Biosensors and Bioelectronics 40 (2013) 121126

Contents lists available at SciVerse ScienceDirect

Biosensors and Bioelectronics

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

Electrochemical detection of Salmonella using gold nanoparticles

S. Afonso a,b,c, Briza Pe rez-Lo pez a,d, Ronaldo C. Faria b, Luiz H.C. Mattoso c, Manuela Andre e s e, Marisa Maltez-da Costa a, Arben Merkoc ndez-Herrero , Artur Xavier Roig-Sague - i a,f,n Herna
a

 Nanobioelectronics & Biosensors Group, Catalan Institute of Nanotechnology, CIN2 (ICN-CSIC), Universitat Autonoma de Barcelona, 08193 Bellaterra, Catalonia, Spain ~ Carlos, CP 676, Sao ~ Carlos, Sa ~ o Paulo, CEP 13565-905, Brazil mica, Universidade Federal de Sao Departamento de Qu c ~ Agropecua ~ o Carlos, Sao ~ Paulo, CEP 13560-970, Brazil ria, Laborato rio Nacional de Nanotecnologia para o Agronego cio, CP 741, Sa Embrapa Instrumentac - ao d LEITAT Technological Center, 08225 Terrasa, Spain e  ria, Universitat  ncia Animal i dels Aliments, Facultat de Veterina Centre Especial de Recerca Planta de Tecnologia dels Aliments (CERPTA), XaRTA, TECNIO, Departament de Cie  Autonoma de Barcelona, 08193 Bellaterra, Barcelona, Spain f ICREA, Barcelona, Spain
b

a r t i c l e i n f o
Available online 16 July 2012 Keywords: Gold nanoparticles Magneto-immunoassay Salmonella Electrochemical detection Label

abstract
A disposable immunosensor for Salmonella enterica subsp. enterica serovar Typhimurium LT2 (S) detection using a magneto-immunoassay and gold nanoparticles (AuNPs) as label for electrochemical detection is developed. The immunosensor is based on the use of a screen-printed carbon electrode (SPCE) that incorporates a permanent magnet underneath. Salmonella containing samples (i.e. skimmed milk) have been tested by using anti-Salmonella magnetic beads (MBs-pSAb) as capture phase and sandwiching afterwards with AuNPs modied antibodies (sSAb-AuNPs) detected using differential pulse voltammetry (DPV). A detection limit of 143 cells mL 1 and a linear range from 103 to 106 cells mL 1 of Salmonella was obtained, with a coefcient of variation of about 2.4%. Recoveries of the sensor by spiking skimmed milk with different quantities of Salmonella of about 83% and 94% for 1.5 103 and 1.5 105 cells mL 1 were obtained, respectively. This AuNPs detection technology combined with magnetic eld application reports a limit of detection lower than the conventional commercial method carried out for comparison purposes in skimmed milk samples & 2012 Elsevier B.V. All rights reserved.

1. Introduction Foodborne disease has been a serious threat to public health for many years and still remains a public health problem (WHO, 2011). Salmonella is one of the most frequently occurring pathogens in food affecting peoples health (Newell et al., 2010). This bacteria is transmitted to humans mainly through the consumption of contaminated food of animal origin such as milk, meat and eggs. According to World Health Organization (WHO) in the United States of America (USA), for instance, around 76 million cases of foodborne diseases, resulting in 325,000 hospitalizations and 5,000 deaths, are estimated to occur yearly (WHO, 2011). In 2011, more than 10 outbreaks comprising hundreds of patients were reported by Centers for Disease Control and Prevention (CDC) originated in the ingestion of Salmonella-contaminated food, leading to medical costs of thousands of dollars (CDC, 2011).

n Corresponding author at: Nanobioelectronics & Biosensors Group, Catalan  Institute of Nanotechnology, CIN2 (ICN-CSIC), Universitat Autonoma de Barcelona, 08193 Bellaterra, Catalonia, Spain. Tel.: 34 935868014; fax: 34 935868020. E-mail address: arben.merkoci@icn.cat (A. Merkoc - i).

The methods recommended by International agencies of food health control and International Organization of Standardization for Salmonella detection in food samples (ICMSF, 2002; ISO, 2002) are the classical culture methods. These methods can give qualitative and quantitative information, however, a pre-treatment of the samples is needed; furthermore they are greatly restricted by the assay time at locations in the food processing or distribution network, to achieve an earlier detection. Furthermore, to perform them, it is necessary to employ highly skilled people and more than three days, which exclude their use in eld applications. The development of new methodologies with faster response time, better sensitivity and selectivity and easy multiplexing is still a challenge for food hygiene inspection. In recent years, new technologies have been developed in order to improve the time of analysis of the traditional culture detection. These technologies are mainly based on polymerase chain reaction (PCR) and immunoassays. Moreover, biosensor technologies have been used as potential alternatives to circumvent the bottlenecks of the standard method because they have rapid response time and furthermore they are sensitive, robust, portable and easy to use (Liebana et al., 2009a; Liebana et al., 2009b; Mata et al., 2010; Salam and Tothill, 2009).

0956-5663/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bios.2012.06.054

122

A.S. Afonso et al. / Biosensors and Bioelectronics 40 (2013) 121126

The electrochemical detection methods possess several advantages such as easy operation, low cost, high sensitivity, simple instrument and suitability for portable devices. Currently, we are observing a noticeable growth in AuNPs as electrochemical label iz et al., 2010). This for immunoassay (De la Escosura-Mun electrochemical approach is based on the adsorption of AuNPs on the surface of the electrotransducer, electrooxidation of the AuNPs to Au(III), and reverse electroreduction to Au(0), which generates cathodic peak constituting the analytical signal. The AuNPs as a label in connection to magnetic particles and screenprinted carbon electrodes (SPCEs) was also shown to be a very iz useful alternative for proteins detection (De la Escosura-Mun et al., 2011). However, this technology has not been used for the screening of pathogenic organisms. Nanomaterials have received special attention in the development of novel biosensing systems (Merkoc - i, 2010). Particularly nanoparticles have shown to bring interesting advantages for iz et al., DNA (Merkoc - i et al., 2005), proteins (De la Escosura-Mun zou and Merkoc 2010) and even cells (Perfe - i., 2012) analysis. Our group has already shown the effectiveness of AuNPs for ICP-MS linked (Merkoc - i et al., 2005a) and electrochemical (Pumera et al., 2005) DNA assays, electrochemical and optical detections of human IgG (Ambrosi et al., 2007), CA 15-3 glycoprotein (mainly used to watch patients with breast cancer) (Ambrosi et al., 2010) iz and even of human tumor HMy2 cells (De la Escosura-Mun et al., 2009). Herein, a rapid and sensitive strategy for Salmonella detection, that takes advantages of AuNPs used as labels and magnetic particles as preconcentrators, is developed and shown to be effective enough even for real sample applications. In this approach the bacteria are captured from the samples (i.e. skimmed milk) and preconcentrated by immunomagnetic separation, followed by labeling with AuNPs modied with a polyclonal anti-Salmonella antibody. Then, the modied MBs are captured by applying a magnetic eld below the SPCE used as transducer for the electrochemical detection. Although other electrochemical biosensing strategies for Salmonella detection based on nanoparticles (Noguera et al., 2011), carbon nanotubes (Zelada-Guillen et al., 2010) etc. have already been developed (see Table SI-1 in Supporting Information Section) the proposed AuNPs electrochemical labeling strategy is previewed to be of special interest for future in eld applications given the robustness of the electrochemical system in general and that of nanoparticles particularly.

capturing antibody, Salmonella and AuNPs modied with antiSalmonella rabbit polyclonal second antibody (MBs-pSAb/S/sSAbAuNPs) and used later during the electrochemical measurements. All glassware used in the synthesis of AuNPs was washed with aqua regia overnight and the rinsed carefully with milli-Q water. 2.2. Reagents and solutions Anti-Salmonella magnetic beads modied with the rst capturing antibody (MBs-pSAb) (Prod. no.1 710.02) was purchased from Dynal Biotech ASA (Oslo, Norway) and Anti-Salmonella rabbit polyclonal second antibody (sSAb) (Prod. no. 01.91.99) was from Biogen scientic (Madrid, Spain). Salmonella enterica subsp. enterica serovar Typhimurium LT2 (CECT 722T) and Escherichia coli n EspaK-12 (CECT 433) strains were purchased from Coleccio ola de Cultivos Tipo (CECT), Bovine serum albumin, Hydrogen n tetrachloroaurate (III) trihydrate (HAuCl4 3H2O, 99.9%), trisodium citrate, were purchased from Sigma-Aldrich (St. Louis, MO). Millipore milli-Q water was obtained from purication system (18.2 M cm). The buffers were prepared in deionized water: PBS buffer 10 mM pH 7.4 with 2.7 mM KCl, and 137 mM NaCl; PBSTween buffer (PBS buffer pH 7.4 with tween 20% (m/ v)). Samples for SEM analysis were prepared by using glutaraldehyde and hexamethyldisilazane (HMDS) microscopy grade solutions, Sigma-Aldrich (Spain). The electrochemical measurements were performed in a 0.2 M HCl solution. Finally, all reagents and other inorganic chemicals were supplied by Sigma-Aldrich or Fluka, unless otherwise stated. 2.3. Bacterial strains, inocula preparation Freeze-dried cultures of Salmonella and E. coli were revived in Tryptone Soy Broth (TSB, Oxoid Ltd., Basingstoke, Hampshire, UK). Stock cultures of both strains were prepared on Tryptone Soy Agar (TSA, Oxoid), incubated at 37 1C for 24 h and stored at 4 1C for a maximum time of 9 weeks. Stock cultures were subcultured into 10 mL of TSB and incubated at 37 1C for 20 h. After incubation, the broth was spread using a disposable loop on TSA plates and incubated at 37 1C for 2024 h. Subsequently, cell suspensions were prepared in 10 mL of PBSTween to obtain 9.9.5 log cells mL 1. Tubes were placed into a boiling water bath (100 1C) for 15 min and they were cooled to room temperature prior to immunological testing. To determine the load of cells before the heat treatment dilutions were prepared in buffered peptone water (Oxoid). Then, 1 mL of these dilutions was placed as duplicate in TSA (Oxoid) and incubated at 37 1C for 24 h. 2.4. Synthesis of gold nanoparticles (AuNPs)

2. Experimental section 2.1. Materials and apparatus All voltammetric experiments were performed using an electrochemical analyzer Autolab 20 (Eco-Chemie, The Netherlands) connected to a personal computer using a software package GPS 4.9 (General Purpose Electrochemical System). A thermoshaker TS1 (Biometra) was used to stir the samples operating at controlled temperature. Transmission Electron Microscope (TEM) images were taken with Jeol JEM-2011 (Jeol Ltd., Japan). Scanning electrochemical microscopy (SEM) images were acquired using a Field Emission-Scanning Electron Microscopy (Merlin, Carl Zeiss). The electrochemical transducers were homemade screenprinted carbon electrode (SPCEs), which are constituted by three electrodes in a single strip: carbon working electrode (WE) with diameter of 3 mm, Ag/AgCl reference electrode (RE) and carbon counter electrode (CE). A magnet (3 mm in diameter), inserted under the WE, was also used to accumulate the complex formed due to magnetic beads modication with anti-Salmonella rst The Turkevich synthesis generates AuNPs of 20 nm (Fig. SI-1). First a solution of 0.508 mL HAuCl4 (1% m/v) in 49.492 milli-Q water was heated at 150 1C and stirred. When the solution was boiling, 5 mL of sodium citrate (40 mmol L 1) were added rapidly. In the next 10 min of heating and stirring the solution changed its color from pale yellow to red; it was stirred for iz et al., 2009a) and another 15 min at 25 1C (De la Escosura-Mun after this step the AuNPs were ready to use. AuNPs were protected from the light and stored at 4 1C. 2.5. Conjugation of anti-Salmonella rabbit polyclonal second antibody with AuNPs First, 100 mL of anti-Salmonella rabbit polyclonal second antibody (sSAb) (1 mg mL 1) was added with gentle stirring in 1.5 mL of colloidal gold suspension with pH adjusted to 9.0 using borate buffer 50 mM. It was incubated for 20 min at 25 1C and 650 rpm

A.S. Afonso et al. / Biosensors and Bioelectronics 40 (2013) 121126

123

and after that, 100 mL BSA 5% (in milli-Q water) was added and incubated again for 20 min at 25 1C and 650 rpm. Finally, the suspension of sSAb-AuNPs was centrifuged for 20 min at 14000 rpm at 4 1C and suspended in 1.5 mL of PBS with 0.3% BSA according to optimization of immunoassay procedure explained below.

2.7. SEM sample preparation for MBs-pSAb and MBs-pSAb/S immunoassay After the incubation of bacteria with MBs-pSAb, as described above, the MBs-pSAb/S were kept in PBS suspension and treated with glutaraldehyde solution followed by sequential dehydration with ethanol and resuspension in HMDS (hexamethyldisilazane) solution. This protocol is well suited for xation of bacteria in suspension. SEM images were acquired after dropping 4 mL of sample onto a 0.5 0.5 mm2 SiO2 wafer.

2.6. The magneto immunoassay The magneto-immunoassay (see schematic presentation in Fig. 1A) was performed by mixing 500 mL of solution of different cells of dead Salmonella (diluted in PBSTween in Eppendorf tubes of 1.5 mL) with 10 mL MBs-pSAb. The mixture was incubated for 30 min at 25 1C with 700 rpm to form the MBs-pSAb/S magneto-immunoconjugate. After this, the MBs-pSAb/S was separated from the supernatant by placing the eppendorf tubes in a magnetic separator for 3 min and then the supernatant was discarded. The washing step was performed for 2 min with PBS Tween at 25 1C (700 rpm) and MBs-pSAb/S were separated from supernatant. After two washing steps, the MBs-pSAb/S magnetoimmunoconjugate was resuspended with 140 mL of AuNPs modied with Salmonella antibody (sSAb-AuNPs) and incubated for 35 min at 25 1C and 700 rpm. Afterwards, the formed MBspSAb/S/sSAb-AuNPs magneto-immunosandwich was magnetically separated again, and two times washing step performed as before. Finally MBs-pSAb/S/sSAb-AuNPs was resuspended in 150 mL PBS and used for further electrochemical analysis.

2.8. Electrochemical measurements The MBs-pSAb/S/sSAb-AuNPs magneto-immunosandwich has been detected by using SPCEs and electrochemical detection based on AuNPs label signal (see schematic in Fig. 1B). An aliquot of 25 mL of MBs-pSAb/S/sSAb-AuNPs magneto-immunosandwich and 25 mL of 0.2 M HCl was inserted onto SPCE surface while applying a magnetic eld below the SPCE. The electrochemical detection using DPV technique with parameters previously optimized by Ambrosi et al. (2007): DPV was performed by scanning from 1.25 to 0 V (step potential 10 mV, modulation amplitude 50 mV, scan rate 33.5 mV s 1).

3. Results and discussion 3.1. SEM characterization of MBs-pSAb and MBs-pSAb/S magnetoconjugates Biological samples often lack the requirements of structure stability and electron conductivity necessary for high magnication SEM images, and it is often necessary to apply metalization procedures that cover the entire sample with a nano/micro layer of conductive material that hide the low rugosity of small particles interacting with the microorganisms surface. To avoid the possible SEM artifacts introduced by the mentioned procedures, we applied another sample preparation protocol that allows a good xation of bacteria and proved to be well suited for SEM analysis. SEM images in Fig. 2 clearly show the immunologic attachment of the bacteria to the MBs-pSAb forming MBs-pSAb/S magneto conjugate. Fig. 2A shows the MBs-pSAb before incubation and Fig. 2B that corresponds to the incubation of MBs-pSAb with 105 cells mL 1 Salmonella, shows aggregates due to interaction between MBspSAb and bacteria. The difference between them is concordant with the good recognition obtained during electrochemical detection of Salmonella, even in the presence of E. coli as interfering bacteria (images are not shown). It is important to point out that the micrographs of Fig. 2B correspond to fragments of bacteria due to the thermal treatment used to kill these.

Fig. 1. Schematic (not in scale) of Salmonella detection. (A) Principle of the assay. In a rst step incubation of Salmonella (S) with magnetic beads (MBs) modied with primary antibodies specic to the bacteria (pSAb) (MBs-pSAb) occurs. During this step Salmonella is captured and remains in the MBs-pSAb/S conjugate. During the second step MBs-pSAb/S conjugate is captured through application of a permanent magnetic eld and washed accordingly. Third step consists in the incubation of MBs-pSAb/S conjugate with gold nanoparticles (AuNPs) modied with secondary antibodies (sSAb-AuNPs) and captured again through application of a permanent magnetic eld and washed accordingly. (B) Electrochemical detection of MBs-pSAb/S/sSAb-AuNPs onto SPCE captured with a magnetic eld (step 5) by using DPV technique. Other experimental conditions as described in the text.

Fig. 2. SEM images of MBs-pSAb: before (A) and after (B) incubation with 105 cells mL 1 of Salmonella. Other experimental conditions as described in the text.

124

A.S. Afonso et al. / Biosensors and Bioelectronics 40 (2013) 121126

3.2. Optimizations of the immunoassay parameters The optimization of labeling parameters of MBs-pSAb/S with sSAb-AuNPs was performed by evaluating incubation times, blocking agent and different concentrations of sSAb-AuNPs (0.3; 0.6; 1.45; 3.6 and 7.21 mmol L 1 sSAb-AuNPs) for 105 cells mL 1 Salmonella, collected in PBSTween. Fig. 3A and B show the optimization of incubation time (from 10 and 30 min, respectively) between MBs-pSAb/S conjugate and sSAb-AuNPs (in PBS or PBSTween with 0%; 0.3% and 1% of blocking agent (BSA)). The volume of MBs-pSAb for the immunoassay was recommended by supplier (10 mL). Although all experiments performed were useful for Salmonella detection an optimum result in terms of nonspecic adsorption, current value and standard deviation was obtained by using PBS with 0.3% BSA with an incubation time of 30 min. (see Fig. 3B). In this assay, BSA was very important to reduce unspecic interaction between MBs-pSAb and sSAbAuNPs. Later, the inuence of the concentration of sSAb-AuNPs in PBS with 0.3% of BSA to be used during its second immunoreaction with bacteria was also evaluated. As shown in Fig. 3C sSAb-AuNPs concentration slightly affects the immunoreactions response, mainly on the reproducibility of method. However, an optimum and reliable signal was achieved for sSAb-AuNPs 1.45 mmol L 1. Thus, PBS with 0.3% BSA was used as buffer in all experiments for labeling the MBs-pSAb/S with sSAb-AuNPs (1.45 mmol L 1 for 30 min). 3.3. Immunosensor response towards Salmonella Fig. 4 shows the Salmonella detection (from sample collected in PBSTween) obtained due to the signal coming from the sSAb-

AuNPs label. The results obtained for the developed immunosensor for increasing concentrations of target (from 102 to 107 cells mL 1) by using DPV technique show a linear response (from 103 to 106 cells mL 1 with r2 0.985). For this assay the current value corresponding to the LOD was estimated by processing ve negative control samples (0 cells mL 1) that were performed in two different single inter-day assay, obtaining a mean value of 0.75 mA (n 5) that corresponds to 143 cells mL 1 with total time of analysis of 1:30 h. The precision of the method was evaluated by testing six different samples with 105 cells mL 1 of Salmonella. The coefcient of variation (CV) obtained was 2.4% (n 6) indicating good reproducibility under the conditions describes. A comparison of these results with those reported previously using other methods, based on nanoparticles, showed an improvement in general of this approach (see Table SI-1 Supporting information). 3.4. Specicity study for the immunoassay approach Once the feasibility of detecting of Salmonella using AuNPs was demonstrated, this assay shows the specicity of the developed immunoassay by using PBSTween and skimmed milk for evaluating the response toward the Salmonella target in the presence of E. coli as possible interference (see Fig. 5). The current values obtained for E. Coli assays (in PBSTween and skimmed milk) show similar values (0.85 and 0.98 mA, respectively) as the control (PBSTween and skimmed milked without bacteria) assays (0.66 and 0.92 mA, respectively). Thus, as expected, the electrochemical signal obtained for E. coli was almost 85% lower than the one corresponding to Salmonella in both performed assays (PBSTween and skimmed milk). However, for E. Coli assay in PBSTween an

Fig. 3. Optimization of the immunoassay approach with sSAb-AuNPs in PBS; PBS BSA 1%; PBS BSA 0.3%; PBSTween (PBST); PBST BSA 1% or PBST BSA 0.3%. This assay was performed with incubation times of 10 min (A) and 30 min (B). Inuence of the different concentration of sSAb-AuNPs diluted in PBS with 0.3% BSA onto the immunoassay response (C). For these assays, 105 cells mL 1 of Salmonella were used. Other experimental conditions as described in the text.

Fig. 4. (A) Typical DPV curves obtained using AuNPs electrochemical detection; (1) immunoassay without Salmonella, (2) 102, (3) 103, (4) 104, (5) 105, (6) 107, (7) 106 cells mL 1 of bacteria. (B) Electrochemical results obtained in the range between 102 and 107 of Salmonella. Dot line represent LOD based of 3 times standard deviation (n 5) of control (A-1) plus average of control. The error bars show standard deviation for n 3. Other experimental conditions as described in the text.

A.S. Afonso et al. / Biosensors and Bioelectronics 40 (2013) 121126

125

4. Conclusions A specic and rapid electrochemical based magneto-immunosensor for Salmonella detection in food samples by using AuNPs has been performed. Salmonella has been captured from the samples of skimmed milk and preconcentrated by immunomagnetic separation, followed by labeling with AuNPs modied with a polyclonal anti-Salmonella antibody. The developed immunosensor is able to detect up to 143 cells mL 1 Salmonella at a rather shorter time (up to 1:30 h). The obtained results are better than those reported previously not only in the response time but also due to the fact that AuNPs are easy to be obtained, modied and detected. The synergy between the immunoassay and magnetic particles has led to an enhancement of the sensitivity and removal of interferences from other species. Finally, this technique of detection is suitable for the rapid and sensitive screeningout of Salmonella in real samples. Furthermore, it could nd several applications in food, medical and environmental elds where a rapid, cost-efcient and easy to use device for in-eld applications is required.

Fig. 5. Specicity study for immunoassay performed in both PBSTween (PBST) and diluted skimmed milk, inoculated with E. coli (107 cells mL 1) and a mix of E. coli and Salmonella (with 107 cells mL 1 of each bacteria). Control samples correspond to immunoassays without bacteria. Other experimental conditions as described in the text.

Acknowledgements We acknowledge MICINN (Madrid) for the projects PIB2010JP00278 and IT2009-0092, and the NATO Science for Peace and Security Programmes support under the project SfP 983807 and co e Tecnolo to the Conselho Nacional de Desenvolvimento Cient Santiago gico (CNPq), Brasil for the scholarship given to Andre Afonso, grant number 200826/2011-5 and also Torres Quevedo rez-Lo pez. scholarship given to Briza Pe

Table 1 Spike and recovery from milk diluted 1/10 in PBSTween articially inoculated with Salmonella. Sample Added Found Recovery (%) Current mean (lA) 7 SD 1.01 7 0.01 1.25 7 0.02

Milk

1.5 103 cells mL 1 1.23 103 cells mL 1 83.0 1.5 105 cells mL 1 1.41 105 cells mL 1 94.0

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.2012.06. 054.

increase of the signal of around 28% was observed. This result explained also by the suppliers can be related to a certain degree of cross reactivity and non-specic binding of the used antibody (Invitrogen, 2011; KPL, 2011). Nevertheless this level of interference does not affect the feasibility of the detection. On the other hand the result obtained for a mixture of both pathogens is similar with that obtained for the sample spiked only with 107 cells mL 1 Salmonella.

References
Ambrosi, A., Airo, F., Merkoc - i, A., 2010. Analytical Chemistry 82, 11511156. eda, M.T., Killard, A.J., Smyth, M.R., Alegret, S., Merkoc Ambrosi, A., Castan - i, A., 2007. Analytical Chemistry 79, 52325240. CDC, 2011. /http://www.cdc.gov/salmonella/index.htmlS. iz, A., Maltez-da Costa, M., Merkoc De la Escosura-Mun - i, A., 2009a. Biosensors and Bioelectronics 24, 24752482. iz, A., Parolo, C., Maran, F., Merkoc De la Escosura-Mun - i, A., 2011. Nanoscale 3, 33503356. iz, A., Parolo, C., Merkoc De la Escosura-Mun - i, A., 2010. Materials Today 13, 2434. nchez-Espinel, C., D az-Freitas, B., Gonza lez-Ferna n iz, A., Sa De la Escosura-Mun dez, A., Maltez-da Costa, M., Merkoc - i, A., 2009. Analytical Chemistry 81, 1026810274. Fung, D.Y.C., 2002. Comprehensive Reviews in Food Science and Food Safety 1, 122. ICMSF, 2002. Microorganisms in Foods 7Microbial Testing in Food Safety Management, 1 ed. Kluwer Academic/Plenum Publishers, New York. Invitrogen, 2011. Dynabeadss Anti-Salmonella. /www.invitrogen.comS. ISO, 2002. Microbiology of Food and Animal Feeding StuffsHorizontal Method for the Detection of Salmonella spp., 4 ed. International Organisation for Standardization, Geneva, Switzerland. KPL, 2011. Afnity Puried Antibody to Salmonella Common Structural Antigen (CSA-1). /www.kpl.comS. Liebana, S., Lermo, A., Campoy, S., Barbe, J., Alegret, S., Pividori, M.I., 2009a. Analytical Chemistry 81, 58125820. Liebana, S., Lermo, A., Campoy, S., Cortes, M.P., Alegret, S., Pividori, M.I., 2009b. Biosensors and Bioelectronics 25, 510513. Mata, D., Bejarano, D., Botero, M.L., Lozano, P., Constanti, M., Katakis, I., 2010. Electrochimica Acta 55, 42614266. Merkoc - i, A., 2010. Biosensors and Bioelectronics 26, 11641177. Merkoc - i, A., Aldavert, M., Marin, S., Alegret, S., 2005. TrAC, Trends in Analytical Chemistry 24, 341349.

3.5. Analysis of Salmonella in real samples The immunosensor was applied to determine the level of Salmonella in skimmed milk. In order to determine the accuracy of the biosensor technology, skimmed milk purchased in local commerce area was spiked with Salmonella at different concentrations. Recoveries of Salmonella in the range of 83% and 94% (see Table 1) were calculated. These results demonstrate that the developed method can be a promising alternative to determine Salmonella in skimmed milk. The obtained detection in real samples (skimmed milk diluted 10 in PBSTween) is much lower than the one obtained for Salmonella detection in liquid samples (i.e. skimmed milk) by using standard commercial methods, resulting in a value of around 106 cells mL 1 (Fung, 2002). The results obtained show that AuNPs based detection technology combined with a magnetic eld application is capable of detecting Salmonella at lower concentration than by using other methods reported in the literature (see Fig. SI-2).

126

A.S. Afonso et al. / Biosensors and Bioelectronics 40 (2013) 121126 eda, M.T., Pividori, M.I., Eritja, R., Merkoc Pumera, M., Castan - i, A., Alegret, S., 2005. Langmuir 21, 96259629. Salam, F., Tothill, I.E., 2009. Biosensors and Bioelectronics 24, 26302636. WHO, 2011. Food Safety and Foodborne Illness. /http://www.who.int/mediacen tre/factsheets/fs237/en/S. Zelada-Guillen, G.A., Bhosale, S.V., Riu, J., Rius, F.X., 2010. Analytical Chemistry 82, 92549260.

Merkoc - i, A., Aldavert, M., Tarrason, G., Eritja, R., Alegret, S., 2005a. Analytical Chemistry 77, 65006503. Newell, D.G., Koopmans, M., Verhoef, L., Duizer, E., Aidara-Kane, A., Sprong, H., Opsteegh, M., Langelaar, M., Threfall, J., Scheutz, F., van der Giessen, J., Kruse, H., 2010. International Journal of Food Microbiology 139, S3S15. Noguera, P.S., Posthuma-Trumpie, G.A., van Tuil, M., van der Wal, F.J., de Boer, A., Moers, A., van Amerongen, A., 2011. Analytical Chemistry 83, 85318536. zou, M., Turner, A., Merkoc Perfe - i, A., 2012. Chemical Society Reviews 41, 26062622.