Due to recent high profile outbreaks of E. coli
O157:H7 in spinach [2
] and Salmonella
Saintpaul in raw produce [3
], the safety of our food has been of highest concern. From the farm where the food is produced, to the handling practices of manufacturers, to our own kitchens, food safety involves all stages of food production and consumption. Therefore, the need not only for the eradication of food-borne pathogens exists, but also their rapid and sensitive detection once they enter the food chain.
Traditional methods of detecting food-borne pathogens include enrichment, plating to selective and/or differential agar, and biochemical/serological confirmation [4
]. This process can take days and is labor intensive. Immunomagnetic separation (IMS) offers advantages over traditional pathogen enrichment processes. Superparamagnetic particles are coated with antibodies against the target of interest, forming immunomagnetic beads (IMB). The specificity of the antibody coupled with the magnetic properties of the bead, allows a target organism to be separated from a food matrix and background microflora, and concentrated into a smaller sample volume [6
IMS also has the advantage of being very versatile. Many methods of detection can be used to quantify and/or identify the captured bacteria. IMB can be plated directly onto a selective medium [7
] or used as a solid support for an ELISA [8
]. In addition, various types of IMB have been used to capture and concentrate target pathogen cells, which are then eluted from the beads and subjected to PCR identification [9
]. Another common method of detection involves adding a secondary antibody to create a sandwich-immunoassay. Variations on this assay are almost limitless. Secondary antibodies labeled with alkaline phosphatase [12
], horseradish peroxidase [13
], and fluorescein isothiocyanate (FITC) [14
] have been utilized with promising results.
In recent years, time-resolved fluorescence (TRF) has also been used quite extensively as a means of pathogen detection. TRF utilizes lanthanide chelate labels which produce intense fluorescent signals with a long half-life (10-3
s). This long decay time coupled with a large Stokes' shift (>200 nm) and narrow emission peak allow the chelate label to be read after nonspecific background has already decayed. This allows for increased sensitivity and enhanced signal-to-noise ratio. IMS combined with TRF has been successful in the detection of 10 cfu/mL of E. coli
O157:H7 in apple cider [15
], 1 cfu/g of E. coli
O157:H7 in ground beef [1
], and 4 cfu/g of both Salmonella
and E. coli
O157 from germinated alfalfa sprouts [16
In this study, we have chosen a combination of IMS and TRF as a sensitive process to detect different strains of E. coli O157. The study mainly focused on the detection efficiency of using IMB of various sizes which consisted of different types of chemical linkage for conjugating the capture antibody to the superparamagnetic particle. Two types of covalent coupling, Schiff-base and tosylation, along with biotin-streptavidin interaction were used to attach anti-E. coli O157:H7 antibodies to different beads. In addition, we have also examined the efficiency of beads with different size but of a similar density in capturing the targeted pathogens. In conjunction with the efficiency study, a cost analysis was also conducted to reveal the economics of using different IMB. The result could be used as a guide for designing the proper choice of IMB to capture E. coli O157:H7 for TRF measurement.
2. Methods and Materials
2.1. Immunomagnetic beads
Two sizes of IMB consisting of anti-E. coli
O157 antibodies attached via different linking chemistries were examined for the capture and detection of E. coli
O157:H7 (Table 1
). IMB from a commercially available stock were also examined. Small beads with 1 μm diameters include P(S/V-COOH) Mag/Encapsulated (IMB-C1; Bangs, Fishers, IN), and Dynabeads MyOne Streptavidin C1 (IMB-S1; Invitrogen Dynal AS, Oslo, Norway) and Dynabeads MyOne Tosylactivated (IMB-T1; Invitrogen). Larger beads having diameters between 2.6 and 2.8 μm include COMPEL™ Magnetic, COOH-modified microspheres (IMB-C; Bangs), and Dynabeads M-280 streptavidin (IMB-S; Invitrogen) and Dynabeads M-280 tosylactivated (IMB-T; Invitrogen). Commercially available Dynabeads anti-E. coli
O157 (IMB-D; Invitrogen) were also used for comparison. Information about these beads is proprietary. However, the size and density were estimated by microscopic examination and discontinuous sucrose density gradient centrifugation, respectively [17
]. In addition, the stock concentration of beads/mL was estimated by manual counting with a Petroff-Hausser counting chamber (Hausser Scientific, Horsham, PA) as described below in Section 2.4.
2.2. Conjugation of antibodies to IMB
Both sizes of carboxylated beads were conjugated with 16 μg/mg of goat anti-E. coli O157:H7 antibodies (KPL, Gaithersburg, MD) using a PolyLink Protein Coupling Kit (Bangs) following the manufacturer's instructions. Water soluble carbodiimide was used to activate the carboxyl groups on the surface of carboxylated beads. This reaction creates an active ester, which binds to the primary amine groups of the antibody. Biotinylated antibodies were conjugated onto beads containing a layer of streptavidin covalently attached to the bead surface. The affinity of streptavidin for biotin represents a strong noncovalent interaction, having a dissociation constant of ∼10-15 M, rivaling that of covalent bonds. Biotinylated goat anti-E. coli O157:H7 antibodies (KPL) were conjugated to both IMB-S and IMB-S1 at concentrations of 6 μg/mg and 20 μg/mg, respectively, as follows: suspensions of 100 μL of beads were mixed with 900 μL phosphate buffered saline (PBS; Sigma Chemical, St. Louis, MO). Six or twenty microliters of biotinylated goat anti-E. coli O157:H7 antibody (KPL, 1 mg/mL stock concentration) were added and allowed to bind at room temperature for 30 minutes with gentle rocking on a Specimix (Barnstead International, Dubuque, IA). After conjugation, the beads were washed 4 times with a solution of 20 mM PBS, 150 mM NaCl, 2 mM EDTA, and 0.5% BSA. After the final washing, the beads were resuspended in 1 mL of the same solution and stored at 2–8°C. Goat anti-E. coli O157:H7 antibody (KPL) was conjugated to IMB-T and IMB-T1 at concentrations of 20 μg/mg and 40 μg/mg, respectively, following the manufacturer's guidelines. Hydroxy groups on tosylactivated beads were treated with p-toluensulphonyl chloride, resulting in a sulphonyl ester which binds to amino or sulfhydryl groups of the antibody. The optional addition of 0.1% BSA after the first 10 minutes of conjugation was applied. The amount of conjugated antibody and linking chemistry of the IMB-D are considered proprietary by the manufacturer.
2.3. Bound protein determination of IMB
The amount of antibody bound to the IMB was determined by absorbance at 280 nm using a Cary 50 spectrometer (Varian, Inc., Palo Alto, CA). Standard curves of absorbance vs. concentration for each type of antibody were generated by measuring the absorbance of various antibody concentrations at 280 nm using a quartz cuvette. The extinction coefficients at 280 nm derived from the standard curves were used for estimating free antibody concentrations after conjugation. The amounts of antibody bound to beads were calculated from the differences between total-applied prior to and free-remained after the conjugation. The calculated concentrations of antibody bound to the beads are shown in Table 2
2.4. Selection of E. coli strains and enumeration
E. coli O157:H7 strains B1409 (human stool sample; Centers for Disease Control and Prevention, Atlanta, GA), SEA 13B 88 (apple cider outbreak; Food and Drug Administration, Rockville, MD), and 380-94 (salami outbreak; Food Safety and Inspection Service, Washington, DC) along with E. coli O157:NM strain MF 13180-NM (FSIS) and non-O157 strain K12 (source unknown) were grown overnight in 25 mL mEC broth (Becton, Dickinson, and Company, Sparks, MD) at 37°C with shaking at 160 rpm (New Brunswick Scientific, Edison, NJ). After overnight enrichment, bacteria in 1 mL of each culture were pelleted by centrifugation (Eppendorf, Westbury, NY) and resuspended in PBS. The bacteria were then diluted 1:100 in PBS and enumerated using a Petroff-Hausser counting chamber (Hausser Scientific). Six microliters of the 1:100 diluted culture were placed onto the counting chamber slide. The slide consists of 25 0.2 mm×0.2 mm squares. The bacteria in five random squares were counted in duplicate, and cell concentration in cells/mL calculated. Serial dilutions of each E. coli strain were prepared in PBS from 1×108 to 1×102 cells/mL following enumeration. Two hundred microliters of the diluted suspensions were subjected to the TRF immunoassay as described in section 2.6.
2.5. Preparation of cell suspension and inoculation of ground beef
For the ground beef experiments, a suspension of 1 mL PBS containing 25 cells was prepared and used to inoculate 25 g of ground beef in a stomacher bag with mesh filter (Fisher Scientific, Pittsburgh, PA). The inoculum was manually massaged into the ground beef followed by the addition of 225 mL of mEC broth. The sample was then mixed in a stomacher (Seward Medical Limited, London, UK) on low for 30 s, followed by enrichment for 24 h at 37°C with shaking at 160 rpm. After enrichment, aliquots of 200 μL from the side of the mesh filter without ground beef were withdrawn for IMS capture and TRF assay as described in section 2.6. A second aliquot of ground beef was inoculated with 1mL PBS to serve as a blank. The blank was run alongside the sample to gauge the growth of background organisms. The post-enriched sample was also plated onto Sorbitol MacConkey agar supplemented with cefixime and tellurite (CT-SMAC; Becton, Dickinson, and Company) to enumerate the growth of O157:H7 in ground beef. The post-enrichment blank was plated on both plate count agar (PCA; Becton, Dickinson, and Company) and CT-SMAC to enumerate background organisms and ensure no indigenous E. coli O157:H7 in the ground beef.
2.6. Immunomagnetic separation and time-resolved fluorescence detection
Immunomagnetic separation (IMS) was carried out using the KingFisher magnetic particle processor (Thermo Fisher, Waltham, MA). The KingFisher automatically transfers beads between binding and washing steps. The entire assay was performed in black 96-well microtiter plates (Nalge Nunc, Rochester, NY). To row A, suspensions of cultured E. coli cells or enriched ground beef samples with a volume of 200 μL were added to 20 μL of beads. To rows B and D, 200 μL of washing buffer diluted 1:25 from wash concentrate (Wallac Oy, Turku, Finland) supplemented with 0.5% Tween 20 (Acros Organics, Fairlawn, NJ) was added. Goat anti-E. coli O157:H7 (KPL) was labeled with europium using a DELFIA® Eu-Labeling Kit (Perkin Elmer LAS, Boston, MA) following the manufacturer's specifications. The labeled stock was stored frozen at -20°C. Prior to the assay, the europium-labeled detection antibody was diluted in assay buffer (Wallac Oy) supplemented with 0.1% Tween 20 to a concentration of 1 μg/mL, filtered through a 0.45 μm syringe filter (Nalge Nunc), and 200 μL added to each well of row C. Lastly, 200 μL of enhancement solution (Wallac Oy) was added to row E. The samples were processed in the KingFisher, which magnetically transfers the beads from row to row. After initial binding of bacteria to beads in row A for 15 min, the bead/bacteria complexes were washed for 1 min in row B. The washed complexes were then transferred to row C and allowed to bind with the europium-labeled detection antibody for 1 h. After another 1 min wash in row D, the complexes were transferred into the enhancement solution in row E. Proprietary chelators in the enhancement solution extracted the europium to form an Eu-chelator that emitted strong fluorescent signal at 615 nm. The fluorescence intensity displayed as counts per second (CPS) was measured using a VICTOR2 1420 multilabel counter (Perkin-Elmer Wallac, Waltham, MA).
2.7. Data analysis
Data were replicated at least three times, with averages and percent error reported. The percent error was calculated by dividing the standard deviation among replicates by the average signal obtained by the TRF assay and multiplying by 100. Microsoft Excel was applied to determine the statistical parameters of variance and standard deviation using Microsoft Excel spreadsheet. Normalized responses per number of applied beads data were calculated by dividing the TRF response by the number of beads used in the assay for further evaluation on the interaction of the bacteria with the beads.
2.8. IMB cost analysis
The cost of each IMB per assay was estimated by taking into consideration the cost of the beads, antibody, and coupling kit (for carboxylated beads only) used in the conjugation of each IMB. For example, IMB-S costs $899/10mL. Biotinylated goat anti-E. coli antibody costs $400/1 mL. In the labeling procedure, 100 μL of IMB are labeled with 6 μL antibody. Therefore, the beads are $8.99 for 100 μL, and the antibody is $0.40 per μL or $2.40 per 6 μL. Adding the cost of the beads and antibody together, the total cost for preparing 1mL of IMB-S is $11.39. Since 20 μL of beads are used per assay, the cost per assay is $0.23.
We have shown in this report and others that larger size IMB are more effective in the capture of a target of interest. While the surface antibody concentration of IMB-S and IMB-T was less than that of the smaller beads, they still elicited a higher response. This is due to the ability of the larger beads to interact with target organism in larger volumes as proposed in our previous study [18
]. IMB-C, on the other hand, generated higher signals than IMB-C1. However, IMB-C also had a higher concentration of antibodies bound to their surface. Therefore, it cannot be exclusively concluded that the size of the carboxylated beads was the only factor in their generation of higher signals than the IMB-C1. In this report, we have also expanded on this idea to include IMB utilizing three different surface chemistries used to link the capture antibody to the magnetic particles. Different strains of E. coli
O157 react differently to the same beads. This could be due to antigenic differences between the strains. For all strains tested with the 1 μm beads, the streptavidin-coated beads elicited the highest response. For the larger beads, IMB-S also provided the highest normalized signal for strains B1409 and O157:NM. All IMB except IMB-D produced similar signals with strain 380-94. Strain SEA 13B 88 was the only strain which saw IMB-S produce the lowest signal. Again, different antigenic features of this strain are the most probable explanation as to the lower capture with the IMB-S. While each type of antibody linkage tested favorably, especially in a real world sample, the streptavidin-coated beads may hold a slight advantage. On the whole, streptavidin-coated particles performed very well in the capture of E. coli
O157:H7 and O157:NM. Perhaps the orientation of the antibodies on the surface of the streptavidin-coated beads allows for enhanced capture. In addition, streptavidin-coated IMB also provide an economic benefit, costing under $0.50 per assay.