transfer resistance ity for the assay. These results are likely due to a larger amount
resulting from the immobilization of bacteria, through the com- of immobilized antibody using this approach, as a consequence of
plex formed with an appropriate antibody, on the surface of AuSPEs the higher packing density of the SAM achieved with the thiolated
without any amplification step. It should be remarked that no antibodies. Furthermore, the 2-pyridyl-disulfide exhibits a higher
significant non-specific bacteria adsorption was detected on the reactivity against antibodies when compared with DTSSP (Chatrathi
ethanolamine–DTSSP-AuSPEs prepared without immobilized anti- et al., 2007), thus giving rise to a larger amount of immobilized
bodies, thus confirming the E. coli cells capture by the antibodies antibodies.
on the electrode surface.
Fig. 2c shows cyclic voltammograms for the 3.2. Optimization of experimental variables
[Fe(CN)6 3− ]/[Fe(CN)6 4− ] redox probe at a bare AuSPE (curve
1), a DTSSP-AuSPE (curve 2), an anti-E. coli-DTSSP-AuSPE (curve Table 1 summarizes the experimental variables checked, the
3), and anti-E. coli-DTSSP-AuSPE after blocking with ethanolamine ranges tested and the selected parameters in the optimization
(curve 4), and an E. coli-anti-E. coli-DTSSP-AuSPE (curve 5). As studies carried out (see Additional information). Concerning the
expected, the electrode modification with a DTSSP SAM resulted anti-E. coli-DTSSP-based immunosensor, the influence of the anti-E.
in a decrease in the peak current and an increase in the separation coli concentration on the EIS responses for a constant E. coli con-
of the peak potentials (curve 2) compared to the voltammetric centration of 1 × 105 cfu mL−1 (Fig. A1a, additional information),
behaviour of the unmodified electrode. Immobilization of anti- the incubation time in the antibody solution (Fig. A1b, additional
 V. Escamilla-Gómez et al. / Biosensors and Bioelectronics 24 (2009) 3365–3371 3369

Table 1
Optimization of experimental variables for both immunosensor configurations.

Electrode configuration Experimental variable Studied range Selected value

−1
[Anti-E. coli] (␮g mL ) 0–100 10
Anti-E. coli-DTSSP-AuSPE tanti-E. coli (h) 0.5–2 1
tE. coli (h) 0.5–4 1

[Sulfo-LC-SPDP] (mM) 0–80 40

tsulfo-LC-SPDP (min) 0–90 60
[Anti-E. coli] (␮g mL−1 ) 50–400 200
Thiolated-anti-E. coli-AuSPE
[DTT] (mM) 0–150 150
tDTT (min) 0–60 45
tthiolated-anti-E. coli (h) 3–24 15

information), and the antibody–bacteria incubation time were eval- of bacteria, and s is the standard deviation for the same sig-
uated. nals expressed in concentration units. The values obtained were
In the case of thiolated antibodies-based immunosensors, the 1.0 × 104 and 2.3 × 104 cfu mL−1 , respectively. Regarding the thio-
parameters investigated and optimized included the thiolation lated antibodies-based configuration, a limit of detection as low
reagent concentration (Fig. A2a, additional information), the thi- as 3.3 cfu mL−1 was obtained, with a linear range from 5 to
olation reaction time (Fig. A2b, additional information), amount of 1.0 × 108 cfu mL−1 (slope = (54,091 ± 3147) , r = 0.992), in spite of
capture antibody (Fig. A2c, additional information), reducing agent the fact that the binding efficiency of anti-bacteria antibodies to
concentration (Fig. A2d, additional information), reduction pro- their antigen is normally low (Ruan et al., 2002). Obviously, the
cess time (Fig. A2e, additional information), and thiolated antibody analytical performance of this latter configuration is notably better.
immobilization time (Fig. A2f, additional information). Negative controls (without immobilized antibodies) were car-
See Additional information for a discussion on the influence of ried out with both configurations. No significant Ret changes was
all these variables on the immunosensors impedimetric responses. observed even for large bacteria concentrations, thus ensuring that
the measured electrochemical signal is specific and not partly due
3.3. Immunosensors analytical performance to the non-specific adsorption of bacteria on the electrode surfaces.
Table 2 compares the analytical characteristics of the thio-
First, the reproducibility was evaluated by constructing different lated antibodies-based immunosensor configuration with those
immunosensors for each of the developed configurations under the reported in literature for other E. coli impedimetric biosensors. As
optimized conditions. Measurements for 1.0 × 105 cfu mL−1 were it can be deduced, the limit of detection achieved in this work is
carried out with the anti-E. coli-DTSSP-based configuration while, much lower than all those reported previously, in spite of no precon-
in the case of the thiolated antibodies-based configuration, the centration or pre-enrichment steps was used. Moreover, the range
bacteria concentration level used was 1.0 × 103 cfu mL−1 . Relative of linearity, covering seven orders of magnitude, was much wider
standard deviation (RSD) values for the impedimetric responses than the best ones reported for other E. coli impedimetric sensors.
were 3.7% (n = 5) and 5.7% (n = 5), respectively, which are fairly good In summary, the thiolated-anti-E. coli-AuSPE immunosensor design
for impedimetric immunosensors. exhibits improved analytical performance with respect to previous
Calibration curves for E. coli, expressed as a function of the loga- impedimetric approaches. Moreover, the simplicity and rapidity of
rithmic value of the E. coli cells concentration (in cfu mL−1 ) were the proposed methodology are additional advantages to be taken
constructed for each configuration (see Additional information). into account.
Regarding the anti-E. coli-DTSSP-immunosensor, Ret varied lin-
early (slope = (158 ± 17) ; r = 0.984) with the logarithmic value 3.4. Selectivity of the developed immunosensors
of the E. coli cells concentration over three orders of mag-
nitude, between 1.0 × 104 and 1.0 × 107 cfu mL−1 . The limits of The selectivity of the immunosensors against two other bacteria
detection and determination were calculated according to the (S. aureus and S. choleraesuis) was evaluated by checking the impedi-
3 × sb /m and 10s criteria, respectively, where m is the slope of the metric responses at the same concentration level than that of E. coli
linear calibration plot, sb was estimated as the standard devia- (1 × 105 and 1.0 × 103 cfu mL−1 for the anti-E. coli-DTSSP-based and
tion (n = 8) of the impedimetric signals obtained in the absence the thiolated antibodies-based configurations, respectively), and

Fig. 3. Immunosensors specificity towards E. coli bacteria: (a) anti-E. coli-DTSSP-AuSPE and (b) thiolated-anti-E. coli-AuSPE configurations. Bacteria concentrations:
1.0 × 105 cfu mL−1 (a) and 1.0 × 103 cfu mL−1 (b). Other conditions as in Fig. 2a and b, respectively.
3370 V. Escamilla-Gómez et al. / Biosensors and Bioelectronics 24 (2009) 3365–3371

Radke and Alocilja (2005)

under the experimental conditions selected above. Fig. 3 shows

Varshney and Li (2007)

the electron-transfer resistance changes after binding with the

Varshney et al. (2007)

Maalouf et al. (2007)
Zhang et al. (2005)
Yang and Li (2005)
different bacteria for both immunosensors. In the case of the anti-

Geng et al. (2008)

Ruan et al. (2002)

Yang et al. (2004)

E. coli-DTSSP-based configuration, Ret increased less than 10 
when S. aureus and S. choleraesuis were assayed, whereas a marked

This work
Reference

increase (237 ) was found after binding with E. coli cells, indicating
that there was no significant cross-reaction of the immunosen-
sor with other bacterial species. These results demonstrated that
the electron-transfer resistance as recorded reflected the interac-
tion between the antibody and the target E. coli cells, therefore
incubation with bacteria solution until solvent evaporation

∼1 h: immunomagnetic concentration of bacteria; 35 min:

∼38 h: preparation of antibodies-modified substrates; 1 h:

∼16 h: preparation of antibodies-modified substrates; 1 h:
showing the specificity of the immunosensor for E. coli. A good

5 min: pre-treatment of the AuSPEs; 15 h: preparation of

Overnight preparation of substrate-attached antibodies;

selectivity was also observed for the thiolated-anti-E coli-AuSPEs

∼36 h: preparation of antibodies-modified chips; 1 h:

∼24 h: preparation of antibodies-modified substrates

antibodies modified AuSPEs; 1 h: binding reaction

∼1 h: immunomagnetic concentration of bacteria (Fig. 3b), where the electron-transfer resistance found in the pres-
ence of E. coli was about one order of magnitude higher than that
found in the presence of S. aureus or S. choleraesuis. Neverthe-
less, the relative Ret provoked by S. aureus or S. choleraesuis with
antibodies–bacteria binding reaction

antibodies–bacteria binding reaction

antibodies–bacteria binding reaction respect to the E. coli response is higher than that obtained with
the anti-E. coli-DTSSP-based configuration. This can be due to the
fact that a blocking step with ethanolamine was involved in this
latter configuration, while no blocking steps were performed for
the thiolated-anti-E. coli-AuSPEs. This fact could permit a small
antibodies–bacteria

unspecific adsorption of S. aureus and S. choleraesuis on the gold

electrode surface, thus giving rise to the Ret signals observed for
measurement

these microorganisms.
Assay time

3.5. Water analysis

–

The usefulness of the thiolated antibodies-based immunosen-

sor to detect E. coli in water samples was investigated. The response
preconcentration steps)

of the developed sensor for river Eresma (Segovia, Spain) and tap
water samples was below its detection limit. Also, plate count-
1.0 × 10 (50 with
L.D. (cfu mL−1 )

ing indicated that there were no faecal coliforms in the analyzed

samples. Therefore, water samples were inoculated with E. coli
4.215 × 103
3

6.0 × 105

1.6 × 102
6

7.4 × 104

at a concentration level of 10 cfu mL−1 , and the assay procedure

6.0 × 10

1.0 × 10

1.0 × 10

described in the Section 2 was applied. Thereafter, a calibration

3.3
10

graph for E. coli cells in both types of water was constructed over
the 10–1.0 × 103 cfu mL−1 range following the same methodology
4.215 × 103 –4.215 × 106

that for standard solutions. Analysis of six inoculated river and

8
4.36 × 10 –4.36 × 10
7

7
7

1.0 × 101 –1.0 × 104

5.0–1.0 × 108
6.0 × 10 –6.0 × 10

3.0 × 10 –3.0 × 10
1.0 × 10 –1.0 × 10

of six inoculated tap water samples yielded E. coli mean con-

Analytical performance of different E. coli impedimetric biosensors reported in the literature.

tents of (12 ± 2) and (11 ± 2) cfu mL−1 , respectively (confidence

L.R. (cfu mL−1 )

interval calculated for ˛ = 0.05). The total time for the assay, once
–

–
4

the immunosensor is prepared, is of approximately 1 h. All these

data show fairly well the usefulness of the developed impedi-
metric immunosensor to be used for the rapid and sensitive
determination of this pathogen microorganism in water sam-
ples.
gold electrode modified with a mercaptoacetic acid (MACA) SAM
Covalent immobilization of antibodies through EDC and NHS on a
Biotinylated polyclonal antibody linked to a mixed SAM on a gold
Antibody immobilization via heterobifunctional cross-linkers on
Covalent immobilization of antibodies onto an indium–tin oxide

Interdigitated array gold microelectrode coupled with magnetic

Interdigitated array gold microelectrode coupled with magnetic

4. Conclusions
Covalent immobilization of antibodies using an epoxysilane

Covalent immobilization of antibodies using an epoxysilane

Covalent immobilization of antibodies using an epoxysilane

The possibility of rapid and selective label-free detection, and

electrode through biotin–neutravidin interaction

quantification of E. coli based on impedimetric monitoring of the

Thiolated antibodies immobilization on AuSPEs

bacteria affinity reaction with covalently immobilized antibodies

L.R., linear range and L.D., limit of detection.

on gold screen-printed electrodes has been demonstrated. Among

the two configurations tested, the one based on the use of immobi-
an interdigitated array gold electrode
interdigitated array microelectrode

lized thiolated antibodies on the AuSPE exhibits a highly improved

nanoparticle–antibody conjugates

nanoparticle–antibody conjugates

analytical performance with respect to conventional bacterial plate

monolayer on ITO substrate

monolayer on ITO substrate

monolayer on ITO substrate

counting and other electrochemical approaches, enabling, with-

out preconcentration or pre-enrichment steps, the detection of a
Immobilization method

few cfu mL−1 E. coli in river and tap water samples. Besides, the
total analysis time of 1 h from sampling to measurement (includ-
ing recognition, washing, and sensing processes), the use of small
amounts of test solution and the low cost of the assay compared
with conventional culturing method for E. coli identification, con-
Table 2

stitute important practical advantages that make the developed

immunosensors powerful tools for E. coli tests.
 V. Escamilla-Gómez et al. / Biosensors and Bioelectronics 24 (2009) 3365–3371 3371

Acknowledgements Geng, P., Zhang, X., Meng, W., Wang, Q., Zhang, W., Jin, L., Feng, Z., Wu, Z., 2008.
Electrochim. Acta 53, 4663–4668.
Gupta, R., Chaudhury, N.K., 2007. Biosens. Bioelectron. 22, 2387–2399.
The financial support of Santander/Complutense Research Kanno, S., Yanagida, Y., Haruyama, T., Kobatake, E., Aizawa, M.J., 2000. Biotechnology
Project PR 27/05-13953, and of the Spanish Ministerio de Cien- 76, 207–214.
cia e Innovación Research Project CTQ2006-02743BQU is gratefully Kharitonov, A.B., Alfonta, L., Katz, E., Willner, I., 2000. J. Electroanal. Chem. 487,
133–141.
acknowledged. V. Escamilla-Gómez acknowledges a pre-PhD fel- Kim, N., Parl, I.S., 2003. Biosens. Bioelectron. 18, 1101–1107.
lowship of the Spanish Ministerio de Ciencia y Tecnología. S. Koubova, V., Brynda, E., Karasova, L., Skvor, J., Homola, J., Dostalek, J., Tobiska, P.,
Campuzano acknowledges a “Juan de la Cierva” contract to the Rosicky, J., 2001. Sens. Actuators B 74, 100–105.
Lo, Y.S., Huefner, N.D., Chan, W.S., Stevens, F., Harris, J.M., Beebe, T.P., 1999. Langmuir
Spanish Ministerio de Educación y Ciencia. 15, 1373–1382.
Maalouf, R., Fournier-Wirth, C., Coste, J., Chebib, H., Saïkali, Y., Vittori, O., Errachid, A.,
Appendix A. Supplementary data Cloarec, J.P., Martelet, C., Jaffrezic-Renault, N., 2007. Anal. Chem. 79, 4879–4886.
Park, I.S., Kim, N., 1998. Biosens. Bioelectron. 13, 1091–1097.
Park, I.S., Kim, W.Y., Kim, N., 2000. Biosens. Bioelectron. 15, 167–172.
Supplementary data associated with this article can be found, in Park, I.S., Kim, D.K., Adanyi, N., Varadi, M., Kim, N., 2004. Biosens. Bioelectron. 19,
the online version, at doi:10.1016/j.bios.2009.04.047. 667–674.
Pei, J., Hu, J.M., Hu, Y., Zeng, Y., 1998. J. Chem. Technol. Biotechnol. 73, 59–63.
Pournaras, A.V., Koraki, T., Prodromidis, M.I., 2008. Anal. Chim. Acta 624, 301–307.
References Radke, S.M., Alocilja, E.C., 2005. Biosens. Bioelectron. 20, 1662–1667.
Ruan, C., Yang, L., Li, Y., 2002. Anal. Chem. 74, 4814–4820.
Chatrathi, M.P., Wang, J., Collins, G.E., 2007. Biosens. Bioelectron. 22, 2932–2938. Sigal, G.B., Bamddad, C., Barberis, A., Strominger, J., Whitesides, G.M., 1996. Anal.
Dan, D., Xiaoxing, X., Shengfu, W., Aidong, Z., 2007. Talanta 71, 1257–1262. Chem. 68, 490–497.
Anon, 2000. Directive 2000/54/EC. Protection of workers from risks related to expo- Subramanian, A., Irudayaraj, J., Ryan, T., 2006. Biosens. Bioelectron. 21, 998–1006.
sure to biological agents at work. Off. J. European Comm., L 262/21–45. Tanaka, T., Nakamura, N., Matsunaga, M., 1999. Electrochim. Acta 44, 3827–3832.
Dong, Y., Shanon, C., 2000. Anal. Chem. 72, 2371–2376. Taylor, A.D., Yu, Q., Chen, S., Homola, J., Jiang, S., 2005. Sens. Actuators B 107, 202–208.
Dudak, F.C., Boyacı, İ.H., 2007. Food Res. Int. 40, 803–807. Varshney, M., Li, Y., 2007. Biosens. Bioelectron. 22, 2408–2414.
Escamilla-Gómez, V., Campuzano, S., Pedrero, M., Pingarrón, J.M., 2007. Electroanal- Varshney, M., Li, Y., Srinivasan, B., Tung, S., 2007. Sens. Actuators B 128, 99–107.
ysis 19, 1476–1482. Yang, L., Li, Y., Erf, G.F., 2004. Anal. Chem. 76, 1107–1113.
Escamilla-Gómez, V., Campuzano, S., Pedrero, M., Pingarrón, J.M., 2008. Anal. Bioanal. Yang, L., Li, Y., 2005. Biosens. Bioelectron. 20, 1407–1416.
Chem. 391, 837–845. Zhang, D., Chen, S., Qin, L., Li, R., Wang, P., Li, Y., 2005. Conf. Proc. IEEE Eng. Med. Biol.
Fung, Y.S., Wong, Y.Y., 2001. Anal. Chem. 73, 5302–5309. Soc. 7, 7111–7113.