- Email: [email protected]
Rapid detection of foodborne contaminants using an Array Biosensor
Sensors and Actuators B 113 (2006) 599–607
Rapid detection of foodborne contaminants using an Array Biosensor Kim E. Sapsford a , Miriam M. Ngundi b , Martin H. Moore b , Michael E. Lassman b , Lisa C. Shriver-Lake b , Chris R. Taitt b , Frances S. Ligler b,∗ a
b
George Mason University, 10910 University Blvd, MS 4E3, Manassas, VA 20110, USA Center for Bio/Molecular Science & Engineering, Naval Research Laboratory, Washington, DC 20375, USA Available online 19 August 2005
Abstract Foodborne contaminants come in a variety of sizes ranging from simple chemical compounds to entire bacterial cells. Due to public health concerns, there is a current need in the food industry for a sensitive, specific and rapid method to monitor for the presence of these toxic species, either as a result of natural or deliberate contamination. The Array Biosensor developed at the NRL encompasses these qualities, including the ability to measure multiple analytes simultaneously on a single substrate. In this study, we demonstrate the Array Biosensor’s ability to measure both large pathogens, such as the bacteria Campylobacter jejuni (C. jejuni), and small toxins, including the mycotoxins ochratoxin A, fumonisin B, aflatoxin B1 and deoxynivalenol. Sandwich immunoassays were used to measure C. jejuni in buffer and a number of food matrices, while competitive immunoassays, taking only 15 min, were developed for the simultaneous detection of multiple mycotoxins. The combination of sandwich and competitive immunoassay formats on a single substrate was demonstrated, allowing the simultaneous detection of both large (C. jejuni) and small (aflatoxin B1 ) food contaminants. © 2005 Elsevier B.V. All rights reserved. Keywords: Array Biosensor; Sandwich immunoassay; Competitive immunoassay; Bacteria; Campylobacter jejuni; Mycotoxins; Multi-analyte
1. Introduction Foodborne contaminants can take a number of forms, ranging from whole bacterial cells or proteins to simple chemical compounds [1–4]. Food contamination, whether accidental or deliberate, can occur anywhere along the food processing line, from the source to the consumer, and is a major concern for the food industry. Traditional methods for the detection of food contaminants range from culture in selective media, followed by numerous biochemical and serological tests for bacterial or viral pathogens, to chromatography techniques for the smaller chemical compounds [1,3]. While these gold standard methods are typically very sensitive, they often require multiple time-consuming steps, ∗
Corresponding author. Tel.: +1 202 404 6002; fax: +1 202 404 8897. E-mail address: [email protected] (F.S. Ligler).
0925-4005/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2005.07.008
such as extraction, sample clean-up or preconcentration for toxins or multiple cultures for pathogens, prior to measurement, resulting in analysis times which can run into days. Furthermore, they are designed to detect and measure only one particular target at a time. In an effort to reduce the analysis time while maintaining sensitivity, researchers have developed a variety of immunoassay-based measurements for pathogen and toxin detection, with the enzyme-linked immunosorbent assay (ELISA) format being the most common [1,3]. ELISAs have been used for a variety of food pathogens and analysis times range from 20 min to 2 days depending on the extent of sample pretreatment required. The Array Biosensor developed at the Naval Research Laboratory (NRL) has successfully been used in the detection of a variety of protein toxins, organic molecules, physiological health markers, a virus and a number of bacteria, initially in buffer but increasingly in food, biological and
600
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
environmental matrices [5–16]. The developed immunoassays are rapid (15–25 min), simple to perform and require little-to-no sample pretreatment prior to analysis, even for more complex sample matrices. In addition, the twodimensional nature of the slide sensing surface facilitates simultaneous analysis of multiple samples for multiple analytes. In this study, the Array Biosensor measures both a large pathogen, the bacteria Campylobacter jejuni (C. jejuni), and small toxins, such as the mycotoxins ochratoxin A, fumonisin B, aflatoxin B1 and deoxynivalenol. In an extension of a previous study [13], sandwich immunoassays were used to measure C. jejuni in an extended range of food matrices including milk, yogurt, turkey ham and turkey sausage. Campylobacter, which is found to inhabit the intestinal tracts of a variety of healthy mammals and birds, is the most common cause of intestinal and diarrheal disease in the US [1,2]. Infection typically results from the consumption of unpasteurized milk and milk products and undercooked poultry with symptoms including diarrhea, fever, abdominal and muscle pain, nausea and headache. In stark contrast to the large bacterial cells of Campylobacter, mycotoxins are small metabolites produced by fungi that grow on a number of agricultural products prior to harvest or during storage [3,4]. Mycotoxicosis results either from inhalation exposure or the ingestion of contaminated foodstuffs. The associated health problems, which range from vomiting to cancer, are dependent upon the specific mycotoxin to which an individual is exposed. In this study, competitive immunoassays taking only 15 min were developed for the simultaneous detection of multiple mycotoxins. Finally, the combination of sandwich and competitive immunoassay formats on a single substrate was demonstrated, which allows the simultaneous detection of both large (C. jejuni) and small (aflatoxin B1 ) food contaminants.
2. Experimental 2.1. Materials Unless otherwise specified, chemicals were of reagent grade and used as received. Borosilicate glass slides from Daigger & Co. Inc. (Vernon Hills, IL; www.daigger. com) were used as waveguides in all the assays described. Poly(dimethyl)siloxane (PDMS), used for making the assay flow cells, was obtained from Nusil Silicone Technology (Carpintera, CA; www.nusil.com). The 3-mercaptopropyltrimethoxy silane (MTS) and N-(␥-maleimidobutyryloxy) succinimide ester (GMBS) were purchased from Fluka Chemical Co. (St. Louis, MO; www.sigmaaldrich.com). EZ-link biotin-LC-NHS, biotin-LC-PEO-amine, biotin-PEO-amine, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), N-hydroxysuccinimide (NHS) and NeutrAvidin were obtained from Pierce (Rockford, IL; www.piercenet.com). The biotin-SP-conjugated rabbit anti-chicken IgY and Cy5chicken IgY, used as positive controls, were purchased
from Jackson ImmunoResearch (West Grove, PA; www. jacksonimmuno.com). The bacterial antigen C. jejuni (ATCC35918) used in the sandwich assay was grown as described, by Dr. Avraham Rasooly (FDA) and used under biosafety II conditions, requiring that the cells were killed prior to transportation to the NRL [13]. The polyclonal antibody against C. jejuni, rabbit anti-C. jejuni, was obtained from Biodesign International (Saco, ME; www. biodesign.com). The mycotoxins, ochratoxin A (OTA), fumonisin B (FB), aflatoxin B1 (AFB1 ) and deoxynivalenol (DON), used in the competitive assays were purchased from Sigma (St. Louis, MO; www.sigmaaldrich.com). Rabbit anti-OTA was purchased from ImmuneChem Pharmaceuticals Inc. (Burnaby, BC, Canada; www.immunechem.com). Monoclonal mouse antibodies against FB, AFB1 and DON were kindly supplied by Dr. C. Maragos of the USDA (Peoria, IL). Fluorescent labeling of the antibodies was achieved using Cy5 bisfunctional dye from Amersham Biosciences (Arlington Heights, IL; www.amershambiosciences.com). Bovine serum albumin (BSA), gelatin, potassium hydroxide (KOH), sodium chloride (NaCl), sodium phosphate monoand di-basic, polyoxyethylenesorbitan monolaurate (Tween 20), dimethyl sulfoxide (DMSO), N,N-dimethylformamide (DMF) and toluene were supplied by Sigma–Aldrich (www.sigmaaldrich.com). Absolute ethanol was obtained from Warner-Graham Co. (Cockeysville, MD). All food was purchased from local grocery stores. A Waring two-speed commercial blender, equipped with a mini-sample container (37–110 ml), used for some of the food preparation, and isopropyl alcohol were purchased from Fisher Scientific (Pittsburgh, PA; www.fisherscientific.com). 2.2. Preparation of biotin-labeled mycotoxins 2.2.1. Ochratoxin A-biotin To a test tube containing a small stir bar, 6.3 mg NHS and 10.3 mg EDC were added and the tube closed with a septum. Next, a solution of 10 mg OTA in 0.4 ml DMF was added to the test tube. After stirring for 1 h, a solution of 22.6 mg biotin-LC-PEO-amine in 1 ml 0.05 M carbonate–bicarbonate buffer (pH 9.5) was added and then stirred for 24 h at 4 ◦ C. The reaction mixture was then transferred to a 500 MWCO dialysis bag and dialyzed against several changes of 1 l phosphate buffered saline (PBS). The biotin-OTA product was then stored in a vial at 4 ◦ C [15]. 2.2.2. Deoxynivalenol-biotin In a 20 ml vial, 10 mg DON was dissolved in 2 ml of acetone. Next, 110 mg of 1,1 -carbonyldiimidazole was added and the reaction stirred for 1 h at room temperature. Water, 40–50 l, was then added followed, by a solution of 28.3 mg biotin-PEO-amine in 1 ml of 0.1 M sodium bicarbonate buffer (pH 8.5) and stirred for 24 h at 4 ◦ C. The reaction mixture was then transferred to a 500 MWCO dialysis bag and dialyzed several times against 1 l PBS. The biotin-DON product was then stored in a vial at 4 ◦ C.
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
2.2.3. Aflatoxin B1 -biotin The route to AFB1 -biotin is a two-step process in which an oxime intermediate of the AFB1 is first prepared. Into a 50 ml flask was placed 31 mg of (aminooxy)acetic acid. Next, 20 mg of AFB1 , dissolved in 1:4:1 of pyridine:methanol:water, was added to the flask. The reaction was refluxed for 2 h, cooled and the solvents removed by rotary evaporation. The resulting AFB1 -(O-carboxy methyl) oxime intermediate was characterized by positive ion LC mass spectrometry equipped with a nano-spray source. For the second part of the synthesis, 15 mg NHS and 30 mg EDC were added to a flask containing a stir bar which was then closed with a septum. Next, a solution of 24.6 mg AFB1 -(O-carboxy methyl) oxime in 3 ml of DMF was added to the flask. After 1 h of stirring, a solution of 50 mg of biotin-LC-PEO-amine in 2 ml of a 0.05 M carbonate–bicarbonate buffer (pH 9.5) was added to the flask and then stirred for 24 h at 4 ◦ C. The reaction mixture was then transferred to a 500 MWCO dialysis bag and dialyzed several times against 1 l PBS. The biotin-AFB1 product was then stored in a vial at 4 ◦ C. 2.2.4. Fumonisin-biotin A solution containing 2 mg FB in 0.1 M sodium bicarbonate buffer (pH 8.5) was prepared and biotin-LC-NHS (dissolved in DMSO) was added to give a final molar ratio of 3 biotin:1 FB. The reaction mix was incubated for 1 h at room temperature before being transferred to a 500 MWCO dialysis bag and dialyzed several times against 1 l PBS. The biotin-FB product was then stored in a vial at 4 ◦ C. All of the resulting biotin-mycotoxins were characterized using mass spectroscopy and quantified using UV–vis spectroscopy. 2.3. Preparation of labeled antibodies Biotinylation of the anti-C. jejuni capture antibody was performed using the EZ-link biotin-LC-NHS at a 5:1 molar ratio (biotin:antibody) according to the manufacturer’s instructions. Labeled antibodies were separated from unincorporated biotin using size exclusion chromatography (BioGel P10, Bio-Rad, Hercules, CA; www.biorad.com). Cy5 labeling of the rabbit anti-C. jejuni, rabbit anti-OTA and monoclonal mouse antibodies to the mycotoxins DON, AFB1 and FB were performed according to the manufacturer’s instructions. Labeled antibodies were separated from unincorporated dye using size exclusion chromatography (BioGel P10). Protein-to-dye ratios were determined using UV–vis spectroscopy. 2.4. Patterning biotinylated capture species using PDMS flow cells Microscope slides, used as waveguides, were cleaned by immersion in a 10% (w/v) KOH in 2-propanol for 30 min at room temperature, followed by rinsing with deionized
601
water and drying with a nitrogen stream. The slides were then immediately immersed in a toluene solution containing 2% MTS for 1 h under nitrogen. The silanized slides were then rinsed with toluene, dried with nitrogen and immersed in 1 mM GMBS, prepared in absolute ethanol, for 30 min at room temperature. The slides were then rinsed with water, incubated in 25 g/ml NeutrAvidin in PBS overnight at 4 ◦ C, before being washed in PBS and either used for patterning or stored in PBS at 4 ◦ C until required. Patterning of the biotinylated capture species was carried out as previously described [5]. Briefly, a 6- or 12-channel patterning PDMS flow cell was clamped onto the NeutrAvidin functionalized waveguide surface. Biotin-conjugated capture species, either anti-C. jejuni (10 g/ml) for the sandwich assay or mycotoxin OTA (10 g/ml), DON (10 g/ml), AFB1 (5 g/ml) and FB (5 g/ml) all in PBS + 0.05% Tween (PBST) for the competitive assays, were introduced into the channels of the flow cell. In the case of the sandwich assays, biotinylated antichicken IgY (10 g/ml in PBST) was introduced into one or two of the channels for use as a positive control. The slides were then incubated overnight at 4 ◦ C. The channels of the flow cell were then rinsed with 1 ml PBST, the slide removed from the PDMS template and placed in PBST blocking solution containing either 1% BSA or 1% gelatin. After 1–2 h, the slides were rinsed with Milli-Q water and assembled in a 6- or 12-channel assay PDMS flow cell. 2.5. Food sample preparation for C. jejuni sandwich assays In a previous study, dose–response curves were obtained for Campylobacter in buffer and a number of spiked food samples [13]. In all cases, bacterial cells were spiked into an initial sample and a 50% serial dilution was carried out with unspiked sample matrix to produce samples for analysis. The spiked samples were typically allowed to stand at room temperature for 2 h prior to analysis. The 2 h incubation period was chosen following dose–response versus incubation time studies carried out previously by Shriver-Lake et al. [12]. For the ground turkey sausage and ham, 10 g of meat and 8 ml of PBST containing 1 mg/ml bovine serum albumin (PBSTB) were placed into a 37–110 ml mini-sample container on a Waring Blender and blended on high for 2 min. The resulting mixture was put into a 50 ml centrifuge tube. The blender container was then rinsed with 2 ml PBSTB and the rinse solution added to the centrifuge tube to yield a 1:1 (w/v) mixture of meat and buffer. The bacterial cells were then spiked into this mixture (assuming total volume of 10 ml) and the sample left at room temperature for 2 h. The mixture was centrifuged at 4000 rpm for 10 min, the supernatant collected and then serially diluted with unspiked supernatant prior to analysis. Carnation nonfat dried milk was prepared per instruction on the box. In a 50 ml centrifuge tube, 45 ml of milk and 5 ml 10× PBSTB was mixed and the sample was then spiked with the bacterial cells for analysis. Serial dilutions were carried
602
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
out using unspiked milk. Vanilla fat free yogurt (100 ml) was diluted 50% with 100 ml PBSTB; the sample was then spiked with the bacterial cells, and serial dilutions were carried out using unspiked yogurt. 2.6. Sandwich and competitive assays All assays were performed using 6- or 12-channel assay PDMS flow cells, whose channels are perpendicular to the stripes of capture species patterned on the waveguides [17]. A sandwich assay was performed for the detection C. jejuni. The antibody functionalized slide was assembled with the assay PDMS flow cell and hooked up to an ISMATEC® multichannel pump (Cole-Parmer Instruments Company, Vernon Hills, IL; www.coleparmer.com) with tubing connected to one end of each channel in the PDMS (outlet). Syringe barrels (1 ml) were then attached at the opposite end of each channel (inlet). The PDMS channels were first washed with 1 ml of PBSTB at 0.5 ml/min, to check for any leaks in the PDMS flow cell. Next, 0.8 ml of sample was applied to each channel at a flow rate of 0.15 ml/min for 15 min; the sample was recycled over the surface by connecting the inlet and outlet of each channel. The sample was then removed and the channels washed with 1 ml of PBSTB at 0.5 ml/min. The fluorescently labeled tracer mixture (0.4 ml:100 ng/ml Cy5-Chicken IgY (positive control) and 10 g/ml Cy5-anti-C. jejuni in PBSTB) was applied to the channels at 0.15 ml/min. The channels were washed with 1 ml of PBSTB, before the assay PDMS flow cell was removed. Finally, the slide was washed with Milli-Q water, dried with nitrogen and imaged on the Array Biosensor. Competitive assays were performed for the simultaneous detection of the mycotoxins OTA, FB, AFB1 and DON. Here, NeutrAvidin slides functionalized with biotinylated analogs of the mycotoxins were assembled with the assay PDMS flow cell and hooked up to the pump as described above for the sandwich assay. The PDMS channels were first washed with 1 ml of PBSTB, at 0.5 ml/min, then 0.8 ml of sample was applied to each channel at a flow rate of 0.15 ml/min; the sample was not recycled over the surface. In the case of the competitive assay, the sample contained both the “free” mycotoxin to be measured and the fluorescent antibody specific to that particular mycotoxin; therefore, the assay is a one-step procedure, rather than the two-step sandwich assay. After the sample had passed over the surface the channels were washed with 1 ml of PBSTB at 0.5 ml/min. Then, the PDMS flow cell was removed and the slide was washed with Milli-Q water, dried with nitrogen and imaged on the Array Biosensor. The possibility of running both competitive and sandwich assay formats simultaneously on the same waveguide was also investigated. In this case, the waveguides were functionalized with capture antibodies to C. jejuni and the biotin analog of AFB1 . The slides were assembled with the assay PDMS flow cell and hooked up to the pump. The PDMS channels were first washed with 1 ml of PBSTB at 0.5 ml/min,
then 0.8 ml of sample (in PBSTB) was applied to each channel at a flow rate of 0.15 ml/min; sample was not recycled over the surface. In this case, the sample contained C. jejuni bacteria, “free” AFB1 and Cy5-anti-AFB1 antibody. After the sample had passed over the surface, the channels were washed with 1 ml of PBSTB at 0.5 ml/min, and the fluorescently labeled tracer (0.4 ml; 10 g/ml Cy5-anti-C. jejuni in PBSTB) was applied to the channels at 0.15 ml/min. The channels were washed with 1 ml of PBSTB, the assay PDMS flow cell removed and the slide washed with Milli-Q water, dried with nitrogen and imaged on the Array Biosensor. 2.7. Slide imaging and data analysis The slide was imaged using a Peltier-cooled CCD, which has previously been described [17]. Briefly, evanescent wave excitation of the surface-bound fluorescent species was achieved using a 635 nm, 12 mW diode laser (Lasermax, Rochester, NY; www.oemlaser.com). Light was launched through a 1 cm focal length lens with a line generator into the end of the waveguide at an appropriate angle. The fluorescence emission was monitored at right angles to the planar surface of the waveguide. A two-dimensional graded index of refraction (GRIN) lens array (Nippon Sheetglass, Summerset, NJ) was used to image the fluorescent pattern onto the Peltier-cooled CCD camera (Spectra Source, Teleris, Westlake Village, CA). Long-pass (Schott 0G-0665, Schott Glass, Duryea, PA; www.schott.com) and band-pass filters (Corion S40-670-S, Franklin, MA; www.corion.com) were mounted on the device scaffolding to eliminate excitation and scattered light prior to CCD imaging. Data were acquired in the form of digital image files in Flexible Image Transport System (FITS) format. To analyze the images, a custom software application was written in LabWindows/CVI (National Instruments; www.ni.com). The program creates a mask consisting of data squares (enclosing the areas where the capture antibody is patterned) and background rectangles which are located on either side of each data square. The average background value is subtracted from the average data square value and net intensity value is calculated and imported into a Microsoft Excel file for data analysis. Limits of detection (LODs) were calculated as the lowest tested concentrations at which the fluorescence signal was at least three standard deviations above the mean fluorescence of the buffer blank.
3. Results and discussion The NRL Array Biosensor has been used in the rapid (25 min or less) identification and quantification of a variety of different target species in a number of complex matrices with little-to-no sample pretreatment or target concentration [5–16]. Typically, detection of the target species involves the use of the sandwich immunoassay format which is appropriate for the larger protein and bacterial pathogens. In an
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
603
Table 1 C. jejuni detection in food and beverage samples
Fig. 1. Sandwich immunoassays. (a) The final CCD image after the slide was exposed to C. jejuni in yogurt, 0–3 × 104 cfu/ml, and a 3 × 104 cfu/ml C. jejuni in PBSTB positive control (PC) and a PBSTB negative control (NC). (b) The resulting dose–response curves for C. jejuni in various sample matrices plotted as normalized net intensity versus bacterial cell concentration; PBSTB (closed circles), yogurt (open circles), milk (closed square), ground turkey sausage (open squares) and ground turkey ham (closed triangles). The net intensity values are an average of 10 or more data squares ± S.D. with a minimum of two separate slides.
extension of a previous study, we have expanded the number of different foods in which we demonstrated detection of C. jejuni. A typical CCD image, taken with the NRL Array Biosensor, is shown in Fig. 1a for the detection of C. jejuni bacteria in yogurt. The CCD image was then converted into intensity values and the resulting normalized dose–response curves are shown in Fig. 1b. A positive control in PBSTB was run as part of the assay and the net intensities for the different food matrices are normalized using this positive control. This allows one to account for slide-to-slide variations in the sandwich assay, independently of variations related to the use of different sample matrices. The dose–response curves and the resulting limits of detection, listed in Table 1, were consistent with previous observations [13] and were dependent on the sample matrix. C. jejuni could be detected at concentrations
Sample matrix
Limit of detection (cfu/ml)
PBSTB Yogurt Milk Ground turkey sausage Ground turkey ham
938 1880 469 469 3750
as low as 4.7 × 102 cfu/ml in milk and ground turkey sausage, but the LOD was found to increase to 3.8 × 103 cfu/ml in ground turkey ham. The higher LODs in certain food matrices may result from (1) an increase in viscosity of the sample with concomitant decrease in target transport to the surface, (2) an interaction of the bacterial cells with the food matrix, making the cells unavailable to interact with the capture antibody or (3) interaction of the food matrix with the capture antibody decreasing interaction with the bacterial cells. Due to the observed variation in the dose–response curves, it is important that a dose–response curve in the matrix of choice be included as part of the assay design if quantification of the analyte is required. While the sandwich immunoassay format works extremely well for species large enough to contain separate recognition sites for both the capture and tracer antibodies, there are a number of foodborne contaminants too small to meet such criteria. In this situation, detection of the contaminant can be achieved using either a competitive or displacement immunoassay. These different immunoassay formats have been demonstrated previously for the detection of TNT, using the Array Biosensor [10]. Mycotoxins are examples of toxic organic compounds often found in agricultural products which are not destroyed by food processing procedures [3,4]. We have recently developed a competitive immunoassay for detection of the mycotoxin ochratoxin A which uses an immobilized biotin analog of OTA, to compete with free OTA present in a solution containing fluorescently labeled anti-OTA antibody [16]. Unlike the sandwich immunoassay, the competitive immunoassay format produces a decreasing fluorescence signal with increasing concentration of the mycotoxin free in solution. One of the advantages of using the planar waveguide system coupled with the PDMS flow cells is the ability to test multiple samples simultaneously for multiple target analytes. Therefore, we wanted to investigate the ability of the NRL Array Biosensor system to carry out simultaneous, multiplexed, competitive immunoassay-based detection of different mycotoxins and also to determine the feasibility of running sandwich and competitive immunoassays formats on a single waveguide surface. To address whether multiple competitive-based assays could be performed on a single substrate, NeutrAvidin slides were patterned with biotinylated analogs of the mycotoxins OTA, AFB1 , FB and DON; a biotinylated anti-chicken IgY lane was also added as a positive control. The slides were then exposed to an antibody mix (Ab-mix), made up of 100 ng/ml Cy5-chicken IgY,
604
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
10 g/ml Cy5-anti-OTA, 20 g/ml Cy5-anti-DON, 2 g/ml Cy5-anti-AFB1 and 0.25 g/ml Cy5-anti-FB in PBSTB. This mix also contained various combinations of the mycotoxins OTA, AFB1 , FB and DON at concentrations known to compete with the immobilized mycotoxins for binding sites on the antibodies. The resulting CCD image is shown in Fig. 2a, with the predicted slide image illustrated in Fig. 2b. Clearly, when no free mycotoxins are present in the assay solution the antibodies bind to the immobilized toxin and generate signals from all mycotoxin spots (lowest lane). However, as mycotoxins are introduced into the Ab-mix solution, a resultant decrease in signal intensity is observed only in the region of the slide functionalized with the biotinylated equivalent of the particular mycotoxin introduced. The captured CCD image (Fig. 2a) is in good agreement with the predicted image in Fig. 2b, demonstrating the specificity of these competitive-based immunoassay formats. While there was found to be some non-specific binding of the labeled antibodies to the NeutrAvidin, as observed by the signal intensity found both in the PBS column and in background regions between data squares, see Ab-mix lane (Fig. 2a), the intensity is lower than the data regions and was easily subtracted during the image analysis. This situation was found for slides blocked either in 1% gelatin or 1% BSA solutions prior to the assay and may have implications for the sensitivity of the assay. The ability to run both sandwich and competitive immunoassays formats on a single waveguide surface is advantageous as it allows the user to monitor for both large and small molecular weight food contaminants simultaneously. C. jejuni and AFB1 were chosen, representing a large bacterial pathogen and a small organic toxin, respectively, to determine if these two immunoassay formats could be run on a single waveguide. Here, the NeutrAvidin waveguide was
assembled in a six-column patterning PDMS flow cell with three columns exposed to biotinylated Rb-anti-C. jejuni and the remaining columns to biotinylated AFB1 . The standard sandwich immunoassay format is a two-step process where the functionalized slide surface is exposed first to the analyte solution for 8–15 min followed by exposure to the antibody tracer for 4 min. The competitive format, on the other hand, is a one-step process where the slide surface is exposed to a mixture of the antibody tracer and free toxin simultaneously for 8 min. When combining immunoassay formats, the C. jejuni and the competitive assay tracer/toxin (AFB1 and Cy5-anti-AFB1 ) mix were combined and incubated together with the waveguide for 8 min. This was followed by incubation of the sandwich assay antibody tracer (Cy5-anti-C. jejuni) for 4 min. Fig. 3a shows the final CCD image of a slide exposed to the sandwich immunoassay for C. jejuni and the competitive immunoassay for AFB1 , run either individually or combined. The top two assays lanes represent the sandwich immunoassay lanes only, with and without the presence of the bacteria. Clearly, when no C. jejuni is present the tracer antibody, Cy5-anti-C. jejuni, does not bind to the surface; however, when bacteria are introduced, a signal is generated only in the Campylobacter-specific region of the slide functionalized and not in the biotin-AFB1 region. Likewise, in the middle two assay lanes, which represent the competitive immunoassay only lanes, signal is generated by the Cy5-anti-AFB1 antibody in the absence of free AFB1 only in the biotin-AFB1 functionalized regions (lane 3). When AFB1 is introduced into the competitive reaction mix (lane 4), the signal is completely lost at the AFB1 data squares due to the free solution toxin saturating the antibody binding sites, therefore, preventing binding of the tracer antibody to the surface immobilized toxin. The bottom two assay lanes
Fig. 2. Multiplexed mycotoxin competitive assays. (a) Slide was patterned with the biotinylated anti-chicken IgY, OTA, DON, AFB1 and FB. The slide was assayed with the mycotoxin analyte(s) in the presence of the antibody mix (Ab-mix). The following concentrations of mycotoxins were assayed: 10 g/ml OTA, 10 g/ml DON, 50 ng/ml AFB1 and 20 g/ml FB. (b) Predicted slide image from the toxin combinations exposed; white squares represent regions that should produce a signal and black squares where low-to-no signal is expected.
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
605
Fig. 3. Sandwich and competitive immunoassays on the same sensing substrate. (a) The final CCD image after the slide was simultaneously exposed to the C. jejuni (5 × 104 cfu/ml) sandwich assay and the AFB1 (1 ng/ml) competitive assay in various combinations. (b) Dose–response curves are shown for C. jejuni presence (closed circles) and absence (open circles) of the AFB1 competitive assay plotted as normalized net intensity versus bacterial cell concentration. (c) Dose–response curves are shown for AFB1 in the presence (closed circles) and absence (open circles) of the C. jejuni sandwich immunoassay plotted as inhibition (%) versus free solution AFB1 concentration.
demonstrate the signals obtained when the sandwich and competitive assays are run together with target present. Here, C. jejuni is present in both lanes, therefore, the sandwich assay, in the anti-C. jejuni functionalized regions, generates signal in both assay lanes. Free AFB1 , on the other hand, is only present in the final assay lane and results in the expected loss of signal in the biotin-AFB1 functionalized regions. Fig. 3b and c show the dose–response curves for the C. jejuni sandwich assay and the AFB1 competitive assay, respectively, obtained either in the presence (closed circles) or absence (open circles) of the other assay format. In both cases, the LODs were actually found to improve (decrease) slightly in the presence of the other assay. C. jejuni sandwich assay alone had an LOD of 1000 cfu/ml [this study, 13] which was found to decrease to 500 cfu/ml in the in presence of the AFB1 competitive assay. Likewise, the AFB1 competitive assay alone had an LOD of 0.9 ng/ml, but was found to decrease to 0.3 ng/ml in the presence of the C. jejuni sandwich assay. These results clearly demonstrate that the sandwich and competitive immunoassay formats can successfully be
run on the same substrate for the detection of both large and small food contaminants simultaneously.
4. Conclusions This study has clearly demonstrated the versatility of the Array Biosensor for the detection of both large and small food contaminants either individually or simultaneously. The bacterial pathogen C. jejuni was measured using a sandwich immunoassay both in buffer and additional complex food matrices with LODs ranging from 500 to 3780 cfu/ml. The mycotoxins OTA, DON, AFB1 and FB were detected simultaneously on a signal substrate using a competitive-based immunoassay format taking only 15 min. The combination of sandwich and competitive immunoassay formats on a single substrate was demonstrated, allowing the simultaneous detection of both large (C. jejuni) and small (AFB1 ) food pathogens with LODs in buffer of 500 cfu/ml and 0.3 ng/ml, respectively. In the case of C. jejuni, as little as 500 cells are known to case human illness and therefore our assay
606
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
LOD currently stands around this desired threshold [18]. On the other hand, the AFB1 LOD, 0.3 ng/ml for the combined assay, is below the current maximum levels set at 2 ng/g (2 ng/ml) [19]. Future studies will include testing the competitive and sandwich/competitive combined immunoassay formats in food samples to assess if assay performance will be affected.
Acknowledgements The authors would like to thank Dr. A. Rasooly (FDA) and Dr. C. Maragos (USDA) for their contribution of materials. This work was supported by funding from the United States Department of Agriculture (USDA) and the Food and Drug Administration (FDA). The views expressed are those of the authors and do not represent those of the US Navy, US Department of Defense or the U.S. Government.
References [1] E.C. Alocilja, S.M. Radke, Market analysis of biosensors for food safety, Biosens. Bioelectron. 18 (2003) 841–846. [2] Y. Mao, C. Zhu, E.C. Boedeker, Foodborne enteric infections, Curr. Opin. Gastroenterol. 19 (2003) 11–22. [3] Council for Agricultural Science and Technology, Mycotoxins: risks in plants, animals and human systems, Report No. 139, Council for Agricultural Science and Technology, Ames, IA, January 2003. [4] B.B. Jarvis, J.D. Miller, Mycotoxins as harmful indoor air contaminants, Appl. Microbiol. Biotechnol. 66 (2005) 367–372. [5] C.A. Rowe, L.M. Tender, M.J. Feldstein, J.P. Golden, S.B. Scruggs, B.D. MacCraith, J.J. Cras, F.S. Ligler, Array Biosensor: simultaneous analysis for bacterial, viral, and toxic analytes in blind samples, Anal. Chem. 71 (1999) 3846–3852. [6] C.A. Rowe, S.D. Balderson, M.J. Feldstein, J.P. Golden, F.S. Ligler, An array immunosensor for simultaneous detection of clinical analytes, Anal. Chem. 71 (1999) 433–439. [7] C.A. Rowe-Taitt, J.J. Cras, C.H. Patterson, J.P. Golden, F.S. Ligler, A ganglioside-based assay for cholera toxin using an Array Biosensor, Anal. Biochem. 281 (2000) 123–133. [8] C.A. Rowe-Taitt, J.P. Golden, M.J. Feldstein, J.J. Cras, K.E. Hoffman, F.S. Ligler, Array Biosensor for detection of biohazards, Biosens. Bioelectron. 14 (2000) 785–794. [9] C.A. Rowe-Taitt, J.W. Hazzard, K.E. Hoffman, J.J. Cras, J.P. Golden, F.S. Ligler, Simultaneous detection of six biohazardous agents using a planar waveguide Array Biosensor, Biosens. Bioelectron. 15 (2000) 579–589. [10] K.E. Sapsford, P.T. Charles, C.H. Patterson Jr., F.S. Ligler, Demonstration of four immunoassay formats using the Array Biosensor, Anal. Chem. 74 (2002) 1061–1068. [11] C.R. Taitt, G.P. Anderson, B.D. Lingerfelt, F.S. Ligler, Nineanalyte detection using an Array Biosensor, Anal. Chem. 74 (2002) 6114–6120. [12] L.C. Shriver-Lake, Y.S. Shubin, F.S. Ligler, Detection of staphylococcal enterotoxin B in spiked food samples, J. Food Prot. 66 (2003) 1851–1856. [13] K.E. Sapsford, A. Rasooly, C.R. Taitt, F.S. Ligler, Detection of Campylobacter and Shigella species in food samples using the Array Biosensor, Anal. Chem. 76 (2004) 433–440. [14] C.R. Taitt, Y.S. Shubin, R. Angel, F.S. Ligler, Detection of Salmonella enterica serovar Typhimurium by using a rapid, arraybased immunosensor, Appl. Environ. Microbiol. 70 (2004) 152–158.
[15] L.C. Shriver-Lake, C.R. Taitt, F.S. Ligler, Applications of Array Biosensor for detection of food allergens, J. AOAC Int. 87 (2004) 1498–1502. [16] M.M. Ngundi, L.C. Shriver-Lake, M.H. Moore, M.E. Lassman, F.S. Ligler, C.R. Taitt, Array Biosensor for detection of ochratoxin A in cereals and beverages, Anal. Chem. 77 (2005) 148–154. [17] M.J. Feldstein, J.P. Golden, C.A. Rowe, B.D. MacCraith, F.S. Ligler, Array Biosensor: optical and fluidics systems, J. Biomed. Microdevices 1 (1999) 139–153. [18] M.H. Kothary, U.S. Babu, Infective dose of foodborne pathogens in volunteers: a review, J. Food Saf. 21 (2001) 49–73. [19] J. Stroka, E. Anklam, New strategies for the screening and determination of aflatoxins and the detection of aflatoxin-producing moulds in food and feed, Trends Anal. Chem. 21 (2002) 90– 95.
Biographies Kim E. Sapsford has been a George Mason University contractor at the Center for Bio/Molecular Science and Engineering at the Naval Research Laboratory in Washington, DC, since 2001. She earned her BSc (Chemistry) and PhD (Biosensors) degrees from the University of East Anglia in the UK. Her research at NRL has been primarily in the field of optical biosensors. Miriam M. Ngundi received her PhD from the State University of New York (Binghamton University) in 2003. Her dissertation work was on the rational design of chemical and biochemical sensors. She is currently a postdoctoral fellow at the Center for Bio/Molecular Science & Engineering at the Naval Research Laboratory. Her current area of research is array biosensors. Martin H. Moore received his BS in education and MS in chemistry from Youngstown State University. In 1998, he accepted a position as a research scientist for Geo-Center’s Inc. to do research at the Naval Research Laboratory (NRL) in Washington, DC, where he became a federal employee in 2001 and is currently employed. Martin Moore has been extensively involved in the development and synthesis of a variety of compounds and materials for use in a number of applications. Michael E. Lassman served as an American Society for Engineering Education Postdoctoral Fellow from 2003 to 2005 at the Naval Research Laboratory’s Center for Bio/Molecular Science and Engineering. He earned his PhD in analytical chemistry from the University of Delaware in 2003 and is now a Senior Research Biochemist at Merck and Co. His current research initiatives focus on using mass spectrometry to solve biological problems. Lisa C. Shriver-Lake received her BS in animal science degree in 1979 from the University of Maryland at College Park and a MS in chemistry from George Mason University, Fairfax, VA, in 1991. In 1985, she accepted a position as a research scientist for Geo-Center’s Inc. to do research at the Naval Research Laboratory (NRL) in Washington, DC, where she became a federal employee in 1987. During the last 19 years, Ms. Shriver-Lake has developed several antibody-based biosensors and applied them for biological warfare detection, environmental applications and food safety. Chris R. Taitt earned her AB in biology from Dartmouth College in 1984, her MS in pomology from Cornell University in 1986 and her PhD in biology from the Johns Hopkins University in 1995. She has been a member of the Center for Bio/Molecular Science and Engineering at the Naval Research Laboratory since 1995. Her research at NRL has been primarily in the field of optical biosensors and has involved biochemical assay development, as well as transitioning of sensors to field use. Her interest in basic biochemistry has influenced her most recent work; she currently leads several projects developing novel recognition
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607 schemes, rapid detection methodologies and amplification methods for rapid detection purposes. Frances S. Ligler is the US Navy’s Senior Scientist for Biosensors and Biomaterials. She earned a BS from Furman University (1971) and both
607
DPhil (1977) and DSc (2000) degrees from Oxford University. She is a member of the National Academy of Engineering and a Presidential Rank Award winner. She has over 220 full-length publications and 24 patents. Her current R&D initiatives focus on optical biosensors, microfluidics and proteomics.
Rapid detection of foodborne contaminants using an Array Biosensor Kim E. Sapsford a , Miriam M. Ngundi b , Martin H. Moore b , Michael E. Lassman b , Lisa C. Shriver-Lake b , Chris R. Taitt b , Frances S. Ligler b,∗ a
b
George Mason University, 10910 University Blvd, MS 4E3, Manassas, VA 20110, USA Center for Bio/Molecular Science & Engineering, Naval Research Laboratory, Washington, DC 20375, USA Available online 19 August 2005
Abstract Foodborne contaminants come in a variety of sizes ranging from simple chemical compounds to entire bacterial cells. Due to public health concerns, there is a current need in the food industry for a sensitive, specific and rapid method to monitor for the presence of these toxic species, either as a result of natural or deliberate contamination. The Array Biosensor developed at the NRL encompasses these qualities, including the ability to measure multiple analytes simultaneously on a single substrate. In this study, we demonstrate the Array Biosensor’s ability to measure both large pathogens, such as the bacteria Campylobacter jejuni (C. jejuni), and small toxins, including the mycotoxins ochratoxin A, fumonisin B, aflatoxin B1 and deoxynivalenol. Sandwich immunoassays were used to measure C. jejuni in buffer and a number of food matrices, while competitive immunoassays, taking only 15 min, were developed for the simultaneous detection of multiple mycotoxins. The combination of sandwich and competitive immunoassay formats on a single substrate was demonstrated, allowing the simultaneous detection of both large (C. jejuni) and small (aflatoxin B1 ) food contaminants. © 2005 Elsevier B.V. All rights reserved. Keywords: Array Biosensor; Sandwich immunoassay; Competitive immunoassay; Bacteria; Campylobacter jejuni; Mycotoxins; Multi-analyte
1. Introduction Foodborne contaminants can take a number of forms, ranging from whole bacterial cells or proteins to simple chemical compounds [1–4]. Food contamination, whether accidental or deliberate, can occur anywhere along the food processing line, from the source to the consumer, and is a major concern for the food industry. Traditional methods for the detection of food contaminants range from culture in selective media, followed by numerous biochemical and serological tests for bacterial or viral pathogens, to chromatography techniques for the smaller chemical compounds [1,3]. While these gold standard methods are typically very sensitive, they often require multiple time-consuming steps, ∗
Corresponding author. Tel.: +1 202 404 6002; fax: +1 202 404 8897. E-mail address: [email protected] (F.S. Ligler).
0925-4005/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2005.07.008
such as extraction, sample clean-up or preconcentration for toxins or multiple cultures for pathogens, prior to measurement, resulting in analysis times which can run into days. Furthermore, they are designed to detect and measure only one particular target at a time. In an effort to reduce the analysis time while maintaining sensitivity, researchers have developed a variety of immunoassay-based measurements for pathogen and toxin detection, with the enzyme-linked immunosorbent assay (ELISA) format being the most common [1,3]. ELISAs have been used for a variety of food pathogens and analysis times range from 20 min to 2 days depending on the extent of sample pretreatment required. The Array Biosensor developed at the Naval Research Laboratory (NRL) has successfully been used in the detection of a variety of protein toxins, organic molecules, physiological health markers, a virus and a number of bacteria, initially in buffer but increasingly in food, biological and
600
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
environmental matrices [5–16]. The developed immunoassays are rapid (15–25 min), simple to perform and require little-to-no sample pretreatment prior to analysis, even for more complex sample matrices. In addition, the twodimensional nature of the slide sensing surface facilitates simultaneous analysis of multiple samples for multiple analytes. In this study, the Array Biosensor measures both a large pathogen, the bacteria Campylobacter jejuni (C. jejuni), and small toxins, such as the mycotoxins ochratoxin A, fumonisin B, aflatoxin B1 and deoxynivalenol. In an extension of a previous study [13], sandwich immunoassays were used to measure C. jejuni in an extended range of food matrices including milk, yogurt, turkey ham and turkey sausage. Campylobacter, which is found to inhabit the intestinal tracts of a variety of healthy mammals and birds, is the most common cause of intestinal and diarrheal disease in the US [1,2]. Infection typically results from the consumption of unpasteurized milk and milk products and undercooked poultry with symptoms including diarrhea, fever, abdominal and muscle pain, nausea and headache. In stark contrast to the large bacterial cells of Campylobacter, mycotoxins are small metabolites produced by fungi that grow on a number of agricultural products prior to harvest or during storage [3,4]. Mycotoxicosis results either from inhalation exposure or the ingestion of contaminated foodstuffs. The associated health problems, which range from vomiting to cancer, are dependent upon the specific mycotoxin to which an individual is exposed. In this study, competitive immunoassays taking only 15 min were developed for the simultaneous detection of multiple mycotoxins. Finally, the combination of sandwich and competitive immunoassay formats on a single substrate was demonstrated, which allows the simultaneous detection of both large (C. jejuni) and small (aflatoxin B1 ) food contaminants.
2. Experimental 2.1. Materials Unless otherwise specified, chemicals were of reagent grade and used as received. Borosilicate glass slides from Daigger & Co. Inc. (Vernon Hills, IL; www.daigger. com) were used as waveguides in all the assays described. Poly(dimethyl)siloxane (PDMS), used for making the assay flow cells, was obtained from Nusil Silicone Technology (Carpintera, CA; www.nusil.com). The 3-mercaptopropyltrimethoxy silane (MTS) and N-(␥-maleimidobutyryloxy) succinimide ester (GMBS) were purchased from Fluka Chemical Co. (St. Louis, MO; www.sigmaaldrich.com). EZ-link biotin-LC-NHS, biotin-LC-PEO-amine, biotin-PEO-amine, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), N-hydroxysuccinimide (NHS) and NeutrAvidin were obtained from Pierce (Rockford, IL; www.piercenet.com). The biotin-SP-conjugated rabbit anti-chicken IgY and Cy5chicken IgY, used as positive controls, were purchased
from Jackson ImmunoResearch (West Grove, PA; www. jacksonimmuno.com). The bacterial antigen C. jejuni (ATCC35918) used in the sandwich assay was grown as described, by Dr. Avraham Rasooly (FDA) and used under biosafety II conditions, requiring that the cells were killed prior to transportation to the NRL [13]. The polyclonal antibody against C. jejuni, rabbit anti-C. jejuni, was obtained from Biodesign International (Saco, ME; www. biodesign.com). The mycotoxins, ochratoxin A (OTA), fumonisin B (FB), aflatoxin B1 (AFB1 ) and deoxynivalenol (DON), used in the competitive assays were purchased from Sigma (St. Louis, MO; www.sigmaaldrich.com). Rabbit anti-OTA was purchased from ImmuneChem Pharmaceuticals Inc. (Burnaby, BC, Canada; www.immunechem.com). Monoclonal mouse antibodies against FB, AFB1 and DON were kindly supplied by Dr. C. Maragos of the USDA (Peoria, IL). Fluorescent labeling of the antibodies was achieved using Cy5 bisfunctional dye from Amersham Biosciences (Arlington Heights, IL; www.amershambiosciences.com). Bovine serum albumin (BSA), gelatin, potassium hydroxide (KOH), sodium chloride (NaCl), sodium phosphate monoand di-basic, polyoxyethylenesorbitan monolaurate (Tween 20), dimethyl sulfoxide (DMSO), N,N-dimethylformamide (DMF) and toluene were supplied by Sigma–Aldrich (www.sigmaaldrich.com). Absolute ethanol was obtained from Warner-Graham Co. (Cockeysville, MD). All food was purchased from local grocery stores. A Waring two-speed commercial blender, equipped with a mini-sample container (37–110 ml), used for some of the food preparation, and isopropyl alcohol were purchased from Fisher Scientific (Pittsburgh, PA; www.fisherscientific.com). 2.2. Preparation of biotin-labeled mycotoxins 2.2.1. Ochratoxin A-biotin To a test tube containing a small stir bar, 6.3 mg NHS and 10.3 mg EDC were added and the tube closed with a septum. Next, a solution of 10 mg OTA in 0.4 ml DMF was added to the test tube. After stirring for 1 h, a solution of 22.6 mg biotin-LC-PEO-amine in 1 ml 0.05 M carbonate–bicarbonate buffer (pH 9.5) was added and then stirred for 24 h at 4 ◦ C. The reaction mixture was then transferred to a 500 MWCO dialysis bag and dialyzed against several changes of 1 l phosphate buffered saline (PBS). The biotin-OTA product was then stored in a vial at 4 ◦ C [15]. 2.2.2. Deoxynivalenol-biotin In a 20 ml vial, 10 mg DON was dissolved in 2 ml of acetone. Next, 110 mg of 1,1 -carbonyldiimidazole was added and the reaction stirred for 1 h at room temperature. Water, 40–50 l, was then added followed, by a solution of 28.3 mg biotin-PEO-amine in 1 ml of 0.1 M sodium bicarbonate buffer (pH 8.5) and stirred for 24 h at 4 ◦ C. The reaction mixture was then transferred to a 500 MWCO dialysis bag and dialyzed several times against 1 l PBS. The biotin-DON product was then stored in a vial at 4 ◦ C.
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
2.2.3. Aflatoxin B1 -biotin The route to AFB1 -biotin is a two-step process in which an oxime intermediate of the AFB1 is first prepared. Into a 50 ml flask was placed 31 mg of (aminooxy)acetic acid. Next, 20 mg of AFB1 , dissolved in 1:4:1 of pyridine:methanol:water, was added to the flask. The reaction was refluxed for 2 h, cooled and the solvents removed by rotary evaporation. The resulting AFB1 -(O-carboxy methyl) oxime intermediate was characterized by positive ion LC mass spectrometry equipped with a nano-spray source. For the second part of the synthesis, 15 mg NHS and 30 mg EDC were added to a flask containing a stir bar which was then closed with a septum. Next, a solution of 24.6 mg AFB1 -(O-carboxy methyl) oxime in 3 ml of DMF was added to the flask. After 1 h of stirring, a solution of 50 mg of biotin-LC-PEO-amine in 2 ml of a 0.05 M carbonate–bicarbonate buffer (pH 9.5) was added to the flask and then stirred for 24 h at 4 ◦ C. The reaction mixture was then transferred to a 500 MWCO dialysis bag and dialyzed several times against 1 l PBS. The biotin-AFB1 product was then stored in a vial at 4 ◦ C. 2.2.4. Fumonisin-biotin A solution containing 2 mg FB in 0.1 M sodium bicarbonate buffer (pH 8.5) was prepared and biotin-LC-NHS (dissolved in DMSO) was added to give a final molar ratio of 3 biotin:1 FB. The reaction mix was incubated for 1 h at room temperature before being transferred to a 500 MWCO dialysis bag and dialyzed several times against 1 l PBS. The biotin-FB product was then stored in a vial at 4 ◦ C. All of the resulting biotin-mycotoxins were characterized using mass spectroscopy and quantified using UV–vis spectroscopy. 2.3. Preparation of labeled antibodies Biotinylation of the anti-C. jejuni capture antibody was performed using the EZ-link biotin-LC-NHS at a 5:1 molar ratio (biotin:antibody) according to the manufacturer’s instructions. Labeled antibodies were separated from unincorporated biotin using size exclusion chromatography (BioGel P10, Bio-Rad, Hercules, CA; www.biorad.com). Cy5 labeling of the rabbit anti-C. jejuni, rabbit anti-OTA and monoclonal mouse antibodies to the mycotoxins DON, AFB1 and FB were performed according to the manufacturer’s instructions. Labeled antibodies were separated from unincorporated dye using size exclusion chromatography (BioGel P10). Protein-to-dye ratios were determined using UV–vis spectroscopy. 2.4. Patterning biotinylated capture species using PDMS flow cells Microscope slides, used as waveguides, were cleaned by immersion in a 10% (w/v) KOH in 2-propanol for 30 min at room temperature, followed by rinsing with deionized
601
water and drying with a nitrogen stream. The slides were then immediately immersed in a toluene solution containing 2% MTS for 1 h under nitrogen. The silanized slides were then rinsed with toluene, dried with nitrogen and immersed in 1 mM GMBS, prepared in absolute ethanol, for 30 min at room temperature. The slides were then rinsed with water, incubated in 25 g/ml NeutrAvidin in PBS overnight at 4 ◦ C, before being washed in PBS and either used for patterning or stored in PBS at 4 ◦ C until required. Patterning of the biotinylated capture species was carried out as previously described [5]. Briefly, a 6- or 12-channel patterning PDMS flow cell was clamped onto the NeutrAvidin functionalized waveguide surface. Biotin-conjugated capture species, either anti-C. jejuni (10 g/ml) for the sandwich assay or mycotoxin OTA (10 g/ml), DON (10 g/ml), AFB1 (5 g/ml) and FB (5 g/ml) all in PBS + 0.05% Tween (PBST) for the competitive assays, were introduced into the channels of the flow cell. In the case of the sandwich assays, biotinylated antichicken IgY (10 g/ml in PBST) was introduced into one or two of the channels for use as a positive control. The slides were then incubated overnight at 4 ◦ C. The channels of the flow cell were then rinsed with 1 ml PBST, the slide removed from the PDMS template and placed in PBST blocking solution containing either 1% BSA or 1% gelatin. After 1–2 h, the slides were rinsed with Milli-Q water and assembled in a 6- or 12-channel assay PDMS flow cell. 2.5. Food sample preparation for C. jejuni sandwich assays In a previous study, dose–response curves were obtained for Campylobacter in buffer and a number of spiked food samples [13]. In all cases, bacterial cells were spiked into an initial sample and a 50% serial dilution was carried out with unspiked sample matrix to produce samples for analysis. The spiked samples were typically allowed to stand at room temperature for 2 h prior to analysis. The 2 h incubation period was chosen following dose–response versus incubation time studies carried out previously by Shriver-Lake et al. [12]. For the ground turkey sausage and ham, 10 g of meat and 8 ml of PBST containing 1 mg/ml bovine serum albumin (PBSTB) were placed into a 37–110 ml mini-sample container on a Waring Blender and blended on high for 2 min. The resulting mixture was put into a 50 ml centrifuge tube. The blender container was then rinsed with 2 ml PBSTB and the rinse solution added to the centrifuge tube to yield a 1:1 (w/v) mixture of meat and buffer. The bacterial cells were then spiked into this mixture (assuming total volume of 10 ml) and the sample left at room temperature for 2 h. The mixture was centrifuged at 4000 rpm for 10 min, the supernatant collected and then serially diluted with unspiked supernatant prior to analysis. Carnation nonfat dried milk was prepared per instruction on the box. In a 50 ml centrifuge tube, 45 ml of milk and 5 ml 10× PBSTB was mixed and the sample was then spiked with the bacterial cells for analysis. Serial dilutions were carried
602
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
out using unspiked milk. Vanilla fat free yogurt (100 ml) was diluted 50% with 100 ml PBSTB; the sample was then spiked with the bacterial cells, and serial dilutions were carried out using unspiked yogurt. 2.6. Sandwich and competitive assays All assays were performed using 6- or 12-channel assay PDMS flow cells, whose channels are perpendicular to the stripes of capture species patterned on the waveguides [17]. A sandwich assay was performed for the detection C. jejuni. The antibody functionalized slide was assembled with the assay PDMS flow cell and hooked up to an ISMATEC® multichannel pump (Cole-Parmer Instruments Company, Vernon Hills, IL; www.coleparmer.com) with tubing connected to one end of each channel in the PDMS (outlet). Syringe barrels (1 ml) were then attached at the opposite end of each channel (inlet). The PDMS channels were first washed with 1 ml of PBSTB at 0.5 ml/min, to check for any leaks in the PDMS flow cell. Next, 0.8 ml of sample was applied to each channel at a flow rate of 0.15 ml/min for 15 min; the sample was recycled over the surface by connecting the inlet and outlet of each channel. The sample was then removed and the channels washed with 1 ml of PBSTB at 0.5 ml/min. The fluorescently labeled tracer mixture (0.4 ml:100 ng/ml Cy5-Chicken IgY (positive control) and 10 g/ml Cy5-anti-C. jejuni in PBSTB) was applied to the channels at 0.15 ml/min. The channels were washed with 1 ml of PBSTB, before the assay PDMS flow cell was removed. Finally, the slide was washed with Milli-Q water, dried with nitrogen and imaged on the Array Biosensor. Competitive assays were performed for the simultaneous detection of the mycotoxins OTA, FB, AFB1 and DON. Here, NeutrAvidin slides functionalized with biotinylated analogs of the mycotoxins were assembled with the assay PDMS flow cell and hooked up to the pump as described above for the sandwich assay. The PDMS channels were first washed with 1 ml of PBSTB, at 0.5 ml/min, then 0.8 ml of sample was applied to each channel at a flow rate of 0.15 ml/min; the sample was not recycled over the surface. In the case of the competitive assay, the sample contained both the “free” mycotoxin to be measured and the fluorescent antibody specific to that particular mycotoxin; therefore, the assay is a one-step procedure, rather than the two-step sandwich assay. After the sample had passed over the surface the channels were washed with 1 ml of PBSTB at 0.5 ml/min. Then, the PDMS flow cell was removed and the slide was washed with Milli-Q water, dried with nitrogen and imaged on the Array Biosensor. The possibility of running both competitive and sandwich assay formats simultaneously on the same waveguide was also investigated. In this case, the waveguides were functionalized with capture antibodies to C. jejuni and the biotin analog of AFB1 . The slides were assembled with the assay PDMS flow cell and hooked up to the pump. The PDMS channels were first washed with 1 ml of PBSTB at 0.5 ml/min,
then 0.8 ml of sample (in PBSTB) was applied to each channel at a flow rate of 0.15 ml/min; sample was not recycled over the surface. In this case, the sample contained C. jejuni bacteria, “free” AFB1 and Cy5-anti-AFB1 antibody. After the sample had passed over the surface, the channels were washed with 1 ml of PBSTB at 0.5 ml/min, and the fluorescently labeled tracer (0.4 ml; 10 g/ml Cy5-anti-C. jejuni in PBSTB) was applied to the channels at 0.15 ml/min. The channels were washed with 1 ml of PBSTB, the assay PDMS flow cell removed and the slide washed with Milli-Q water, dried with nitrogen and imaged on the Array Biosensor. 2.7. Slide imaging and data analysis The slide was imaged using a Peltier-cooled CCD, which has previously been described [17]. Briefly, evanescent wave excitation of the surface-bound fluorescent species was achieved using a 635 nm, 12 mW diode laser (Lasermax, Rochester, NY; www.oemlaser.com). Light was launched through a 1 cm focal length lens with a line generator into the end of the waveguide at an appropriate angle. The fluorescence emission was monitored at right angles to the planar surface of the waveguide. A two-dimensional graded index of refraction (GRIN) lens array (Nippon Sheetglass, Summerset, NJ) was used to image the fluorescent pattern onto the Peltier-cooled CCD camera (Spectra Source, Teleris, Westlake Village, CA). Long-pass (Schott 0G-0665, Schott Glass, Duryea, PA; www.schott.com) and band-pass filters (Corion S40-670-S, Franklin, MA; www.corion.com) were mounted on the device scaffolding to eliminate excitation and scattered light prior to CCD imaging. Data were acquired in the form of digital image files in Flexible Image Transport System (FITS) format. To analyze the images, a custom software application was written in LabWindows/CVI (National Instruments; www.ni.com). The program creates a mask consisting of data squares (enclosing the areas where the capture antibody is patterned) and background rectangles which are located on either side of each data square. The average background value is subtracted from the average data square value and net intensity value is calculated and imported into a Microsoft Excel file for data analysis. Limits of detection (LODs) were calculated as the lowest tested concentrations at which the fluorescence signal was at least three standard deviations above the mean fluorescence of the buffer blank.
3. Results and discussion The NRL Array Biosensor has been used in the rapid (25 min or less) identification and quantification of a variety of different target species in a number of complex matrices with little-to-no sample pretreatment or target concentration [5–16]. Typically, detection of the target species involves the use of the sandwich immunoassay format which is appropriate for the larger protein and bacterial pathogens. In an
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
603
Table 1 C. jejuni detection in food and beverage samples
Fig. 1. Sandwich immunoassays. (a) The final CCD image after the slide was exposed to C. jejuni in yogurt, 0–3 × 104 cfu/ml, and a 3 × 104 cfu/ml C. jejuni in PBSTB positive control (PC) and a PBSTB negative control (NC). (b) The resulting dose–response curves for C. jejuni in various sample matrices plotted as normalized net intensity versus bacterial cell concentration; PBSTB (closed circles), yogurt (open circles), milk (closed square), ground turkey sausage (open squares) and ground turkey ham (closed triangles). The net intensity values are an average of 10 or more data squares ± S.D. with a minimum of two separate slides.
extension of a previous study, we have expanded the number of different foods in which we demonstrated detection of C. jejuni. A typical CCD image, taken with the NRL Array Biosensor, is shown in Fig. 1a for the detection of C. jejuni bacteria in yogurt. The CCD image was then converted into intensity values and the resulting normalized dose–response curves are shown in Fig. 1b. A positive control in PBSTB was run as part of the assay and the net intensities for the different food matrices are normalized using this positive control. This allows one to account for slide-to-slide variations in the sandwich assay, independently of variations related to the use of different sample matrices. The dose–response curves and the resulting limits of detection, listed in Table 1, were consistent with previous observations [13] and were dependent on the sample matrix. C. jejuni could be detected at concentrations
Sample matrix
Limit of detection (cfu/ml)
PBSTB Yogurt Milk Ground turkey sausage Ground turkey ham
938 1880 469 469 3750
as low as 4.7 × 102 cfu/ml in milk and ground turkey sausage, but the LOD was found to increase to 3.8 × 103 cfu/ml in ground turkey ham. The higher LODs in certain food matrices may result from (1) an increase in viscosity of the sample with concomitant decrease in target transport to the surface, (2) an interaction of the bacterial cells with the food matrix, making the cells unavailable to interact with the capture antibody or (3) interaction of the food matrix with the capture antibody decreasing interaction with the bacterial cells. Due to the observed variation in the dose–response curves, it is important that a dose–response curve in the matrix of choice be included as part of the assay design if quantification of the analyte is required. While the sandwich immunoassay format works extremely well for species large enough to contain separate recognition sites for both the capture and tracer antibodies, there are a number of foodborne contaminants too small to meet such criteria. In this situation, detection of the contaminant can be achieved using either a competitive or displacement immunoassay. These different immunoassay formats have been demonstrated previously for the detection of TNT, using the Array Biosensor [10]. Mycotoxins are examples of toxic organic compounds often found in agricultural products which are not destroyed by food processing procedures [3,4]. We have recently developed a competitive immunoassay for detection of the mycotoxin ochratoxin A which uses an immobilized biotin analog of OTA, to compete with free OTA present in a solution containing fluorescently labeled anti-OTA antibody [16]. Unlike the sandwich immunoassay, the competitive immunoassay format produces a decreasing fluorescence signal with increasing concentration of the mycotoxin free in solution. One of the advantages of using the planar waveguide system coupled with the PDMS flow cells is the ability to test multiple samples simultaneously for multiple target analytes. Therefore, we wanted to investigate the ability of the NRL Array Biosensor system to carry out simultaneous, multiplexed, competitive immunoassay-based detection of different mycotoxins and also to determine the feasibility of running sandwich and competitive immunoassays formats on a single waveguide surface. To address whether multiple competitive-based assays could be performed on a single substrate, NeutrAvidin slides were patterned with biotinylated analogs of the mycotoxins OTA, AFB1 , FB and DON; a biotinylated anti-chicken IgY lane was also added as a positive control. The slides were then exposed to an antibody mix (Ab-mix), made up of 100 ng/ml Cy5-chicken IgY,
604
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
10 g/ml Cy5-anti-OTA, 20 g/ml Cy5-anti-DON, 2 g/ml Cy5-anti-AFB1 and 0.25 g/ml Cy5-anti-FB in PBSTB. This mix also contained various combinations of the mycotoxins OTA, AFB1 , FB and DON at concentrations known to compete with the immobilized mycotoxins for binding sites on the antibodies. The resulting CCD image is shown in Fig. 2a, with the predicted slide image illustrated in Fig. 2b. Clearly, when no free mycotoxins are present in the assay solution the antibodies bind to the immobilized toxin and generate signals from all mycotoxin spots (lowest lane). However, as mycotoxins are introduced into the Ab-mix solution, a resultant decrease in signal intensity is observed only in the region of the slide functionalized with the biotinylated equivalent of the particular mycotoxin introduced. The captured CCD image (Fig. 2a) is in good agreement with the predicted image in Fig. 2b, demonstrating the specificity of these competitive-based immunoassay formats. While there was found to be some non-specific binding of the labeled antibodies to the NeutrAvidin, as observed by the signal intensity found both in the PBS column and in background regions between data squares, see Ab-mix lane (Fig. 2a), the intensity is lower than the data regions and was easily subtracted during the image analysis. This situation was found for slides blocked either in 1% gelatin or 1% BSA solutions prior to the assay and may have implications for the sensitivity of the assay. The ability to run both sandwich and competitive immunoassays formats on a single waveguide surface is advantageous as it allows the user to monitor for both large and small molecular weight food contaminants simultaneously. C. jejuni and AFB1 were chosen, representing a large bacterial pathogen and a small organic toxin, respectively, to determine if these two immunoassay formats could be run on a single waveguide. Here, the NeutrAvidin waveguide was
assembled in a six-column patterning PDMS flow cell with three columns exposed to biotinylated Rb-anti-C. jejuni and the remaining columns to biotinylated AFB1 . The standard sandwich immunoassay format is a two-step process where the functionalized slide surface is exposed first to the analyte solution for 8–15 min followed by exposure to the antibody tracer for 4 min. The competitive format, on the other hand, is a one-step process where the slide surface is exposed to a mixture of the antibody tracer and free toxin simultaneously for 8 min. When combining immunoassay formats, the C. jejuni and the competitive assay tracer/toxin (AFB1 and Cy5-anti-AFB1 ) mix were combined and incubated together with the waveguide for 8 min. This was followed by incubation of the sandwich assay antibody tracer (Cy5-anti-C. jejuni) for 4 min. Fig. 3a shows the final CCD image of a slide exposed to the sandwich immunoassay for C. jejuni and the competitive immunoassay for AFB1 , run either individually or combined. The top two assays lanes represent the sandwich immunoassay lanes only, with and without the presence of the bacteria. Clearly, when no C. jejuni is present the tracer antibody, Cy5-anti-C. jejuni, does not bind to the surface; however, when bacteria are introduced, a signal is generated only in the Campylobacter-specific region of the slide functionalized and not in the biotin-AFB1 region. Likewise, in the middle two assay lanes, which represent the competitive immunoassay only lanes, signal is generated by the Cy5-anti-AFB1 antibody in the absence of free AFB1 only in the biotin-AFB1 functionalized regions (lane 3). When AFB1 is introduced into the competitive reaction mix (lane 4), the signal is completely lost at the AFB1 data squares due to the free solution toxin saturating the antibody binding sites, therefore, preventing binding of the tracer antibody to the surface immobilized toxin. The bottom two assay lanes
Fig. 2. Multiplexed mycotoxin competitive assays. (a) Slide was patterned with the biotinylated anti-chicken IgY, OTA, DON, AFB1 and FB. The slide was assayed with the mycotoxin analyte(s) in the presence of the antibody mix (Ab-mix). The following concentrations of mycotoxins were assayed: 10 g/ml OTA, 10 g/ml DON, 50 ng/ml AFB1 and 20 g/ml FB. (b) Predicted slide image from the toxin combinations exposed; white squares represent regions that should produce a signal and black squares where low-to-no signal is expected.
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
605
Fig. 3. Sandwich and competitive immunoassays on the same sensing substrate. (a) The final CCD image after the slide was simultaneously exposed to the C. jejuni (5 × 104 cfu/ml) sandwich assay and the AFB1 (1 ng/ml) competitive assay in various combinations. (b) Dose–response curves are shown for C. jejuni presence (closed circles) and absence (open circles) of the AFB1 competitive assay plotted as normalized net intensity versus bacterial cell concentration. (c) Dose–response curves are shown for AFB1 in the presence (closed circles) and absence (open circles) of the C. jejuni sandwich immunoassay plotted as inhibition (%) versus free solution AFB1 concentration.
demonstrate the signals obtained when the sandwich and competitive assays are run together with target present. Here, C. jejuni is present in both lanes, therefore, the sandwich assay, in the anti-C. jejuni functionalized regions, generates signal in both assay lanes. Free AFB1 , on the other hand, is only present in the final assay lane and results in the expected loss of signal in the biotin-AFB1 functionalized regions. Fig. 3b and c show the dose–response curves for the C. jejuni sandwich assay and the AFB1 competitive assay, respectively, obtained either in the presence (closed circles) or absence (open circles) of the other assay format. In both cases, the LODs were actually found to improve (decrease) slightly in the presence of the other assay. C. jejuni sandwich assay alone had an LOD of 1000 cfu/ml [this study, 13] which was found to decrease to 500 cfu/ml in the in presence of the AFB1 competitive assay. Likewise, the AFB1 competitive assay alone had an LOD of 0.9 ng/ml, but was found to decrease to 0.3 ng/ml in the presence of the C. jejuni sandwich assay. These results clearly demonstrate that the sandwich and competitive immunoassay formats can successfully be
run on the same substrate for the detection of both large and small food contaminants simultaneously.
4. Conclusions This study has clearly demonstrated the versatility of the Array Biosensor for the detection of both large and small food contaminants either individually or simultaneously. The bacterial pathogen C. jejuni was measured using a sandwich immunoassay both in buffer and additional complex food matrices with LODs ranging from 500 to 3780 cfu/ml. The mycotoxins OTA, DON, AFB1 and FB were detected simultaneously on a signal substrate using a competitive-based immunoassay format taking only 15 min. The combination of sandwich and competitive immunoassay formats on a single substrate was demonstrated, allowing the simultaneous detection of both large (C. jejuni) and small (AFB1 ) food pathogens with LODs in buffer of 500 cfu/ml and 0.3 ng/ml, respectively. In the case of C. jejuni, as little as 500 cells are known to case human illness and therefore our assay
606
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607
LOD currently stands around this desired threshold [18]. On the other hand, the AFB1 LOD, 0.3 ng/ml for the combined assay, is below the current maximum levels set at 2 ng/g (2 ng/ml) [19]. Future studies will include testing the competitive and sandwich/competitive combined immunoassay formats in food samples to assess if assay performance will be affected.
Acknowledgements The authors would like to thank Dr. A. Rasooly (FDA) and Dr. C. Maragos (USDA) for their contribution of materials. This work was supported by funding from the United States Department of Agriculture (USDA) and the Food and Drug Administration (FDA). The views expressed are those of the authors and do not represent those of the US Navy, US Department of Defense or the U.S. Government.
References [1] E.C. Alocilja, S.M. Radke, Market analysis of biosensors for food safety, Biosens. Bioelectron. 18 (2003) 841–846. [2] Y. Mao, C. Zhu, E.C. Boedeker, Foodborne enteric infections, Curr. Opin. Gastroenterol. 19 (2003) 11–22. [3] Council for Agricultural Science and Technology, Mycotoxins: risks in plants, animals and human systems, Report No. 139, Council for Agricultural Science and Technology, Ames, IA, January 2003. [4] B.B. Jarvis, J.D. Miller, Mycotoxins as harmful indoor air contaminants, Appl. Microbiol. Biotechnol. 66 (2005) 367–372. [5] C.A. Rowe, L.M. Tender, M.J. Feldstein, J.P. Golden, S.B. Scruggs, B.D. MacCraith, J.J. Cras, F.S. Ligler, Array Biosensor: simultaneous analysis for bacterial, viral, and toxic analytes in blind samples, Anal. Chem. 71 (1999) 3846–3852. [6] C.A. Rowe, S.D. Balderson, M.J. Feldstein, J.P. Golden, F.S. Ligler, An array immunosensor for simultaneous detection of clinical analytes, Anal. Chem. 71 (1999) 433–439. [7] C.A. Rowe-Taitt, J.J. Cras, C.H. Patterson, J.P. Golden, F.S. Ligler, A ganglioside-based assay for cholera toxin using an Array Biosensor, Anal. Biochem. 281 (2000) 123–133. [8] C.A. Rowe-Taitt, J.P. Golden, M.J. Feldstein, J.J. Cras, K.E. Hoffman, F.S. Ligler, Array Biosensor for detection of biohazards, Biosens. Bioelectron. 14 (2000) 785–794. [9] C.A. Rowe-Taitt, J.W. Hazzard, K.E. Hoffman, J.J. Cras, J.P. Golden, F.S. Ligler, Simultaneous detection of six biohazardous agents using a planar waveguide Array Biosensor, Biosens. Bioelectron. 15 (2000) 579–589. [10] K.E. Sapsford, P.T. Charles, C.H. Patterson Jr., F.S. Ligler, Demonstration of four immunoassay formats using the Array Biosensor, Anal. Chem. 74 (2002) 1061–1068. [11] C.R. Taitt, G.P. Anderson, B.D. Lingerfelt, F.S. Ligler, Nineanalyte detection using an Array Biosensor, Anal. Chem. 74 (2002) 6114–6120. [12] L.C. Shriver-Lake, Y.S. Shubin, F.S. Ligler, Detection of staphylococcal enterotoxin B in spiked food samples, J. Food Prot. 66 (2003) 1851–1856. [13] K.E. Sapsford, A. Rasooly, C.R. Taitt, F.S. Ligler, Detection of Campylobacter and Shigella species in food samples using the Array Biosensor, Anal. Chem. 76 (2004) 433–440. [14] C.R. Taitt, Y.S. Shubin, R. Angel, F.S. Ligler, Detection of Salmonella enterica serovar Typhimurium by using a rapid, arraybased immunosensor, Appl. Environ. Microbiol. 70 (2004) 152–158.
[15] L.C. Shriver-Lake, C.R. Taitt, F.S. Ligler, Applications of Array Biosensor for detection of food allergens, J. AOAC Int. 87 (2004) 1498–1502. [16] M.M. Ngundi, L.C. Shriver-Lake, M.H. Moore, M.E. Lassman, F.S. Ligler, C.R. Taitt, Array Biosensor for detection of ochratoxin A in cereals and beverages, Anal. Chem. 77 (2005) 148–154. [17] M.J. Feldstein, J.P. Golden, C.A. Rowe, B.D. MacCraith, F.S. Ligler, Array Biosensor: optical and fluidics systems, J. Biomed. Microdevices 1 (1999) 139–153. [18] M.H. Kothary, U.S. Babu, Infective dose of foodborne pathogens in volunteers: a review, J. Food Saf. 21 (2001) 49–73. [19] J. Stroka, E. Anklam, New strategies for the screening and determination of aflatoxins and the detection of aflatoxin-producing moulds in food and feed, Trends Anal. Chem. 21 (2002) 90– 95.
Biographies Kim E. Sapsford has been a George Mason University contractor at the Center for Bio/Molecular Science and Engineering at the Naval Research Laboratory in Washington, DC, since 2001. She earned her BSc (Chemistry) and PhD (Biosensors) degrees from the University of East Anglia in the UK. Her research at NRL has been primarily in the field of optical biosensors. Miriam M. Ngundi received her PhD from the State University of New York (Binghamton University) in 2003. Her dissertation work was on the rational design of chemical and biochemical sensors. She is currently a postdoctoral fellow at the Center for Bio/Molecular Science & Engineering at the Naval Research Laboratory. Her current area of research is array biosensors. Martin H. Moore received his BS in education and MS in chemistry from Youngstown State University. In 1998, he accepted a position as a research scientist for Geo-Center’s Inc. to do research at the Naval Research Laboratory (NRL) in Washington, DC, where he became a federal employee in 2001 and is currently employed. Martin Moore has been extensively involved in the development and synthesis of a variety of compounds and materials for use in a number of applications. Michael E. Lassman served as an American Society for Engineering Education Postdoctoral Fellow from 2003 to 2005 at the Naval Research Laboratory’s Center for Bio/Molecular Science and Engineering. He earned his PhD in analytical chemistry from the University of Delaware in 2003 and is now a Senior Research Biochemist at Merck and Co. His current research initiatives focus on using mass spectrometry to solve biological problems. Lisa C. Shriver-Lake received her BS in animal science degree in 1979 from the University of Maryland at College Park and a MS in chemistry from George Mason University, Fairfax, VA, in 1991. In 1985, she accepted a position as a research scientist for Geo-Center’s Inc. to do research at the Naval Research Laboratory (NRL) in Washington, DC, where she became a federal employee in 1987. During the last 19 years, Ms. Shriver-Lake has developed several antibody-based biosensors and applied them for biological warfare detection, environmental applications and food safety. Chris R. Taitt earned her AB in biology from Dartmouth College in 1984, her MS in pomology from Cornell University in 1986 and her PhD in biology from the Johns Hopkins University in 1995. She has been a member of the Center for Bio/Molecular Science and Engineering at the Naval Research Laboratory since 1995. Her research at NRL has been primarily in the field of optical biosensors and has involved biochemical assay development, as well as transitioning of sensors to field use. Her interest in basic biochemistry has influenced her most recent work; she currently leads several projects developing novel recognition
K.E. Sapsford et al. / Sensors and Actuators B 113 (2006) 599–607 schemes, rapid detection methodologies and amplification methods for rapid detection purposes. Frances S. Ligler is the US Navy’s Senior Scientist for Biosensors and Biomaterials. She earned a BS from Furman University (1971) and both
607
DPhil (1977) and DSc (2000) degrees from Oxford University. She is a member of the National Academy of Engineering and a Presidential Rank Award winner. She has over 220 full-length publications and 24 patents. Her current R&D initiatives focus on optical biosensors, microfluidics and proteomics.