International Journal of Food Microbiology 189 (2014) 89–97

Contents lists available at ScienceDirect

International Journal of Food Microbiology journal homepage: www.elsevier.com/locate/ijfoodmicro

Development of a rapid capture-cum-detection method for Escherichia coli O157 from apple juice comprising nano-immunomagnetic separation in tandem with surface enhanced Raman scattering Roya Najafi b,1, Shubhasish Mukherjee c,2, Jim Hudson Jr. c, Anup Sharma d, Pratik Banerjee a,b,⁎ a

Division of Epidemiology, Biostatistics, and Environmental Health, School of Public Health, The University of Memphis, Memphis, TN 38152, USA Laboratory of Food Microbiology and Immunochemistry, Department of Food & Animal Sciences, Alabama A&M University, Huntsville, AL 35762, USA c HudsonAlpha Institute for Biotechnology, Huntsville, AL 35806, USA d Department of Physics, Alabama A&M University, Huntsville, AL 35762, USA b

a r t i c l e

i n f o

Article history: Received 18 March 2014 Received in revised form 25 July 2014 Accepted 30 July 2014 Available online 7 August 2014 Keywords: Escherichia coli O157 Rapid detection Nanoparticles Immunomagnetic separation (IMS) Surface enhanced Raman spectroscopy (SERS) Apple juice

a b s t r a c t A combined capture and detection method comprising of nano-immunomagnetic separation (NIMS) and surface enhanced Raman spectroscopy (SERS) was developed to detect Escherichia coli O157 from liquid media including apple juice. The capture antibodies (cAbs) were immobilized on magnetite–gold (Fe3O4/ Au) magnetic nanoparticles (MNPs) which were used for separation and concentration of the E. coli O157 cells from model liquid food matrix. The capture efficiency (CE) for E. coli O157 using MNP was found to be approximately 84–94%. No cross reactivity was observed with background non-target organisms. There was a significant difference in the mean CE of bacteria captured by MNP and commercially sourced immunomagnetic microbeads (p b 0.05). For the detection of target pathogen, SERS labels were prepared by conjugating gold nanoparticles with Raman reporter molecules and the detector antibody (dAb). Au-Raman labeldAb was interacted with gold coated MNP-cAb-E. coli O157 complex. The ability of this immunoassay to detect E. coli O157 in apple juice was investigated. We have successfully applied the synthesized Fe3O4/Au nanoclusters to E. coli O157 detection in apple juice using the SERS method. The lowest detectable bacterial cell concentration in apple juice was 102 CFU/mL with a total analysis time of less than an hour. This method presents a convenient way of preconcentration, separation, and detection of low levels of target pathogen from liquid food matrix. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Food safety is a major public health concern worldwide. According to the Centers for Disease Control and Prevention (CDC, USA), foodborne pathogens cause 9.4 million illnesses, 55,961 hospitalization and 1351 deaths annually in the United States (Scallan et al., 2011). Rapid and sensitive detection of pathogens in food products is a critical step in the food production chain to ensure food safety. Conventional detection and identification strategies for microbial pathogens include bacterial colony counting methods, immunology-based methods involving antigen/ antibody interactions, and the polymerase chain reaction (PCR) involving nucleic acid analysis. These methods are sensitive, inexpensive and can give both qualitative and quantitative information of the target microorganisms, but they are time consuming, labor intensive, and most of these ⁎ Corresponding author at: Division of Epidemiology, Biostatistics, and Environmental Health, School of Public Health, The University of Memphis, 338 Robison Hall, Memphis, TN 38152, USA. Tel.: +1 901 678 4443. E-mail address: [email protected] (P. Banerjee). 1 Present address: Qualitest Pharmaceuticals, Huntsville, Alabama 35811, USA. 2 Present address: Avanti Polar Lipids, Inc., Alabaster, Alabama 35007, USA.

http://dx.doi.org/10.1016/j.ijfoodmicro.2014.07.036 0168-1605/© 2014 Elsevier B.V. All rights reserved.

methods need initial enrichment for detection of pathogens (Velusamy et al., 2010). Given the perishable nature of most food items, a quick microbial test result will minimize product hold-time at the production facility enabling maximum utilization of the shelf-life of the product, which is a desirable aspect from food producers' point of view. Thus, rapid, sensitive, selective bacterial isolation and detection methods for food products are critically needed to ensure safety of food products (Bhunia, 2008; Mandal et al., 2011). For microbiological analysis of food materials, sample preparation is often the most critical step. Sample preparation involves multiple time-consuming steps such as stomaching, filtration and centrifugation in order to lower the so-called “food matrix effect” on the capture of bacteria (Bhunia, 2008). Furthermore, one of the challenges in detection of pathogens is that the number of target pathogens may be very low (10 to 1000 bacterial cells) in a given sample. So, there is a need for an efficient pathogen separation and concentration method to detect them successfully, and to avoid false-negative results. Immunomagnetic separation (IMS), an antibody-based method, is a widely used, sensitive, and simple sample preparation and concentration technique that has been successfully applied for isolation of pathogens from the food matrix (Bhunia, 2008; Dwivedi and Jaykus, 2011; Xiong

90

R. Najafi et al. / International Journal of Food Microbiology 189 (2014) 89–97

et al., 2014). This separation method eliminates multiple sample preparation steps such as, filtration or centrifugation, and can be applied directly to capture the target organisms from a complex matrix. In traditional IMS, paramagnetic metallic or polystyrene microbeads of varying diameters (~1–3 μm, referred as “MMB” hereafter) coated with organism specific antibodies are utilized to immunomagnetically capture bacterial cells from food samples, which are subsequently enumerated and identified by agar plating based methods, or by a follow-up analytical method, such as PCR or ELISA (IMS-PCR or IMS-ELISA) (Bhunia, 2008; Tu et al., 2009). Despite the advantages of this method, limitations such as nonspecific adsorption of non-target organisms on the immunomagnetic microbeads, poor limit of detection, post-IMS detection/identification time and costs attributed to relatively high reagent volumes are some of the major drawbacks that preclude widespread application of IMS methods in food microbiology (Liberti et al., 1996; Tomoyasu, 1998; Varshney et al., 2005). For improvement of IMS methods to counteract some of the problems mentioned above, magnetic nanoparticles (MNPs) have been used for target analyte capture (Gilmartin and O'kennedy, 2012; Perez-Lopez and Merkoci, 2011; Wang H. et al., 2011; Wang Y. et al., 2011). Nanoparticles with higher surface/volume ratio as compared to commonly used MMB present more contact surface area for attaching to the target molecules such as bacterial cells. Also the reaction kinetics of magnetic nanoparticles is faster than MMBs (Liberti et al., 1996; Varshney et al., 2005). The performance of the assay can be improved by using nanoparticles (Nps) as labels or transducers in a single assay (Guven et al., 2011; Naja et al., 2007; Penn et al., 2003; Temur et al., 2010; Wang H. et al., 2011; Wang Y. et al., 2011; Xu et al., 2005). Optical sensing methods such as surface enhanced Raman spectroscopy (SERS) are emerging and promising techniques for the detection of foodborne pathogens because of their sensitivity as well as the simplicity of interpreting the generated data (Bantz et al., 2011; Bhunia, 2008; Dutta et al., 2009). Unique physicochemical properties of noble metal Nps provide optical biosensing methods with a wide range of potential applications such as multiplexing or label free detection. Some other specific advantages include reducing reagent volumes and detection time, offering high sensitivity, lower cost, and eliminating the need for skilled personnel (Perez-Lopez and Merkoci, 2011). Gold (Au) Nps can significantly amplify the Raman scattering efficiencies of adsorbed molecules due to their unique optical properties (Zhou et al., 2010). SERS strongly increases Raman signals from the molecules attached to nanometer-sized metallic structures (Campion and Kambhampati, 1998; Haynes et al., 2005). Due to the high Raman signal enhancement, SERS allows for the acquisition of amplified spectra from single bacterial cells in seconds, even without the need for labeling or sample amplification (Premasiri et al., 2005). In food microbiological applications, IMS employing noble metal coated MNP followed by SERS based detection can significantly reduce the over-all assay time, as it does not require any plating of captured bacterial cells, and the post-IMS detection by this method takes less time than IMS-PCR or IMS-ELISA (Chu et al., 2008; Guven et al., 2011; Wang H. et al., 2011; Wang Y. et al., 2011). In this paper, we report the development of a combined capture and detection approach for Escherichia coli O157 assay utilizing nano-immunomagnetic separation (NIMS) and SERS. For capture and separation of the target microorganisms, magnetic iron Nps (‘core’) were coated with gold Nps (‘shell’) and capture antibodies in one nanocomposite. These nanocomposites were utilized in the separation and concentration of the target cells from samples. Subsequently, the detector antibodies and Raman reporter molecules conjugated on gold Nps were used for a rapid SERS based detection and enumeration of the captured cells. The assay performance of the newly developed NIMS was evaluated in terms of sensitivity, selectivity, reagent utilization, and assay time. The assay validation was done using artificially E. coli O157:H7 spiked apple juice as model food-pathogen combination.

2. Materials and methods 2.1. Reagents and materials All the chemical and analytical reagents used in this study were obtained from Sigma-Aldrich (St. Louis, MO), if not mentioned otherwise. Purified mouse anti-E. coli O157 monoclonal antibody (clone 9/ 140.156.152.11) was obtained from EMD Millipore (Billerica, MA). Dynabeads MAX E. coli O157 and Dynal magnet for magnetic separation were purchased from Life Technologies (Carlsbad, CA). Malachite green isothiocyanate (MGITC) was obtained from Fisher Scientific (Pittsburg, PA). Ultrapure water was used throughout the work. 2.2. Bacterial cultures Bacterial cultures (target bacteria: E. coli O157:H7 EDL933, ATCC 43895; and the other non-O157 bacteria: Listeria monocytogenes Scott A, Bacillus cereus A926, Enterococcus faecalis 102StE3, Staphylococcus epidermidis 101StE1, Micrococcus luteus 4698, and Streptococcus mutans [“Gram-positive group”]; and Citrobacter freundii, Enterobacter cloacae, Escherichia vulneris 301ECD32, Hafnia alvei, Klebsiella oxytoca 503ECE45, Pseudomonas aeruginosa ATCC 10145, and Salmonella Enteritidis 13096 [“Gram-negative group”]) were subcultured twice from −80 °C frozen stocks in brain–heart infusion (BHI) broth (Difco Laboratories) at 37 °C. For routine experiments, one hundred microliters of the bacterial culture was added to 10 mL of BHI broth and incubated overnight at 37 °C. Fresh overnight grown cultures (1 mL each) were centrifuged (13,000 ×g for 10 min) at room temperature, and the cell pellets were washed twice in filter (pore size of 0.45 μm) sterilized PBS. 2.3. Synthesis of gold-coated magnetic nanoclusters The iron oxide/gold (Fe3O4/Au) nanoclusters were prepared by synthesizing iron oxide Nps (the core) by oxidizing Fe(OH)2 to Fe3O4 nanoparticles with PEI on the surface according to the method described earlier (Sugimoto and Matijevic, 1980). The 2 nm colloidal gold nanoparticles were synthesized by using a method described earlier (Jana et al., 2001). These iron oxide Nps and the gold seed colloids were utilized in the synthesis of the ‘core–shell’ iron oxide/gold (Fe3O4/Au) nanoclusters following the procedure outlined by Goon et al. (2009) with some modifications. Briefly, gold seed colloidal solution (94 mL) was stirred with 1.09 mL of Fe3O4 Nps (20 nm) for 1.5 h. Then the Fe3O4/Au seed particles were mixed with 80 mL of PEI (0.5 mM) and heated at 60 °C for 3 h. This step was followed by rinsing the solution three times with water and removing dissociated Au seeds through magnetic separation. The resulting nanoparticles were dispersed in ultrapure water. Finally, the magnetic particle/Au seed aqueous solution was stirred with NaOH aqueous solution (110 mL, 0.01 M), and subsequently, NH2OH∙HCl and HAuCl4 were added for five iterations. For the first iteration, 0.75 mL of NH2OH∙HCl (0.2 M) and 0.5 mL of HAuCl4 (25.4 mM) were added and mixed for 10 min. Then, for the second, third, fourth and fifth iterations, 0.5 mL of NH2OH∙HCl and 0.25 mL of HAuCl4 were added resulting in the formation of Fe3O4/Au nanoclusters. 2.4. Synthesis of antibody-conjugated Fe3O4/Au nanocluster (MNP-cAb) and Raman nanoprobes (RNP-dAb) The conjugation of capture antibody on Fe3O4/Au nanoclusters was done by a method described earlier by Zhou et al. (2010). The concentration of the capture antibody used for the conjugation was 1.15 mg/mL. The obtained antibody conjugated nanocluster was resuspended in PBS and kept at 4 °C. For the synthesis of Raman nanoprobes (RNPdAbs) containing malachite green isothiocyanate (MGITC) as Raman labels and the detector antibody we followed a method described by Qian et al. (2008). The antibody concentration on the SERS nanoprobes was quantified using a NanoDrop spectrophotometer.

R. Najafi et al. / International Journal of Food Microbiology 189 (2014) 89–97

91

2.5. Characterization of synthesized nanoparticles

2.7. Immunoassay procedure

Electron microscopy was used to characterize different nanoparticles synthesized in this study. For scanning electron microscopy (SEM), nanoparticle samples were placed on the SEM holder. The sample was dried and sputter coated with a thin layer of gold. SEM images were obtained by using a JOEL JSM 6335F scanning electron microscope (JOEL, Tokyo, Japan) at an accelerating voltage of 25 keV, in the high vacuum/secondary electron imaging mode of operation. Using the energy dispersive spectrometer (EDS) feature of the SEM instrument, the elemental analysis of the samples was performed to evaluate the extent of gold-coating on the iron MNPs. For EDS analyses, samples were not sputter coated with gold. For transmission electron microscopy (TEM), approximately 10 μL of nanoparticle solution was directly onto a TEM grid. The samples on the grid were then allowed to air dry for 10 min. TEM micrographs were obtained using a JEM 1200EX (JEOL) with an accelerating voltage of 100 keV.

The immunoassay was based on a sandwich-type immunoassay. Fig. S1 (in the Supplementary information) is a schematic depiction of the immunoassay using MNP-cAb mediated NIMS followed by SERS based detection using Raman nanoprobes (RNP-dAbs). First, immunomagnetic separation and washing of nano-aggregate of MNPcAb were performed as described in the previous section. Then, RNPdAbs were added to the MNP-cAb nano-aggregate and incubated for 25 min at room temperature (~23 °C). The composite was again subjected to magnetic separation, and then the aggregate (RNP-dAbs bound to MNP-cAbs through E. coli O 157:H7) was washed three times (Zhou et al., 2010). After the final washing, the sample was resuspended in PBS for SERS measurement.

2.6. Capture efficiency (CE) and the specificity of antibody-conjugated nanocluster In order to determine the capture efficiency of MNP-cAb, overnight grown E. coli O157 cells were prepared in BHI. Then, MNP-cAb was mixed with various concentrations of E. coli (101–108 CFU/mL) and incubated for 25 min at room temperature (~23 °C) in a shaker incubator. We used two different reaction volumes, 5 μL of the MNP to 250 μL of bacteria, and 20 μL of the MNP to 1 mL of bacteria. Next, the MNPcAbs (with potentially captured bacteria) were separated magnetically using an external magnetic particle concentrator (MPC, DynaMag; Life Technologies, Carlsbad, CA). The magnetically separated MNP-cAbs were washed in PBST three times, and then resuspended again in PBS. This was immediately followed by plating of bacteria captured MNPcAbs on sorbitol–MacConkey agar supplemented with cefixime–tellurite (CT) supplement (CT-SMAC; BBL/Difco, Sparks, MD). The plates were incubated in a microbiological incubator (Labline, Fisher Scientific, Pittsburg, PA) for 18–24 h at 37 °C in order to enumerate bacteria. The CE was calculated according to a method described earlier (Perez and Mascini, 1998) utilizing the following equation:

CE% ¼

Cb  100 Co

where Co is the total number of cells in the sample (CFU/mL), and Cb is the number of bound cells (CFU/mL). We compared the CE of the MNPcAb with Dynabeads MAX MMB (Life Technologies, Carlsbad, CA). While performing IMS using MMBs, in one set of experiments, we followed manufacturer's recommended protocol using 20 μL of MMB per 1 mL of E. coli O157:H7 suspensions. In another set, we used 5 μL of the MMB with 250 μL of E. coli suspension (to replicate MNP-cAb mediated IMS, described above). The specificity of the MNP-cAb was evaluated by testing with a mixture of Gram-positive and Gram-negative non-O157 bacteria following a method described earlier (Banerjee and Bhunia, 2010) with some modifications. Briefly, the MNP-cAb was added to a mixture of nonO157 bacteria containing “Gram-positive group” and “Gram-negative group” (described in Section 2.2) where the final concentration of each of the cultures in the non-O157 bacteria mix was 1 × 106 CFU/ mL. Following NIMS and washing (3 times), MNP-cAbs were plated on BHI agar. For CE enumeration in the presence of non-target flora, varying concentrations of E. coli O157:H7 (1 × 105–1 × 101 CFU/mL, final concentrations) were separately added to non-O157 bacteria mix (described above). Following NIMS and washing, selective plating (CT-SMAC) was done to enumerate the captured bacteria. The CE was calculated as above.

2.8. SERS measurement SERS experiments were performed using the portable Enwave Optronics EZ Raman I system (Enwave Optronics, Inc., Irvine, CA) which houses a frequency stabilized diode laser with an excitation wavelength of 785 nm and a power of 40 mW. The system uses a fiber optic probe with a 7 mm focal point, a 100 μm diameter, and a 0.3 numerical aperture. This probe transmits the Raman signal to a spectrophotometer equipped with a thermoelectrically cooled (−50 °C) CCD detector. After placing the sample on the glass slide, the laser beam was placed orthogonal to the sample surface area, and spectra were recorded in the spectral range (Raman shift) of 250 cm−1 to 2350 cm−1. The acquisition time for each recorded spectrum was 60 s. For the instrument control and initial acquisition of spectral data, the in-built RamanReader data collection software (including TimeChart and TimeTrend functions) (Enwave Optronics) was utilized. The raw spectral data was exported to Microsoft Excel for plotting and reporting. 2.9. Validation studies in apple juice Pasteurized apple juice purchased from a local store was used directly at room temperature (∼23 °C). E. coli O157:H7 prepared in different concentrations (101–108 CFU/mL) as stated above was added to aliquots of apple juice in microfuge tubes. After thoroughly mixing the MNP-cAbs for 25 min at room temperature by slowly rotating the microfuge tubes, NIMS was carried out. The samples were washed thrice with PBST to remove any non-specific bindings and were redispensed in 0.5 mL of PBS. Next, 0.1 mL of these samples was used for a SERS analysis and was also plated on CT-SMAC. 2.10. Statistical analysis Experiments were performed in duplicates. The average values (three independent experiments) and standard error of means (Std. E.M.) were determined for CFU/mL values (as obtained by MNP-based capture and MMB-based capture). The data analysis was done by performing ANOVA (SAS, version 9.4, Cary, NC., USA) and to find out differences in CE values among different volumes of MNP or MMB, Tukey's test was performed on a completely randomized design. The limit for statistical significance was set at p b 0.05. 3. Results and discussion 3.1. Characterization of gold-coated iron nanoclusters and SERS nanoprobes Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) were employed to examine the morphology and intrinsic structure of the synthesized Fe3O4/Au nanoclusters. The visualization of the structural property (i.e., the shape) is critical to ensure the success of synthesis process. The shape of the Nps is reported to be an important parameter for improved SERS signals (Tamer et al., 2011).

92

R. Najafi et al. / International Journal of Food Microbiology 189 (2014) 89–97

Fig. 1a and b shows the SEM and TEM of the synthesized Fe3O4 nanoparticles, respectively. It is evident that the Fe3O4 nanoparticles have a cuboid shape. Fig. 1c and d shows the SEM and TEM of the Fe3O4/Au nanocluster. After coating the Fe3O4 nanoparticles with gold seeds, a rough gold nanoshell was formed. Gold nanoparticles (indicated by “➙”) can be seen on the surface of the iron nanoparticles (core) in Fig. 1d. TEM and SEM results of both Fe3O4 nanoparticles and the Fe3O4/Au nanoclusters showed the difference in the morphology and structure of Fe3O4 nanoparticles after being coated with gold seeds. The results showed that a single Fe3O4 nanoparticle has a cuboid morphology with smooth surfaces. After the nanoparticles were coated with gold seeds, the TEM images revealed that many small gold nanoparticles had attached to the surface of PEI-coated Fe3O4 and produced a rough surface. The TEM image also revealed that gold nanoparticles were only on the surface of the Fe3O4 nanoparticles; this was based on the fact that PEI was coated homogeneously on the Fe3O4 surfaces, and Au was easily deposited onto those surfaces. The energy dispersive X-ray spectroscopy (EDS) results also verified gold coating on the iron nanoparticles (data not shown). Our nanoparticle characterization results were in agreement with the previous research done by Zhou et al. (2010), where the uniform distribution of gold on Fe3O4 nanoparticles was attributed to the homogenous attachment and growth of Au seeds onto the nanoparticles. It is also well documented that cuboid gold nanoparticles exhibited significantly stronger SERS response as compared to spherical-shaped gold nanoparticles (Gou and Murphy, 2005; Narayanan et al., 2008; Tamer et al., 2011). In order to prepare SERS nanoprobes, gold nanoparticles with the size of 60 nm were functionalized with Raman dye and antibody. Gold nanoparticles with the size of 60–80 nm have shown more efficiency for SERS at red (630–650 nm) and near-infrared (780 nm) excitations as reported previously (Krug et al., 1999). Hence in the present study we used 60 nm gold nanoparticles at a 780 nm excitation. In order to ensure that the gold nanoparticles are covered with dye (MGITC), Raman spectroscopy (SERS) was performed. Fig. 2 shows the Raman signal of MGITC and SERS nanoprobes composed of gold nanoparticles that were coated with MGITC and an antibody. The presence of antibody on the surface of gold nanoparticles was examined using a NanoDrop spectrophotometer. The concentration of antibody in the gold nanoparticles and SERS nanoprobes measured at 280 nm was 1.06 mg/mL, which is almost the same as the concentration of antibody used for conjugation (1.15 mg/mL). Based on the NanoDrop results, modification of the Au nanoparticles with antibody was successful. It is also apparent in

Fig. 2 that the peak intensities of the pure MGITC (spectra a) at regions of 1180, 1370, and 1620 cm−1 are enhanced (spectra b) using SERS nanoprobes. This characteristic signal enhancement is attributed to the presence of gold as a SERS substrate, which is in agreement with previously reported findings (Neng et al., 2010; Pettinger et al., 2004, 2005). The differences in the two spectra can be attributed to the Raman intensities of two different entities represented in Fig. 2, MGITC only (spectra a) and MGITC + Au + dAb (spectra b). The presence of additional moieties (dAb) seems to cause minor alteration of the spectral character of spectra b as compared to spectra a in some spectral regions. For example, appearance of a peak in the 530 cm−1 region in spectra b can be attributed to S–S structure. Similarly different peaks are related to different species, such as 725 cm−1 (adenine) and 1230 cm−1 (amide III), as reported previously (Huang et al., 2010; Lu et al., 2011; Maquelin et al., 2002). 3.2. CE of the MNP-cAbs The bacterial concentrations (CFU/mL) for both the control and test samples were calculated to determine the CEs of NIMS and IMS (MMBmediated) methods for different bacterial concentrations. Fig. 3 shows the relationship between bacterial cell concentration and CEs using MNP-cAbs and MMBs for the immunoreaction time of 25 min in four different target concentrations. For the four different concentrations of E. coli O157, the CE values of MNP-cAbs were found to be between 84% and 94%. The overall percentage of captured bacteria for E. coli O157 was found to be approximately 85%. Based on the E. coli cell concentration shown in Fig. 3, there was not a significant difference between the bacteria captured by the Ab-coated nanocluster (NIMS) and that captured by the anti-E. coli O157 MMBs (traditional IMS) except for 108 CFU/mL. However, it is noteworthy that for NIMS, we used 5 μL of the MNP-cAbs with 250 μL of cell-suspension; while for traditional IMS 20 μL of the MMBs was reacted with 1 mL sample volume. Our objective of utilizing lower reagent and sample volumes was to evaluate if a nanoparticle-based IMS method can be realized with lesser amounts of test sample and reagents as compared to a microbead-based IMS method. Since the antibody-containing reagents are expensive, hence reduction in reagent volume would reduce the cost per assay. For traditional IMS format, when the volume of MMB was reduced to 5 μL per 250 μL sample volumes, CE values were found to be significantly (p b 0.05) lower across all four test concentrations (Fig. 3). For 7.3 × 102 and 7.5 × 104 CFU/mL of E. coli, the mean values of CE

Fig. 1. SEM and TEM images showing surface geometry of nanoparticles. The top two panels show Fe3O4 nanoparticles (a, b), and bottom two panels show Fe3O4/Au nanocluster (c, d). Gold nanoparticles coated on Fe3O4 nanoparticles are shown by “➙”.

R. Najafi et al. / International Journal of Food Microbiology 189 (2014) 89–97

Another observation from their study is that the CEs of MNP did not change across various concentrations of E. coli O157 in this test indicating that the concentration of bacteria possibly is not a limiting factor. It is evident from our results that in NIMS the volume of antibodyconjugated capture particles needed is significantly less than MMB based separations. Representative TEM images of MNP and MMB mediated immunomagnetic separation of E. coli O157 cells from a sample volume of 250 μL with 5 μL of magnetic particles are depicted in Fig. 4a and b, respectively. Also, it can be seen in Fig. 4a and c that the nanoparticles seem to form a long chain like structure. We could also identify discrete nanoparticles which are involved in bacterial capture as seen in Fig. 4d. The better CE values of MNPs obtained at lower sample and reagent volumes can be partially attributed to the fact that several MNPs are associated with one single bacterial cell (Fig. 4a, c and d). While in the same condition, MMBs captured less number of cells, and several beads can be found to form clusters that are not associated with significant amount of captured cells (Fig. 4b). Prior studies show that antibody conjugated nanoparticles cover the surface of bacterium due to their small size (Zhao et al., 2004). So the number of nanoparticles which is required for the capture of E. coli O157 can be one of the factors affecting the CE. The small size and large surface area of Fe3O4/ Au nanocluster-Ab produce faster binding kinetics to the bacterial cells, especially in the food matrix (Qian et al., 2008). These could be some plausible factors attributing to the higher CE values obtained by us using MNP as compared to MMB. The reported capture efficiencies of E. coli employing MNP by different groups varied from 55 to 97%. For example, the CE value of gold coated magnetic iron oxide nanospheres (with an average diameter of 15.5 ± 3 nm) was 55% (Guven et al., 2011); in another study, amine-functionalized magnetic nanoparticles (20 nm in diameter) resulted in a CE of 80% (Cheng et al., 2009). Using iron oxide magnetic nanoparticles (145 nm in average diameter) Varshney et al. (2005) reported CE values of 96% or higher for E. coli O157:H7. Our results indicate that the CE of MNP made in this study was in the range of 84–94%, which was either comparable or higher than those reported in these previous studies. The specificity of synthesized MNP-cAbs in capturing different concentrations of E. coli O157 from non-O157 bacteria mix was evaluated. We have found no significant difference in CE values with or without the presence of background flora (p N 0.05, data not shown). Also, we observed no non-specific binding (of bacteria from both the “Gram-positive group” and “Gram-negative group” tested), as after performing three washings and plating of MNP-cAbs, we found no colonies on BHI

1000 units

(b)

(a)

Raman Shift (cm - 1 ) Fig. 2. SERS spectra for MGITC. (a) MGITC only, and (b) SERS nanoprobe (gold nanoparticles) mediated signal enhancement of MGITC (MGITC + Au + dAb).

for MNP-cAbs were 85 and 86%, respectively. Higher concentrations (7.8 × 106 and 6 × 108 CFU/mL) of E. coli resulted in the mean values of CE of 85 and 84%, respectively. For MMB-based capture, two different volumes, 5 μL per 250 μL (denoted as MMB-5) sample and 20 μL per 1 mL (denoted as MMB-20) sample were used. For MMB-20, the CE was found to be within the range of 85–77%. However, for MMB-5, the obtained CE values reduced to 59 to 36%, which were significantly lower than CE values obtained from MNP (except 108 CFU/mL). So based on these results, there was a significant (p b 0.05) difference between the CE of MNP and MMBs (for 5 μL of capture particles per 250 μL sample volume) across all bacterial concentrations tested. When MMB volume was increased to 20 μL (per 1 mL sample), the resulting CEs were found to be not significantly (p N 0.05) different from MNP (5 μL) CEs. We also noted that increasing the volume of MNP to 20 μL (per 1 mL sample) significantly improved the CE for lower concentrations (102 and 104 CFU/mL), but not for higher concentrations (106 and 108 CFU/mL, Fig. 3). Our results of MNP volume and CE were in agreement with the previous study by Varshney et al. (2005), where it was reported that a higher amount (10 versus 20 μL) of nanocluster-Ab conjugates lead to higher binding of bacteria (Varshney et al., 2005).

120 100

MNP-cAbs -20

MNP-cAbs -5

MMB-20

MMB-5

A

A B

Capture Efficiency (%)

93

B

B

B

A

AB

A

B

A B

80 C

C

60

C C

40 20 0

10^2

10^4

10^6

Concentration of E. coli O157:H7 (CFU/mL)

10^8

Fig. 3. CE of MNP and MMB in different E. coli O157:H7 concentrations. MNP-cAb-20 or MNP-cAb-5 and MMB-cAb-20 or MMB-cAb-5 represents the two different volumes (5 and 20 μL, respectively) of MNP and MMB tested in this study. The X-axis legends are (units in CFU/mL): 102, 7.3 × 102; 104, 7.5 × 104; 106, 7.8 × 106; and 108, 6 × 108. Values are presented as mean ± Std. E.M. of three experiments done in duplicate. Columns, mean; and bars, Std. E.M. Columns with different letters indicate significant differences (p b 0.05).

94

R. Najafi et al. / International Journal of Food Microbiology 189 (2014) 89–97

(a)

(c)

(b)

(d)

Fig. 4. TEM images of nanoparticle and microbead mediated captured E. coli O157:H7. The images are taken after immunomagnetic separation of cells from a sample volume of 250 μL with 5 μL of magnetic particles, nanoparticles (a) and microbeads (b). In some cases, nanoparticles formed “chain-like” nanocluster structures attached to captured bacterial cells (c), while discrete nanoparticles are also found to be involved in bacterial capture, as these particles can be seen attached to the cell surfaces as shown in (d). Captured E. coli O157:H7 is shown by “➨” while MNPs or MMBs are indicated by “➙”.

agar plates (data not shown). In a previous study, Guven et al. (2011) reported that washing three times post-IMS with PBST resulted in an optimal capture efficiency and also eliminated non-specific bindings. In our study we directly used this protocol and could eliminate attachment of non-target organisms. 3.3. Detection of E. coli O157 in apple juice using SERS-based sandwich immunoassay The newly developed NIMS method was validated in the food system by artificially spiking different concentrations of E. coli O157 in commercial apple juice. The target E. coli O157 acted as a bridge between MNP-cAbs and the Au-Raman label nanoparticles (Fig. S1). After collecting the sample by using an external magnet, SERS analysis of the magnet-separated sample was performed. The SERS spectra for E. coli O157 in apple juice were obtained using Raman-labeled gold nanoparticles (Fig. 5a). The results display typical SERS responses from the immunoassay for various concentrations of E. coli O157 (101 through 107 CFU/mL) in apple juice. Based on the SERS result, it is apparent that E. coli O157 could be detected within 5 min through the Raman intensity of the reporter molecule (MGITC). The incremental intensities of Raman spectra are directly related to the increasing concentrations of E. coli cells, which are consistent with the Raman shift signature of the MGITC reporter at 1180, 1370, and 1620 cm− 1 peak densities (Fig. 5b). The differential peak intensities as a function of bacterial cell concentrations are attributed to the fact that higher signal intensity depends on the amount of SERS nanoprobes that are attached

to the target pathogen. The average peak heights at 1620 cm−1 peak density were plotted against target concentrations of E. coli (Log CFU/mL) in Fig. 5c. An R2 value of 0.96 was obtained by performing a linear regression analysis of the peak heights versus Log CFU/mL of target E. coli (Fig. 5c). Therefore, the higher the captured cell concentrations, the higher is the amount of magnetically separated gold-Raman dye attributing to a higher signal. Various intensities of Raman spectra obtained for different concentrations of E. coli using a SERS based sandwich immunoassay were shown in previous research (Guven et al., 2011). The SERS based limit of detection (LOD) of their study using water samples varies between 27 and 451 CFU/mL. In another study, using silica-coated MNP and mercaptobenzoic acid (MBA) as Raman reporter, similar concentration-dependent SERS signal acquisition was reported for Staphylococcus aureus and Salmonella Typhimurium (Wang H. et al., 2011). The lowest detectable concentration of target bacteria using their method was 103 CFU/mL in a 0.5 mL total sample volume. The LOD of our method was 102 CFU/mL, and we observed no cross reactivity with other organisms tested. Therefore, the SERS result of the present study was comparable to previous studies mentioned above. Application of Fe3O4/Au nanocluster mediated NIMS followed by SERS-based detection eliminates many disadvantages of traditional IMS and ELISA. For example, following MMB mediated traditional IMS, the captured aggregates are enumerated either by plating or by a follow-up analytical method, such as PCR or ELISA. Plating followed by incubation usually takes 18–48 h or longer, while IMS-ELISA or IMSPCR may take between 2–4 h (Ozalp et al., 2013; Wang et al., 2013a, b). On the contrary, SERS-based detection following NIMS is very

R. Najafi et al. / International Journal of Food Microbiology 189 (2014) 89–97

95

(a) 1370 cm - 1

1620 cm - 1

Intensity (a.u.)

1180 cm - 1

h

a

Raman Shift (cm - 1 )

(b)

1370 cm - 1

1620 cm - 1

Intensity (a.u.)

1180 cm - 1

h g f e d c b a

Raman Shift (cm - 1 )

(c)

Fig. 5. SERS spectra for E. coli O157:H7 capture in apple juice using MNP. (a) Stacked spectra of different concentrations of E. coli O157:H7-nanoparticle aggregates containing Raman reporter: No E. coli O157:H7 (spectrum a); 101 CFU/mL (spectrum b); 102 CFU/mL (spectrum c); 103 CFU/mL (spectrum d); 104 CFU/mL (spectrum e); 105 CFU/mL (spectrum f); 106 CFU/mL (spectrum g); and 107 CFU/mL (spectrum h). (b) Stacked spectra showing E. coli O157:H7 concentration dependent Raman shift signature of the MGITC reporter at vibrational frequency zones of 1180, 1370, and 1620 cm−1. (c) Plot representing SERS intensity at 1620 cm−1 vs. Log CFU/mL of target E. coli O157:H7. The spectra represent the mean two independent experiments done in duplicate.

rapid (less than 10 min). NIMS takes the advantages of the traditional IMS as the Fe3O4/Au nanoclusters can easily be captured by an external magnet due to their unique magnetic properties (Zhou et al., 2010). Moreover, the use of a Raman tag leads to a better signal to noise ratio for direct detection in food matrices (Wang et al., 2010; Wang Y. et al., 2011; Zhou et al., 2010) without the need of lengthy incubation. Several E. coli O157:H7 outbreaks occurred in the last two decades involving consumption of unpasteurized apple juice or ciders in North America (Besser et al., 1993; CDC, 1997; Cody et al., 1999; Hilborn et al., 2000; Tamblyn et al., 1999). In the United States, FDA regulation requires a “5-Log pathogen reduction performance standard” for apple juice and

apple cider or if sold raw (un-pasteurized), must carry a warning label showing that the product may have potential harmful bacteria (FDA, 2001; Mak et al., 2001). With all these actions, there has still been some E. coli O157:H7 outbreak related to apple juice. So, this is one of the reasons for choosing apple juice as a model beverage in this study. In addition, as a food matrix, apple juice is relatively low in suspended matters with minimal insoluble particulates, so it could be easily used in the developed method for the first trial. The total analysis time of the developed immunoassay was less than 60 min, including 25 min for bacteria capture, 25 min for interaction between the captured bacteria and the SERS labeled antibody and less

96

R. Najafi et al. / International Journal of Food Microbiology 189 (2014) 89–97

than total 10 min for washing steps and SERS measurement. The analysis time of this method was comparable to that of the SERS based sandwich immunoassay developed previously (Guven et al., 2011) and the SERS based detection platform as reported by Temur et al. (2010). Our results indicate that multifunctional Fe3O4/Au nanocluster can be successfully applied to detect E. coli O157:H7 by obtaining the SERS signal in apple juice containing the target pathogen. The Fe3O4 core inside this nanocomposite can be conveniently used due to its strong magnetic properties, which, in turn, can lead to the rapid separation and concentration of the target analyte. The rough gold shell on the outer part of the nanocomposite results in stability and bacterial detection via a SERS method (Zhou et al., 2010). 4. Conclusions E. coli O157:H7 is a major foodborne pathogen that has been a serious concern in the food industry. Thus, the development of a rapid, sensitive and selective detection method for this pathogen is needed. A new immunoassay has been developed, which is a combination of NIMS and SERS. Based on the results of food sample analysis, the developed immunoassay can be used for the rapid detection of E. coli O157: H7 in apple juice. The Fe3O4/Au nanocluster is effective in capturing, concentrating, and detecting E. coli O157:H7 via surface enhanced Raman spectroscopy (SERS). The present method is simple and does not require time-consuming sample preparation and enrichment. Another advantage of this newly developed method is that it holds the potential of multiplexing. Just by incorporating multiple Raman probes on the same assay platform, a multiplexed immunoassay can be developed for simultaneous detection of multiple pathogens in food matrices. The developed immunoassay may also be applicable in particulate foods, such as ground beef, which will be an extension of the present study, and will be conducted in the future. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijfoodmicro.2014.07.036. Acknowledgments This research was partly supported by grants from the USDA–NIFA (ALAX-012-0210 and 2010-38821-21448) and by start-up funds from The University of Memphis. The authors thank Dr. Josh Herring for SEM, and Dr. Omar Skalli and Lou G Boykins for TEM analyses. References Banerjee, P.,Bhunia, A.K., 2010. Cell-based biosensor for rapid screening of pathogens and toxins. Biosens. Bioelectron. 26, 99–106. Bantz, K.C.,Meyer, A.F.,Wittenberg, N.J.,Im, H.,Kurtulus, O.,Lee, S.H.,Lindquist, N.C.,Oh, S.H., Haynes, C.L., 2011. Recent progress in SERS biosensing. Phys. Chem. Chem. Phys. 13, 11551–11567. Besser, R.E.,Lett, S.M.,Weber, J.T., Doyle, M.P.,Barrett, T.J.,Wells, J.G.,Griffin, P.M., 1993. An outbreak of diarrhea and hemolytic uremic syndrome from Escherichia coli O157:H7 in fresh-pressed apple cider. J. Am. Med. Assoc. 269, 2217–2220. Bhunia, A.K., 2008. Biosensors and bio-based methods for the separation and detection of foodborne pathogens. Adv. Food Nutr. Res. 54, 1–44. Campion, A.,Kambhampati, P., 1998. Surface enhanced Raman scattering. Chem. Soc. Rev. 27, 241–250. CDC, 1997. Outbreaks of Escherichia coli O157:H7 infection and cryptosporidiosis associated with drinking unpasteurized apple cider—Connecticut and New York, October 1996. Morb. Mortal. Wkly Rep. 46. Cheng, Y., Liu, Y., Huang, J., Li, K., Zhang, W., Xian, Y., Jin, L., 2009. Combining biofunctional magnetic nanoparticles and ATP bioluminescence for rapid detection of Escherichia coli. Talanta 77, 1332–1336. Chu, H., Huang, Y., Zhao, Y., 2008. Silver nanorod arrays as a surface-enhanced Raman scattering substrate for foodborne pathogenic bacteria detection. Appl. Spectrosc. 62, 922–931. Cody, S.H.,Glynn, M.K.,Farrar, J.A.,Cairns, K.L.,Griffin, P.M.,Kobayashi, J.,Fyfe, M., Hoffman, R.,King, A.S.,Lewis, J.H.,Swaminathan, B.,Bryant, R.G.,Vugia, D.J., 1999. An outbreak of Escherichia coli O157:H7 infection from unpasteurized commercial apple juice. Ann. Intern. Med. 130, 202–209. Dutta, R.K.,Sharma, P.K.,Pandey, A.C., 2009. Surface enhanced Raman spectra of E. coli cells using ZnO nanoparticles. Dig. J. Nanomater. Biostruct. 4, 83–87.

Dwivedi, H.P., Jaykus, L.-A., 2011. Detection of pathogens in foods: the current state-ofthe-art and future directions. Crit. Rev. Microbiol. 37, 40–63. FDA, 2001. Hazard Analysis and Critical Control Point (HAACP); Procedures for the Safe and Sanitary Processing and Importing of Juice; Final Rule. Federal Register, pp. 6137–6202. Gilmartin, N.,O'Kennedy, R., 2012. Nanobiotechnologies for the detection and reduction of pathogens. Enzym. Microb. Technol. 50, 87–95. Goon, I.Y., Lai, L.M.H.,Lim, M., Munroe, P., Gooding, J.J., Amal, R., 2009. Fabrication and dispersion of gold-shell-protected magnetite nanoparticles: systemic control using polyethyleneimine. Chem. Mater. 21, 673–681. Gou, L., Murphy, C.J., 2005. Fine-tuning the shape of gold nanorods. Chem. Mater. 17, 3668–3672. Guven, B., Basaran-Akgul, N.,Temur, E.,Tamer, U., Boyaci, I.H., 2011. SERS-based sandwich immunoassay using antibody coated magnetic nanoparticles for Escherichia coli enumeration. Analyst 136, 740–748. Haynes, C.L.,McFarland, A.D.,Van Duyne, R.P., 2005. Surface-enhanced Raman spectroscopy. Anal. Chem. 77, 338–346. Hilborn, E.D.,Mshar, P.A.,Fiorentino, T.R.,Dembek, Z.F.,Barrett, T.J.,Howard, R.T.,Cartter, M.L., 2000. An outbreak of Escherichia coli O157:H7 infections and haemolytic uraemic syndrome associated with consumption of unpasteurized apple cider. Epidemiol. Infect. 124, 31–36. Huang, W.E., Li, M., Jarvis, R.M., Goodacre, R., Banwart, S.A., 2010. Shining light on the microbial world the application of Raman microspectroscopy. Adv. Appl. Microbiol. 70, 153–186. Jana, N.R., Gearheart, L., Murphy, C., 2001. Evidence for seed-mediated nucleation in the formation of gold nanoparticles from gold salts. Chem. Mater. 13, 2313–2322. Krug, J.T., Wang, G.D., Emory, S.R., Nie, S.M., 1999. Efficient Raman enhancement and intermittent light emission observed in single gold nanocrystals. J. Am. Chem. Soc. 121, 9208–9214. Liberti, P.A., Chiarappa, J.N., Hovsepian, A.C., Rao, C.G., 1996. Bioreceptor ferrofluids: novel characteristics and their utility in medical applications. In: Pelizzetti, E. (Ed.), Fine Particles Science and Technology. Springer, Netherlands, pp. 777–790. Lu, X., Al-Qadiri, H.M., Lin, M., Rasco, B.A., 2011. Application of mid-infrared and Raman spectroscopy to the study of bacteria. Food Bioprocess Technol. 4, 919–935. Mak, P.P., Ingham, B.H., Ingham, S.C., 2001. Validation of apple cider pasteurization treatments against Escherichia coli O157:H7, Salmonella, and Listeria monocytogenes. J. Food Prot. 64, 1679–1689. Mandal, P.K., Biswas, A.K., Choi, K., Pal, U.K., 2011. Methods for rapid detection of foodborne pathogen: an overview. Am. J. Food Technol. 6, 87–102. Maquelin, K., Kirschner, C., Choo-Smith, L.P., van den Braak, N., Endtz, H.P., Naumann, D., Puppels, G.J., 2002. Identification of medically relevant microorganisms by vibrational spectroscopy. J. Microbiol. Methods 51, 255–271. Naja, G., Bouvrette, P., Hrapovic, S., Luong, J.H.T., 2007. Raman-based detection of bacteria using silver nanoparticles conjugated with antibodies. Analyst 132, 679–686. Narayanan, R.,Lipert, R.J.,Porter, M.D., 2008. Cetyltrimethylammonium bromide-modified spherical and cube-like gold nanoparticles as extrinsic Raman labels in surfaceenhanced Raman spectroscopy based heterogeneous immunoassays. Anal. Chem. 80, 2265–2271. Neng, J.,Harpster, M.H.,Zhang, H.,Mecham, J.O.,Wilson, W.C.,Johnson, P.A., 2010. A versatile SERS-based immunoassay for immunoglobulin detection using antigen-coated gold nanoparticles and malachite green-conjugated protein A/G. Biosens. Bioelectron. 26, 1009–1015. Ozalp, V.C., Bayramoglu, G., Arica, M.Y., Oktem, H.A., 2013. Design of a core–shell type immuno-magnetic separation system and multiplex PCR for rapid detection of pathogens from food samples. Appl. Microbiol. Biotechnol. 97, 9541–9551. Penn, S.G.,He, L.,Natan, M.J., 2003. Nanoparticles for bioanalysis. Curr. Opin. Chem. Biol. 7, 609–615. Perez, F.G., Mascini, M., 1998. Immunomagnetic separation with mediated flow injection analysis amperometric detection of viable Escherichia coli O157. Anal. Chem. 70, 2380–2386. Perez-Lopez, B.,Merkoci, A., 2011. Nanomaterials based biosensors for food analysis applications. Trends Food Sci. Technol. 22 (11), 625–639. Pettinger, B., Ren, B., Picardi, G., Schuster, R., Ertl, G., 2004. Nanoscale probing of adsorbed species by tip-enhanced Raman spectroscopy. Phys. Rev. Lett. 92, 096101. Pettinger, B., Ren, B., Picardi, G., Schuster, R., Ertl, G., 2005. Tip‐enhanced Raman spectroscopy (TERS) of malachite green isothiocyanate at Au (III): bleaching behavior under the influence of high electromagnetic fields. J. Raman Spectrosc. 36, 541–550. Premasiri, W.R.,Moir, D.T.,Klempner, M.S.,Krieger, N.,Jones, G.,Ziegler, L.D., 2005. Characterization of the surface enhanced raman scattering (SERS) of bacteria. J. Phys. Chem. B 109, 312–320. Qian, X., Peng, X.H., Ansari, D.O., Yin-Goen, Q., Chen, G.Z., Shin, D.M., Yang, L., Young, A.N., Wang, M.D., Nie, S., 2008. In vivo tumor targeting and spectroscopic detection with surface-enhanced Raman nanoparticle tags. Nat. Biotechnol. 26, 83–90. Scallan, E.,Hoekstra, R.M.,Widdowson, M.A.,Hall, A.J.,Griffin, P.M., 2011. Foodborne illness acquired in the United States. Emerg. Infect. Dis. 17, 1339–1340. Sugimoto, T., Matijevic, E., 1980. Formation of uniform spherical magnetite particles by crystallization from ferrous hydroxide gels. J. Colloid Interface Sci. 74, 227–243. Tamblyn, S., deGrosbois, J., Taylor, D., Stratton, J., 1999. An outbreak of Escherichia coli O157:H7 infection associated with unpasteurized non-commercial, custom-pressed apple cider—Ontario, 1998. Can. Commun. Dis. Rep. 25, 117–120. Tamer, U., Boyacı, İ.H., Temur, E., Zengin, A., Dincer, İ., Elerman, Y., 2011. Fabrication of magnetic gold nanorod particles for immunomagnetic separation and SERS application. J. Nanoparticle Res. 13, 3167–3176. Temur, E., Boyacı, I.H., Tamer, U., Unsal, H., Aydogan, N., 2010. A highly sensitive detection platform based on surface-enhanced Raman scattering for Escherichia coli enumeration. Anal. Bioanal. Chem. 397, 1595–1604.

R. Najafi et al. / International Journal of Food Microbiology 189 (2014) 89–97 Tomoyasu, T., 1998. Improvement of the immunomagnetic separation method selective for Escherichia coli O157 strains. Appl. Environ. Microbiol. 64, 376–382. Tu, S.I.,Reed, S.,Gehring, A.,He, Y.,Paoli, G., 2009. Capture of Escherichia coli O157:H7 using immunomagnetic beads of different size and antibody conjugating chemistry. Sensors 9, 717–730. Varshney, M., Yang, L., Su, X.L., 2005. Magnetic nanoparticle–antibody conjugates for the separation of Escherichia coli O157:H7 in ground beef. J. Food Prot. 68, 1804–1811. Velusamy, V., Arshak, K., Korostynska, O., Oliwa, K., Adley, C., 2010. An overview of foodborne pathogen detection: in the perspective of biosensors. Biotechnol. Adv. 28, 232–254. Wang, Z., Miu, T., Xu, H., Duan, N., Ding, X., Li, S., 2010. Sensitive immunoassay of Listeria monocytogenes with highly fluorescent bioconjugated silica nanoparticles probe. J. Microbiol. Methods 83, 179–184. Wang, H., Li, Y.,Wang, A., Slavik, M., 2011. Rapid, sensitive, and simultaneous detection of three foodborne pathogens using magnetic nanobead-based immunoseparation and quantum dot-based multiplex immunoassay. J. Food Prot. 74, 2039–2047. Wang, Y., Ravindranath, S., Irudayaraj, J., 2011. Separation and detection of multiple pathogens in a food matrix by magnetic SERS nanoprobes. Anal. Bioanal. Chem. 399, 1271–1278.

97

Wang, Z., Wang, J., Yue, T., Yuan, Y., Cai, R., Niu, C., 2013a. Immunomagnetic separation combined with polymerase chain reaction for the detection of Alicyclobacillus acidoterrestris in apple juice. PLoS One 8, e82376. Wang, Z.,Yue, T.,Yuan, Y.,Cai, R.,Niu, C.,Guo, C., 2013b. Development and evaluation of an immunomagnetic separation-ELISA for the detection of Alicyclobacillus spp. in apple juice. Int. J. Food Microbiol. 166, 28–33. Xiong, Q., Cui, X., Saini, J.K., Liu, D., Shan, S., Jin, Y., Lai, W., 2014. Development of an immunomagnetic separation method for efficient enrichment of Escherichia coli O157:H7. Food Control 37, 41–45. Xu, S., Ji, X., Xu, W., Zhao, B., Dou, X., Bai, Y., Ozaki, Y., 2005. Surface-enhanced Raman scattering studies on immunoassay. J. Biomed. Opt. 10, 031112. Zhao, X., Hilliard, L.R., Mechery, S.J., Wang, Y., Bagwe, R.P., Jin, S., Tan, W., 2004. A rapid bioassay for single bacterial cell quantitation using bioconjugated nanoparticles. Proc. Natl. Acad. Sci. U. S. A. 101, 15027–15032. Zhou, X., Xu, W., Wang, Y., Kuang, Q., Shi, Y., Zhong, L., Zhang, Q., 2010. Fabrication of cluster/shell Fe3O4/Au nanoparticles and application in protein detection via a SERS method. J. Phys. Chem. C 114, 19607–19613.