Rapid Detection and Molecular Typing of Foodborne Pathogens

NON-TECHNICAL SUMMARY: Foodborne diseases cause approximately 76 million illnesses with 5000 deaths annually in the US. Of these, L. monocytogenes causes about 2,500 illnesses and 500 deaths while E. coli O157:H7 causes 73,500 illnesses and 61 deaths annually. Salmonella enterica causes gastroenteritis in humans and nearly 1.4 million people in the US are infected annually, resulting in approximately 16,000 hospitalizations and over 500 deaths. The economic impact of salmonellosis in the US is substantial, with annual costs of 0.5 to 2.3 billion dollars. Immunologically challenged individuals, children and pregnant women, are most susceptible to L. monocytogenes infection while E. coli affects mostly the children. Furthermore, these three are among the top five foodborne pathogens that are of significant concern and currently the US government has implemented a zero tolerance policy for these pathogens in ready-to-eat foods. Much of these foodborne illnesses could be avoided if sensitive and dependable accurate detection tools are available. Furthermore, the "bottleneck" of a rapid detection method is sample preparation. Often the concentration of the target organisms are very low and due to the complex nature of food, it is difficult to extract bacterial cells or toxins from food without exhaustive sample preparation. We propose to aid in the sample preparation strategy by improving our sample preparation device, PED system and by developing improved medium for enrichment of multipathogens and thereby facilitating detection by biosensors. A majority of the biosensor tools reported in the literature are capable of detecting pure bacterial cultures in an artificial media; however, their usefulness with real-world food samples has not been thoroughly investigated yet. From our ongoing project we have shown that fiber-optic, cell-based, and optical light scattering sensors can be used with food samples and these sensors will be modified to allow detection of multipathogens on a single device to provide total microbiological safety of a product in a cost effective manner.

<P>
APPROACH: Sample preparation and media development: In the last 5 years, we have designed and built a simple, inexpensive, hand-held sample preparation device called, PED (Pathogen Enrichment Device) to allow bacterial separation from complex food matrices. Our goal is to continue to improve the device to accommodate larger sample sizes (25-65 g) and also to separate bacteria from more complex samples like milk, soil and fecal maters for application with multiplex biosensor platforms. Fiber optic biosensor: The fiber optic sensor instrument is equipped with four sample testing channels thus would allow the detection of four separate pathogens from a given sample. The optical fibers will be coated with pathogen specific antibody or receptors to capture target pathogens and exposed to samples. Subsequently, pathogen-specific second antibody previously conjugated with a fluorophor (Cy5 or Alexa-Flour 647) will be added. The formation of antigen-antibody sandwich will emit fluorescence light that will generate evanescent wave, which will be detected by a laser detector. Cell-based sensor: As part of infection process, pathogens/toxins interact with mammalian cells in body resulting in cell damage. As a result, intracellular enzyme such as alkaline phosphatase is released, which could be sensitively detected in a plate reader. Recently, we have demonstrated that collagen matrix can be used to immobilize mammalian cells in a 96-well plate and this could be used to test a large number of samples at a time. We plan to continue this research by optimizing the assay by testing with samples of food, water and beverages artificially spiked with various pathogens and toxins. Beside food safety applications, this technology could potentially be used by first responders in food defense to test for the presence of potential toxic agents in food or water. Microfluidic biochips: Microfabricated electronic devices, such as a semiconductor chips are often referred to as biochips and is used for monitoring the presence of target pathogen in a sample. In the past, we have used antibody for capture of bacteria on the chip; however to improve sensitivity and specificity we plan to use specific receptor such as Hsp60 that is commonly used by pathogens during infection of the host to capture and detect on biochip. Optical light scattering sensor: We have developed a laser-based sensor called BARDOT (BActerial Rapid Detection using Optical Technology) for detection and identification of bacterial colonies growing on a Petri dish. BARDOT was able to discriminate different bacterial species within a genus. Our goal is to create a robust scatter image library of bacterial colonies from different genera including Listeria, Staphylococcus, Salmonella, Vibrio and Escherichia, Campylobacter, Enterococcus and others. In addition, we plan to optimize procedure for detection of foodborne bacterial pathogens from food using BARDOT. <P>

PROGRESS: 2006/10 TO 2007/09 <BR>
OUTPUTS: Improving sensitivity and specificity of biosensor tools has continued to be a major focus of our group. In addition, detection of multipathogens using a single sensor platform is of major importance because it can provide safety assessment of a product very rapidly and can save money for pathogen testing. In support of multipathogen detection, a multiplex enrichment broth, SEL, has been formulated that allows simultaneous growth of Salmonella, E. coli and Listeria. This media is currently being evaluated for detection of Listeria monocytogenes, Escherichia coli O157:H7 and Salmonella enterica from various artificially inoculated food samples using fiber optic-based immunosensors. A cell-based sensor that employs collagen-trapped mammalian cells for onsite detection of Listeria and Bacillus toxins is shown to be promising and the results could be obtained in 4-6 h with inoculated food samples. An automated laser light scattering system has been built that allows rapid (in seconds) identification of bacterial colonies growing on the plate without destroying the sample. The identification success rate is 99.6% for Listeria, 95.8% for Staphylococcus spp., 94% for Vibrio spp., 87% for E. coli virotypes and 78% for Salmonella serovars. This system has been currently evaluated for its ability to detect and identify pathogens from food samples. TARGET AUDIENCES: Food scientists interested in food safety and rapid biosensor detection
<P>
IMPACT: 2006/10 TO 2007/09<BR>
The biosensors tools currently under development would reduce assay steps and thus facilitate improved and specific detection of three major pathogens, Listeria monocytogenes, E. coli O157:H7 and Salmonella from food in a cost-effective manner.

<P>

PROGRESS: 2005/10/01 TO 2006/09/30<BR>
Improving sensitivity, specificity and the applicability of biosensors with food samples were the major focus for the last year. Fiber optic - based immunosensor was able to detect E. coli O157:H7 at an initial contamination level of 1 cfu/g of ground beef after 4-6 h and Listeria monocytogenes with a sensitivity of 100,000 cells/ml in meat samples after 20 h enrichment employing the automated RATOR system. A laser light scattering system was used to demonstrate its ability to distinguish bacterial colonies grown on solid agar plates. This noninvasive simple system was able to differentiate different species of Listeria with 90-99% accuracy and was also able to distinguish closely related bacterial strains in minutes. In a blind test, this system was able to detect viable and stress exposed L. monocytogenes accurately from meat samples. This system was also demonstrated to be useful for identification of bacterial colonies belonging to genus, Salmonella, Escherichia, and Vibrio.
<P>
IMPACT: 2005/10/01 TO 2006/09/30<BR>
The light scattering sensor can identify bacterial colonies growing on solid agar plates in a noninvasive manner in seconds thus would reduce assay time and cost of bacterial detection/identification from food.
<P>

PROGRESS: 2004/10/01 TO 2005/09/30<BR>
A fiber optic sensor has been developed for Listeria monocytogenes with a sensitivity of 1000 cells/ml in a pure culture setup. An automated microfluidic controlled fiber optic system called RATOR is also shown to be able to detect Listeria from hotdog samples. A cell-based sensor that employs collagen trapped mammalian cells for onsite detection of Listeria and Bacillus toxins is shown to be promising and the results could be obtained in 4-6 h. A laser light scattering system is being used to demonstrate its ability to distinguish bacterial colonies grown on solid agar plates. This noninvasive simple system is able to differentiate different species of Listeria with 90-99% accuracy and is also able to distinguish closely related bacterial strains in minutes. This system is currently being evaluated for its ability to differentiate Salmonella species.
<P>
IMPACT: 2004/10/01 TO 2005/09/30<BR>
The biosensors tools currently under development would reduce assay steps and thus facilitate improved and specific detection of Listeria monocytogenes from food.
<P>

PROGRESS: 2003/10/01 TO 2004/09/29<BR>
Improving sensitivity and specificity of biosensor tools has continued to be a major focus. A Fiber-Optic sensor has been developed for Listeria monocytogenes. This sensor is sensitive and can detect ~1000 cells/ml in a pure culture setup. Sensitivity diminishes when the cells are stressed or present with natural microflora in food. A buffered selective sample enrichment step helps resuscitation of stressed cells and eliminates interference with other microflora. Food containing an initial load of 10-1000 cells/g could be detected in less than 24 h from the point of food sampling. A two-step detection system specific for pathogenic L. monocytogenes has also been developed. In step one, Listeria cells from enriched ready-to-eat food samples are captured on antibody-coated immunobeads and tested for their ability to kill mammalian cells. This assay is sensitive and extremely specific for cytopathogenic L. monocytogenes and results could be obtained in less than 28 h starting with the food sample. A laser light scattering system has also been developed to distinguish bacterial colonies grown on solid agar plates. This system is able to differentiate closely related bacterial species in minutes. Further testing with contaminated food products is in progress.

IMPACT: 2003/10/01 TO 2004/09/29<BR>
The detection tools developed here are sensitive and specific and will enable us to detect only the pathogenic Listeria monocytogenes in 24 h from ready-to-eat food products where this organism is a major concern. Early detection would reduce ware-house holding time for products, and prevent potential foodborne Listeria monocytogenes related outbreaks and mortality.

<P>
PROGRESS: 2002/10/01 TO 2003/09/30<BR>
We are continuing our efforts in developing biosensor tools for the detection of Listeria monocytogenes. Antibodies are critical for immunosensor applications. We are using a L. monocytogenes-reactive monoclonal antibody (MAb-C11E9) and polyclonal antibodies for our biosensor assays. These antibodies were characterized in detail to determine their reactivity spectrum with strains of L. monocytogenes and L. innocua (a nonpathogenic species). About 88% of L. monocytogenes strains showed strong reactions with C11E9 while only 23% of L. innocua gave equivalent results. This antibody was used in a resonant mirror immunosensor and the antibody was able to detect surface protein extracts from L. monocytogenes and L. innocua and showed no reaction with other Listeria species. However, this biosensor instrument was unable to detect whole cells of Listeria because of the configuration of the sensing layer that was not suitable for detection of intact bacterial cells. In the fiber optic sensor, rabbit polyclonal antibody was used to capture bacteria on the fiber wave-guide and the fluorescent-labeled C11E9 was used to detect bacteria. Using this setup the sensitivity limit of this sensor was determined to be 1000 cells/ml. This sensor could detect L. monocytogenes in the presence of other Listeria species, or other common food contaminants (Enterococcus faecalis, Escherichia coli, Salmonella typhimurium). Furthermore, this sensor was able to detect L. monocytogenes from naturally contaminated or spiked hotdog at cell concentrations of 10, 100 or 1000 cfu/ml after 20 h of enrichment.

IMPACT: 2002/10/01 TO 2003/09/30<BR>
Fiber optic sensor shows promise in detecting low levels of Listeria monocytogenes from spiked hotdog samples after 20 hour of enrichment. This method would allow early and sensitive detection of L. monocytogenes from processed products thus reduce holding time for processed products and prevent potential foodborne Listeria infection. Efficacy of this sensor to detect L. monocytogenes from other meat products is currently under investigation.
<P>
PROGRESS: 2001/10/01 TO 2002/09/30<BR>
Developing a sensitive detection tool for viable and virulent Listeria monocytogenes is the major focus of this project. (1) We are continuing to improve our cell-based sensor that measures L. monocytogenes interaction with mammalian cells. We were able to grow mammalian cells (Ped-2E9, Caco-2, Vero and RAW) on the interdigitated microsensor electrode (IME)-chip; and demonstrated that Ped-2E9 cell death on the chip could be detected within two hours. We also have demonstrated that all 13 serotypes of L. monocytogenes were capable of adhering and invading Caco-2 cells suggesting that Caco-2 cells could be used as a potential candiate cell line for use in the cell-based sensor. (2) We have completed our enzyme-fluorescence-based cytoxicity assay to measure Listeria interaction with Ped-2E9 cells. Data showed that L. monocytogenes at 10 million/ml could be detected in 1 hour. (3) In the antibody-coupled fiber optic biosensor, we are evaluating the cross reaction of fiber optic sensor with non-Listeria organisms and optimizing conditions for detection of L. monocytrogenes cells only. (4) In the past year we have initiated a project on detection of Listeria by using a surface plasmon resonance immunosensor (IAsys sensor). This optical sensor measures the binding event of antigens to the antibody on a sensing layer at real-time. We found out that this system is unable to detect intact bacterial cells, however can detect soluble antigens such as toxins or surface proteins. We have demonstrated that L. monocytogenes proteins at 5 microgram quantities could be detected in less than 15 minutes. Although the testing is very expensive but shows great promise in real-time detection of L. monocytogenes. (5) In addition, TaqMan PCR system was used to confirm the presence of toxin genes (Stx I and II) in food and environmental E. coli isolates. All food isolates were confirmed to be O157:H7 containing stx I or stx II genes. None of the environmental isolates contained stx genes, however 15% of 604 isolates were identified as O157:H7. This data indicates that TaqMan system is capable of generating false results and may not be suitable for testing environmental E. coli isolates.

IMPACT: 2001/10/01 TO 2002/09/30<BR>
Cell-based sensor shows promise in the developemnt of Listeria monocytogenes detection tool. Mammalian cell-based enzyme fluorescence assay can detect L. monocytogenes at levels of 10 million in one hour. Surface plasmon resonance immunosensor can detect soluble L. monocytogenes antigens in 15 minutes but is unable to detect intact cells.

<P>

PROGRESS: 2000/10/01 TO 2001/09/30<BR>
One of our goals is to develop methods that would be able to detect very low numbers of Listeria monocytogenes rapidly from food. Two-step methods were developed: First, concentration of Listeria cells from food and second specific detection by biosensor-based probes such as (a) interdigitated microsensor electrode (IME)-chip, (b) enzyme-fluorescence assay to measure Listeria interaction with animal cells, and (c) antibody-coupled fiber optic biosensor. In addition, various genomic typing methods (amplified fragment length polymorphism (AFLP) and repetitive PCR (Rep-PCR) methods were used to generate genomic fingerprint patterns for E. coli O157:H7 strains. [1] Immunoseparation and cytotoxicity assay with Ped-2E9 cell: Immunobeads (magnetic and nonmagnetic) were used to capture Listeria cells from naturally contaminated or spiked hotdog samples after a selective enrichment step, and bead-captured bacterial cells were directly tested in a cytotoxicity assay. Results showed that L. monocytogenes at a concentration of about 1-100 CFU/100 ml of food extract could be captured in 12-18 h and subsequently be detected in additional 1-2 h. Advantages of this current assay; (i) it is specific, (ii) it detects only viable cells and (iii) most importantly, it detects virulent Listeria. [2] Interdigitated microsensor electrode (IME)-chip based cytotoxicity assay: Mammalian cells (Ped-2E9) exposed to L. monocytogenes for 1-2 h was placed on the IME-chip, and the Ped-2E9 damage was measured by an Impedance Analyzer. The data showed that IME-chip is capable of distinguishing Listeria induced damaged Ped-2E9 cells from the healthy ones. This IME-chip based cytotoxicity assay could be used in conjunction with immunobead separation method for direct detection of L. monocytogenes. [3] Fiber optic biosensor: Polystyrene wave-guides were immobilized with anti-Listeria antibody and were used to capture Listeria cells. Specific detection was achieved by fluorescein-labeled anti-Listeria mAb and a laser detector. Using a pure culture of L. monocytogenes, we were able to detect as low as ~38 CFU/ml with the fiber optic probe. [4]Fluorescence based cytotoxicity assay using mammalian Ped-2E9 cells: Various potential foodborne contaminants such as E. coli O157:H7, Salmonella, Citrobacter, Bacillus, Staphylococcus, Corynaebacterium and L. monocytogenes were examined with Ped-2E9 cells in a fluorescence-based cytotoxicity assay. Results showed that Citrobacter, Bacillus, Corynaebacterium and L. monocytogenes produced positive cytotoxicity, whereas other test organisms did not. Selective enrichment media was used to inhibit non-Listeria organisms prior to the assay and the data showed that Ped-2E9 -based cytotoxicity assay could be used for specific detection of L. monocytogenes. [5] Genomic fingerprinting of E. coli O157:H7: Preliminary results indicated that genomic fingerprint patterns generated by AFLP and Rep-PCR could effectively separate E. coli O157:H7 strains from other E. coli serotypes. Furthermore, these methods appeared to be more sensitive than the Ribotyping and pulsed-field gel electrophoresis (PFGE) methods.

IMPACT: 2000/10/01 TO 2001/09/30<BR>
A rapid Listeria detection method was developed that could sensitively detect 1-100 CFU of L. monocytogenes cells in 14-20 h. Advantages of this current assay; (i) it is specific for L. monocytogenes, (ii) it detects only viable cells and (iii) most importantly it detects virulent Listeria. Currently, we are testing this method with naturally contaminated hotdog and other food samples.
<P>

PROGRESS: 1999/10/01 TO 2000/09/30<BR>
Direct detection of Listeria monocytogenes: We have developed fiber optic biosensor to directly detect Listeria from food samples. Polystyrene fibers (wave-guide) were coated with Listeria species specific polyclonal antibody and exposed to different Listeria species. Subsequently fibers were exposed to L. monocytogenes-specific monoclonal antibody (C11E9)-conjugated to a fluorescene dye (Cy5) and signals were obtained from a laser photodiode detector (Analyte 2000). Preliminary data showed that signals generated by L. monocytogenes is about 2.5 times higher than the nonpathogenic L. innocua. Ability of this fiber optic biosensor to detect very low number of L. monocytogenes cells is in progress. Indirect detection of Listeria monocytogenes: L. monocytogenes positively interact with hybrid B-lymphocytes (Ped-2E9) from humans and mice. Sensitive measurement of this interaction is currently used for detection of Listeria in 1-2 h. An Interdigitated Microsensor Electrode (IME)-chip was employed to measure Listeria interaction with Ped-2E9 cells using an impedance analyzer. Preliminary results indicated that this technology could be used for sensitive detection of L. monocytogenes. Also, we measured alkaline phosphatase (AP) release from infected Ped-2E9 cells using a substrate (methylumbelliferyl phosphate) that produces fluorescence end product. We have observed that fluorescence based detection is at least 10 times more sensitive than the previously described colorimetric based assay. Additionally, we found that a panel of Gram-negative foodborne pathogen did not show any cytotoxicity effect on Ped-2E9 cells. Several Gram-positive bacterial strains will also be tested with Ped-2E9 cells. Based on this preliminary result it appears that Listeria interaction with Ped-2E9 cell is specific and could be used to detect L. monocytogenes only. Genomic typing: Genetic diversity and virulence gene expressions of 30 Listeria monocytogenes isolates from human patients and foods originated from two different geographic locations without any epidemiological relations were analyzed. All strains contained virulence genes, inlA, inlB, actA, hlyA, plcA and plcB with expected product size in PCR assay except for the actA gene. Some strains produced actA gene product of 268 and others 385 bp. Phenotypically all were hemolytic but showed variable expressions of phospholipase activity. Ribotyping classified isolates into 12 different groups, which consisted primarily of clinical or food isolates or both. Cluster analysis also indicated possible existence of clones of L. monocytogenes that are maintained in food or human hosts or are evenly distributed in nature. Two isolates (F1 from food and CHL1250 from patient) had unique ribotype patterns that were not previously reported in the RiboPrinterO database. This study indicates distribution of diverse L. monocytogenes strains in clinical and food environments and some of which carry a remarkably high genetic homogeneity in spite of their origins from two different geographic locations and environments.

IMPACT: 1999/10/01 TO 2000/09/30<BR>
Various biosensor based methods employed in this project show great promise for direct or indirect detection of Listeria monocytogenes from food. These assays would reduce time for overall microbial analysis of ready-to-eat food products thus save lives and reduce economic losses. Genetically diverse L. monocytogenes strains exist in clinical and food environments and some of which carry a remarkably high genetic homogeneity in spite of their origins from different geographic locations or environments.

<P>

PROGRESS: 1998/10/01 TO 1999/09/30<BR>
1. Listeria monocytogenes positively interact with hybrid B-lymphocytes from humans and mice, which could be used as an indicator for L. monocytogenes cytotoxicity in 5-6 h. Addition of dithiothreitol (0.5 mM) to the cell suspension can expedite L. monocytogenes mediated lymphocyte death in 2 h or less thus allowing rapid detection of L. monocytogenes. Monoclonal antibody (C11E9) specific for L. monocytogenes are harvested in large quantities from hybridoma C11E9 cells and purified using Protein-A affinity chromatography. These antibodies will be used to capture Listeria cells from meat samples with antibody coated magnetic beads. Furthermore, experiments are currently underway to conjugate this antibody to fiber optic cables for sensitive detection of Listeria cells. 2. We have isolated several enterohemorrhagic E. coli (EHEC) strains from naturally contaminated ground beef (an ongoing project from Alabama A&M University). Thirteen percent beef samples are found to be positive for E. coli O157:H7. About 76% of those isolates showed 0-45% Vero cell toxicity by lactate dehydrogenase (LDH) assay and 24% exhibited 45-83% toxicity. The cytotoxicity patterns were comparable to those with EHEC obtained from various outbreaks. Additionally, these isolates were shown to be more resistant to multiple antibiotics than the previously reported outbreak strains. 3. EHEC isolates were also examined for genomic fingerprint profiles using randomly amplified polymorphic DNA fragmentation (RAPD) assay. RAPD indicated that isolates from same sample, as expected, had similar RAPD patterns while very distinct patterns for isolates from different samples. RiboPrinter analysis of selected strains confirmed the isolates to be E. coli O157:H7.
<P>
IMPACT: 1998/10/01 TO 1999/09/30<BR>
1. Dithiothreitol (DTT) can enhance L. monocytogenes mediated lymphocyte toxicity thus can significantly reduce the cytotoxicity assay time from 6 h to 2 h or less. 2. About 13% of beef samples were found to be positive for E. coli O157:H7 strains. 3. These E. coli isolates showed enhanced antibiotic resistance patterns than the previously published reports.