Anal. Chem. 2007, 79, 140-146
Combination of Immunosensor Detection with Viability Testing and Confirmation Using the Polymerase Chain Reaction and Culture Brandy Johnson-White, Baochuan Lin, and Frances S. Ligler*
Center for Bio/Molecular Science & Engineering, Naval Research Laboratory, Washington, D.C. 20375-0001
Rapid and accurate differential determination of viable versus nonviable microbes is critical for formulation of an appropriate response after pathogen detection. Sensors for rapid bacterial identification can be used for applications ranging from environmental monitoring and homeland defense to food process monitoring, but few provide viability information. This study combines the rapid screening capability of the array biosensor using an immunoassay format with methods for determination of viability. Additionally, cells captured by the immobilized antibodies can be cultured following fluorescence imaging to further confirm viability and for cell population expansion for further characterization, e.g., strain identification or antibiotic susceptibility testing. Finally, we demonstrate analysis of captured bacteria using the polymerase chain reaction (PCR). PCR results for waveguide-captured cells were 3 orders of magnitude more sensitive than the fluorescence immunoassay and can also provide additional genetic information on the captured microbes. These approaches can be used to rapidly detect and distinguish viable versus nonviable and pathogenic versus nonpathogenic captured organisms, provide culture materials for further analysis on a shorter time scale, and assess the efficacy of decontamination or sterilization procedures. Several techniques are available for the identification and quantification of microorganisms for applications ranging from environmental monitoring and homeland defense to food process monitoring. A major challenge for developing microbial diagnostic methods is distinguishing viable and nonviable microbes.1 This capability is essential for determination of a threat as well as development of an appropriate threat response. Traditional microbial techniques employing selective agars or selective motility often require greater than 24 h for identification of microorganisms and nonviable bacteria are not identified. Polymerase chain reaction (PCR)-based methods are more rapid and versatile (completed in hours), but are limited by the number of targets that can be distinguished, and are not suitable for determining the viability of identified microbes.2 * Corresponding author. E-mail:
[email protected]. Phone: 202-4046002. Fax: 202-404-8897. (1) Nocker, A.; Camper, A. K. Appl. Environ. Microbiol. 2006, 72, 1997-2004. (2) Yang, S.; Rothman, R. E. Lancet Infect. Dis. 2004, 4, 337-348.
140 Analytical Chemistry, Vol. 79, No. 1, January 1, 2007
Immunosensors provide rapid screening methods for microbial detection (completed in minutes). Like other immunoassays, however, they detect viable and nonviable bacteria, fragments, and intact cells indiscriminately, making it impossible to formulate a response based solely on a positive indication. The presence of nonviable cells is inevitable for samples collected following food or water processing or other decontamination procedures. Simple detection methods without differentiation between viable and nonviable microorganisms can lead to false alarms with potentially costly outcomes. Viability determination must also be included in order to devise an appropriate threat response. The Naval Research Laboratory (NRL) Array Biosensor employing immunoassays for microorganism detection has been extensively described elsewhere.3-7 The sensor uses antibodies immobilized in a patterned array onto the surface of a planar waveguide (microscope slide) and can be used to interrogate multiple samples rapidly for multiple analytes including bacteria and proteins simultaneously in a variety of matrixes.3-7 Assays can be completed in ∼15 min. This system has demonstrated potential utility for both surveillance and point-of-care diagnostics. As with other immunoassay-based methods, lack of viability confirmation remains a primary obstacle limiting the application of this technique for obtaining decision-quality information for realtime food and water monitoring. Culture, PCR analysis, and viability determination of antibody captured cells from fiber optics has been demonstrated previously.8-10 The surface area covered by the capture antibody on the NRL array sensor waveguide is significantly less than that covered by the capture antibody of the fiber-optic biosensor. The format of the array sensor waveguide, however, has the added benefit that multiple targets can be captured simultaneously and that solid media can be used for enrichment. (3) Taitt, C. R.; Golden, J. P.; Shubin, Y. S.; Shriver-Lake, L. C.; Sapsford, K. E.; Rasooly, A.; Ligler, F. S. Microb. Ecol. 2004, 47, 175-185. (4) Shriver-Lake, L. C.; Breslin, K. A.; Charles, P. T.; Conrad, D. W.; Golden, J. P.; Ligler, F. S. Anal. Chem. 1995, 67, 2431-2435. (5) Kulagina, N. V.; Lassman, M. E.; Ligler, F. S.; Taitt, C. R. Anal. Chem. 2005, 77, 6504-6508. (6) Rowe, C. A.; Tender, L. M.; Feldstein, M. J.; Golden, J. P.; Scruggs, S. B.; MacCraith, B. D.; Cras, J. J.; Ligler, F. S. Anal. Chem. 1999, 71, 38463852. (7) Ngundi, M. M.; Shriver-Lake, L. C.; Moore, M. H.; Lassman, M. E.; Ligler, F. S.; Taitt, C. R. Anal. Chem. 2005, 77, 148-154. (8) Zhao, W. T.; Yao, S. J.; Hsing, I. M. Biosens. Bioelectron. 2006, 21, 11631170. (9) Simpson, J. M.; Lim, D. V. Biosens. Bioelectron. 2005, 21, 881-887. (10) Tims, T. B.; Lim, D. V. J. Microbiol. Methods 2003, 55, 141-147. 10.1021/ac061229l Not subject to U.S. Copyright. Publ. 2007 Am. Chem. Soc.
Published on Web 12/02/2006
The research presented here combines the rapid screening capability of an immunosensor with the capacity for viability testing. Additionally, we demonstrate that cells captured by immobilized antibodies can be cultured to provide confirmation of viability in a shorter time frame than traditional methods and to provide samples for further forensics investigations. Finally, we demonstrate the potential for genetic analysis of captured bacteria. Genetic confirmation of identity using the PCR methods showed sensitivity up to 3 orders of magnitude greater than that of the fluorescence immunoassay and can also provide further information about the captured microbes. The results presented demonstrate that this approach can be used to rapidly detect and distinguish viable versus nonviable and pathogenic versus nonpathogenic organisms, provide culture materials for further analysis, and assess the effect of decontamination. MATERIALS AND METHODS Materials. Dibasic and monobasic sodium phosphate, phosphate-buffered saline (PBS) powder packs (rehydrated to yield 10 mM PBS pH 7.4), phosphate-buffered saline with Tween-20 (PBST) powder packs (rehydrated to yield 10 mM PBS pH 7.4 with 0.05% Tween-20), low biotin bovine serum albumin (BSA), Tween-20, and HEPES were obtained from Sigma-Aldrich (St. Louis, MO). Escherichia coli ATCC 35218 and Bacillus subtilis (formerly Bacillus globigii) ATCC 49760 were obtained from and propagated as directed by American Type Culture Collection (Manassas, VA). Rabbit polyclonal antibody to E. coli was obtained from Abcam Inc. (Cambridge, MA). Rabbit and goat polyclonal antibodies to B. globigii were gifts from Naval Medical Research Center (NMRC, Silver Springs, MD). Biotin-SP-conjugated rabbit antibody to chicken IgY (IgG) and Cy5-conjugated chicken IgY (IgG) were obtained from Jackson ImmunoResearch (West Grove, PA). Capture antibodies were biotinylated using an excess of biotinLC-NHS ester (Pierce Chemicals, Rockford, IL) as described previously.11 The tracer antibodies were fluorescently labeled with Cy5 (GE HealthSciences, Piscataway, NJ) as directed except that 3 mg of protein was labeled with the amount of the dye intended for 1 mg. Biotinylated or fluorescently labeled protein was separated from excess dye or biotin by gel chromatography using a Bio-Gel P-10 (Medium) column (BioRad; Hercules, CA). Immunoarray Biosensor. Immobilization of NeutrAvidin biotin-binding protein onto waveguide surfaces has been described elsewhere.12 Glass microscope slides (Daigger, Wheeling, IL), were cleaned by immersion in 10% potassium hydroxide/methanol solution for 30 min at room temperature.13 After rinsing and drying, the clean waveguides were incubated in a nitrogen glove bag for 1 h in a 2% 3-mercaptopropyltriethoxysilane toluene solution (Pierce Chemicals), after which they were rinsed in toluene and dried. Maleimidobutyryloxysuccinimide ester (GMBS) was applied to the surface of the slides as a cross-linker by incubating in waveguides in a 2.1 mM GMBS solution in ethanol for 30 min. After rinsing in deionized water, the slides were (11) Johnson-White, B.; Buquo, L.; Zeinali, M.; Ligler, F. S. Anal. Chem. 2006, 78, 853-857. (12) Rhodehamel, E. J.; Harmon, S. M. In Bacteriological Analytical Manual Online, U.S. Food and Drug Administration; U.S. Food and Drug Administration, 2001. (13) Nutrition, C. f. F. S. A., 2001; Vol. 2004.
Figure 1. Schematic of the Naval Research Laboratory Array Biosensor. Capture molecules are patterned in rows oriented along the short axis of the waveguide. Samples are flowed perpendicularly to the patterned capture molecule array. Interrogation is accomplished through excitation of tracer fluorescence using a diode laser and an image is collected using a CCD camera.
incubated overnight at 4 °C in 30 µg/mL NeutrAvidin (Pierce Chemicals) in PBS, rinsed with either PBS or HEPES (10 mM pH 7.4), and stored in the same at 4 °C until use. Six-channel poly(dimethylsiloxane) (PDMS) patterning gaskets, which form a fluid tight seal with the glass waveguide surfaces, were used for immobilization of biotinylated capture antibodies. The flow channels of the patterning gaskets were oriented along the short axis of the slide (Figure 1). Biotinylated capture antibody in PBST or HEPES-T (HEPES with 0.05% Tween20) was injected into the channel and incubated for 2 h at room temperature. The E. coli antibody was used at a concentration of 50 µg/mL while the antibodies against B. globigii and rabbit antibody to chicken IgY (IgG) were used at 10 µg/mL. Antibody solutions were flushed from the patterning template using PBST or HEPES-T and slides were blocked with BSA (10 mg/mL in sodium phosphate buffer or HEPES) for 30 min at room temperature. Blocked slides were stored dry at 4 °C until use. For assaying, a PDMS gasket was applied to the waveguide surface with flow channels oriented perpendicularly to those of the patterning gasket, allowing each of the patterned “rows” to be exposed to each of six sample “lanes” (Figure 1). Samples in PBSTC or HEPES-TC (PBST or HEPES-T with 25% cranberry juice cocktail11) were introduced to each lane in 0.8 mL at a flow rate of 0.1 mL/min (total time 8 min). Tracer solutions of 0.3 mL containing 40 µg/mL Cy5-labeled rabbit antibody to E. coli, 10 µg/mL each of the Cy5-labeled antibodies against B. globigii, and 50 ng/mL chicken IgY (IgG) in PBST or HEPES-T were flowed across the waveguide surface at 0.06 mL/min (total time 5 min). Imaging of the waveguides was accomplished with a CCD camera and using 635-nm laser excitation of the Cy5 labels via total internal reflectance (Figure 1).3 Data analysis was accomplished using automated software described elsewhere.6 For viability determination using a Live/Dead BacLight Bacterial Viability Kit (Molecular Probes, Inc., Eugene, OR), the use of PBS and PBST was found to cause binding of the dyes to unexpected areas of the waveguide surface. Replacing all phosphate buffers with HEPES buffer following NeutrAvidin immobilization drastically reduced nonspecific staining of the surface. PBST was replaced with 10 mM HEPES buffer at pH 7.4 containing 0.05% Tween-20 (HEPES-T). A standard BacLight solution was prepared according to the manufacturer’s directions using a 1:1 ratio of SYTO 9 and propidium iodide in DMSO. This solution was used Analytical Chemistry, Vol. 79, No. 1, January 1, 2007
141
at 3 µL/mL buffer. Staining was accomplished by injecting BacLight dye in HEPES with 1 mg/mL BSA into the channels of the PDMS gasket after the steps for the assay had been completed. The solution was incubated for 15 min, the lanes were flushed with HEPES, and images of the waveguide surfaces were immediately collected using a CCD camera with excitation by a tunable argon ion laser (488- and 514-nm lines; Ion Laser Technology, Salt Lake City, UT) with an optical setup nearly identical to that of the array sensor.3 The use of HEPES buffers in place of PBS did not impact the Cy5 fluorescence immunoassay. Limits of detection for fluorescence-based assays were determined at a threshold of three standard deviations above the mean of the fluorescence intensities of the negative controls. Sterilization Considerations. The major difficulty encountered in the translation of the immunosensor into a system allowing viability determination, PCR verification, and further sample analysis through providing sample cultures was contamination of the sensor components (PDMS gaskets, tubing, and waveguides) by common environmental bacteria. Culturing of bacteria by recirculation of broth media through the system tubing and PDMS flow channels was abandoned due to the additional potential for contamination. The primary source of contamination was found to be the PDMS gaskets themselves. Sterilization of the PDMS to acceptable levels required soaking in a 20% bleach solution for a minimum period of 1 h followed by rinsing in water and finally rinsing in 70% ethanol/water. Bleach at a concentration of 20% was also used to thoroughly rinse the tubing of the peristaltic pump as well as the polystyrene assembly used to mount the waveguides and gaskets. Syringes and syringe needles were rinsed with the bleach solution followed by rinsing with 70% ethanol. All water, buffers, and juice and all other glass- and plasticwear used were autoclaved. Waveguides were briefly rinsed (
