Biosensors and Bioelectronics 57 (2014) 143–148
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
In-situ fluorescent immunomagnetic multiplex detection of foodborne pathogens in very low numbers Il-Hoon Cho a, Lisa Mauer b, Joseph Irudayaraj a,b,n a Bindley Bioscience and Birck Nanotechnology Center, Department of Agricultural & Biological Engineering, Purdue University, West Lafayette, IN 47907, United States b Department of Food Science, Purdue University, West Lafayette, IN 47907, United States
art ic l e i nf o
a b s t r a c t
Article history: Received 24 December 2013 Received in revised form 1 February 2014 Accepted 5 February 2014 Available online 19 February 2014
Consumption of foods contaminated with pathogenic bacteria is a major public health concern. Foods contain microorganisms, the overwhelming majority of which are nonpathogenic, some are responsible for food spoilage, and some cause serious illness leading to death or a variety of diseases in humans. The key challenge in food safety is to rapidly screen foods to determine the presence of pathogens so that appropriate intervention protocols can be pursued. A simple fluorometric immunological method in combination with a magnetic concentration step was developed for rapid detection of target bacteria with high sensitivity and specificity in less than 2 h without enumeration. The method constitutes performing an in-situ immunoassay on a magnetic bead through the formation of a sandwich complex of the target bacteria and the probe (detection antibody—denatured BSA labelled with fluorophores) followed by the release of fluorophores by means of enzymatic digestion with proteinase K. The limit of detection (LOD) was o 5 CFU/mL of the tested pathogens (Escherichia coli O157:H7, Salmonella typhimurium, and Listeria monocytogenes) in buffer. When the pathogens were inoculated in foods (spinach, chicken, and milk), the LOD was under 5 CFU/mL for E. coli O157:H7, S. typhimurium and L. monocytogenes. Furthermore, the method was highly specific in detecting the target pathogens in a multiplex format. The developed in-situ fluorescent immunomagnetic sensor approach offers distinct advantages because it is rapid, highly sensitive, and easy to use and could therefore be potentially used as a pathogen screening tool. & 2014 Elsevier B.V. All rights reserved.
Keywords: Foodborne pathogens Rapid detection Signal enhancement Magnetic separation Fluorescent detection Highly sensitive
1. Introduction The annual burden of foodborne illness, particularly United States, as estimated by the Centers for Disease Control and Prevention, is 48 million cases of foodborne illness, 128,000 hospitalizations, and 3000 deaths along with a cost of $152 billion in medical expenses, lost productivity and business, lawsuits, and compromised branding (Scallan et al., 2011; Zhang et al., 2009). Efforts to improve the safety of the food supply are highlighted by the FDA Food Safety Modernization Act (Drew and Clydesdale, 2013). This regulation requires enhanced detection of pathogens in both imported and domestically produced foods coupled with effective intervention methods to achieve food safety. However, mandated testing has not yet been implemented, in part due to the lack of technology capable of meeting the requirements. A working premise is that rapid and n Corresponding author at: Purdue University, Department of Agricultural and Biological Engineering, 225 South University Street, West Lafayette, IN 47907, United States. Tel.: þ1 765 494 0388; fax: þ 1 765 496 1115. E-mail address: [email protected] (J. Irudayaraj).
http://dx.doi.org/10.1016/j.bios.2014.02.012 0956-5663 & 2014 Elsevier B.V. All rights reserved.
comprehensive pathogen detection methods will enable implementation of effective strategies to avoid, or when needed intervene in, the distribution of foods that have been contaminated with foodborne pathogens. The complexity of food matrices, the wide variety of microorganisms therein, and the varying growth and replication traits of target pathogenic microorganisms that, when present, are often in much smaller numbers than the nonpathogenic microorganisms, pose significant challenges for pathogen detection. Conventional methods used to identify pathogens in contaminated foodstuffs include: colony culture (Reissbrodt, 2004; Velusamy et al., 2010), polymerase chain reaction (PCR) (Malorny et al., 2008; Navas et al., 2006), and immunoassays (e.g., enzyme-linked immunosorbent assays (ELISA) and immunochromatographic assays) (Delehanty and Ligler, 2002; Seo et al., 2010), often in combination with selective enrichment. However, these conventional methods are time-consuming, laborious, and/or lack in sensitivity in part due to the requirement of cultivation steps. Thus, alternate approaches are needed for rapid detection. Biosensors have been developed and documented to improve the limit of detection (LOD) and/or
144
I.-H. Cho et al. / Biosensors and Bioelectronics 57 (2014) 143–148
time to result, based on electrochemical (Patel et al., 2011; Setterington and Alocilja, 2012; Smietana et al., 2011) and optical (Li et al., 2011; Smietana et al., 2011; Subramanian et al., 2006a, 2006b) techniques, including surface enhanced Raman spectroscopy (SERS) (Craig et al., 2013; Wang et al., 2010, 2011b) methods. However, a need still exists to develop highly sensitive easily implementable biosensors for onsite detection of pathogens in complex food matrices. To address this need, an in-situ immunomagnetic bead with enhanced fluorescence approach was developed and applied for the detection of three prevalent foodborne pathogens, Escherichia coli O157:H7, Salmonella typhimurium, and Listeria monocytogenes. Factors critical to the suitability of detection methods and sensors involved in food analysis were investigated, including sensitivity, limit of detection (LOD), detection time, and ease of use.
2. Materials and methods 2.1. Materials Monoclonal antibodies specific to E. coli O157:H7, S. typhimurium, and L. monocytogenes and a polyclonal antibody raised from rabbit specific to L. monocytogenes were purchased from Abcam (Cambridge, MA). Polyclonal antibodies raised from goat which are reactive to E. coli O157:H7 and S. typhimurium were purchased from KPL (Gaithersburg, MD). Dynabeads (M-270, NH2-functionalized), protein ladders (pre-stained) for SDS-PAGE, and two maleimidefunctionalized Alexa Fluor™ dyes (532 and 647) were obtained from Life Technologies (Carlsbad, CA). E. coli O157:H7, S. typhimurium, and L. monocytogenes were obtained from the culture collection at the Center for Food Safety Engineering consortium at Purdue University. Casein (sodium salt), Tween 20, tetramethyl benzidine (TMB), ethanolamine, urea, bovine serum albumin (BSA), Sephadex G-15, dialysis tubing, and proteinase K were obtained from Sigma (St. Louis, MO). 2-(N-morpholino) ethanesulfonic acid (MES), 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), succinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxy-(6-amidocaproate) (LC-SMCC), 2-iminothiolane, dithiothreitol (DTT), fluorescein5-maleimide (maleimide-FITC), and mouse anti-goat IgG labeled with horseradish peroxidase (HRP) were purchased from Pierce (Rockford, IL). Source of the antibodies and bacteria used in this study are provided in Supporting information (S-5). 2.2. Preparation of bacteria E. coli O157:H7, S. typhimurium, and L. monocytogenes were respectively cultured in 500 mL of Luria-Bertani (LB) media at 37 1C with shaking, for 18 h and then harvested by centrifugation (5000 rpm, 20 min) and stored in 50 mL of sterilized PBS (phosphate buffered saline, 10 mM phosphate and 150 mM NaCl). The concentration of bacteria in these stock solutions was determined in triplicate by enumeration on plate count agar following incubation at 37 1C for 24 h. 2.3. Preparation of immuno-magnetic beads For coupling antibodies to the beads, monoclonal antibodies specific to E. coli O157:H7, S. typhimurium, and L. monocytogenes were respectively immobilized onto amine-functionalized beads as described elsewhere (Cho and Irudayaraj, 2013). Briefly, the beads (1 mg) were washed with 0.1 M MES buffer, pH 4.5, and then reacted with the corresponding antibody (10 mg) by adding 50 mM EDC at room temperature for 2 h. For blocking the EDC-activated sites of antibody and the bead surfaces, 1 M ethanolamine (12.2 mL), pH 8.5 and 0.5% (w/v) casein dissolved in PBS buffer was
sequentially reacted for 15 min and 1 h, respectively. After washing with 0.5% (w/v) casein-PBS three times, the conjugates for detection of the three pathogens were stored at 4 1C until use. 2.4. Conjugation of antibody with fluorescent dBSA BSA (0.25 mg) was treated with 1 mL of 8 M urea in de-ionized water (DIW) for 30 min followed by the addition of 20 mM of DTT in DIW for 30 min at 37 1C with gentle shaking to ensure complete denaturation. The urea and DTT were removed by gel filtration with a Sephadex G-15 column (volume 10 mL) and the protein fractions identified by Bradford assay were immediately mixed with the three maleimide-functionalized dyes (i.e., FITC, Alexa Fluor 532 and 647) respectively and maintained at room temperature for 1 h. The excess dyes were consecutively removed by dialysis and gel filtration on Sephadex G-15 chromatography. The dBSA bearing fluorescent dyes were coupled to 50-fold molar excess of LC-SMCC for 1 h followed by purified on Sephadex G-15 column. The fluorescent dBSA was coupled to antibody treated with 10-fold molar excess of iminothiolane (reagent used for thiolation of antibody) at room temperature for 2 h. 2.5. SDS-PAGE analysis Protein samples, i.e., native BSA, anti-L. monocytogenes antibody and the fluorescent dBSA conjugate, and protein marker were treated with sample buffer without mercaptoethanol (non-reducing condition) and loaded on an 8% polyacrylamide gel to verify the conjugation efficiency of the synthesized IFdBSA conjugate. For electrophoretic separation, two voltages (80 and 120 V) were sequentially applied for 20 and 60 min respectively using a MiniProtein Tetra Cell apparatus (Bio-Rad, Hercules, CA). The separation gel was sequentially immersed in staining solution including 0.1% Coomassie blue R-250 for 15 min and destaining solution for 1 h for developing the image. 2.6. In-situ fluorescent bead assays E. coli O157:H7, S. typhimurium, and L. monocytogenes samples serially diluted in PBS (0, 1 101, 1 102, 1 103, 1 104, and 1 105 cells/mL) from stock solution (3 1010, 1 1011, and 5 109 cells/mL for E. coli O157:H7, S. typhimurium, and L. monocytogenes respectively) were first reacted with the corresponding beads (50 μL) with gentle mixing at room temperature for 20 min followed by separation of the supernatant using a magnet (AM10026, Life Technologies). The supernatant was discarded by pipetting and the bead was carefully washed with sterilized PBS. For probe attachment, 0.01 μg/mL of the fluorescent dBSA conjugate (200 μL) for each bacteria in 0.5% casein-PBS containing 0.1% (v/v) Tween 20 (casein-PBS-Tw) was mixed with the bead and incubated at room temperature for 20 min. The beads were washed with casein-PBS-Tw three times and then proteinase K solution (200 μL) diluted 100 times in reaction buffer (20 mM Tris–HCl, pH 8.0, 20 mM NaCl and 10 mM CaCl2) was mixed with the beads. This enzymatic digestion was maintained at room temperature for 30 min. The fluorescent signals obtained from the supernatant after the digestion step were measured by a fluorescence spectrophotometer (Cary Eclipse, Agilent Technologies) at different excitation and emission wavelengths (i.e., 495/519 nm, 531/554 nm, and 650/668 nm for FITC, Alexa532, and Alexa647, respectively). 2.7. Multiplex assays Bacteria were prepared in two concentrations (0 and 100 cells/mL) by using the blank, uninoculated samples, for the zero dose and serial
I.-H. Cho et al. / Biosensors and Bioelectronics 57 (2014) 143–148
dilutions of the stock solutions for the 100 cells/mL samples. Counts were verified by enumeration on plate count agar as described above. For multiplex assays, a cocktail consisting of the three pathogens was mixed and reacted with the three different immunobeads at the same time. The respective fluorescence signal was obtained at the corresponding wavelengths: the fluorophores used were, FITC for E. coli O157:H7, Alexa 532 for S. typhimurium and Alexa 647 for L. monocytogenes). The final volume was kept the same as in the single pathogen assay (200 μL), and the analytical procedure was the same as described above. 2.8. Food sample tests To test the feasibility of the methods for pathogen detection in whole foods, the bead assay was applied to the three different food matrices: spinach wash for E. coli O157:H7, boneless chicken for S. typhimurium, and 2% reduced fat milk for L. monocytogenes using uninoculated foods as controls and also artificially inoculating each bacteria into the corresponding sample medium. Bacteria concentrations used, as enumerated by colony counting were: 0, 3 and 22 CFU/ mL for E. coli O157:H7, 0, 5 and 39 CFU/mL for S. typhimurium and 0, 4 and 30 CFU/mL for L. monocytogenes. Standard sampling procedures were used (Mohammed et al., 2008; Muldoon et al., 2012): a 10 g sample of chicken was mixed with 10 mL of sterilized PBS and inoculated with 1 mL of prepared S. typhimurium; 10 mL of spinach wash was inoculated with 1 mL of E. coli O157:H7; and 10 mL of milk was inoculated with 1 mL of L. monocytogenes. After inoculation and 2 h of incubation at room temperature, 1 mL of the bacterial solution was used and the in-situ bead assays were performed in triplicate, as described above. 2.9. Biosafety concerns All experiments were conducted in BSL-2 approved facilities with proper safety training in place and approvals.
3. Results and discussion 3.1. Analytical concept The proposed in-situ fluorescent bead strategy for the rapid detection of pathogenic bacteria consists of an immunomagnetic bead as a solid support for capturing bacteria and a probe antibody-dBSA conjugate labeled with fluorophores (IFdBSA), as depicted in Fig. 1.
145
The entire procedure of the immunoassays can be performed on a magnetic bead coated with capture antibody, which is specific to the target bacteria. Upon incubation and binding of the beads to the pathogens in the food sample, the bead complex is isolated by a simple magnetic separation step and the target pathogen detected by binding with the IFdBSA conjugate followed by enzymatic digestion for fluorometric signal measurement. This in-situ fluorescent immuno-bead assay could be performed in one tube which enables us to not only detect the target bacteria with simple methods but also in very low numbers due to fluorescence enhancement from the large number of fluorescent signal tracers labeled with dBSA used as a carrier of fluorophores. In addition, the large surface area of the bead facilitates a higher degree of capture efficiency of the target bacteria compared with common immunoassays for pathogen detection (see Fig. S1 in Supporting information), enabling a rapid and sensitive method for the detection of bacteria (Setterington et al., 2011). To maximize the fluorescent signal from the assay, it is crucial for the carrier protein (i.e., dBSA) to bear high numbers of fluorophores. BSA is known as a plasma carrier and transport protein in blood and is a cysteine-rich protein with 35 sulfhydryls (17 intramolecular cysteine disulfides) out of 583 total amino acids of the protein (Liu et al., 2012; Meng et al., 2011), has higher portion of the reactive groups compared to other proteins, and was utilized as a fluorophore carrier in this study. The binding sites of BSA could be exposed by mild treatment using chaotropic agents such as urea or guanidine hydrochloride in aqueous environment due to its destabilizing hydrophobic interaction inside protein (Lim et al., 2009) followed by the cleavage of covalent disulfide bonds with the reduction by DTT. The reduced free sulfhydryl groups allow fluorescent dyes with thiol-reactive groups (e.g., maleimide and iodoacetyl derivatives) to bind so that numerous fluorophores could be functionalized to the interior of the dBSA. The degree of fluorescence on BSA were estimated by considering several factors (see S-2 in Supporting information). Even though antibodies that are directly labeled with aminereactive fluorophores can be used for the detection of bacteria as a probe, the number of fluorophores per antibody molecule is nearly in the range between 3 and 9 (Ogawa et al., 2009; Vira et al., 2010), which limits the level of detection. It should be noted that the fluorescent signals can be quenched among other signals when fluorophores are positioned within the energy transfer distance (Forster Resonance Energy Transfer distance is approximately within 10 nm), to result in a decrease in fluorescent signals. Since the dimension of BSA is approximately 140 40 40 Å3, even though it is totally denatured by urea and
Fig. 1. (a) Scheme of in-situ fluorescent immunoassay on magnetic beads for the detection of pathogenic bacteria. Target bacterium captured by antibody on magnetic bead can be detected by antibody-dBSA conjugate labeled with fluorophores (IFdBSA). (b) Fluorescence images of bacteria used as targets and stained by the corresponding IFdBSA probe. (c) Fluorescence image of bacterium (Escherichia coli O157:H7) captured by the immunomagnetic bead. Red color of bacteria indicates lifetime of FITC. The structural image of BSA is borrowed from protein data bank. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
146
I.-H. Cho et al. / Biosensors and Bioelectronics 57 (2014) 143–148
DTT, the possibility of self-quenching could not be ruled out given the structural configuration. To circumvent this drawback each fluorophore labeled at the intra-molecular site has to be liberated from the fluorescent dBSA conjugate by means of proteasemediated enzymatic digestion. Once amino acid fragments containing the labeled fluorophore are released into the solution, the fluorescent signals can be noted with clarity because of minimal self-quenching since the fluorophores are no longer in close proximity with each other (Jones et al., 1997). Proteinase K known as a strong endopeptidase in a broad range (Betzel et al., 1993) was employed for the dissociation of fluorophores from the IFdBSA conjugate and its enzymatic efficiency on the reduction of selfquench were tested (see Fig. S2 in Supporting information). 3.2. Characterization of the conjugate To verify the biophysical state of the synthesized IFdBSA conjugate, SDS-PAGE analysis was performed on a 8% polyacrylamide
gel under non-reducing condition (i.e., without breaking other disulfide bonds by extra reduction). From the results shown in Fig. 2., it is obvious that the two types of IFdBSA conjugates exist and are positioned at over 200 kDa (lane IV), the one with a 1:1 ratio (i.e., one fluorescent dBSA per one antibody) that presents over half a portion of the total number of conjugates and the other is 1:2 ratio (two fluorescent dBSA per one antibody). Even though a large portion of the free antibody (anti-L. monocytogenes antibody was used in this analysis) and small fluorescent dBSA still exists, approximate efficiency of this conjugate is around 40%. It can also be noted that most of the native BSA in lane II are labeled with a large number of sulfhydryl reactive fluorophores after denaturation, with increasing molecular weight as shown in lane IV. The IFdBSA conjugate characterized by its molecular state by electrophoresis was used as a probe for fluorometric detection of bacteria in combination with the immuno-magnetic bead. 3.3. Performance of the in-situ fluorescent bead assays
Fig. 2. SDS-PAGE analysis of the synthesized IFdBSA conjugate [Lane I (protein ladder), Lane II (native BSA), Lane III (anti-Listeria monocytogenes antibody) and Lane IV (IFdBSA conjugate solution)].
To verify its usability, the IFdBSA conjugate was applied for in-situ fluorescent immunoassays as a probe for the detection of target bacteria in combination with magnetic beads which are used as a solid support. To maximize the analytical performance, some of the key factors (e.g., amount of capture antibody and IFdBSA conjugate and enzymatic digestion time, etc.) were empirically optimized prior to performance the analyses (i.e., 0.01 μg/mL of IFdBSA was used, see Fig. S3 in Supporting information). Bacteria to be tested in solution were first captured and concentrated by the 2.8 μm sized immunomagnetic bead by using a magnet and reacted with IFdBSA in sequence. For fluorometric detection of target bacteria different fluorophores were employed, e.g., FITC (Ex/Em: 495/519 nm), Alexa532 (Ex/Em: 531/554 nm) and Alexa647 (Ex/Em: 650/668 nm) for E. coli O157:H7, S. typhimurium and L. monocytogenes, respectively. As shown in Fig. 1, the emission fluorescent colors from bacteria to be tested are
Fig. 3. Performance tests for the in-situ fluorescent immunobead assays against E. coli O157:H7 (a), Salmonella typhimurium (b) and L. monocytogenes (c), respectively. The fluorescent signals in the Y-axis obtained from the assay were plotted against the bacteria concentration (0, 101, 102, 103, and 104 cells/mL) diluted from stock solution. The assays were performed in triplicate and the limit of detection (LOD) for each assay was calculated by adding three times the standard deviation to the mean of the zero dose of bacteria.
I.-H. Cho et al. / Biosensors and Bioelectronics 57 (2014) 143–148
green (FITC), yellow (Alexa532) and red (Alexa647), which can be released into solution after enzymatic digestion of the conjugates by using the proteinase K. Signal intensities were then obtained from the liberated fluorophores by fluorescence spectroscopy. The in-situ bead assays proposed are able to detect extremely low concentrations of bacteria, regardless of the bacteria species (Fig. 3). The fluorescent intensity from samples containing 10 CFU/mL was significantly higher than the uncontaminated controls. To compare the detection sensitivity of the method, the limit of detection (LOD) was determined by multiplying the standard deviation at zero dose (0 cells/mL) by three and adding that value to the mean at zero dose (Binder and Archimbaud, 2000; Cho et al., 2006) to result in an estimated LOD of o5 cells/mL for E. coli O157:H7, S. typhimurium, and L. monocytogenes. It is evident (Fig. 3) that the analytical sensitivity of this fluorescent immuno-bead assay is much higher, approximately over 104-fold increase in sensitivity, than conventional colorimetric immunoassays such as ELISA (105–106 cells/mL LOD is common) (Galikowska et al., 2011; Kim et al., 1999). Such high level of sensitivity was possible because of the high fluorescent signals obtained from the IFdBSA conjugate, in other words, bearing several
Fig. 4. Multiplex analyses of the in-situ fluorescent immunobead assay against three bacteria (i.e., E. coli O157:H7, S. typhimurium and L. monocytogenes) in singular and mixed types of bacteria samples. These tests were performed at a low concentration ( 100 cells/mL) of the pathogens.
147
tracers inside the dBSA which can be liberated by proteinase Kmediated digestion, resulting in an extremely high fluorescent signal to detect low number of cells.
3.4. Multiplex detection To verify the durability of the bead assays in a food matrix, high specificity toward target bacteria is required because the structure of the bacteria gives rise to analogous antigenic sites (e.g., K-, O-, and H-antigens) which exist at bacteria surfaces. By using a monoclonal antibody as a capture agent, the unique antigenic site of the target bacteria could be recognized so that the cross-reactivity of the assay is minimized as demonstrated by multiplex analysis (Fig. 4). Multiplex testing was performed at a low concentration of the bacteria (i.e., 100 cells/mL) diluted from each stock solution in order to assess the robustness of the assay. It is obvious that multiplex assays of in-situ bead experiments have high specificity to each target bacteria with regard to both singular and multiplex analyses without cross-reactivity (Fig. 4). For the detection of E. coli O157:H7, the assay did not show crossreactivity with other bacteria, i.e., S. typhimurium and L. monocytogenes. For S. typhimurium and L. monocytogenes analyses, the assay was only specific to the corresponding target agent without crossreactivity. Furthermore, the signal intensities obtained from each target bacteria in singular assays are nearly identical to those from the mixed bacteria experiment, implying that this assay is specific and the interference of other pathogens may not hinder the LOD. Although several multiplex immunomethods based on magnetic beads have been reported for multiplex pathogen detection (Charlermroj et al., 2013; Wang et al., 2011a), these methodologies employ other fluorescent tracers such as quantum dots for antibody labeling and lack the sensitivity to detect extremely low concentrations (e.g., under 10 cells/mL) of bacteria. However, the novel fluorometric assay proposed in this study utilizes dBSA as a fluorophore carrier where many signal tracers can be functionalized, enabling a high degree of sensitivity because of the enhancement without the need for further signal amplification steps.
Fig. 5. Performance metrics of the proposed assay in food matrices. (a) Spinach wash for E. coli O157:H7, (b) chicken meat for S. typhimurium, and (c) 2% reduced milk for L. monocytogenes. Error bars denoting the standard deviation are marked at the corresponding concentrations.
148
I.-H. Cho et al. / Biosensors and Bioelectronics 57 (2014) 143–148
3.5. Performance of the biosensor in food samples
References
An extremely low concentration of E. coli O157:H7 (o5 CFU/ mL) in spinach wash could be detected by the bead assay (Fig. 5) employing the same analytical procedure as described above. A very high LOD was also demonstrated when the assay was applied to both S. typhimurium (45 CFU/mL) and L. monocytogenes (o5 CFU/mL). Since a signal-to-noise ratio is an important in parameter in bacteria analysis, the fluorescence intensities were able to be represented as S/N ratio (see Fig. S4 in Supporting information). The sensitivity of the assay applied to inoculated food samples was similar to the LOD found in the buffer systems. Therefore, this assay is expected to be robust and has broad applicability. Furthermore in this assay a pre-enrichment step is likely not necessary due to its high sensitivity.
Betzel, C., Singh, T., Visanji, M., Peters, K., Fittkau, S., Saenger, W., Wilson, K., 1993. J. Biol. Chem. 268 (21), 15854–15858. Binder, R., Archimbaud, Y., 2000. Regul. Toxicol. Pharmacol. 31 (2), S23–S26. Charlermroj, R., Himananto, O., Seepiban, C., Kumpoosiri, M., Warin, N., Oplatowska, M., Gajanandana, O., Grant, I.R., Karoonuthaisiri, N., Elliott, C.T., 2013. PLoS One 8 (4). Cho, I.H., Irudayaraj, J., 2013. Int. J. Food Microbiol. 164 (1), 70–75. Cho, J.H., Han, S.M., Paek, E.H., Cho, I.H., Paek, S.H., 2006. Anal. Chem. 78 (3), 793–800. Craig, A.P., Franca, A.S., Irudayaraj, J., 2013. Annu. Rev. Food Sci. Technol. 4, 369–380. Delehanty, J.B., Ligler, F.S., 2002. Anal. Chem. 74 (21), 5681–5687. Drew, C.A., Clydesdale, F.M., 2013. Crit. Rev. Food Sci. Nutr. (just-accepted) Galikowska, E., Kunikowska, D., Tokarska-Pietrzak, E., Dziadziuszko, H., Łoś, J.M., Golec, P., Węgrzyn, G., Łoś, M., 2011. Eur. J. Clin. Microbiol. Infect. Dis. 30 (9), 1067–1073. Jones, L.J., Upson, R.H., Haugland, R.P., Panchuk-Voloshina, N., Zhou, M., Haugland, R.P., 1997. Anal. Biochem. 251 (2), 144–152. Kim, J.W., Jin, L.Z., Cho, S.H., Marquardt, R.R., Frohlich, A.A., Baidoo, S.K., 1999. J. Sci. Food Agric. 79 (11), 1513–1518. Li, S., Huang, J., Cai, L., 2011. Nanotechnology 22 (42), 425502. Lim, W.K., Rösgen, J., Englander, S.W., 2009. Proc. Nat. Acad. Sci. 106 (8), 2595–2600. Liu, J.Y., Liu, Y., Gao, M.X., Zhang, X.M., 2012. Proteomics 12 (14), 2258–2270. Malorny, B., Löfström, C., Wagner, M., Krämer, N., Hoorfar, J., 2008. Appl. Environ. Microbiol. 74 (5), 1299–1304. Meng, F., Yao, D., Shi, Y., Kabakoff, J., Wu, W., Reicher, J., Ma, Y., Moosmann, B., Masliah, E., Lipton, S.A., 2011. Mol. Neurodegener. 6, 34. Mohammed, Z., Souna, E., Anthony, T., 2008. Springer: New York, NY, USA. Muldoon, M.T., Allen, A.-C.O., Gonzales, V., Sutzko, M., Lindpaintner, K., 2012. J. AOAC Int. 95 (3), 850–859. Navas, J., Ortiz, S., Lopez, P., Jantzen, M.M., Lopez, V., Martinez-Suarez, J.V., 2006. Foodbourne Pathog. Dis. 3 (4), 347–354. Ogawa, M., Kosaka, N., Choyke, P.L., Kobayashi, H., 2009. Cancer Res. 69 (4), 1268–1272. Patel, M., Gonzalez, R., Halford, C., Lewinski, M.A., Landaw, E.M., Churchill, B.M., Haake, D.A., 2011. J. Clin. Microbiol. 49 (12), 4293–4296. Reissbrodt, R., 2004. Int. J. Food Microbiol. 95 (1), 1–9. Scallan, E., Hoekstra, R.M., Angulo, F.J., Tauxe, R.V., Widdowson, M.A., Roy, S.L., Jones, J.L., Griffin, P.M., 2011. Emerg. Infect. Dis. 17 (1), 7–15. Seo, S.-M., Cho, I.-H., Jeon, J.-W., Cho, H.-K., Oh, E.-G., Yu, H.-S., Shin, S.-B., Lee, H.-J., Paek, S.-H., 2010. J. Food Prot. 73 (8), 1466–1473. Setterington, E.B., Alocilja, E.C., 2012. Biosensors 2 (1), 15–31. Setterington, E.B., Cloutier, B.C., Ochoa, J.M., Cloutier, A.K., Patel, P.J., Alocilja, E.C., 2011. Int. J. Food Saf. Nutr. Public Health 4 (1), 83–100. Smietana, M., Bock, W.J., Mikulic, P., Ng, A., Chinnappan, R., Zourob, M., 2011. Opt. Express 19 (9), 7971–7978. Subramanian, A., Irudayaraj, J., Ryan, T., 2006a. Biosens. Bioelectron. 21 (7), 998–1006. Subramanian, A., Irudayaraj, J., Ryan, T., 2006b. Sens. Actuators, B 114 (1), 192–198. Velusamy, V., Arshak, K., Korostynska, O., Oliwa, K., Adley, C., 2010. Biotechnol. Adv. 28 (2), 232–254. Vira, S., Mekhedov, E., Humphrey, G., Blank, P.S., 2010. Anal. Biochem. 402 (2), 146–150. Wang, H., Li, Y., Wang, A., Slavik, M., 2011a. J. Food Prot. 74 (12), 2039–2047. Wang, Y., Lee, K., Irudayaraj, J., 2010. J. Phys. Chem. C 114 (39), 16122–16128. Wang, Y., Ravindranath, S., Irudayaraj, J., 2011b. Anal. Bioanal. Chem. 399 (3), 1271–1278. Zhang, D., Zhang, H., Yang, L., Guo, J., Li, X., Feng, Y., 2009. J. Food Saf. 29 (3), 348–363.
4. Conclusions The proposed in-situ fluorescent immunoassay strategy in combination with a magnetic concentration step was conceived and validated in this study. The method developed enabled the three pathogens, i.e., E. coli O157:H7, S. typhimurium and L. monocytogenes, to be detected with simple steps and could be completed within 2 h. In addition to the ease of implementation and rapid response, the bead assays also show high sensitivity and specificity to target bacteria, in addition to multiplexing. A LOD of o5 CFU/mL was established for the pathogens interrogated in buffer conditions. The LOD in food matrices was also comparable to that in buffer ( o5 CFU/mL) was demonstrated. We expect the simplicity, analytical sensitivity, and robustness of the biosensors will enable effective implementation in various sectors of the food chain as well as in critical check points to enhance food security.
Acknowledgements This effort was supported by funds from USDA-ARS project number 1935-42000-049-00D in conjunction with the Center for Food Safety Engineering at Purdue University. Dr. Andrew Ghering from USDA-ARS provided helpful insights on sampling, and Zhongwu Zhou and Yi Cui at Purdue University are acknowledged for their assistance with fluorescence imaging. 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.2014.02.012.
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
In-situ fluorescent immunomagnetic multiplex detection of foodborne pathogens in very low numbers Il-Hoon Cho a, Lisa Mauer b, Joseph Irudayaraj a,b,n a Bindley Bioscience and Birck Nanotechnology Center, Department of Agricultural & Biological Engineering, Purdue University, West Lafayette, IN 47907, United States b Department of Food Science, Purdue University, West Lafayette, IN 47907, United States
art ic l e i nf o
a b s t r a c t
Article history: Received 24 December 2013 Received in revised form 1 February 2014 Accepted 5 February 2014 Available online 19 February 2014
Consumption of foods contaminated with pathogenic bacteria is a major public health concern. Foods contain microorganisms, the overwhelming majority of which are nonpathogenic, some are responsible for food spoilage, and some cause serious illness leading to death or a variety of diseases in humans. The key challenge in food safety is to rapidly screen foods to determine the presence of pathogens so that appropriate intervention protocols can be pursued. A simple fluorometric immunological method in combination with a magnetic concentration step was developed for rapid detection of target bacteria with high sensitivity and specificity in less than 2 h without enumeration. The method constitutes performing an in-situ immunoassay on a magnetic bead through the formation of a sandwich complex of the target bacteria and the probe (detection antibody—denatured BSA labelled with fluorophores) followed by the release of fluorophores by means of enzymatic digestion with proteinase K. The limit of detection (LOD) was o 5 CFU/mL of the tested pathogens (Escherichia coli O157:H7, Salmonella typhimurium, and Listeria monocytogenes) in buffer. When the pathogens were inoculated in foods (spinach, chicken, and milk), the LOD was under 5 CFU/mL for E. coli O157:H7, S. typhimurium and L. monocytogenes. Furthermore, the method was highly specific in detecting the target pathogens in a multiplex format. The developed in-situ fluorescent immunomagnetic sensor approach offers distinct advantages because it is rapid, highly sensitive, and easy to use and could therefore be potentially used as a pathogen screening tool. & 2014 Elsevier B.V. All rights reserved.
Keywords: Foodborne pathogens Rapid detection Signal enhancement Magnetic separation Fluorescent detection Highly sensitive
1. Introduction The annual burden of foodborne illness, particularly United States, as estimated by the Centers for Disease Control and Prevention, is 48 million cases of foodborne illness, 128,000 hospitalizations, and 3000 deaths along with a cost of $152 billion in medical expenses, lost productivity and business, lawsuits, and compromised branding (Scallan et al., 2011; Zhang et al., 2009). Efforts to improve the safety of the food supply are highlighted by the FDA Food Safety Modernization Act (Drew and Clydesdale, 2013). This regulation requires enhanced detection of pathogens in both imported and domestically produced foods coupled with effective intervention methods to achieve food safety. However, mandated testing has not yet been implemented, in part due to the lack of technology capable of meeting the requirements. A working premise is that rapid and n Corresponding author at: Purdue University, Department of Agricultural and Biological Engineering, 225 South University Street, West Lafayette, IN 47907, United States. Tel.: þ1 765 494 0388; fax: þ 1 765 496 1115. E-mail address: [email protected] (J. Irudayaraj).
http://dx.doi.org/10.1016/j.bios.2014.02.012 0956-5663 & 2014 Elsevier B.V. All rights reserved.
comprehensive pathogen detection methods will enable implementation of effective strategies to avoid, or when needed intervene in, the distribution of foods that have been contaminated with foodborne pathogens. The complexity of food matrices, the wide variety of microorganisms therein, and the varying growth and replication traits of target pathogenic microorganisms that, when present, are often in much smaller numbers than the nonpathogenic microorganisms, pose significant challenges for pathogen detection. Conventional methods used to identify pathogens in contaminated foodstuffs include: colony culture (Reissbrodt, 2004; Velusamy et al., 2010), polymerase chain reaction (PCR) (Malorny et al., 2008; Navas et al., 2006), and immunoassays (e.g., enzyme-linked immunosorbent assays (ELISA) and immunochromatographic assays) (Delehanty and Ligler, 2002; Seo et al., 2010), often in combination with selective enrichment. However, these conventional methods are time-consuming, laborious, and/or lack in sensitivity in part due to the requirement of cultivation steps. Thus, alternate approaches are needed for rapid detection. Biosensors have been developed and documented to improve the limit of detection (LOD) and/or
144
I.-H. Cho et al. / Biosensors and Bioelectronics 57 (2014) 143–148
time to result, based on electrochemical (Patel et al., 2011; Setterington and Alocilja, 2012; Smietana et al., 2011) and optical (Li et al., 2011; Smietana et al., 2011; Subramanian et al., 2006a, 2006b) techniques, including surface enhanced Raman spectroscopy (SERS) (Craig et al., 2013; Wang et al., 2010, 2011b) methods. However, a need still exists to develop highly sensitive easily implementable biosensors for onsite detection of pathogens in complex food matrices. To address this need, an in-situ immunomagnetic bead with enhanced fluorescence approach was developed and applied for the detection of three prevalent foodborne pathogens, Escherichia coli O157:H7, Salmonella typhimurium, and Listeria monocytogenes. Factors critical to the suitability of detection methods and sensors involved in food analysis were investigated, including sensitivity, limit of detection (LOD), detection time, and ease of use.
2. Materials and methods 2.1. Materials Monoclonal antibodies specific to E. coli O157:H7, S. typhimurium, and L. monocytogenes and a polyclonal antibody raised from rabbit specific to L. monocytogenes were purchased from Abcam (Cambridge, MA). Polyclonal antibodies raised from goat which are reactive to E. coli O157:H7 and S. typhimurium were purchased from KPL (Gaithersburg, MD). Dynabeads (M-270, NH2-functionalized), protein ladders (pre-stained) for SDS-PAGE, and two maleimidefunctionalized Alexa Fluor™ dyes (532 and 647) were obtained from Life Technologies (Carlsbad, CA). E. coli O157:H7, S. typhimurium, and L. monocytogenes were obtained from the culture collection at the Center for Food Safety Engineering consortium at Purdue University. Casein (sodium salt), Tween 20, tetramethyl benzidine (TMB), ethanolamine, urea, bovine serum albumin (BSA), Sephadex G-15, dialysis tubing, and proteinase K were obtained from Sigma (St. Louis, MO). 2-(N-morpholino) ethanesulfonic acid (MES), 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), succinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxy-(6-amidocaproate) (LC-SMCC), 2-iminothiolane, dithiothreitol (DTT), fluorescein5-maleimide (maleimide-FITC), and mouse anti-goat IgG labeled with horseradish peroxidase (HRP) were purchased from Pierce (Rockford, IL). Source of the antibodies and bacteria used in this study are provided in Supporting information (S-5). 2.2. Preparation of bacteria E. coli O157:H7, S. typhimurium, and L. monocytogenes were respectively cultured in 500 mL of Luria-Bertani (LB) media at 37 1C with shaking, for 18 h and then harvested by centrifugation (5000 rpm, 20 min) and stored in 50 mL of sterilized PBS (phosphate buffered saline, 10 mM phosphate and 150 mM NaCl). The concentration of bacteria in these stock solutions was determined in triplicate by enumeration on plate count agar following incubation at 37 1C for 24 h. 2.3. Preparation of immuno-magnetic beads For coupling antibodies to the beads, monoclonal antibodies specific to E. coli O157:H7, S. typhimurium, and L. monocytogenes were respectively immobilized onto amine-functionalized beads as described elsewhere (Cho and Irudayaraj, 2013). Briefly, the beads (1 mg) were washed with 0.1 M MES buffer, pH 4.5, and then reacted with the corresponding antibody (10 mg) by adding 50 mM EDC at room temperature for 2 h. For blocking the EDC-activated sites of antibody and the bead surfaces, 1 M ethanolamine (12.2 mL), pH 8.5 and 0.5% (w/v) casein dissolved in PBS buffer was
sequentially reacted for 15 min and 1 h, respectively. After washing with 0.5% (w/v) casein-PBS three times, the conjugates for detection of the three pathogens were stored at 4 1C until use. 2.4. Conjugation of antibody with fluorescent dBSA BSA (0.25 mg) was treated with 1 mL of 8 M urea in de-ionized water (DIW) for 30 min followed by the addition of 20 mM of DTT in DIW for 30 min at 37 1C with gentle shaking to ensure complete denaturation. The urea and DTT were removed by gel filtration with a Sephadex G-15 column (volume 10 mL) and the protein fractions identified by Bradford assay were immediately mixed with the three maleimide-functionalized dyes (i.e., FITC, Alexa Fluor 532 and 647) respectively and maintained at room temperature for 1 h. The excess dyes were consecutively removed by dialysis and gel filtration on Sephadex G-15 chromatography. The dBSA bearing fluorescent dyes were coupled to 50-fold molar excess of LC-SMCC for 1 h followed by purified on Sephadex G-15 column. The fluorescent dBSA was coupled to antibody treated with 10-fold molar excess of iminothiolane (reagent used for thiolation of antibody) at room temperature for 2 h. 2.5. SDS-PAGE analysis Protein samples, i.e., native BSA, anti-L. monocytogenes antibody and the fluorescent dBSA conjugate, and protein marker were treated with sample buffer without mercaptoethanol (non-reducing condition) and loaded on an 8% polyacrylamide gel to verify the conjugation efficiency of the synthesized IFdBSA conjugate. For electrophoretic separation, two voltages (80 and 120 V) were sequentially applied for 20 and 60 min respectively using a MiniProtein Tetra Cell apparatus (Bio-Rad, Hercules, CA). The separation gel was sequentially immersed in staining solution including 0.1% Coomassie blue R-250 for 15 min and destaining solution for 1 h for developing the image. 2.6. In-situ fluorescent bead assays E. coli O157:H7, S. typhimurium, and L. monocytogenes samples serially diluted in PBS (0, 1 101, 1 102, 1 103, 1 104, and 1 105 cells/mL) from stock solution (3 1010, 1 1011, and 5 109 cells/mL for E. coli O157:H7, S. typhimurium, and L. monocytogenes respectively) were first reacted with the corresponding beads (50 μL) with gentle mixing at room temperature for 20 min followed by separation of the supernatant using a magnet (AM10026, Life Technologies). The supernatant was discarded by pipetting and the bead was carefully washed with sterilized PBS. For probe attachment, 0.01 μg/mL of the fluorescent dBSA conjugate (200 μL) for each bacteria in 0.5% casein-PBS containing 0.1% (v/v) Tween 20 (casein-PBS-Tw) was mixed with the bead and incubated at room temperature for 20 min. The beads were washed with casein-PBS-Tw three times and then proteinase K solution (200 μL) diluted 100 times in reaction buffer (20 mM Tris–HCl, pH 8.0, 20 mM NaCl and 10 mM CaCl2) was mixed with the beads. This enzymatic digestion was maintained at room temperature for 30 min. The fluorescent signals obtained from the supernatant after the digestion step were measured by a fluorescence spectrophotometer (Cary Eclipse, Agilent Technologies) at different excitation and emission wavelengths (i.e., 495/519 nm, 531/554 nm, and 650/668 nm for FITC, Alexa532, and Alexa647, respectively). 2.7. Multiplex assays Bacteria were prepared in two concentrations (0 and 100 cells/mL) by using the blank, uninoculated samples, for the zero dose and serial
I.-H. Cho et al. / Biosensors and Bioelectronics 57 (2014) 143–148
dilutions of the stock solutions for the 100 cells/mL samples. Counts were verified by enumeration on plate count agar as described above. For multiplex assays, a cocktail consisting of the three pathogens was mixed and reacted with the three different immunobeads at the same time. The respective fluorescence signal was obtained at the corresponding wavelengths: the fluorophores used were, FITC for E. coli O157:H7, Alexa 532 for S. typhimurium and Alexa 647 for L. monocytogenes). The final volume was kept the same as in the single pathogen assay (200 μL), and the analytical procedure was the same as described above. 2.8. Food sample tests To test the feasibility of the methods for pathogen detection in whole foods, the bead assay was applied to the three different food matrices: spinach wash for E. coli O157:H7, boneless chicken for S. typhimurium, and 2% reduced fat milk for L. monocytogenes using uninoculated foods as controls and also artificially inoculating each bacteria into the corresponding sample medium. Bacteria concentrations used, as enumerated by colony counting were: 0, 3 and 22 CFU/ mL for E. coli O157:H7, 0, 5 and 39 CFU/mL for S. typhimurium and 0, 4 and 30 CFU/mL for L. monocytogenes. Standard sampling procedures were used (Mohammed et al., 2008; Muldoon et al., 2012): a 10 g sample of chicken was mixed with 10 mL of sterilized PBS and inoculated with 1 mL of prepared S. typhimurium; 10 mL of spinach wash was inoculated with 1 mL of E. coli O157:H7; and 10 mL of milk was inoculated with 1 mL of L. monocytogenes. After inoculation and 2 h of incubation at room temperature, 1 mL of the bacterial solution was used and the in-situ bead assays were performed in triplicate, as described above. 2.9. Biosafety concerns All experiments were conducted in BSL-2 approved facilities with proper safety training in place and approvals.
3. Results and discussion 3.1. Analytical concept The proposed in-situ fluorescent bead strategy for the rapid detection of pathogenic bacteria consists of an immunomagnetic bead as a solid support for capturing bacteria and a probe antibody-dBSA conjugate labeled with fluorophores (IFdBSA), as depicted in Fig. 1.
145
The entire procedure of the immunoassays can be performed on a magnetic bead coated with capture antibody, which is specific to the target bacteria. Upon incubation and binding of the beads to the pathogens in the food sample, the bead complex is isolated by a simple magnetic separation step and the target pathogen detected by binding with the IFdBSA conjugate followed by enzymatic digestion for fluorometric signal measurement. This in-situ fluorescent immuno-bead assay could be performed in one tube which enables us to not only detect the target bacteria with simple methods but also in very low numbers due to fluorescence enhancement from the large number of fluorescent signal tracers labeled with dBSA used as a carrier of fluorophores. In addition, the large surface area of the bead facilitates a higher degree of capture efficiency of the target bacteria compared with common immunoassays for pathogen detection (see Fig. S1 in Supporting information), enabling a rapid and sensitive method for the detection of bacteria (Setterington et al., 2011). To maximize the fluorescent signal from the assay, it is crucial for the carrier protein (i.e., dBSA) to bear high numbers of fluorophores. BSA is known as a plasma carrier and transport protein in blood and is a cysteine-rich protein with 35 sulfhydryls (17 intramolecular cysteine disulfides) out of 583 total amino acids of the protein (Liu et al., 2012; Meng et al., 2011), has higher portion of the reactive groups compared to other proteins, and was utilized as a fluorophore carrier in this study. The binding sites of BSA could be exposed by mild treatment using chaotropic agents such as urea or guanidine hydrochloride in aqueous environment due to its destabilizing hydrophobic interaction inside protein (Lim et al., 2009) followed by the cleavage of covalent disulfide bonds with the reduction by DTT. The reduced free sulfhydryl groups allow fluorescent dyes with thiol-reactive groups (e.g., maleimide and iodoacetyl derivatives) to bind so that numerous fluorophores could be functionalized to the interior of the dBSA. The degree of fluorescence on BSA were estimated by considering several factors (see S-2 in Supporting information). Even though antibodies that are directly labeled with aminereactive fluorophores can be used for the detection of bacteria as a probe, the number of fluorophores per antibody molecule is nearly in the range between 3 and 9 (Ogawa et al., 2009; Vira et al., 2010), which limits the level of detection. It should be noted that the fluorescent signals can be quenched among other signals when fluorophores are positioned within the energy transfer distance (Forster Resonance Energy Transfer distance is approximately within 10 nm), to result in a decrease in fluorescent signals. Since the dimension of BSA is approximately 140 40 40 Å3, even though it is totally denatured by urea and
Fig. 1. (a) Scheme of in-situ fluorescent immunoassay on magnetic beads for the detection of pathogenic bacteria. Target bacterium captured by antibody on magnetic bead can be detected by antibody-dBSA conjugate labeled with fluorophores (IFdBSA). (b) Fluorescence images of bacteria used as targets and stained by the corresponding IFdBSA probe. (c) Fluorescence image of bacterium (Escherichia coli O157:H7) captured by the immunomagnetic bead. Red color of bacteria indicates lifetime of FITC. The structural image of BSA is borrowed from protein data bank. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
146
I.-H. Cho et al. / Biosensors and Bioelectronics 57 (2014) 143–148
DTT, the possibility of self-quenching could not be ruled out given the structural configuration. To circumvent this drawback each fluorophore labeled at the intra-molecular site has to be liberated from the fluorescent dBSA conjugate by means of proteasemediated enzymatic digestion. Once amino acid fragments containing the labeled fluorophore are released into the solution, the fluorescent signals can be noted with clarity because of minimal self-quenching since the fluorophores are no longer in close proximity with each other (Jones et al., 1997). Proteinase K known as a strong endopeptidase in a broad range (Betzel et al., 1993) was employed for the dissociation of fluorophores from the IFdBSA conjugate and its enzymatic efficiency on the reduction of selfquench were tested (see Fig. S2 in Supporting information). 3.2. Characterization of the conjugate To verify the biophysical state of the synthesized IFdBSA conjugate, SDS-PAGE analysis was performed on a 8% polyacrylamide
gel under non-reducing condition (i.e., without breaking other disulfide bonds by extra reduction). From the results shown in Fig. 2., it is obvious that the two types of IFdBSA conjugates exist and are positioned at over 200 kDa (lane IV), the one with a 1:1 ratio (i.e., one fluorescent dBSA per one antibody) that presents over half a portion of the total number of conjugates and the other is 1:2 ratio (two fluorescent dBSA per one antibody). Even though a large portion of the free antibody (anti-L. monocytogenes antibody was used in this analysis) and small fluorescent dBSA still exists, approximate efficiency of this conjugate is around 40%. It can also be noted that most of the native BSA in lane II are labeled with a large number of sulfhydryl reactive fluorophores after denaturation, with increasing molecular weight as shown in lane IV. The IFdBSA conjugate characterized by its molecular state by electrophoresis was used as a probe for fluorometric detection of bacteria in combination with the immuno-magnetic bead. 3.3. Performance of the in-situ fluorescent bead assays
Fig. 2. SDS-PAGE analysis of the synthesized IFdBSA conjugate [Lane I (protein ladder), Lane II (native BSA), Lane III (anti-Listeria monocytogenes antibody) and Lane IV (IFdBSA conjugate solution)].
To verify its usability, the IFdBSA conjugate was applied for in-situ fluorescent immunoassays as a probe for the detection of target bacteria in combination with magnetic beads which are used as a solid support. To maximize the analytical performance, some of the key factors (e.g., amount of capture antibody and IFdBSA conjugate and enzymatic digestion time, etc.) were empirically optimized prior to performance the analyses (i.e., 0.01 μg/mL of IFdBSA was used, see Fig. S3 in Supporting information). Bacteria to be tested in solution were first captured and concentrated by the 2.8 μm sized immunomagnetic bead by using a magnet and reacted with IFdBSA in sequence. For fluorometric detection of target bacteria different fluorophores were employed, e.g., FITC (Ex/Em: 495/519 nm), Alexa532 (Ex/Em: 531/554 nm) and Alexa647 (Ex/Em: 650/668 nm) for E. coli O157:H7, S. typhimurium and L. monocytogenes, respectively. As shown in Fig. 1, the emission fluorescent colors from bacteria to be tested are
Fig. 3. Performance tests for the in-situ fluorescent immunobead assays against E. coli O157:H7 (a), Salmonella typhimurium (b) and L. monocytogenes (c), respectively. The fluorescent signals in the Y-axis obtained from the assay were plotted against the bacteria concentration (0, 101, 102, 103, and 104 cells/mL) diluted from stock solution. The assays were performed in triplicate and the limit of detection (LOD) for each assay was calculated by adding three times the standard deviation to the mean of the zero dose of bacteria.
I.-H. Cho et al. / Biosensors and Bioelectronics 57 (2014) 143–148
green (FITC), yellow (Alexa532) and red (Alexa647), which can be released into solution after enzymatic digestion of the conjugates by using the proteinase K. Signal intensities were then obtained from the liberated fluorophores by fluorescence spectroscopy. The in-situ bead assays proposed are able to detect extremely low concentrations of bacteria, regardless of the bacteria species (Fig. 3). The fluorescent intensity from samples containing 10 CFU/mL was significantly higher than the uncontaminated controls. To compare the detection sensitivity of the method, the limit of detection (LOD) was determined by multiplying the standard deviation at zero dose (0 cells/mL) by three and adding that value to the mean at zero dose (Binder and Archimbaud, 2000; Cho et al., 2006) to result in an estimated LOD of o5 cells/mL for E. coli O157:H7, S. typhimurium, and L. monocytogenes. It is evident (Fig. 3) that the analytical sensitivity of this fluorescent immuno-bead assay is much higher, approximately over 104-fold increase in sensitivity, than conventional colorimetric immunoassays such as ELISA (105–106 cells/mL LOD is common) (Galikowska et al., 2011; Kim et al., 1999). Such high level of sensitivity was possible because of the high fluorescent signals obtained from the IFdBSA conjugate, in other words, bearing several
Fig. 4. Multiplex analyses of the in-situ fluorescent immunobead assay against three bacteria (i.e., E. coli O157:H7, S. typhimurium and L. monocytogenes) in singular and mixed types of bacteria samples. These tests were performed at a low concentration ( 100 cells/mL) of the pathogens.
147
tracers inside the dBSA which can be liberated by proteinase Kmediated digestion, resulting in an extremely high fluorescent signal to detect low number of cells.
3.4. Multiplex detection To verify the durability of the bead assays in a food matrix, high specificity toward target bacteria is required because the structure of the bacteria gives rise to analogous antigenic sites (e.g., K-, O-, and H-antigens) which exist at bacteria surfaces. By using a monoclonal antibody as a capture agent, the unique antigenic site of the target bacteria could be recognized so that the cross-reactivity of the assay is minimized as demonstrated by multiplex analysis (Fig. 4). Multiplex testing was performed at a low concentration of the bacteria (i.e., 100 cells/mL) diluted from each stock solution in order to assess the robustness of the assay. It is obvious that multiplex assays of in-situ bead experiments have high specificity to each target bacteria with regard to both singular and multiplex analyses without cross-reactivity (Fig. 4). For the detection of E. coli O157:H7, the assay did not show crossreactivity with other bacteria, i.e., S. typhimurium and L. monocytogenes. For S. typhimurium and L. monocytogenes analyses, the assay was only specific to the corresponding target agent without crossreactivity. Furthermore, the signal intensities obtained from each target bacteria in singular assays are nearly identical to those from the mixed bacteria experiment, implying that this assay is specific and the interference of other pathogens may not hinder the LOD. Although several multiplex immunomethods based on magnetic beads have been reported for multiplex pathogen detection (Charlermroj et al., 2013; Wang et al., 2011a), these methodologies employ other fluorescent tracers such as quantum dots for antibody labeling and lack the sensitivity to detect extremely low concentrations (e.g., under 10 cells/mL) of bacteria. However, the novel fluorometric assay proposed in this study utilizes dBSA as a fluorophore carrier where many signal tracers can be functionalized, enabling a high degree of sensitivity because of the enhancement without the need for further signal amplification steps.
Fig. 5. Performance metrics of the proposed assay in food matrices. (a) Spinach wash for E. coli O157:H7, (b) chicken meat for S. typhimurium, and (c) 2% reduced milk for L. monocytogenes. Error bars denoting the standard deviation are marked at the corresponding concentrations.
148
I.-H. Cho et al. / Biosensors and Bioelectronics 57 (2014) 143–148
3.5. Performance of the biosensor in food samples
References
An extremely low concentration of E. coli O157:H7 (o5 CFU/ mL) in spinach wash could be detected by the bead assay (Fig. 5) employing the same analytical procedure as described above. A very high LOD was also demonstrated when the assay was applied to both S. typhimurium (45 CFU/mL) and L. monocytogenes (o5 CFU/mL). Since a signal-to-noise ratio is an important in parameter in bacteria analysis, the fluorescence intensities were able to be represented as S/N ratio (see Fig. S4 in Supporting information). The sensitivity of the assay applied to inoculated food samples was similar to the LOD found in the buffer systems. Therefore, this assay is expected to be robust and has broad applicability. Furthermore in this assay a pre-enrichment step is likely not necessary due to its high sensitivity.
Betzel, C., Singh, T., Visanji, M., Peters, K., Fittkau, S., Saenger, W., Wilson, K., 1993. J. Biol. Chem. 268 (21), 15854–15858. Binder, R., Archimbaud, Y., 2000. Regul. Toxicol. Pharmacol. 31 (2), S23–S26. Charlermroj, R., Himananto, O., Seepiban, C., Kumpoosiri, M., Warin, N., Oplatowska, M., Gajanandana, O., Grant, I.R., Karoonuthaisiri, N., Elliott, C.T., 2013. PLoS One 8 (4). Cho, I.H., Irudayaraj, J., 2013. Int. J. Food Microbiol. 164 (1), 70–75. Cho, J.H., Han, S.M., Paek, E.H., Cho, I.H., Paek, S.H., 2006. Anal. Chem. 78 (3), 793–800. Craig, A.P., Franca, A.S., Irudayaraj, J., 2013. Annu. Rev. Food Sci. Technol. 4, 369–380. Delehanty, J.B., Ligler, F.S., 2002. Anal. Chem. 74 (21), 5681–5687. Drew, C.A., Clydesdale, F.M., 2013. Crit. Rev. Food Sci. Nutr. (just-accepted) Galikowska, E., Kunikowska, D., Tokarska-Pietrzak, E., Dziadziuszko, H., Łoś, J.M., Golec, P., Węgrzyn, G., Łoś, M., 2011. Eur. J. Clin. Microbiol. Infect. Dis. 30 (9), 1067–1073. Jones, L.J., Upson, R.H., Haugland, R.P., Panchuk-Voloshina, N., Zhou, M., Haugland, R.P., 1997. Anal. Biochem. 251 (2), 144–152. Kim, J.W., Jin, L.Z., Cho, S.H., Marquardt, R.R., Frohlich, A.A., Baidoo, S.K., 1999. J. Sci. Food Agric. 79 (11), 1513–1518. Li, S., Huang, J., Cai, L., 2011. Nanotechnology 22 (42), 425502. Lim, W.K., Rösgen, J., Englander, S.W., 2009. Proc. Nat. Acad. Sci. 106 (8), 2595–2600. Liu, J.Y., Liu, Y., Gao, M.X., Zhang, X.M., 2012. Proteomics 12 (14), 2258–2270. Malorny, B., Löfström, C., Wagner, M., Krämer, N., Hoorfar, J., 2008. Appl. Environ. Microbiol. 74 (5), 1299–1304. Meng, F., Yao, D., Shi, Y., Kabakoff, J., Wu, W., Reicher, J., Ma, Y., Moosmann, B., Masliah, E., Lipton, S.A., 2011. Mol. Neurodegener. 6, 34. Mohammed, Z., Souna, E., Anthony, T., 2008. Springer: New York, NY, USA. Muldoon, M.T., Allen, A.-C.O., Gonzales, V., Sutzko, M., Lindpaintner, K., 2012. J. AOAC Int. 95 (3), 850–859. Navas, J., Ortiz, S., Lopez, P., Jantzen, M.M., Lopez, V., Martinez-Suarez, J.V., 2006. Foodbourne Pathog. Dis. 3 (4), 347–354. Ogawa, M., Kosaka, N., Choyke, P.L., Kobayashi, H., 2009. Cancer Res. 69 (4), 1268–1272. Patel, M., Gonzalez, R., Halford, C., Lewinski, M.A., Landaw, E.M., Churchill, B.M., Haake, D.A., 2011. J. Clin. Microbiol. 49 (12), 4293–4296. Reissbrodt, R., 2004. Int. J. Food Microbiol. 95 (1), 1–9. Scallan, E., Hoekstra, R.M., Angulo, F.J., Tauxe, R.V., Widdowson, M.A., Roy, S.L., Jones, J.L., Griffin, P.M., 2011. Emerg. Infect. Dis. 17 (1), 7–15. Seo, S.-M., Cho, I.-H., Jeon, J.-W., Cho, H.-K., Oh, E.-G., Yu, H.-S., Shin, S.-B., Lee, H.-J., Paek, S.-H., 2010. J. Food Prot. 73 (8), 1466–1473. Setterington, E.B., Alocilja, E.C., 2012. Biosensors 2 (1), 15–31. Setterington, E.B., Cloutier, B.C., Ochoa, J.M., Cloutier, A.K., Patel, P.J., Alocilja, E.C., 2011. Int. J. Food Saf. Nutr. Public Health 4 (1), 83–100. Smietana, M., Bock, W.J., Mikulic, P., Ng, A., Chinnappan, R., Zourob, M., 2011. Opt. Express 19 (9), 7971–7978. Subramanian, A., Irudayaraj, J., Ryan, T., 2006a. Biosens. Bioelectron. 21 (7), 998–1006. Subramanian, A., Irudayaraj, J., Ryan, T., 2006b. Sens. Actuators, B 114 (1), 192–198. Velusamy, V., Arshak, K., Korostynska, O., Oliwa, K., Adley, C., 2010. Biotechnol. Adv. 28 (2), 232–254. Vira, S., Mekhedov, E., Humphrey, G., Blank, P.S., 2010. Anal. Biochem. 402 (2), 146–150. Wang, H., Li, Y., Wang, A., Slavik, M., 2011a. J. Food Prot. 74 (12), 2039–2047. Wang, Y., Lee, K., Irudayaraj, J., 2010. J. Phys. Chem. C 114 (39), 16122–16128. Wang, Y., Ravindranath, S., Irudayaraj, J., 2011b. Anal. Bioanal. Chem. 399 (3), 1271–1278. Zhang, D., Zhang, H., Yang, L., Guo, J., Li, X., Feng, Y., 2009. J. Food Saf. 29 (3), 348–363.
4. Conclusions The proposed in-situ fluorescent immunoassay strategy in combination with a magnetic concentration step was conceived and validated in this study. The method developed enabled the three pathogens, i.e., E. coli O157:H7, S. typhimurium and L. monocytogenes, to be detected with simple steps and could be completed within 2 h. In addition to the ease of implementation and rapid response, the bead assays also show high sensitivity and specificity to target bacteria, in addition to multiplexing. A LOD of o5 CFU/mL was established for the pathogens interrogated in buffer conditions. The LOD in food matrices was also comparable to that in buffer ( o5 CFU/mL) was demonstrated. We expect the simplicity, analytical sensitivity, and robustness of the biosensors will enable effective implementation in various sectors of the food chain as well as in critical check points to enhance food security.
Acknowledgements This effort was supported by funds from USDA-ARS project number 1935-42000-049-00D in conjunction with the Center for Food Safety Engineering at Purdue University. Dr. Andrew Ghering from USDA-ARS provided helpful insights on sampling, and Zhongwu Zhou and Yi Cui at Purdue University are acknowledged for their assistance with fluorescence imaging. 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.2014.02.012.
