DEVELOPMENT OF A BIOSENSOR FOR THE RAPID DETECTION OF SALMONELLA TYPHIMURIUM IN MILK Except where reference is made to the work of others, the work described in this dissertation is my own or was done in collaboration with my advisory committee. This dissertation does not include proprietary or classified information. Shuqing Zhang Certificate of Approval: Tung-Shi Huang Sondra Jean Weese, Chair Assistant Professor Professor Nutrition and Food Science Nutrition and Food Science Shelly R. McKee Stephen L. McFarland Associate Professor Dean Poultry Science Graduate School DEVELOPMENT OF A BIOSENSOR FOR THE RAPID DETECTION OF SALMONELLA TYPHIMURIUM IN MILK Shuqing Zhang A Dissertation Submitted to the Graduate Faculty of Auburn University in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy Auburn, Alabama December 16, 2005 DEVELOPMENT OF A BIOSENSOR FOR THE RAPID DETECTION OF SALMONELLA TYPHIMURIUM IN MILK Shuqing Zhang Permission is granted to Auburn University to make copies of this dissertation at its discretion, upon request of individuals or institutions and at their expense. The author reserves all publication rights. Signature of Author Date of Signature iii iv DISSERTATION ABSTRACT DEVELOPMENT OF A BIOSENSOR FOR THE RAPID DETECTION OF SALMONELLA TYPHIMURIUM IN MILK Shuqing Zhang Doctor of Philosophy, December 16, 2005 (M.S., Capital University of Medical Sciences, 1998) (B.S., Xinxiang Medical College, 1995) 88 Typed Pages Directed by Sondra Jean Weese Foodborne illnesses are still serious problems in the United States. Methods for rapid detection of foodborne pathogens are needed to protect public health. The recent advances in biosensor technology enabled the rapid detection of pathogens. The purpose of this research was to develop a biosensor for the rapid detection of Salmonella in milk. In the first study, a 55 kD purified outer membrane protein (OMP) of S. enterica Typhimurium was used to produce monoclonal and polyclonal antibodies. The S. enterica Typhimurium specific monoclonal antibody (mAb) 1B4 was immobilized on a sensor platform as a bacterial capture agent, and the captured S. enterica Typhimurium was visualized by a light microscope. The developed biosensor has a detection limit of 10 2 cfu/ml. v In the second study, one step enrichment in conjunction with a centrifugation technique was investigated. In order to increase the relatively low number of Salmonella in milk, 7 non-selective and 5 selective enrichment media were compared. Among these media, brain heart infusion broth (BHI) and brilliant green broth (BG) were the most efficient. These media were able to increase Salmonella concentration from 0.01 cfu/ml to above 10 2 cfu/ml in 6 h. The centrifugation technique used in this study was able to concentrate bacteria up to 50 times. In conjunction with centrifugation, enrichment in BG or BHI for 4 h could increase Salmonella concentration from 0.01 cfu/ml to a detectable level. The milk content, excessive amounts of other bacteria, and enrichment broth content did not affect the detection sensitivity and specificity of the developed biosensor. The developed biosensor was found to be a rapid and sensitive method for the detection of S. enterica Typhimurium in milk. Total time needed for the detection of 0.1 cfu/ml S. enterica Typhimurium in milk was 5 h. vi ACKNOWLEDGMENTS I wish to express my deepest gratitude to Dr. Jean Weese, my major advisor, for her guidance, support, and encouragement throughout the course of the study. I would like to thank Dr. Tung-Shi Huang and Dr. Shelly R. McKee for their constructive suggestions. I also would like to acknowledge Roger Bridgman and Patricia West for their technical assistance during this study. Finally, I express gratefulness to my parents, my parents-in-law, my husband, Fei Ding, and my sons, Andrew and Daniel, for their understanding, encouragement and patience while I spent seemingly endless nights and week-ends on this research. vii Style manual or journal used: Journal of Food Science Computer software used: Microsoft Word 2004 and Microsoft Excel 2004 viii TABLE OF CONTENTS LIST OF TABLES ?????????????????????????? x LIST OF FIGURES ?????????????????????????... xi CHAPTER I. INTRODUCTION ????????????????????.. 1 CHAPTER II. LITERATURE REVIEW?????????????????... 5 Statement of Research Objectives?????????????????? 26 CHAPTER III. Development and characterization of monoclonal and polyclonal antibodies against Salmonella enterica Typhimurium for biosensor application???????????????????????.. 27 Abstract ???????????????????????????... 27 Introduction ??????????????????????????. 28 Materials and Methods ?????????????????????? 31 Results and Discussion ?????????????????????? 36 Conclusion .??????????????????????????.. 41 CHAPTER IV. Rapid detection of S. enterica Typhimurium in milk using a biosensor ????????????????????. 47 Abstract ???????????????????????????... 47 Introduction ??????????????????????????. 49 Materials and Methods ?????????????????????? 51 Results and Discussion ?????????????????????? 55 ix Conclusion .??????????????????????????.. 59 CHAPTER V. OVERALL CONCLUSION ????????????????. 66 CHAPTER VI. REFERENCES ????????????????????? 68 LIST OF TABLES 1. Specificity of mAbs and pAbs for S. enterica Typhimurium assayed by indirect ELISA ?????????????????????.. 45 2. The binding of Salmonella enterica Typhimurium on mAb 1B4 immobilized biosensor platforms at various mAb concentrations?????????????????????. 45 3. Numbers of bacteria bound to immuniosensors at various S. enterica Typhimurium concentrations as well as mixed with E. coli O157:H7????????????????? 46 4. Efficacy of one step enrichment at 2, 4, and 6 h incubation period before or after centrifugation????????????????????? 63 5. Bacterial concentration efficacy of centrifugation????????????... 64 6. Comparison of four different sensor platforms?????????????? 65 x xi LISTE OF FIGURES 1. SDS-PAGE of purified S. enterica Typhimurium outer membrane protein ??????????????????????.. 42 2. Detection sensitivity of mAb 1B4 to Salmonella enterica Typhimurium by indirect ELISA ????????????????????????.. 43 3. Photograph of biosensor captured S. enterica Typhimurium at 10 8 cfu/ml (A), 10 7 cfu/ml (B), 10 6 cfu/ml (C), and blank (D) ?????????????? 44 4. Comparion of different preenrichment and enrichment media for the recovery of inoculated S. enterica Typhimurium in milk ??????? 60 5. Detection sensitivity of indirect sandwich ELISA using mAb 1B4 as capture antibody ????????????????????. 61 6. Light microscopic images of captured S. enterica Typhimurium in milk, enrichment broth, and centrifuged samples???????????? 62 1 CHAPTER I. INTRODUCTION Foodborne illnesses affect millions of American each year. The presence of Salmonella in foods is a great concern to the food industry, the public, and the regulatory agencies. In the United States, the number of cases of foodborne illnesses due to Salmonella is the highest of any reported gastrointestinal infection (CDC 2005). Each year, an estimated 1.4 million people contract salmonellosis in the United States (Mead and others 1999). Outbreaks of salmonellosis associated with milk continue to occur (Ryan and others 1987, Olsen and others 2004). The medical cost of all foodborne illnesses in the US is 5 ? 6 billion dollars each year and salmonellosis alone accounts for 1 billion dollars. Rapid and sensitive methods for the detection of relatively low incidence of Salmonella in foods, including milk, are of urgent need. Several methods have been developed to detect the presence of potentially low levels of Salmonella in foods. These methods include culturing techniques, bacteriophages (Goodridge and Griffiths 2002; Hirsh and Martin 1983, 1984), nucleic acid-based methods (Fitts and others 1983, Fitts 1985, Fluit and others 1993, McElroy and others 1996, Kwang and others 1996), immunoassays (Mattingly 1984, Ibrahim and others 1986, Hadfield and others 1987), and biosensors (Ye and others 1997, Babacan and others 2002, Bokken and others 2003, Kramer and Lim 2004, Taitt and others 2004). Standard culture methods for Salmonella detection involve multi-steps which are labor intensive and time-consuming, taking 4 to 5 days for detection and confirmation (FDA 1992). Furthermore, it has been reported that some of the routinely used selective 2 enrichment broths are inhibitory towards Salmonella (Van der Zee 1994). The nucleic acid-based methods, and bacteriophage methods have shown considerable success by reducing the detection time to 2-3 days, but none of them has been identified as fast, specific, and user-friendly for routine use in screening food samples. Immunoassay tests have been regarded as an effective tool for microorganism detection and identification. Several formats of immunoassays have been developed, including radioimmunoassay (RIA), fluorescent immunoassay (FIA), and enzyme linked immunosorbent assay (ELISA). Among these, ELISA was regarded as the most promising one. Several types of enzyme-linked immunosorbent assays (ELISAs) have been developed, using either polyclonal antibodies (pAb) or monoclonal antibodies (mAb) that are able to detect most Salmonella serotypes. The majority of the developed antibodies are specific to either O-antigen or H-antigen (NG and others 1996, Wang and others 1996, Jaradat and Zawistowski 1996). Some of these methods produce false- positive results due to the cross-reactivity with other enterobacteria (Curiale and others 1990). Recently, the use of biosensors to detect microorganisms has attracted the public?s attention. Most of the currently developed biosensors for bacterial pathogen detection are based on the specific antigen-antibody binding reactions, where the antibody is immobilized on the sensor platform to capture the bacteria that are of interest. Then the detection is measured through electrochemical, optical, or piezoelectrical signals. These biosensors include piezoelectrical biosensors (Prusak-Sachaczewski and others 1990, Si and others 1996, Ye and others 1997, Babacan and others 2002), fiber optic biosensors (Kramer and Lim 2004), optical surface plasmon resonance biosensors (Bokken and 3 others 2003), bienzyme electrochemical biosensors (Yang and others 2001). The detection limit of these biosensors is high, ranging from 10 5 cfu/ml to 10 7 cfu/ml. Therefore, more sensitive biosensors are in urgent need to detect relatively low numbers of pathogens in foods. The advancement in biosensor technology requires highly specific antibodies to reduce detection interference. Salmonella are antigenically complex and serotypes have been differentiated by somatic lipopolysaccharide (LPS) or flagellar protein antigens (Le Minor 1984). The outer membrane proteins (OMPs) of Salmonella are also known to have an important role in evoking immune responses (Kim and others 1991; Meenakshi and others 1999). There appears to be a potential to develop species or serotype specific antibodies from Salmonella OMPs. Due to the low incidence of Salmonella in foods, preenrichment, enrichment, and sometimes even a post-enrichment were involved in rapid detection methods (Flowers and others 1988). One step enrichment has been investigated prior to PCR, ELISA, and biosensor assays for the detection of pathogens in foods. Different enrichment media have been investigated on their efficacy of rapid culturing of pathogens in foods (Kwang and others 1996, Eyigor and others 2002, Fluit and others 1993, Prusak-Sochaczewski and Luong 1989). The total detection time required by their methods was more than 24 h. Lack of unanimity on the selection of media for one step enrichment is still a problem. Methods for separation and concentration of Salmonella from foods using immunomagnetic separation (IMS), filtration, or centrifugation have been investigated. IMS uses antibody coated magnetic particles to separate and concentrate microbes from foods (Fluit and others 1993, Cerro and others 2002). The membrane filtration technique 4 for the capture of microbes has been investigated since 1951 (Goetz and Tsuneshi). The hydrophobic grid membrane filter (HGMF) technique has been used for bacterial enumeration since 1974 (Sharpe and Michaud). Centrifugation has been studied as a means of concentration of microbes by researchers and regarded as an effective way of concentrating microbes (Hawa and others 1984, McElroy and others 1995, McElroy and others 1996, Cerro and others 2002). In this study, purified Salmonella OMP were used as antigens for the production of specific pAb and mAb. The Salmonella specific mAb was used to construct a novel biosensor, which combines an antibody immobilized sensor platform with microscopic imaging systems to accurately and sensitively detect Salmonella in milk. One step enrichment in conjunction with centrifugation was investigated to quickly culture and concentrate Salmonella to a detectable level. 5 CHAPTER II. LITERATURE REVIEW Salmonella general characteristics Foodborne illnesses are a worldwide serious public health problem. The Centers for Disease Control and Prevention (CDC) estimates that 76 million persons contract foodborne illnesses, 325,000 are hospitalized and 5,000 die each year due to foodborne illnesses in the United States (Mead and others, 1999). Foodborne illnesses have been estimated to cost as much as $23 billion annually in this country (Jones and Gerber, 2001). The annual cost of foodborne illnesses caused by the 4 most common bacterial pathogens alone (Salmonella, Shigella, Campylobacter, and E. coli) is $6.9 billion (Allos and others, 2004). In 2004, a total of 15,806 laboratory-diagnosed cases of infections in the FoodNet surveillance areas were reported and the top nine pathogens were identified as Salmonella, 6,464; Campylobacter, 5,665; Shigella, 2,231; Cryptosporidium, 613; STEC O157, 401; Yersinia, 173;Vibrio, 124; Listeria, 120; and Cyclospora, 15 (CDC, 2005). These data indicates that Salmonella is still the leading cause of foodborne illnesses in the US. Each year, approximately 40,000 cases of salmonellosis are reported in the United States. Because many milder cases are not reported, the actual number of infections may be thirty or more times greater which would be about 2 to 4 million each year (CDC 2005). It is estimated that approximately 600 people die each year with acute salmonellosis. Salmonella is a group of bacteria and has over 2300 serotypes. In 2004, 6 five serotypes accounted for 56% of infections. They are as follows: Typhimurium, 20%; Enteritidis, 15%; Newport, 10%; Javiana, 7%; and Heidelberg, 5% (CDC, 2005). Salmonella has been known to cause gastrointestinal illness in humans for over 100 years. They were discovered by an American scientist named Salmon, for whom they are named. They are gram-negative, non-spore forming rods, motile (with few exceptions), and microscopic living creatures. They are widely distributed in nature. Humans and animals are their primary reservoirs. They live in the intestinal tract of humans and other animals, including birds. They can pass from the feces of humans and animals to other people or other animals. Salmonellosis results mainly from the ingestion of Salmonella contaminated foods. The infective dose varies with different strains, host health and age with the elderly, infants and immunocompromised people being more easily infected. Contaminated foods are often of animal origin, such as beef, poultry, milk, or eggs, but all foods, including vegetables can become contaminated. S. Typhi and the paratyphoid bacteria cause septicemic and typhoid or typhoid- like fever. Other forms of salmonellosis generally produce milder symptoms, which include diarrhea, fever, and abdominal cramps 12 to 72 hours after ingestion of contaminated foods. The illness usually lasts 4 to 7 days, and most people recover without treatment. However, in some people, such as infants, the elderly, and the infirm, the diarrhea may be so severe that the patient needs to be hospitalized. In these cases, the Salmonella infection may spread from the intestines to the blood stream, and then to other body sites and result in death. A small number of people who are infected with Salmonella will develop Reiter?s syndrome, which includes pains in joints, irritation of 7 the eyes, and painful urination. It can last for months or years, and can lead to chronic arthritis. Salmonellosis outbreaks Large outbreaks of salmonellosis typically occur at banquets or similar circumstances. However, the two largest outbreaks occurred under unusual circumstances. The largest outbreak occurred in 1994, which involved more than 224,000 people (Jay 2000). The vehicle food was ice cream produced from contaminated milk. The contamination occurred in milk transported in a tanker truck which previously hauled liquid eggs. S. Enteritidis was identified as the etiological agent. The second largest outbreak occurred in 1985 which involved around 200,000 people (Ryan and others 1987). The vehicle food was milk pasteurized in a Chicago dairy plant. S. Typhimurium was identified as the etiological agent. The U.S. Food and Drug Administration (USFDA) inspectors discovered that modification of the pasteurization equipment, which was to facilitate the running off of raw milk, resulted in the cross contamination of the pasteurized milk with raw milk. Currently developed methods 1. Standard culture methods Standard culture methods for the detection of Salmonella in foods include: preenrichment (16-20 h) to allow the resuscitation and multiplication of sub-lethally damaged cells; selective enrichment (18-48 h) to increase the ratio of Salmonella to other organisms; and plating on selective/differential media (24-48 h) to enable the recognition 8 of Salmonella colonies; and subsequent serological or biochemical identification (4-48 h) of suspected colonies (Litchfield, 1973). 1.1. Sampling methods Because interpretations of a large consignment of food are based on a relatively small sample of the lot, representative sampling is essential and the adequacy and condition of the sample are of great importance when pathogens are sparsely distributed in the foods. The use of sterile sampling equipment and aseptic techniques should under no circumstances be compromised. Samples should be delivered to the laboratory promptly maintaining as nearly as possible the original storage condition. Usually a 25 g analytical unit is diluted at a 1:9 sample/broth ratio and mixed in a homogenizer, stomacher or pulsifier. Samples may be composit. The mixing processes are used to ensure that the bacteria are completely released from the food matrix and adequately dispersed before the pre-enrichment stage. 1.2. Preenrichment media The first step of conventional culture methods for the detection of Salmonella is preenrichment. One of its primary functions is to rehydrate bacteria that have been dehydrated during the food processing steps. Preenrichment also repairs damaged cells to a stable physiological condition. The preenrichment medium serves as a nutrient source for the proliferation of the rejuvenated cells. Because non-selective, highly nutritious media are used at this stage, there is a risk of overgrowth of other contaminants which could actually inhibit the growth of Salmonella. Therefore, the timing and temperature are important at this stage to ensure recover of Salmonella. Usually an 18-24 h incubation at 37 ?C is used. 9 Preenrichment media vary widely in nutritive values. Lactose broth (LB), trypticase soy broth (TSB), reconstituted nonfat dry milk, brain heart infusion broth (BHI) and buffered peptone water (BPW) are considered to be nutritionally complex and are recommended by various organizations (Andrews 1985). However, some research data indicate that less nutritive broth, such as M-9 broth, would favor the recovery of sublethally injured Salmonella. 1.3. Selective enrichment media The second step of conventional methods for the detection of Salmonella is selective enrichment. Normally, Salmonella is only a small portion among competing microorganisms in foods and can be lost during selective enrichment. Therefore, successful selective enrichment is the decisive step in the detection of Salmonella. Numerous selective enrichments have been developed and great efforts have been made in recent years to improve media and procedures for selective enrichment. Media containing either selenite or tetrathionate plus brilliant green and bile salt or malachite green in combination with high amounts of magnesium chloride are the major selective enrichment media for Salmonella. These media can be distinguished into three groups: selenite broth (Leifson 1936), tetrathionate broth (Muller 1923, Kauffmann 1930), and the Rappaport family. Selenite broth, originated by Leifson (1936), is used as a selective enrichment for the cultivation of Salmonella that may be present in small numbers, competing with other microbes. Enzymatic digestion of casein and enzymatic digestion of animal tissue are used as nitrogen and vitamin sources. Selenite is the selective agent, which inhibits the growth of coliform bacteria and enterococci in the first 6-12 hours of incubation, its 10 inhibitory effect then gradually declines. Lactose is the fermentable carbohydrate. A rise in pH decreases selectivity of selenite. The acid produced by lactose fermentation helps to maintain a neutral pH. Sodium phosphate is also used as a buffering agent to maintain neutral pH of the broth. Salmonella, Proteus and Pseudomonas are only slightly inhibited by selenite. Selenite cystine broth is based on the formulation of selenite broth with the addition of L-cystine, which is a reducing agent. Tetrathionate is produced from thiosulfate by adding iodine to the medium. Tetrathionate and thiosulfate suppress coliform and other enteric bacteria, whereas tetrathionate-reducing bacteria, such as Salmonella and Proteus, can grow in this medium. Calcium carbonate neutralizes the acidic products which are formed from the decomposition of tetrathionate. Bile salts inhibit bacteria which do not normally inhabit in the intestine. Brilliant green suppresses the Gram-positive bacteria. Novobiocin may be added at 40 mg/l to suppress the growth of Proteus (Jeffries 1959). Rappaport broth was proposed by Rappaport and others (1956) for the selective enrichment of Salmonella (with exception of S. Typhosa). The original Rappaport medium contains high amounts of malachite green and magnesium chloride and has a very low pH of 5.2. Malachite green and magnesium chloride inhibit the growth of the other intestinal microorganisms, but do not affect the proliferation of most Salmonella. Magnisium chloride was shown to counteract the toxic effect of the dye for Salmonella. S. Typhosa and Shigella are inhibited by this broth. Vassiliadis and others (1976) modified the medium by reducing the concentration of the dye. Rappaport-Vassiliadis medium (RV) has been regarded as a useful media beside selenite and tetrathionate broths. Based on the completion of AOAC precollaborative and collaborative studies, 11 Rappaport-Vassiliadis medium is being recommended for the analysis of high microbial and low microbial load foods by FDA (2003). FDA also recommends that Rappaport- Vassiliadis medium replaces selenite cystine broth for the analysis of almost all foods (FDA 2003). Vassiliadis and coworker (1976) recommended incubation of RV at 43 ?C to achieve maximum selectivity. However, any deviation above 43 ?C may be lethal for Salmonella. For this reason, the ISO recommendation to incubate at 42 ?C is followed by most researchers (Busse 1995). By-passing the preenrichment step and using direct selective enrichment for the analysis of certain foods, the time needed to complete the conventional culture method would be reduced by one whole day. However, the recovery of Salmonella from raw meats and dried egg albumen was significantly decreased by by-passing the preenrichment step (Andrews 1985). Motility enrichment is an old technique (Craigie 1931) and was originally performed in U-tubes where one side was inoculated and after a certain incubation period fast moving bacteria could be isolated from the other side. Goosens and others (1984) replaced U-tubes with Petri dishes. Motility enrichment on Modified Semisolid Rappaport-Vassiliadis medium (MSRV) was introduced by De Smedt and others (1986) and has been regarded as a very effective procedure for the isolation of Salmonella from foods (Boer 1998). In this media, motility of competitive bacteria is inhibited while Salmonella move into fresh media leaving their competitors behind. 1.4. Plating media Selective plating media for the isolation of Salmonella can be divided into three groups according to the selective agents used. These groups are the bile salt agars, the 12 brilliant green agar (BGA), and bismuth sulfite agar (BSA). The bile salt agars include deoxycholate citrate agar (DCA) originally proposed by Leifson (1935), xylose lysine desoxycholate citrate agar (XLD) (Taylor 1965), Salmonella-Shigella agar (SS) (Pollock and Dahlgren 1974), and Hektoen Enteric agar (HE) (King and Metzger 1968). These media contain various amounts of citrate and bile salts as selective substances. Thiosulphate is added to most of these agars both as a source of hydrogen sulfide to detect hydrogen sulfide positive colonies and as a selective agent (Busse 1995). Tate and others (1990) used XLT4, which is essentially a modification of XLD with the addition of a surfactant, Tergitol, and found it to give improved recovery of Salmonella from poultry. Recently, a new plating medium belonging to the bile salt family has been proposed by Rambach (1990). Rambach agar is characterized by a new chromogenic indicator system. Salmonella is detected by its ability to ferment propylene glycol and produce acid, appearing as bright red colonies. In addition, chromogenic substrate X-Gal (5-bromo-4-chloro-3-indolyl-?-D-galactopyranoside) is included to differentiate Salmonella from other ?-D-galactosidase positive enterobacteria, such as E. coli and other coliform bacteria, which ferment lactose and appear as blue colonies. Proteus spp. are negative for both reactions and their colonies are colorless, while Citrobacter are positive for both reactions and their colonies are violet from the combination of colors resulting from the two reactions. Studies indicate that most Salmonella ferment propylene glycol except for S. Typhi and occasional representatives of other serotypes (Gruenenwald and others 1991). On BGA, Salmonella is detected by its inability to ferment lactose or lactose and sucrose, appearing as pink colonies. The mannitol lysine crystal violet brilliant green agar 13 (MLCB) was introduced by Inoue and others (1968) and its selective components are crystal violet and brilliant green. Bismuth sulphite agar, originally formulated by Wilson and Blair (1927), contains ammonium bismuth citrate and sodium sulphite. Salmonella is detected by H 2 S formation and bismuth reduction. Rainbow agar Salmonella (Biolog) can isolate and differentiate the widest range of Salmonella species, including S. Typhi, and can detect the weakest to the strongest H 2 S producers. 1.5. Biochemical and serological confirmation tests A range of biochemical tests are used to confirm the suspected colonies on selective agars. Of these, triple sugar iron agar (TSI) and lysine iron agar (LIA) are the most commonly used for the identification of Salmonella in mixed cultures, while urea broth is used for the identification of Salmonella in pure cultures (Moats 1981, USFDA 2003). Serological tests may be used to confirm the biochemically screened isolates. Serological polyvalent flagellar (H) tests, serological somatic (O) tests, and Spicer- Edwards serological tests are recommended by USFDA (2003). 1.6. Conclusion of standard culture methods Although the conventional methods are effective and have been used for many years for isolating and identifying bacterial pathogens in foods, the procedures are labor intense. These methods provide a theoretical level of sensitivity of one Salmonella cell per 25 g of food, but detection can be prevented by the presence of other microbes that can compete with Salmonella during cultural enrichment. The requirement for time- consuming enrichment steps and biochemical characterization has lengthened most analysis time to several days. Therefore, these methods are inadequate for making timely 14 assessments on the microbiological safety of foods. In order to overcome these drawbacks, a variety of so called rapid methods have been investigated. 2. Rapid culture methods Researchers have attempted to reduce the incubation periods for both the preenrichment and selective enrichment. Through comparing the efficiency of preenriching a variety of foods and feed ingredients for 6 and 24 hr in 7 media, D?Aoust and Maishment (1979) found that none of the 7 preenrichment media were adequate when incubated for only 6 hr. Investigating the effect of reducing the incubation period for selective enrichments, researchers (D?Aoust and others 1983, Rappold and others 1984) found that selective enrichment for 8 h and overnight resulted in similar recovery of Salmonella. Therefore, it appears to have some potential for reducing the selective enrichment, but not the preenrichment period of conventional culture methods. Combining preenrichment and selective enrichment into a single step has been attempted in order to reduce the incubation period of these two steps. Sveum and Hartman (1977) tested enrichment of samples in a media with the addition of timed- release capsules of selective agents. Results indicated that a 24 h incubation period of the capsule media was inadequate when compared to conventional culture methods. Pignato and others (1995) investigated the use of Salmosyst broth in combination with Rambach agar for rapid detection of Salmonella in foods. Heat injured Salmonella was preenriched in the Salmosyst preenrichment broth (SPB) for 6 h and followed by selective enrichment for 18 h with the addition of selective agent to SPB. Their results indicated that the use of 15 Salmosyst broth was even better than conventional methods for the recovery of heat injured Salmonella. 3. Bacteriophage In the reporter bacteriophage technology, DNA carrying a reporter gene, such as luciferase (lux and luc) gene, ice nucleation (inaW) gene and the E. coli ?-galactosidase (lacZ) gene, is introduced into a target bacterium via a bacteriophage. Once the bacterium-specific phage invade host cells, the reporter gene is expressed with phage DNA, allowing bacterial cells to be rapidly identified. Since bacteriophages can not replicate without host cells, expression of the reporter gene is indicative of the presence of the infected organism (Goodridge and Griffiths, 2002). Bacteriophages can also be used to detect specific bacteria without reporter genes. Felix O1 bacteriophage lyses most Salmonella at concentrations ranging from 10 9 to 10 12 plaque-forming units. Detection of Salmonella was achieved at cell concentrations of ? 4 ? 10 4 cfu/ml by using Felix O1 phage and high performance liquid chromatography (Hirsh and Martin, 1983). Because some Salmonella strains did not support bacteriophage replication, the technique was modified by the same group of researchers (Hirsh and Martin, 1984). A large-pore charge-modified filter was used to concentrate cells, which were then eluted into enrichment broth containing brilliant blue dye. After 16 h incubation, centrifugation and resuspension, the cells were inoculated on a nutrient agar plate to which Felix O1 phage had been added. The presence of lacunae after 6 h was indicative of a positive result. However, the application of this method to food is yet to be assessed. One possible reason is that it is difficult to genetically modify bacteriophages and the bacterial hosts 16 have not been genetically characterized. Another potential problem is that an individual bacteriophage can not detect all isolates of a bacterial species. 4. Nucleic acid-based methods Nucleic acid based methods include pulsed-field gel electrophoresis (PFGE), DNA hybridization (DNAH), and polymerase chain reaction (PCR). PFGE has been used in the subtyping of bacteria, in which restriction patterns of whole bacterial gene sequences are analyzed and compared. The bacteria are immobilized by mixing the bacterial suspension with melted agarose. Then the bacteria are lysed to release DNA and the released bacterial DNA is cut into a moderate number of DNA fragments by a selected enzyme. The DNA fragments are loaded onto agarose gel and separated by pulsed-field electrophoresis, in which the orientation of the electrical field alternates. Several factors, such as recent point mutation, deletion, insertion, and loss or acquisition of plasmids might result in differences in PFGE banding patterns. PFGE alone is unable to replace conventional phage typing of Salmonella (Lukinmaa and others 2004). Some Salmonella strains can not be typed by PFGE since the DNA is degraded during electrophoresis (Liesegang and Tschape 2002). The essential principle of DNAH and PCR methods is the specific formation of double stranded nucleic acid molecules from two complementary, single stranded molecules under defined physical and chemical conditions. Typically, a DNAH assay consists of the following steps. The bacteria are applied to a solid support such as a nitrocellulose membrane filter and are then lysed to release their DNA, and the DNA is denatured into separate strands and fixed to the solid support. The membrane is then 17 hybridized with a labeled DNA probe. Labeled nucleotides are added to the system for synthesis to occur. These nucleotides can be labeled with a radioisotope, or with a non- isotopic reporter molecule. These hybrids contain one radiolabeled strand that can then be detected by autoradiography (Fitts and others 1983, Fitts 1985). Because of safety concerns and disposal problems associated with radiolabels, they have been replaced with non-isotopic labels (King and others 1989). The ribosomal RNA (rRNA) hybridization has several advantages over DNAH: more sensitive (more copies of rRNA per cell) and no denaturation is required before hybridization (single strand of rRNA) (King and others 1989). Because the detection limit of DNAH is 10 4 -10 5 cfu/ml, selective enrichment is needed to bring the bacterial number above the detection limit. This step may take 18-48 h (Swaminathan 1994). Using a PCR method, specific DNA-fragment is amplified during a three-step cyclic process: first, the target DNA is denatured at high temperatures (95-100 ?C), second, two synthetic oligonucleotides (primers) are annealed at opposite strands at a temperature (37-55 ?C) that allows hybridization to the correct target, finally polymerization is performed with the oligonucleotides as primers for the DNA polymerase and the target DNA as a template. When this is repeated over and over with newly synthesized DNA as template in addition to the original target DNA, an exponential amplification of the DNA fragment is obtained. Many different applications of PCR have been described and some of these applications use the whole gene sequences rather than a selected fragment as a target. Random amplified polymorphic DNA (RAPD) is based on the amplification of DNA with arbitrarily developed primers. RAPD is criticized for its lack of reproducibility 18 (Lukinmaa and others 2004). The rep-PCR is based on the amplification of the repetitive sequences of bacterial genomes. Restriction fragment length polymorphism PCR (RFLP- PCR) utilizes a restriction enzyme to cut the amplified known sequence and the restriction fragment profiles are compared between different strains. Amplified fragment length polymorphism PCR (AFLP-PCR) is a recently adopted PCR-based typing technique. Two enzymes are used in this method, one frequently cutting and the other one rarely cutting. The disadvantage of AFLP-PCR is that incomplete digestion of DNA may result in an aberrant AFLP pattern. The real time PCR (rt-PCR) technique combines amplification and detection in one single step (Eyigor and others 2002). The detection sensitivity of rt-PCR was determined as low as 6 cfu/ml from artificially spiked poultry intestine samples after 18 h enrichment in tetrathionate broth by Eyigor and coworkers (2002). Nucleic acid based methods have been commonly used in many research laboratories, the majority of these methods have not been applied in routine food microbiology. Major problems are the detection of dead cells and the presence of enzyme inhibitors in food samples (Olsen, et al, 1995). 5. Immunoassay Immunoassays are powerful analytical tools that permit the specific and rapid detection of antigens to which antibodies can be produced. Application of immunoassays in Salmonella detection could provide a considerable savings in time, effort and materials over conventional cultural methods. 19 Antigens used for antibody production include lipopolysaccharide (LPS, O antigen), flagellin (H antigen), fimbriae, outer membrane protein (OMP), Capsular antigen (Vi of a few serovars including S. typhi and K-antigen of other Salmonella), and whole inactivated bacteria. Variation of H-antigen, O-antigen, and capsular antigen is the basis of Salmonella serotyping (Mortimer and others 2004). The LPS of Salmonella has three components: a lipid moiety, a core oligosaccharide, and the O-specific side chain. LPS affects virulence by the amount produced, the O side chain length and composition, and the degree of glycosylation. Flagella are antigenic surface structures of Salmonella. The LPS antigen had been used to produce polyclonal and monoclonal antibodies for the detection of Salmonella (Tsang and others 1987, Luk and Lindberg 1991, Lu and others 1991, Choi and Ng 1992, Wang and others 1996). LPS has been used for serological identification of Salmonella infected chicken (van Ziderveld and others 1992) and cattle (Veling and others 2000). Flagellin is the major structural protein of flagella and carries the serotype- specific H-antigenic determinants. They elicit an early short-lived humoral response in the host upon infection, making them ideal for the detection of infections in livestock through screening anti-salmonella antibodies of the infected individuals. These H antigens probably induce the production of serotype-specific antibodies in the infected host. Antibodies against the conserved N- and C-terminal fagellin domains give rise to cross-reactions between Salmonella serotypes and other Enterbacteriaceae (Ibrahim and others 1985). Antibodies against H-antigen have been developed for the detection of salmonella in foods (Swaminathan and Ayres 1980, Wyatt and others 1993). 20 Fimbriae or pili are thin proteinaceous and filamentous surface organelles produced by most Salmonella and some other enteric bacteria. Fimbriae are an important factor in bacterial survival and persistence in the host. They mediate the adhesion of bacteria to epithelial cell surfaces (Kisiela and others 2003). Fimbriae are composed of protein subunits called fimbrins and auxiliary proteins (Peralta and others 1994). The expression of fimbriae by certain strains of salmonella was first described in 1958 (Duguid and Gillies), but only recently has their potential as diagnostic antigens been considered. It is believed that Salmonella only express fimbriae in close association with epithelial and other cell types, but rarely express fimbriae in extracellular environments (Thorns, 1995). Therefore this antigen is probably suitable for serologically screening of antibodies produced by Salmonella infected live hosts (Thorn and others 1996) but not suitable for identifying Salmonella in food. Although much effort has gone into the characterization of fimbriae expressed by Salmonella, their application to the detection of specific Salmonella has not been fully exploited. One major reason for this has been a lack of understanding of the role of these antigens in the lifecycle of the bacterium. Four basic types of immunoassay have been developed, including agglutination (Barrow, 1994), radioimmunoassay (RIA) (van Vunaki, 1980), fluorescent immunoassay (FIA), and enzyme immunoassay (EIA) (Candish, 1991). The agglutination method has been used to detect whole cells by adding specific antisera to a suspension of bacteria at approximately 10 9 cfu/ml. The antibody binds to the bacterial surface antigens forming a network of linkages, which result in visual precipitation of the bacteria. Pure bacteria cultures at high concentrations are required and often many different organisms can cause agglutination. Therefore, this method is 21 extremely insensitive and nonspecific (Candish 1991). Agglutination tests have also been used for the serological identification of infected chicken (van Zijderveld and others 1992) and cattle (Veling and others 2000). Latex agglutination using different colored beads coated with antibodies specific to different antigens was applied to Salmonella groupings (Hadfield and others 1987). The first RIA was developed by Yalow and Berson (1959) for clinical studies. RIA is sensitive, which can detect nanograms to picograms of antigens. 125 I, 14 C and 3 H are the commonly used radioactive labels. RIA was used for the detection of Salmonella in foods by Ibrahim and others (1985, 1986). RIA has safety problems due to the use of radioactive labels. Also, the need for expensive equipment for measuring radioactivity and the chemical instability of certain radioactive labels are a limitation of RIA (Ibrahim 1986). Because of the disadvantages of RIA, effort has been attempted to develop alternative labels. Fluorescent molecule and enzymes are proven to be most promising labels. The labeling of antibody with fluorescein isothiocyanate (FITC) was first demonstrated and applied for identifying bacteria in clinical specimens by Coons and others (1942). The first FIA for Salmonella detection was reported by Thomason and others (1957). In their study, the fluorescein labeled antibodies were demonstrated to be able to stain O, Vi, and H antigens of S. typhi. Later, an indirect FIA (using polyclonal anti-salmonella O antibody and fluorescent goat anti-rabbit serum) for the detection of Salmonella in foods was developed (Georgala and Boothroyd 1964, Haglund and others 1964, Silliker and others 1966). Although the FIA technique had been tested in a wide range of foods and feeds, it never enjoyed widespread use. The false positive results 22 (non-specific staining), false negative results (obscuring of fluorescing cells by food particles), high cost of instrumentation, requirement of confirmation of FIA positive results by culture methods, and the limitation of sample throughput have limited its application (Blackburn 1993). FIA in combination with flow cytometry for the detection of Salmonella in milk had been developed (McClelland and Pinder 1994). In order to reduce the high level of background labeling of milk micelles, a milk clearing solution was used to coagulate the particulate matter and then centrifugation was used to remove the coagulant. The clearing solution and centrifugation were effective for whole and reduced fat pasteurized milk but not the ultrahigh-temperature milk (McClelland and Pinder 1994). The use of enzymes as a label in immunoassays was first introduced by van Weenman and Schuurs (1971) and Engvall and Perlmann (1971).The first EIA for Salmonella detection was reported by Krysinski and Heimsch (1977) using a polyclonal antibody specific to flagella of S. Typhimurium. Since then, EIA techniques have shown tremendous potential for the detection of pathogens in foods and are the most widely used format for immunoassays. By Enzyme-linked immunosorbent assay (ELISA), one of the reactants is absorbed on the surface of a solid phase such as polyvinyl chloride microplate. Krysinski and Heimsch (1977) were able to detect S. Typhimurium in artificially inoculated foods and feeds in 48 h with a detection sensitivity of 10 5 cfu/ml. False positive results were observed in their study. Efforts have been attempted to reduce the cross-reactivity of the polyclonal antibodies by fractioning the IgGs using gel filtration and affinity chromatography (Swaminathan and Ayres 1980). 23 Several ELISA formats, which include sandwich ELISA, indirect ELISA, direct ELISA, and competitive ELISA, have been applied in foods (Swaminathan and Feng 1994). Each format of ELISA has its advantages and disadvantages. The competitive ELISA is rapid, involving only one incubation step and one washing step. However, a purified antigen is required to perform the assay. Also, food ingredients may interfere with the assay since the enzyme conjugated antibody reacts directly with the sample. These problems can be avoided in non-competitive ELISA (sandwich, indirect, direct ELISA). Non-competitive ELISA also has the possibility of amplification of the detection signal through binding of several enzyme-labeled antibody molecules to one antigen molecule. Sandwich ELISA is suitable for analyzing food samples because it can avoid the competitive binding of food ingredients and other microbes with the analyst. Also, it is adaptable to a kit format for use. ELISA is undoubtedly a useful format of immunoassay in food microbiology due to its simplicity and ability to simultaneously test multiple samples. The major problems with ELISA for Salmonella detection is the inadequate specificity of the polyclonal antibodies. The development of monoclonal antibodies could lead to improvements in the specificity of ELISA. ELISA involves several steps of incubation which prolonged the detection time. The detection sensitivity of the most developed ELISA is about 10 5 cfu/ml (Candish 1991). Preenrichment or selective enrichment of food samples are needed to increase the bacteria number in order to reach the detection limit. 24 6. Bacterial concentration methods The requirement to detect low numbers of Salmonella in foods has limited the application of rapid detection methods. Preenrichment, selective enrichment, and sometimes even a post-enrichment are normally involved before applying rapid detection methods. As a consequence, research effort has been put on the development of Salmonella concentration methods (Blackburn 1993). Membrane filtration techniques using hydrophobic grid-membrane filter (HGMF), immunomagnetic separation (IMS), and centrifugation have been used by various researchers to separate and concentrate microbes from foods. Membrane filtration was first used for the capture of bacteria from water samples (Goetz and Tsuneshi 1951). A wide variety of applications has been discovered in the past several decades. The concept of using HGMF as a bacterial enumeration tool was introduced by Sharp and Michaud (1974). HGMF is essentially a conventional membrane imprinted with hydrophobic grids to form individual bacterial growth compartments. The wax grids limit the growth of bacteria within the individual compartments. Because more than one bacteria could be seeded in one compartment during the filtration, the bacterial number on the HGMF is determined by the most probable number (MPN) method. The original HGMF technique for enumeration of bacteria in foods involves a complicated and lengthy incubation procedure, which include preenrichment by placing the filter on a nonselective medium and followed by selective enrichment by transferring the filter to an appropriate selective medium (Brodsky and others 1982). Wu and Fung (2004) have simplified the HGMF bacterial enumeration technique by using HGMF in combination with the thin agar layer (TAL) methods. 25 IMS technique uses antibody coated magnetic particles to capture, separate, and concentrate microbes from foods. It has been proposed as a method of selective enrichment by several researchers (Luk and Lindberg 1991, Mansfield and Forsythe 1993). In their methods, IMS only replaced the selective enrichment of the standard culture method. 7. Immunosensor Recently, the use of biosensors to detect microorganisms has attracted the public?s attention. The currently developed biosensors for bacterial pathogen detection are mostly based on the antigen-antibody specific binding reaction where the antibody is immobilized on a sensor platform to capture the bacteria that are of interest. Then the detection is measured electrochemically, optically, or piezoelectricly. These biosensors include piezoelectric biosensors (Prusak-Sachaczewski and others 1990, Si and others 1996, Ye and others 1997, Babacan and others 2002), fiber optic biosensor (Kramer and Lim 2004), optical surface plasmon resonance biosensor (Bokken and others 2003), bienzyme electrochemical biosensor (Yang and others 2001). The detection limit of these biosensors is high, ranging from 10 5 cfu/ml to 10 7 cfu/ml. Therefore, more sensitive biosensors are in urgent need to detect relatively low numbers of pathogens in foods. 26 STATEMENT OF RESEARCH OBJECTIVES The objective of this research was to develop a sensitive biosensor using Salmonella specific mAbs for the rapid detection of Salmonella in milk. The specific objectives for the first study were to: 1. develope S. enterica Typhimurium specific mAb and pAb using purified outer membrane proteins as an immunogen. 2. characterize the developed S. enterica Typhimurium specific mAb and pAb. 3. develop a biosensor using S. enterica Typhimurium specific mAb as a capture antibody and visualize the captured S. enterica Typhimurium by a light microscope. The specific objectives for the second study were to: 1. select an effective media for rapid culturing of low numbers of inoculated S. enterica Typhimurium in milk. 2. evaluate the efficacy of centrifugation as an bacterial concentrating method 3. compare different sensor platforms 4. develop an indirect sandwich ELISA to verify the performance of the biosensor. 5. evaluate the performance of the biosensor for the detection of inoculated S. enterica Typhimurium in milk.. 27 CHAPTER III DEVELOPMENT AND CHARACTERIZATION OF MONOCLONAL AND POLYCLONAL ANTIBODIES AGAINST SALMONELLA ENTERICA TYPHIMURIUM FOR BIOSENSOR APPLICATION ABSTRACT For the construction of a biosensor which can specifically detect Salmonella enterica Typhimurium, monoclonal antibody (mAb) 1B4 and polyclonal antibody (pAb) S48 were developed against purified outer membrane protein (OMP) of Salmonella enterica Typhimurium. Gel electrophoresis data indicated that the apparent molecular weight of OMP was 55 kDa. The mAb 1B4 strongly reacted to Salmonella enterica Typhimurium and Salmonella enterica Paratyphi and had no cross reaction to other gram negative and gram positive bacteria tested. The pAb S48 also showed high reactivity to Salmonella enterica Typhimurium, but had cross reactivity with other tested bacteria. Due to its high specificity and sensitivity against S. enterica Typhimurium, mAb 1B4 was immobilized to the biosensor surface. The detection of Salmonella by the biosensor was photographed by a light microscope. The detection limit of the biosensor was determined to be 1?10 2 cfu/ml in bacterial culture. The presence of excess amounts of E. coli did not interfere with the detection limit of the biosensor constructed in this study. This is the first time the 55 kDa antigenic component purified in this study has been 28 reported. The biosensor (an immunosensor platform in combination with microscopic imaging systems) could accurately and sensitively detect Salmonella within hours. Keywords: Salmonella enterica Typhimurium, biosensor, monoclonal antibody INTRODUCTION The presence of Salmonella in foods is a great concern to the food industry, the public, and the regulatory agencies. In the United States, the number of cases of foodborne illnesses due to Salmonella is the highest of any reported gastrointestinal infections. Although Salmonella has more than 2,300 serotypes, Salmonella enterica serotype Typhimurium is the most common serotype in the U.S. isolated from people with salmonellosis, a disease caused by the infection of Salmonella (CDC 2005). Several methods have been developed to detect the presence of potentially low levels of Salmonella in foods. These methods include culturing techniques, bacteriophages (Goodridge and Griffiths 2002; Hirsh and Martin 1983, 1984), DNA hybridization (Fitts and others 1983, Fitts 1985), polymerase chain reaction (PCR, Fluit and others 1993, McElroy and others 1996, Kwang and others 1996), immunoassays (Mattingly 1984, Ibrahim and others 1986, Hadfield and others 1987), and biosensors (Ye and others 1997, Babacan and others 2002, Bokken and others 2003, Kramer and Lim 2004, Taitt and others 2004). Standard culturing methods for Salmonella detection involve multi-steps which are labor intensive and time-consuming, taking 4 to 5 days for detection and confirmation (FDA 1992). Furthermore, it has been reported that some of 29 the routinely used selective enrichment broths are inhibitory towards Salmonella (Van der Zee 1994). The PCR, DNA hybridization, and bacteriophage methods have shown considerable success by reducing the detection time to 2-3 days, but none of them has been identified as fast, specific, and user-friendly for routine use in screening food samples. Therefore, development of rapid detection methods is necessary for the food industry to quickly respond to food and food product contamination. These methods would also allow regulatory agencies to deal with outbreaks or to enforce food regulation laws. Since the first enzyme immunoassay for Salmonella detection was reported in 1977 (Krysinski and Heimsch), immunochemical techniques have been shown to be stable and reliable methods for microorganism detection or identification. Several enzyme-linked immunosorbent assays (ELISAs) have been developed, using either polyclonal antibodies (pAb) or monoclonal antibodies (mAb) that are able to detect most Salmonella serotypes. The majority of developed antibodies are specific to the Salmonella outer core lipopolysaccharide (LPS, O-antigen) or flagella (H-antigen) (NG and others 1996, Wang and others 1996, Jaradat and Zawistowski 1996). Although immunoassays reduce the total assay time by 1 or 2 days, most of the detection kits actually only replace the agar-plating stage of the culturing assay and even introduce a third, ?post enrichment? broth stage before the sample can be assayed by ELISA. In addition, the immunochemically based methods produces many false-positive results due to the cross-reactivity with other enterobacteria (Curiale and others 1990). Recently, the use of biosensors to detect microorganisms has hit the public. The currently developed biosensors for bacterial pathogen detection are mostly based on the 30 antigen-antibody specific binding reactions, where antibody is immobilized on sensor platform to capture the bacteria that are of interest. Then the detection is measured electrochemically, optically, or piezoelectrically. These biosensors include piezoelectric biosensors (Prusak-Sachaczewski and others 1990, Si and others 1996, Ye and others 1997, Babacan and others 2002), fiber optic biosensor (Kramer and Lim 2004), optical surface plasmon resonance biosensor (Bokken and others 2003), bienzyme electrochemical biosensor (Yang and others 2001). The detection limit of these biosensors is high, ranging from 10 5 cfu/ml to 10 7 cfu/ml. Therefore, more sensitive biosensors are in urgent need to detect relatively low numbers of pathogens in foods. The advancement in biosensor technology requires the highly specific antibody to reduce interference of the detection. Salmonella are antigenically complex, serotypes have been differentiated by somatic lipopolysaccharide (LPS) or flagellar protein antigens (Le Minor 1984). The outer membrane proteins (OMPs) of Salmonella are also known to have an important role in evoking immune responses (Kim and others 1991; Meenakshi and others 1999). There appears to be a potential to develop species or serotype specific antibodies from Salmonella OMPs. Therefore, the objective of this study was to develop Salmonella-specific antibodies to build a novel biosensor, combining antibody immobilized sensor platform with microscopic imaging systems, to accurately and sensitively detect Salmonella within hours. 31 MATERIALS AND METHODS Bacteria culture Salmonella enterica Typhimurium ATCC13311, E. coli 43895, E. coli 932, and E. coli 48-2 were obtained from James Barbaree and Stuart Price laboratories at Auburn University, Auburn, AL. Salmonella enterica serovars Paratyphi, Enteritidis, Mission, and Montevideo, E. coli O157:H7 204P, L. monocytogenes G3982 4b, and L. innocua were provided by the Center for Food Safety, University of Georgia (Griffin, GA). Listeria monocytogenes Scott A and Staphylococcus aureus ATCC 12600 were purchased from the American Type Culture Collection (ATCC), Rockville, MD. Salmonella spp., E. coli spp., and Staphylococcus aureus were grown in Trypticase ? Soy Broth (TSB, Difco Laboratories, Detroit, MI) while Listeria spp. were grown in TSB with the addition of 0.6% yeast extract (TSBYE). The bacterial cultures were incubated in a gyratory water bath at 37 ?C at 100 rpm for 16 hr. After incubation the cultures were washed twice with phosphate buffered saline (PBS, pH 7.2) by centrifugation at 5,000 ? g for 20 min. The bacteria were resuspended in PBS and bacterial concentration was estimated from the absorbance of the bacterial suspensions at 640 nm using a standard curve. The bacterial concentrations were confirmed by plating Salmonella spp., E. coli spp., and Staphylococcus aureus on Trypticase ? Soy Agar (TSA) and Listeria spp. on TSA with addition of 0.6% yeast extract (TSAYE). Antigen preparation The outer membrane proteins (OMPs) were prepared as described by Meenakshi and colleagues (1999) with modifications. After S. Typhimurium was grown in TSB and 32 held in a gyratory water bath (100 rpm) at 37 ?C for 16 h, the bacteria were washed twice with normal saline solution through centrifugation at 5,000 ? g for 10 min at 4 ?C. The cells were suspended in 10 mM Hepes buffer (pH 7.4) and sonicated at 20 kHz five times on ice for 1 min each time at 1 min intervals. The sonicated samples were centrifuged for 1,700 ? g for 20 min to remove the intact cells and debris. The supernatant, which was designated as sonicate extract, was collected and centrifuged for 100,000 ? g at 4 ?C for 60 min. The supernatant was discarded, and the clear gel-like pellet was collected and resuspended in 2% (w/v) sodium lauryl sarcosinate detergent in 10 mM Hepes buffer followed by incubation at room temperature for 1 hr. The detergent insoluble fraction was collected by centrifuging for 100,000 ? g at 4 ?C for 60 min, and the pellet was washed twice with deionized (DI) water through centrifugation for 100,000 ? g at 4 ?C for 60 min. The OMPs were solubilized by adding 6 mol/l guanidine thiocyanate (Sigma Chemical Co., St. Louis, MO) to the pellet. The solution was mixed at room temperature for 60 min. Insoluble material was removed by ultracentrifugation for 300,000 ? g at 4 ?C and the remaining solution was dialyzed against deionized (DI) water overnight at 4 ?C with several changes of DI water (Kerr and others 1992). Protein content was measured by the Bradford method using a protein assay kit (Bio-Rad, Hercules, CA). The purity of OMPs was determined by gel electrophoresis. The OMPs were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) using a Mini-Protein II electrophoresis unit with stacking (4% T, 2.67% C) and separating (12% T, 2.67% C) acrylamide gels. Protein bands were stained with Coomassie Brilliant Blue R-250 and the apparent molecular weights of protein bands were calculated using a calibrating curve constructed from molecular weight standards (Sigma). 33 Monoclonal antibody (mAb) production The purified OMPs of S. Typhimurium were emulsified with RIBI?s adjuvant system (RAS-R700) (Corixa, Hamilton, MT) at 100 ?g/ml for mice immunization. Each BALB/cAnNHsd female mouse (Harlan Sprague Dawley Inc., Indianapolis, IN) was immunized with 200 ?l of the OMPs-adjuvant emulsion through the injection of 100 ?l aliquots into each of two legs subcutaneous sites on the ventral side near the axillary and inguinal lymphatics. Three mice were immunized and the same RAS-R700 was used for the initial immunization and booster. Mice were boosted 4 times after the initial immunization with antigen-adjuvant emulsion at 3-week intervals. Blood was collected 7 days after each booster started at the second booster and the titers of the serum were determined. The procedures described by Kohler and Milstein (1975) were followed with modifications. Spleen cells were harvested 5 days after the final booster to fuse with murine myeloma cells at a ratio of 2:1 (spleen cells/myeloma cells) using 50% polyethylene glycol (PEG). The fused cells were suspended in hypoxanthine-aminopterin-thymidine (HAT) selective medium (Dulbecco's modified Eagle medium containing 200 ?M hypoxanthine, 0.8 ?M aminopterin, and 32 ?M thymidine) and seeded into 96-well cell culture plates. After 10 to 14 days, the supernatant of the growth positive wells were screened for production of anti-Salmonella Typhimurium antibodies using indirect ELISA. The hybridoma secreting antibodies that only reacted with S. Typhimurium were selected for expansion and cloning using limiting dilution methods (Harlow and Lane 1988). The mAbs was produced from selected cell lines through mouse ascites fluid and purified by affinity chromatography using the Biological Duo-Flow System (Bio-Rad) with a protein A affinity column. The purified 34 antibody was dialyzed against 0.01 M phosphate-buffered saline (PBS, pH 7.0) at 4 ?C with three changes of the dialysis buffer. The concentration of IgG was determined by Bradford method using a protein assay kit (Bio-Rad) and the purity of the antibody was checked by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The isotype of mAb was determined by a Mouse Monoclonal Antibody Isotyping Kit (Sigma). The purified IgG was stored at -20 ?C in antibody buffer (0.1 M Tris + 2 mM MgCl 2 + 20 mM Glycine + 30 mM sodium azide, pH 8.0) with the addition of 50% glycerol. Polyclonal antibody (pAb) production New Zealand white rabbits about 3 kg each were used to produce antiserum against S. Typhimurium. The purified OMPs of S. Typhimurium were emulsified with RAS for rabbit (RAS-R730) at 285 ?g/ml. Each rabbit was immunized with 1 ml of the OMPs-adjuvant emulsion through an intradermal injection of 50 ?l aliquot into each of six sites on the back close to the spinal cord and 300 ?l aliquot intramuscular into each hind leg. The same dosage of OMPs-adjuvant emulsion was used to boost the rabbit at 4- week intervals after initial immunization. Rabbit blood was collected from the central ear artery 7 days after each boost and antiserum was collected from the supernatant by centrifuging the blood for 5,000 ? g at 4 o C for 20 min. The polyclonal antibodies in the antiserum were collected from 20-50% saturated ammonium sulfate precipitation and re- suspended for dialysis in PBS for 24 hr with 3 changes of dialysis buffer. The partially purified antibodies were further purified by affinity chromatography using a protein A affinity column (Sigma), where the protein only binds with IgG and its subclasses. The 35 bound antibodies were eluted with 0.1 M citrate buffer, pH 5.0. The titer and specificity of the purified polyclonal antibodies were determined by indirect ELISA. Indirect ELISA for specificity tests of pAb and mAb Each well of a 96-well assay plate (polystyrene plate, Costar, Cambridge, MA) was coated with 100 ?l of 10 8 cfu/ml of formalin inactivated or live bacteria at 37 ?C for 1 h. The plate was washed 3 times with 200 ?l of PBS containing 0.1 % Tween-20 and 0.02% sodium azide (PBST), and then the unbound areas of the wells were blocked with 200 ?l of 1% bovine serum albumin (BSA) in PBS at 37 ?C for 1 h. Following the addition of 100 ?l appropriately diluted Abs to each well, the plate was incubated at 37 ?C for 1 h and then washed three times with PBST. After adding 100 ?l/well of alkaline phosphatase conjugated goat anti-rabbit /rabbit anti-mouse IgG (Sigma) at 3,000 times dilution in PBS, the plate was incubated at 37 ?C for 1 h. The plate was washed 4 times and 100 ?l of p-nitrophenyl phosphate (p-npp, Sigma) solution was added to each well. The plate was incubated at room temperature in the dark for color development. Absorbance at 405 nm was measured by a microplate reader (ThermoLabsystems, Helsinki, Finland) and recorded for analysis. Preparation of biosensor platform and antibody immobilization The sensor platform was prepared by cutting microscope cover glass (0.17 mm thickness) into 5 mm squares. The squares were ultrasonically cleaned in DI water and in 95% ethanol. The polished surface of each square was coated with a 140 nm gold by using a Pelco SC-6 sputter coater (Ted Pella Inc., Redding, CA) to provide a favorable layer for antibody immobilization. The monoclonal antibody (1B4) was coated on the sensor platform as the receptor to capture the bacteria. The biosensor platform was 36 constructed by incubating 100 ?l of 24 ?g/ml mAb 1B4 on the gold coated platform at room temperature for 1 h in a 6 ? 6 mm 2 square well of 96-well Uniplate ? (Whatman Inc, Clifton, NY). After washing 3 times with PBST, the platform was incubated with 1% BSA for 30 min to block unbound areas. After washing 3 times with PBS and then 3 times with DI water, the platform was used immediately or dried at room temperature and stored at 4 ?C till use. Sensor-bacteria binding and light microscopic imaging The biosensor platforms were incubated with 100 ?l of 0, 10, 10 2 , 10 3 , 10 4 , 10 5 , 10 6 , 10 7 , 10 8 cfu/ml Salmonella Typhimurium suspension at room temperature for 1 h, and washed 4 times with DI water. After the immunosensors were dried at room temperature, the bacteria captured by the biosensor were fixed with OsO 4 for 1 h and photographed using a Nikon Eclipse L150 Industrial light microscope (Nikon Instruments Inc., Melville, NY) with 1,000 times magnification. The experiment was repeated five times, and the images of the bacteria on two representative microscope fields at 0.025 mm 2 from each immunosensor were recorded as a data set. The bacterial number on each field was counted and the detection of Salmonella was defined as positive when one or more bacteria were observed on all the images of a data set. RESULTS AND DISCUSSION SDS-PAGE patterns of OMPs The 10% polyacrylamide gel contained 0.1% SDS is optimal for OMPs separation on SDS-PAGE. The apparent molecular weights of protein bands on the gel were calculated through a calibrating curve constructed from the molecular weight standards. 37 From the protein patterns, the purity of OMPs extract is high and the apparent molecular weight of the major protein of the extracted OMPs was about 55 kDa (Figure 1). Few other proteins had been observed on the gel and the contents were very low compared to the major protein. It indicated that there were high probabilities to obtain very specific pAb or mAb against Salmonella by using OMPs as antigen. Compared to Kerr and others (1992) results, who produced mAbs against a 36 kD OMP of S. enterica Enteritidis, this is the first time reporting of the use of the 55 kD OMP to produce mAbs against S. enterica Typhimurium. Sensitivity test of mAb 1B4 and pAb by indirect ELISA One fusion resulted in about 700 growth-positive hybridomas and 120 showed positive reactivity with S. enterica Typhimurium cells as determined by indirect ELISA. Among these positive hybridomas, only two hybridoma cell lines 1B4 and 7B10 secreted antibodies specific to S. enterica Typhimurium and S. paratyphi. After single cell cloning of the 1B4, the cells were used to produce acities from mouse for mAb production. After purification by affinity protein A column, the purity of mAb and pAb was checked by SDS-PAGE with 12% gel and showed very high (data not shown). The isotype of mAb 1B4 was determined to be IgG2a. The binding efficiency of mAb 1B4 to Salmonella enterica Typhimurium decreased rapidly with the decrease of bacterial concentration from 1?10 8 to 1?10 6 cfu/ml with the highest ELISA reading decreasing from above 4.0 to 0.5. The binding efficiency was then decreased slowly from 1?10 6 to 1?10 2 cfu/ml and leveled off from 1?10 2 to 0 cfu/ml (Figure 2). There was a significant difference of ELISA readings between 0 cfu/ml and 1?10 3 cfu/ml (p<0.05). Therefore, the detection limit of the mAb 38 1B4 based indirect ELISA was 1?10 3 cfu/ml. pAb S48 showed similar sensitivity as mAb 1B4 when tested against different concentrations of Salmonella enterica Typhimurium (data not shown). E. coli O157:H7 was added to different concentrations of Salmonella suspensions at 10 8 cfu/ml to evaluate the effect of binding efficiency of mAb 1B4 against Salmonella enterica Typhimurium. The results showed that the bindings between 1B4 and Salmonella were not affected by other bacterium (Figure 2). Specificity of mAb 1B4 and pAb S48 The mAb 1B4 and pAb S48 were then tested for specificity with 13 different bacterial species or strains by indirect ELISA. The bacteria were coated to microtiter plate at 1?10 8 cfu/ml. Both mAb and pAb were diluted to 1.0 ?g/ml and applied 100 ?l to each well based on the results of checker board titration of these two antibodies (results not shown). Among the 13 tested bacteria, the mAb 1B4 only reacted with S. enterica Typhimurium and S. enterica Paratyphi and showed high reactivity. The pAb S48 is able to react with all tested Salmonella except Montevideo, and this antibody also has slight cross-reactivity with E. coli O157:H7 (Table 1). These results indicated that mAb 1B4 had higher binding specificity against Salmonella and lower background signal than those in pAb. By comparing the mAbs 1B4 produced from S. enterica Typhimurium OMPs in this study to the only mAbs developed from S. enterica Enteritidis OMPs by Kerr and others (1992), 1B4 was highly specific to S. enterica Typhimurium and Paratyphi and had no cross reactions with other bacteria. The mAbs developed by Kerr and others reacted to a wide variety of Salmonella and other strains of Enterobacteria, which resulted in false-positive reactions during sample testing. Based on the specificity, mAb 1B4 was selected as the capture antibody to be immobilized onto the biosensor platform 39 to bind Salmonella cells. This would result in reducing the cross reactivity with other bacteria to lower the false positive reaction and increasing detection accuracy. In the sandwich test, the pAb is better than mAb 1B4 used as the second antibody to conjugate with enzymes for enhancing detection signal due to its high reactivity. Detection of S. enterica Typhimurium by biosensor The mAb 1B4 was immobilized onto the surface of gold coated sensor platform at different concentrations to determine the bacterial capture efficiencies. On each sensor, 100 ?l of S. enterica Typhimurium suspension at 1?10 8 cfu/ml were added, and the numbers of captured bacteria at different antibody concentrations were calculated from the light microscopic images (Table 2). There were 467 bacteria been detected on the sensor coated with 1B4 at 24 ?g/ml, which was the highest number of bacteria captured. The number of captured bacteria decreased with the decreasing of antibody concentration from 24 ?g/ml to 0.75 ?g/ml. The mAb concentration and captured bacteria number had shown a high second order polynomial relationship (r 2 = 0.9549). No bacteria had been observed on the sensors coated with mAb 1B4 at 0.375 and 0 ?g/ml. Based on these results, the mAb antibody concentration of 24 ?g/ml was selected for the following studies. The effect of detection sensitivity and specificity of 1B4 to S. enterica Typhiumrium by other nonspecific bacteria on sensor platform was also performed by adding E. coli O157:H7 to Salmonella suspensions at 10 8 cfu/ml. The results showed that the same bacterial numbers were captured on both sensors treated with S. enterica Typhiumrium only and S. enterica Typhiumrium combined with1?10 8 cfu/ml E. coli O157:H7. The population of captured bacteria was about tenfold decrease when the 40 concentrations of S. enterica Typhiumrium suspensions were in tenfold decrease from 1?10 8 to 1?10 6 cfu/ml. Then the captured bacterial numbers drop dramatically to single digit when bacteria concentrations were reduced from 1?10 6 to 1?10 2 cfu/ml. No bacteria were observed on the biosensor in the treatment of E. coli O157:H7 at 1?10 8 cfu/ml without Salmonella (Table 3, Figure 3). The results also showed that the binding of S. enterica Typhiumrium to mAb 1B4 was not affected by E. coli O157:H7. Therefore, the binding specificity of mAb 1B4 to S. enterica Typhiumrium is high. The bacteria on sensors at 10 cfu/ml of S. enterica Typhiumrium treatments, about 50% of the treated sensors were shown positive and only one bacterium was observed on each sensor. Therefore, the detection sensitivity of the sensor was limited to 100 cfu/ml. In contrast, ELISA-based methods using LPS-specific antibodies reported by other researchers have a detection limit of 10 6 -10 7 cfu/ml in 17 to 29 hr (Choi and others 1992, NG and others 1996). Electron microscopy (EM) is the other technique that can directly visualize bacteria more detail than does by light microscope, but it requires more complicate sample preparations and trained personnel to operate the EM (Skepper 2000). 41 CONCLUSION The mAb 1B4 developed in this study had high specificity and high sensitivity to S. enterica Typhiumrium. The 55 kDa OMP purified in this study is the antigen which was first time being reported. The biosensor (an immunosensor combined with microscopic imagining systems) developed in this study can specifically detect S. enterica Typhiumrium in 1 h with a sensitivity of 1?10 2 cfu/ml. The whole detection is rapid, cheap, and simple. MW (kD) 116 97 84 66 55 45 36 Std 1 2 Figure 1. SDS-PAGE of purified S. enterica Typhimurium outer membrane protein. Std: molecular weight standard; 1: S. enterica Typhimurium outer membrane proteins (OMP); 2: Sonicate extract of S. enterica Typhimurium 42 0.0 1.0 2.0 3.0 4.0 5.0 012345678 Concentration of Salmonella enterica Typhimurium (log cfu/ml) O .D . 405n m Salmonella only Salmonella mixed with 10^8 cfu/ml E. coli 0.10 0.15 012345 Fig. 2. Detection sensitivity of mAb 1B4 to Salmonella enterica Typhimurium by indirect ELISA 43 A B C D Fig. 3. Photograph of biosensor captured S. enterica Typhimurium at 10 8 cfu/ml (A), 10 7 cfu/ml (B), 10 6 cfu/ml (C), and blank(D) 44 Table 1. Specificity of mAbs and pAbs for S. enterica Typhimurium assayed by indirect ELISA (n=6, mean ? SD) Absorbance 405nm Bacteria mAb 1B4 pAb S. enterica Typhimurium 4.214 ? 0.089 4.248 ? 0.051 S. enterica Paratyphi 3.788 ? 0.116 3.972 ? 0.063 S. enterica Enteritidis 0.023 ? 0.003 0.430 ? 0.040 S. enterica Mission 0.010 ? 0.003 2.519 ? 0.098 S. enterica Montevideo 0.038 ? 0.005 0.185 ? 0.006 E. coli 43895 0.024 ? 0.004 0.140 ? 0.012 E. coli 932 0.043 ? 0.005 0.177 ? 0.050 E. coli 48-2 0.031 ? 0.001 0.125 ? 0.002 E. coli O157:H7 204P 0.029 ? 0.001 0.391 ? 0.008 Staphylococcus aureusATCC12600 0.092 ? 0.002 0.185 ? 0.006 Listeria innocua 0.011 ? 0.002 0.129 ? 0.006 L. monocytogenes Scott A 0.012 ? 0.003 0.097 ? 0.011 L. monocytogenes G3982 4b 0.022 ? 0.001 0.087 ? 0.080 Table 2. The binding of Salmonella enterica Typhimurium* on mAb 1B4 immobilized biosensor platforms at various mAb concentrations. mAb Concentration (?g/ml) Salmonella 24 12 6 3 1.5 0.75 0.375 0 Bound bacteria 467 378 300 214 45 1 0 0 *The bacterial concentration was 1 x 10 8 cfu/ml. 45 Table 3. Numbers of bacteria bound to biosensors at various S. enterica Typhimurium concentrations as well as mixed with E. coli O157:H7. Bacterial solution S. enterica Typhimurium S. enterica Typhimurium S. enterica Typhimurium mixed (cfu/ml) with 1?10 8 cfu/ml E. coli O157:H7 1?10 8 467 467 1?10 7 46 40 1?10 6 8 6 1?10 5 3 4 1?10 4 2 2 1?10 3 2 2 1?10 2 1 1 1?10 1 0 1 0 0 0 Number in this table is the mean of six replications 46 47 CHAPTER IV RAPID DETECTION OF SALMONELLA IN MILK USING A BIOSENSOR ABSTRACT In the previous study, a Salmonella Typhimurium-specific biosensor with a detection limit of 1?10 2 cfu/ml in pure culture and in a mixture with E. coli was developed. Due to the low number of Salmonella in foods, a one step enrichment in different non-selective enrichment or selective enrichment media with different incubation time, in combination with separation and concentration of bacteria by centrifugation, was evaluated for the rapid detection of S. enterica Typhimurium in milk. Among the seven preenrichment media (tryptic soy broth, half strength tryptic soy broth, lactose broth, nutrient broth, universal preenrichment broth, buffered peptone water, and brain heart infusion broth) tested, S. enterica Typhimurium grew best in brain heart infusion broth. Among the five selective enrichment media (selenite cystine borth, selenite broth, Rappaport-Vassiliadis R10 broth, Salmosyst broth, and brilliant green broth), Salmonella Typhimurium grew best in brilliant green broth. Therefore, brain heart infusion broth and brilliant green broth were selected for further study. S. enterica Typhimurium inoculated milk (0.1, 1, 10 cfu/ml) was enriched (1:9 milk to broth ratio) in the two media for 2, 4, and 6 h. Enrichment in the two media gave similar results. In combination with centrifugation, the bacterial number was able to reach the biosensor detection limit in 4 h for milk samples inoculated with 0.1 cfu/ml Salmonella 48 Typhimurium and in 2 h for milk samples inoculated with 1 cfu/ml S. enterica Typhimurium. Centrifugation was able to increase bacterial concentration to biosensor detection limits without enrichment for milk containing 10 cfu/ml Salmonela Typhimurium. Different sensor platforms, including glass, polystyrene, polyvinyl chloride, and glass coated with gold were compared in this study. Among these platforms, gold showed better consistency regarding sensitivity and more even distribution of captured bacteria. The surface of the gold platform was further processed with a self assembly monolayer (SAM) using 1mM C 11 H 22 O 2 S and 1 mM C 18 H 17 N 3 .HCl. Compared with the gold coated platform, SAM did not improve the detection sensitivity. The light microscopic images showed that milk content did not interfere with the detection of Salmonella using the developed biosensor. Keywords: Salmonella Typhimurium, enrichment, centrifugation, biosensor 49 INTRODUCTION Outbreaks of salmonellosis associated with milk continue to occur (Ryan and others 1987, Jay 2000, Olsen and others 2004). Several rapid Salmonella detection methods have been developed for the detection of Salmonella in foods. These methods include polymerase chain reaction (PCR), DNA hybridization (DNAH), enzyme-linked immunosorbent assay (ELISA), and biosensors. The detection sensitivity of these rapid methods is about 10 5 to10 7 cfu/ml in foods. Due to the low incidence of Salmonella in foods, preenrichment, enrichment, and sometimes even a post-enrichment were involved in the rapid detection methods (Flowers and others 1988). One step enrichment has been investigated prior to PCR, ELISA, and biosensor assays for the detection of pathogens in foods. Kwang and others (1996) used 4-6 h enrichment in a non-selective medium, trypticase soy broth (TSB), for the detection of inoculated S. enterica Typhimurium in ground beef before conducting PCR. One step enrichment in tetrathionate broth (TT) for 18 h was employed by Eyigor and others (2002) for the detection of inoculated S. enterica Enteritidis in poultry samples before PCR. Buffered peptone water (BPW), lactose broth (LB), and selenite cystine broth (SC) were used in one step enrichment (6 h) for the detection of inoculated Salmonella in chicken meat before conducting PCR (Fluit and others 1993). Preenrichment in nutrient broth (NB) followed by enrichment in Rappaport- Vassiliadis (RV) medium was used before ELISA for the detection of S. enterica Typhimurium (Prusak-Sochaczewski and Luong 1989). The total detection time required by their method was about 24 h. Preenrichment in BPW and then enrichment in Salmosyst before ELISA (Salmonella-TEK test kit) was regarded as the quickest method 50 by the author, which took 31 h for the total analysis time, by van Poucke (1990). One step enrichment in non-selective medium, brain heart infusion broth (BHI) for 14 to 16 h before immunoassay was evaluated by NG and others (1996). The total assay time required by their method was 30 h and the detection sensitivity of their developed sandwich ELISA was 10 5 to 10 7 cfu/ml. The preenrichment media recommended by US Food and Drug Administration (USFDA 2003) for Salmonella isolation and identification in various foods include brilliant green water (BGW), lactose broth (LB), nutrient broth (NB), skim milk, brain heart infusion broth (BHI), trypticase soy broth (TSB) and universal preenrichment broth (UPB). LB has been recommended for the preenrichment of liquid milk (skim milk, 2% fat milk, whole, and butter milk) for 24 ? 2 h by the USFDA (2003). Lack of unanimity on the selection of media for one step enrichment is still a problem. Methods for separation and concentration of Salmonella from foods using immunomagnetic separation (IMS), filtration, or centrifugation have been investigated. IMS uses antibody coated magnetic particles to separate and concentrate microbes from foods. This technique has been investigated before performing PCR (Fluit and others 1993, Cerro and others 2002), used as an alternative to enrichment broth in conventional culture method (Mansfield and Forsythe 1993), and used in between preenrichment and enrichment steps (Cudjoe and Krona 1997). The membrane filtration technique for the capture of microbes has been investigated since 1951 (Goetz and Tsuneshi). The hydrophobic grid membrane filter (HGMF) technique has been investigated for bacterial enumeration since 1974 (Sharpe and Michaud). Centrifugation has been studied as a means of concentrating microbes by researchers and is regarded as an effective way of 51 concentrating microbes (Hawa and others 1984, McElroy and others 1995, McElroy and others 1996, Cerro and others 2002). Although various biosensors have been developed and regarded as highly sensitive with a detection limit of about 10 3 to 10 7 cfu/ml (Bokken and others 2003, Taitt and others 2004, Yang and others 2001), few of them could be successfully used for the detection of pathogens in foods, due to the complex nature of the food matrix (Prusak- Sochaczewski and others 1990). The objectives of this study were to select either a preenrichment or an enrichment media that could be used in one step enrichment for rapid culturing of inoculated S. enterica Typhimurium in milk; to evaluate centrifugation as a microorganism concentration method; to evaluate the performance of different sensor platforms; to evaluate the detection sensitivity of the developed biosensor for the detection of inoculated S. enterica Typhimurium in milk; and to compare biosensor and indirect sandwich ELISA. MATERIAL AND METHODS Milk inoculation and one step enrichment Pasteurized Vitamin D whole milk (Barbar?s?) was purchased from a local grocery store in Auburn. Milk was cooked in a water bath at 98 ?C for 5 min to kill background microorganisms. Non-selective enrichment media, which included tryptic soy broth (TSB), half strength TSB (1/2 TSB), lactose broth (LB), nutrient broth (NB), universal preenrichment broth (UPB), buffered peptone water (BPW), and brain heart infusion broth (BHI); and five enrichment broth, which included brilliant green broth (BG), selenite broth (SB), selenite cystine broth (SC), Rappaport-Vassiliadis R10 broth 52 (RV), and Salmosyst broth, were prepared according to manufacturer?s recommendation. For comparing the efficiency of different non-selective enrichment and selective enrichment media, washed overnight culture of Salmonella Typhimurim was diluted in phosphate buffered saline (PBS) to 10 4 cfu/ml. Ten ml of the diluted bacterial culture was inoculated into 90 ml of milk in a sterile milk dilution bottle and mixed well to give bacterial concentration of 10 3 cfu/ml in milk. Then, 2 ml of the inoculated milk were added to 18 ml of each of the seven preenrichment and enrichment media (1:9 milk to broth ratio) and incubated in a gyratory water bath at 37 ?C for 6 h with constant shaking. After incubation the bacterial number was checked by plating on bismuth sulfite agar. The background bacterial number of milk was checked by plating on bismuth sulfite agar and plate count agar. After the selection of the two optimal non-selective enrichment and selective enrichment media, direct enrichment in the selected media, in combination with centrifugation, was used to determine the incubation time needed to bring bacterial concentration to the detection limit of the biosensor. S. enterica Typhimurium was inoculated into milk at 10, 1, and 0.1 cfu/ml. Then 25 ml of the inoculated milk was added to 225 ml of each of the selected media and incubated at 37 ?C for 2, 4, and 6 h. The bacteria number in the media was checked after 2, 4, and 6 h incubation. Centrifugation Centrifugation was used to concentrate bacteria from one step enriched media. Fifty ml of enriched milk at each incubation period (2, 4, and 6 h) were added into a 50 ml sterile polypropylene conical centrifuge tubes (Fisher Scientific, Suwanee, GA) and centrifuged at 5,000 ? g for 10 min at 4 ?C. The supernatant was aspirated and the pellet 53 was resuspended in 1 ml PBS. The bacterial concentration in the resuspended pellet was checked by plating on bismuth sulfite agar. Indirect sandwich ELISA Each well of a 96-well polystyrene microtiter plate (Costar, Cambridge, MA) was coated with 100 ?l of 6 ?g/ml mAb 1B4 and incubated at 37 ?C for 1hr. The plate was washed 3 times with 200 ?l of PBST, then the unbound areas of the wells were blocked with 200 ?l of 1% BSA-PBS at 37 ?C for 1 h. After the plate was washed once with PBST, 100 ?l of inoculated milk containing Salmonella Typhimurium 10 time serial diluted from 10 2 to 10 8 cfu/ml were added to each well and incubated at 37 ?C for 1 h. The plate was washed three times followed by adding 100 ?l of pAb S48 (10 ?g/ml) to each well and incubating at 37 ?C for 1 hr. After the plate was washed three times, 100 ?l of 1:3000 diluted goat anti-rabbit IgG alkaline phosphatase conjugate (GAR-AP) was added to each well and incubated at 37 ?C for 1 hr. After the plate was washed 4 times, 100 ?l of p-nitrophenyl phosphate (p-npp, Sigma Chemical Co, St. Louis, MO) solution (4 mg/ml) was added to each well and incubated at room temperature in darkness for 15 min. Absorbance was measured at 405 nm by a microplate reader (ThermoLabsystems, Helsinki, Finland). Different sensor platform comparison Four sensor platforms, which include glass, polystyrene, polyvinyl chloride, and gold, were compared. Glass platforms were prepared by cutting microscope cover glass (0.17 mm thickness) into 5 ? 5 mm 2 . Polystyrene platforms were prepared from flat polystyrene plate (Costar). Polyvinyl chloride platforms were cut from the bottom of microtiter plate (Costar). Gold platforms were prepared as described in the previous 54 study. Briefly, 100 nm of gold was sputter-coated on the surface of microscope cover glass, which was cut into 5 ? 5 mm 2 . The mAb 1B4 was immobilized on the four sensor platforms by incubating each of the sensor platforms in 100 ?l of 24 ?g/ml 1B4 solution at room temperature for 2 h. Then the sensor platform was washed 3 times with PBST, and incubated with 1% BSA-PBS for 30 min to block unbound areas. After washing 3 times with PBS, the platforms were either used immediately, or dried and stored at 4 ?C till use. Preparation of self-assembly monolayer (SAM) The gold platform was immersed in 1 mM 3-Mercaptopropionic acid 2-ethylhexyl ester (C 11 H 22 O 2 S) dissolved in deoxygenated anhydrous ethanol (SAM solution) at room temperature for 16 h to form SAM on the surface. After the platform was removed from the SAM solution, it was thoroughly rinsed in anhydrous ethanol and then dried in a stream of nitrogen. The dried platform was immediately immersed in 1 mM 1-Ethyl-3-(3- dimethylaminopropyl)-carbodiimide (EDC, C 8 H 17 N 3 .HCl) dissolved in acetonitrile at room temperature for 8 h to chemically activate the tail groups (carboxy groups) of SAM on the gold surface. After the platform was removed from EDC solution and rinsed thoroughly in acetonitrile and dried in a stream of nitrogen, mAb 1B4 was immobilized on SAM surface using the procedures described above. Detection of S. enterica Typhimurium in inoculated milk using developed biosensor Each of the prepared sensor platforms was incubated with 100 ?l of milk inoculated with S. enterica Typhiumrium from 10 1 to 10 8 cfu/ml in a square well microtiter plate at room temperature for 1 h. To evaluate the effect of the background bacteria on the detection of S. enterica Typhimurium using the biosensor, E coli was 55 inoculated into milk at 10 8 cfu/ml before inoculating different concentrations of Salmonella into milk. The unbound bacteria were removed by washing the sensor platform three times with PBST then four times with DI water. After the biosensors were dried at room temperature, the bacteria captured by the biosensor were fixed with OsO 4 for 1 hr and photographed using a Nikon Eclipse L150 Industrial light microscope (Nikon Instruments Inc., Melville, NY) with 1,000 times magnification. Each image represents 0.025 mm 2 area of the biosensor. The experiment was repeated five times, and the number of the captured bacteria on the representative microscopic images was recorded. The bacterial number on each field was counted and the detection of Salmonella was defined as positive when one or more bacteria were observed on each of the images. RESULTS AND DISCUSSION Media comparison In order to rapidly bring low Salmonella concentration in milk to the detectable level, one step enrichment in different non-selective and selective media was compared. The non-selective enrichment media included tryptic soy broth (TSB), half strength TSB (1/2 TSB), lactose broth (LB), nutrient broth (NB), universal preenrichment broth (UPB), buffered peptone water (BPW), and brain heart infusion broth (BHI). The selective enrichment media included brilliant green broth (BG), selenite broth (SB), selenite cystine broth (SC), Rappaport-Vassiliadis R10 broth (RV), and Salmosyst broth. Barbar?s? Vitamin D whole milk was inoculated with S. enterica Typhimurium at 10 3 cfu/ml. Due to the 1:9 ratio of milk to broth, the initial bacterial concentration in broth was 10 2 cfu/ml before incubation. The bacterial number in the media was checked by 56 plating on bismuth sulfide agar after 6 h incubation. Among the seven non-selective enrichment media, Salmonella grew best in BHI, in which the Salmonella number increased from 1 ? 10 2 cfu/ml to 3.7 ? 10 6 cfu/ml after 6 h incubation (Figure1). Among the five selective enrichment media, Salmonella grew best in BG, in which Salmonella number increased from 1 ? 10 2 cfu/ml to 4.8? 10 6 cfu/ml after 6 h incubation. Based on this result, BHI and BG were selected for further study. Efficacy of one step enrichment in BG and BHI BG, selective medium, and BHI, non-selective medium, have similar efficacy for the enrichment of Salmonella in this study (Table 1). When S. enterica Typhimurium was inoculated in milk at 0.1 cfu/ml, the initial bacterial number in broth was 0.01 cfu/ml due to 1:9 ratio of milk to broth. Enrichment in BG and BHI for 4 h increased Salmonella number from 0.01 cfu/ml in broth to less than 10 cfu/ml in broth. Six hour enrichment in broth increased Salmonella number to over 10 2 cfu/ml, which is the detection limit of the developed biosensor as determined in Chapter IV. When S. enterica Typhimurium was inoculated in milk at 1.0 cfu/ml, Salmonella number increased from 0.1 cfu/ml in broth to over 10 2 cfu /ml in broth within 4 h enrichment in both BG and BHI. When S. enterica Typhimurium was inoculated in milk at 10 cfu/ml, Salmonella number increased from 1 cfu/ml in broth to over 10 3 cfu/ml in 4 h enrichment in both BG and BHI. Concentration of Salmonella from one step enrichment media by centrifugation Fifty ml of enrichment media at 2, 4, and 6 h incubation was centrifuged and resuspended in 1 ml PBS. Theoretically, Salmonella should be concentrated 50 times. By comparing the bacterial concentration before centrifugation and after centrifugation, we found that Salmonella concentration increased 30 to 50 times by centrifugation (Table 2). 57 The difference between the theoretical and the actual concentration times may be due to the loss of bacteria during aspiration of the supernatant and/or inaccurate pellet resuspension volume. Comparison of different sensor platform The four different sensor platforms, gold, polyvinyl chloride (PVC), polystyrene (PS), and glass, were used in this study. When different concentrations of Salmonella (10 to 10 8 cfu/ml) were tested using 4 different sensor platforms, the gold platform performed better than the other three types of platforms (Table 3). PVC and PS, which are the commonly used solid phases in immunoassays, did not show better antibody immobilization capability than the gold platform. In the PVC platform, bacteria tended to bind to the edge of the platform, which made it difficult to focus and visualize the captured bacteria. The SAM did not increase the detection sensitivity of the gold platform. Indirect sandwich ELISA Indirect sandwich ELISA was tested against Salmonella inoculated milk with or without an excessive amount of E. coli (10 8 cfu/ml). The ELISA response (O.D. 405 nm reading) decreased rapidly (from above 2 to 0.6) when inoculated Salmonella concentration decreased from 10 8 to 10 5 cfu/ml (Figure 2). ELISA response gradually decreased from 0.6 to around 0.4 when Salmonella concentration decreased from 10 5 to 10 2 cfu/ml. The detection limit of ELISA, defined as the mean O.D. value of the negative control (milk or E. coli inoculated milk without Salmonella) plus 2 standard deviations, was 10 2 cfu/ml. The presence of 10 8 cfu/ml E. coli in milk did not interfere with the detection of Salmonella by indirect sandwich ELISA. 58 Sensitivity of Biosensor The biosensor was used to test against Salmonella inoculated milk with or without excessive amounts of E. coli (10 8 cfu/ml). The number of the captured Salmonella dramatically decreased when inoculated Salmonella concentration decreased from 10 8 to 10 5 cfu/ml, then gradually decreased when inoculated Salmonella concentration decreased from 10 5 to 10 2 cfu/ml (Table 3). These results have shown good correlation with the results of the indirect sandwich ELISA. Although small particles of milk content sometimes were visible under the light microscope, their shape and size were distinctly different from captured Salmonella. Therefore, these particles would not affect the enumeration of Salmonella from the biosensor. Excessive amount of E. coli did not interfere with the detection sensitivity of Salmonella by the biosensor. When the biosensor was used to test the enriched samples, 6 h incubation was needed to bring Salmonella concentration from 0.1 cfu/ml to the detectable level (> 10 2 cfu/ml) of the biosensor. However, when the biosensor was used to test the enriched plus centrifuged samples, 4 h incubation was needed to bring Salmonella concentration from 0.1 cfu/ml to the detectable level of the biosensor. The ingredients of the culture media and milk content did not interfere with the detection of Salmonella by the biosensor. Centrifugation step removed most of the milk content and resulted in clean background on the biosensor. Since 1 h was needed to perform biosensor testing, the entire time needed for detecting 0.1 cfu/ml Salmonella in milk was 5 h when centrifugation was employed to concentrate Salmonella from enrichment broth. 59 CONCLUSION Both BG and BHI are efficient for the enrichment of Salmonella prior to detecting Salmonella using the developed biosensor. Six hour enrichment in either BG or BHI would be sufficient for the detection of 0.1 cfu/ml Salmonella in milk by the developed biosensor. The centrifugation technique used in this study was able to concentrate Salmonella up to 50 times. Through a one step 4 h enrichment in BG/BHI followed by centrifugation, the biosensor was able to detect 0.1 cfu/ml Salmonlla in milk in 5 h. The indirect ELISA results correlated well with the biosensor results. 0 10 20 30 40 50 TS B 1/ 2 T S B L B NB UP B BP W BH I SC S B RV S a l m o syst BG Preenrichment and enrichment media B a ct er i a l n u m b er ( x 10 5 cf u / m l ) Figure 1. Comparion of different preenrichment and enrichment media for the recovery of inoculated S. enterica Typhimurium in milk after 6 h enrichment. TSB: Trypticase soy broth, ? TSB: half strength TSB, LB: lactose broth, NB: nutrient broth, UPB: universal preenrichment media, BPW: buffered peptone water, BHI: brain heart infusion broth, SC: selenite broth, SB: selenite broth, RV: Rappaport-Vassiliadis R10 broth, Salmosyst: salmosyst broth, and BG: brilliant green broth. 60 0.00 0.50 1.00 1.50 2.00 2.50 02468 Concentration of S. enterica Typhimurium (log 10 cfu/ml) O . D . 40 5 n m Salmonella in milk Salmonella in mik with E. coli Figure 2. Detection sensitivity of indirect sandwich ELISA using mAb 1B4 as capture antibody. E. coli was inoculated at 10 8 cfu/ml in milk. 61 A B C D Figure 3. Light microscopic images of captured S. enterica Typhimurium in milk, enrichment broth, and centrifuged samples. A: inoculated milk, B: enrichment broth, C: centrifuged sample, D: milk negative control. 62 Table 1. Efficacy of one step enrichment at 2, 4, and 6 h incubation period before or after centrifugation. Bacterial concentration (cfu/ml) Inoculation Media dose (cfu/ml) Without centrifugation With centrifugation 2 h 4 h 6 h 2 h 4 h 6 h 0.1 BG ND* 8 5.7?10 2 ND 2.7?10 2 2.0?10 4 BHI ND 3 1.6?10 2 ND 1.6?10 2 6.3?10 3 1.0 BG 4 1.5?10 2 1.1?10 4 1.2?10 2 7.8?10 2 3.8 ? 10 5 BHI 2 1.6?10 2 1.0?10 4 1.1?10 2 6.2?10 2 3.8 ? 10 5 10 BG 25 1.8?10 3 8.1?10 4 8.3?10 2 7.1?10 4 4.0 ? 10 6 BHI 11 1.4?10 3 6.5?10 4 5.6?10 2 6.1?10 4 3.2 ? 10 6 ND*: not detectable using pour or spread plate method 63 Table 2. Bacterial concentration efficacy of centrifugation. Enrichment Time Inoculation dose Media 2 h 4 h 6 h 0.1 cfu/ml BG ND* 34 35 BHI ND 50 38 1.0 cfu/ml BG 30 52 38 BHI 50 39 38 10 cfu/ml BG 33 40 48 BHI 53 43 49 Concentration cofactor shown in this table was calculated through dividing the bacterial concentration after centrifugation by the bacterial concentration before centrifugation. ND*: Because the bacterial concentration before centrifugation was not detectable using pour or spread plate methods, the ratio of the bacterial concentration after centrifugation and the bacterial concentration before centrifugation could not be calculated. 64 Table 3. Comparison of four different sensor platforms Sensor platform S. enterica Typhimurium (cfu/m) Gold Glass PS PVC SAM 1?10 8 467 295 360 237 430 1?10 7 46 20 38 33 48 1?10 6 8 2 10 11 7 1?10 5 3 2 5 5 3 1?10 4 2 1 1 0 1 1?10 3 2 0 1 0 1 1?10 2 1 0 1 0 1 1?10 1 0 0 0 0 0 0 0 0 0 0 0 Bacterial number shown in this table is mean of six replications. PS: polystyrene, PVC: polyvinyl chloride 65 66 CHAPTER V. OVERALL CONCLUSIONS Development of a sensitive and specific biosensor requires highly specific and sensitive monoclonal antibodies. Most of the currently developed antibodies are against either O-antigen or H-antigen. They suffer from different degrees of cross reactions with other enteric bacteria. False positive and false negative results have been reported using the O- or H- antigen specific antibodies. Therefore, the development of highly specific antibodies against other antigens needs to be investigated. The 55 kD OMP of S. enterica Typhimurium was extracted and used as an antigen for the production of Salmonella specific antibodies. This is the first time the use of the 55 kD OMP of S. enterica Typhimurium for specific antibody production has been reported. The mAb 1B4 developed in this study has high specificity and high sensitivity to S. enterica Typhiumrium. The biosensor (an immunosensor combined with microscopic imagining systems) developed in this study can specifically detect S. enterica Typhiumrium in 1 h with a sensitivity of 1?10 2 cfu/ml. Therefore, the entire detection method is rapid, cheap, and simple. In order to culture the relatively low number of Salmonella in milk, 7 different non-selective and 4 selective enrichment media were investigated. Among these, BHI and BG were selected as the best non-selective and selective enrichment media respectively. A six hour enrichment in either BG or BHI was able to rapidly increase S. enterica 67 Typhimurium concentration from 0.01 cfu/ml to over 10 2 cfu/ml in broth, which could be detected by the biosensor. The centrifugation technique used in this study was able to concentrate bacteria 50 times. In conjunction with centrifugation, 4 h enrichment was sufficient to bring 0.1 cfu/ml inoculated S. enterica Typhimurium to detectable level. Unlike other biosensors, the detection sensitivity and specificity of this biosensor was not affected by milk content. Further work should focus on the detection of stressed Salmonella in milk and other foods, such as meat, poultry and vegetables. Naturally contaminated foods should be tested using the developed biosensor. The test results of the biosensor should be confirmed by traditional culture methods to verify accuracy. Automation of the detection process using the developed biosensor should be investigated. 68 CHAPTER VI. REFERENCES Allos BM, Moore MR, Griffin PM, Tauxe RV. 2004. Surveillance for sporadic foodborne disease in the 21st century: the FoodNet perspective. Clin Infect Dis 38(Suppl 3):S115-- 20. Anonymous. 2004. Preliminary FoodNet data on the incidence of infection with pathogens transmitted commonly through food ? selected sites, United States, 2003. Morb Mortal Wkly Rep 53:338-343. Andrew WH. 1985. A review of culture methods and their relation to rapid methods for the detection of Salmonella in foods. Food Tech 39:77-82. Barrow PA. 1994. Serological diagnosis of Salmonella serotype enteritidis infections in poultry by ELISA and other tests. Int J Food Microl 21:55-68. Babacan S, Pivarnik P, Letcher S, Rand A. 2002. Piezoelectric flow injection analysis biosensor for the detection of Salmonella Typhimurium. J Food Sci 67:314-320. Bokken GCAM, Corbee RJ, van Knapen F, Bergwerff. 2003. Immunochemical detection of Salmonella group B, D and E using an optical surface plasmon resonance biosensor. FEMS Microbiol Letters 222:75-82. Blackburn CdeW.1993. Rapid and alternative methods for the detection of salmonellas in foods. J Appl Bacteriol 75:199-214. Boer E de. 1998. Update on dedia for isolation of Enterobacteriaceae from foods. Int J Food Microbiol 45:53-53. Brodsky MH, Entis P, Entis MP, Sharpe AN, Jarvis GA. 1982. Determination of aerobic plate and yeast and mold counts on foods using an automated hydrophobic grid- membrane filter technique. J Food Prot 45:301-4. Busse M. 1995. Media for salmonella. Intl J Food Microbiol 26:117-131. Candlish AAG. 1991. Immunological methods in food microbiology. Food Microbiol 8:1-14. 69 Cerro Adel, Soto SM, Landeras E, Gonzalez-Hevia MA, Guijarro JA, Mendoza MC. 2002. PCR-based procedures in detection and DNA-fingerprinting of Salmonella form samples of animal origin. Food Microbiol 19:567-75. CDC. 2005. Preliminary FoodNet data on the incidence of infection with pathogens transmitted commonly through food?10 sites, United States, 2004. MMWR 54(14):352- 356. Choi D, Tsang RSW, Ng MH. 1992. Sandwich capture ELISA by a murine monoclonal antibody against a genus-specific LPS epitope for the detection of different common serotypes of salmonellas. J Appl Bacteriol 72:134-8. Coon AH, Creech HJ, Jones RN, Berliner E. 1942. The demonstration of pneumococcal antigen in tissues by the use of fluorescent antibody. J Immunol 45:159-70. Craigie J. 1931. Studies on serological reactions of flagella of B. typhosus. J Immunol 21:417-511. Cudjoe KS, Krona R. Detection of Salmonela from raw food samples using Dynabeads ? anti-Salmonella and a conventional reference method. Int J Food Microbiol 37:55-62. Curiale MS, McIver D, Weathersby S, Planar C. 1990. Detection of salmonella and other enterobacteriaceae by commercial Deoxyrobonucleic Acid Hybridization and Enzyme Immunoassay Kits. J Food Prot 53:1037-46. D?Aoust JY, Maishment C. 1979. Preenrichment conditions for effective recovery of Salmonella in foods and feed ingredients. J Food Prot 42:153-7. D?Aoust JY, Sewell A, Boville A. 1983. Rapid cultural methods for detection of Salmonella in foods and feed ingredients. J Food Prot 46:851-6 De Smedt J, Bolderdijk R, Rappold H, Lautenschlaeger D. 1986. Rapid Salmonella detection in foods by motility enrichment on a Modified Semi-solid Rappaport- Vassiliadis medium. J Food Prot 49:510-4. Duguid JP, GilliesRR. 1958. Fimbriae and haemagglutinating activity in Salmonella, Klebsiella, Proteus and Chrombacterium. J Pathol Bacteriol 75:519-20. El-Gazzar RE, Marth EH. 1992. Foodborne disease: investigative procedures and economic assessment. J Environ Health 55:24-27. Engvall E, Perlmnn. 1971. Enzyme-linked immunosorbent assay (ELISA). Quantitative assay of immunoglobulin G. Immunochemistry 8:871-4. 70 Eyigor A , Carli KT, Unal CB. 2002. Implementation of real-time PCR to tetrathionate broth enrichment step of Salmonella detection in poultry. Lett Appl Microbiol 34:37-41. Fitts R. 1985. Development of a DNA-DNA hybridization test for the presence of Salmonella in foods. Food Tech 39:95-101. Fitts R, Diamond M, Hamilton C, Neri M. 1983. DNA-DNA hybridization assay for detection of salmonella in foods. Appl Environ Microbiol 46:1146-51. Flowers RS, Klatt MJ, Robinson BJ, Mattingly JA. 1988. Evaluation of abbreviated enzyme immunoassay method for detection of Salmonella in low-moisture foods. J Assoc Off Anal Chem 71:341-3. Fluit AC, Widjojoatmodjo MN, Box ATA, Torensma R, Verhoef J. 1993. Rapid detection of Salmonella in poultry with the magnetic immuno-polymerase chain reaction assay. Appl Environ Microbiol 59(5):1342-6. Georgala DL, Boothroyd M 1964. A rapid immunofluorescence technique for detecting salmonella in raw meat. J Hyg62:319-327. Goetz A, Tsuneshi N. 1951. Applications of molecular filter membranes to the bacteriological analysis of water. J Am Water Works Assn 43:943-5. Goodridge L, Griffiths M. 2002. Reporter bacteriophage assays as a means to detect foodborne pathogenic bacteria. Food Res Int 35: 863-870. Goossens H, Wauters G, De Boeck M, Jansen M, Butzler JP. 1984. Semisolid selective- motility enrichment medium for the isolation of salmonella from fecal specimens. J Clin Microbiol 19:940-1. Gruenenwald R, Henderson RW, Yappow S. 1991. Use of Rambach propylene glycol containing agar for identification of Salmonella spp. J Clin Microbiol 29:2354-6. Hadfield SC, Lane A, McIllmurray MB. 1987. A novel coloured latex test for the detection and identification of more than one antigen. J Immunol Methods 97:153-8. Haglund JR, Ayres JC, Paton AM, Kraft AA, Quinn LY. 1964. Detection of Salmonella in eggs and egg products with fluorescent antibody. Appl Microbiol 12:447-450. Harlow E, Land D. 1988. Antibodies-A Laboratory Manual. Cod Spring Harbor, NY: Cold Spring Harbor Laboratory. p139-243. Hawa SG, Morrison GJ, Fleet GH. 1984. Method to rapidly enumerate Salmonella on chicken carcasses. J Food Prot 47:932-6. 71 Hirsh DC, Martin LD. 1983. Detection of Salmonella spp. in milk by using Felix-O1 bacteriophage and high-pressure liquid chromatography. Appl Environ Microbiol 6:1243- 5. Hirsh DC, Martin LD. 1984. Rapid detection of Salmonella in certified raw milk by using charge-modified filters and Felix-O1 bateriophage. J Food Prot 47:388-390 Ibrahim GF, Fleet GH, Lyons MJ, Walker RA. 1985. Immunological relationships between Salmonella flagellins and between these and flagellins from other species of Enterobacteriaceae. Med Microbiol Immunol 174:101-113. Ibrahim GF, Lyons MJ, Walker RA, Fleet GH. 1985. Rapid detection of salmonellae by immunoassays with titanous hydroxide as the solid phase. Appl Environ Microbiol 50:670-5. Ibrahim GF, Lyons MJ, Walker RA, Fleet GH. 1986. Rapid detection of salmonellae in foods using immunoassay systems. J Food Prot 49:92-8. Ibrahim GF. 1986. A review of immunoassays and their application to Salmonella detection in foods. J Food Prot 49:299-310. Inoue T, Takagi S, Ohnishi A, Tamura K, Suzuki A. 1968. Food-borne disease Salmonella isolation medium (MLCB). The 66 th Annual meeting of the Society Veterinary science of Japan. Jaradat ZW, Zawistowski J. 1996. Production and characterization of monoclonal antibodies against the O-5 antigen of Salmonella typhimurium lipopolysaccharide. Appl Environ Microbiol 62(1):1-5. Jay JM. 2000. Foodborne gastroenteritis caused by Salmonella and Shigella. In: Heldman DR, editor. Modern food microbiology. 6 th ed. Gathersburg, Maryland: Aspen Publishers. P511-530. Jeffries L. 1959. Novobiocin ? tetrathionate broth: A medium of improved selectivity for the isolation of salmonellae from faeces. J Clin Path 12:568-71. Jones TF, Gerber DE. 2001. Perceived etiology of food borne illness among public health personnel. Emerg Infect Dis 7:904-905. Kauffmann F. 1930. Ein kombiniertes Anreicherungsverfahren f?r Typhus- und Paratyphusbazillen. - Zbl. Bakt. I. Orig. 119: 148-152. Kerr S, Ball HJ, Mackis DP, Pollock DA, Finlay DA. 1992. Diagnostic application of monoclonal antibodies to outer membrane protein for rapid detection of salmonella. J Appl Bacteriol 72:302-8. 72 Kim CJ, Nagaraja KV, Pomeroy BS. 1991. Enzyme-linked immunosorbent assay for the detection of Salmonella enteritidis infection in chickens. Am J Vet Res 52(7):1069-74. King S, Metzger WJ. 1968. A new plating medium for the isolation of enteric pathogens. Appl Microbiol 16:577-8. King W, Raposa S, Warshaw J, Johnson A, Halbert D, Klinger JD. 1989. A new colorimetric nucleic acid hybridization assay for Listeria in foods. Int J Food Microbiol 8:225-32. Kisiela D, Kuczkowski M, Wieliczko A, Sambor I, Mazurkiewicz M, Ugorski M. 2003. Comparison of SefA, FimA and AgfA fimbrial proteins of Salmonella Enteritidis in their abilities to elicit humoral immune response in hens. Bull Vet Inst Pulawy 47:95-105. Kohler G., Milstein C. 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature. 256:495-497. Kramer MF, Lim DV. 2004. A rapid and automated fiber optic-based biosensor assay for the detection of Salmonella in spent irrigation water used in the sprouting of sprout seeds. J Food Prot 67:46-52. Krysinski EP, Heimsch RC. 1977. Use of enzyme labeled antibodies to detect Salmonella in foods. Appl and Environ Microbiol. 33:947-54. Kwang J, Littledike ET, Keen JE. 1996. Use of polymerase chain reaction for Salmonella detection. Letters Appl Microbiol 22:46-51. Krysinski EP, Heimsch RC. 1977. Use of enzyme-labeled antibodies to detect Salmonella in foods. Appl Environ Microbiol 33:947-954. Laemmli UK. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-5. Le Minor L. 1984. Salmonella Liginieres. In: Krieg NR, Holt JG, editors. Bergey?s Manual of Systematic Bacteriology. Baltimore, MD: Williams and Wilkins. P 427-458. Leifson E. 1936. Now cuolture media based on sodium deoxycholate for the isolation of intestinal pathogens and for the enumeration of the colon bacillus in milk and water. J Pathol Bacteriol 40:581-599. Leifson E. 1936. New selenite enrichment media for the isolation of typhoid and paratyphoid bacilli. Am J Hyg 24:423-432. 73 Liesegang A, Tschape H. 2002. Modified pulsed-field gel electrophoresis method for DNA degradation ? sensitive Salmonella and Escherichia coli strains. Int J Med Microbiol 291:645-8. Litchfield JH. 1973. Salmonella and the food industry ? Methods for isolation, identification and enumeration. CRC Crit Rev Food Tech 3:415-456. Lu GZ Tsang RSW, Chau PY, Choi D, Law D, Ng MH. 1991. Characterization and application of a murine monoclonal antibody that rects specifically with the serogroup D1 Salmonella. FEMS Microbiol Lett 80:135-140. Luk JMC, Lindberg AA. 1991. Anti-Salmonella lupopolysaccharide monoclonal antibodies: characterization of Salmonella BO-, CO-, DO-, and EO-specific clones and their diagnostic usefulness. J Clin MIcrobiol 29:2424-2433. Lukinmma S, Nakari UM, Eklund M, Siitonen A. 2004. Application of molecular genetic methods in diagnostics and epidemiology of food-borne bacterial pathogens. Acta Pathologica, Microbiologica et Immunologica Scandinavica 112:908-929. Mattingly JA. 1984. An enzyme immunoassay for the detection of all Salmonella using a combination of a myeloma protein and a hybridoma antibody. J Immunol Methods 73:147-156. McElroy AP, Cohen ND, hargis BM. 1995. Evaluation of a centrifugation method for the detection of Salmonella enteritidis in experimentally contaminated chicken eggs. 58:931- 3. McElroy AP, Cohen ND, Hargis BM. 1996. Evaluation of the Polymerase chain reaction for the detection of Salmonella enteritidis in experimentally challenged hens. J Food Prot. 59:1273-8. McClelland RG, Pinder AC. 1994. Detection of Salonella typhimurium in dairy products with flow cytometry and monoclonal antibodies. Appl Environ Microbiol 60:4255-4262. Mead PS, L Slutsker, V Dietz, LF. McCaig, JS. Bresee, C Shapiro, PM. Griffin, and RV. Tauxe. 1999.Food-related illnesses and death in the United States. Emerg Infect Dis 5:607-25. Meenakshi M, Bakshi CS, Butchaiah G, Bansal MP, Siddiqui MZ, Singh VP. 1999. Adjuvanted outer membrane protein vaccine protects poultry against infection with Salmonellla enteritidis. Vet Res Comm 23:81-90. Moats WA. 1981. Update on Salmonella in foods: selective plating media and other diagnostic media. J Food Prot 44:375-380. 74 Mortimer CKB, Peters TM, Gharbia SE, Logan JMJ, Arnold C. 2004. Towards the development of a DNA-sequence based approach to serotyping of Salmonella enterica. BMC Microbiol 4:31-40. Muller L.1923. Un nouveau milieu d'enrichissement pour la recherche du bacille typhique et des paratyphiques. Comp rend Soc Biol 89:434-7. NG SP, Tsui CO, Roberts D, Chau PY, NG MH. 1996. Detection and serogroup differentiation of Salmonella spp. in food within 30 hours by enrichment-immunoassay with a T6 monoclonal antibody capture enzyme-linked immunosorbent assay. Appl Environ Microbiol 62:2294-302. Olsen SJ, MacKinon LC, Goulding JS, Bean NH, Slutsker L. 2000. Surveillance for foodborne disease outbreaks, United States, 1993-1997. Morb Mortal Wkly Rep 49:1-51. Olsen SJ, Ying M, Davis MF, Deasy M, Holland B, Iampietro L, Baysinger CM, Sassano F, Polk LD, Gormley B, Hung MJ, Pilot K, Orsini M, Van Duyne S, Rankin S, Genese C, Bresnitz E, Smucker J, Moll M, Sobel J. 2004. Multidrug-resistant Salmonella Typhimurium infection from milk contaminated after pasteurization. Emerg Infect Dis 10:932-5. Olsen J E, Aabo S, Hill W, Notermans S, Wernars K, Granum PE, Popovic T, Rasmussen HN, Olsvik O. 1995. Probes and polymerase chain reaction for detection of food-borne bacterial pathogens. Int J Food Microbiol 28:1-78. Peralta RC, Yokoyama H, Ikemori Y, Kuroki M, Kodoma Y. 1994. Passive immunization against experimental salmonellosis in mice by orally administered hen egg- yolk antibodies specific for 14-kD fimbriae of Salmonella Enteritidis. J Med Microbiol 41:29-35. Pignato S, Marino AM, Emanuele MC, Iannotta V, Caracappa S, Giammanco G. 1995. Evaluation of new culture media for rapid detection and isolation of Salmonellae in foods. Appl Environ Microbiol 61: 1996-9. Pollock HM, Dahlgren BJ. 1974. Clinical evaluation of enteric media in the primary isolation of Salmonella and Shigella. Appl Microbiol 27:197-201. Prusak-Sochaczewski E, Luong JHT. 1989. An improved ELISA method for the detection of Salmonella typhimurium. J Appl Bacteriol 66:127-135. Prusak-Sochaczewski E, Luong JHT, Guilbault GG. 1990. Development of a piezoelectric immunosensor for the detection of Salmonella typhimurium. Enzyme Microb Technol 12:173-7. 75 Rambach A. 1990. New plate medium for facilitated differentiation of Salmonella spp. from Proteus spp. and other enteric bacteria. Appl Environ Microbiol 56:301-3. Rappaport F, Konforte N, navon B. 1956. A new enrichment medium for certain salmonellae. J Clin Pathol 9:261-6. Rappaport F, Kronforti N. 1959. Selective enrichment medium for paratyphoid bacteria. inhibitory and growth promoting factors. Appl Microbiol 7:63-6. Rappold H, Bolderdijk RF, De Smidt JM. 1984. Rapid cultural method to detect Salmonella in foods. J Food Prot 47:46-8. Ryan CA, Nickels MK, Hargrett-Bean NT, Potter ME, Endo T, Mayer L. 1987. Massive outbreaks of antimicrobial-resistant salmonellosis traced to pasteurized milk. JAMA 258:3269-74. Sharpe AN, Michaud GL. 1974. hydrophobic grid-membrane filters: new approach to microbiological enumeration. Appl Microbiol 28: 223-5. Si S, Ren F, Cheng W. 1997. Preparation of a piezoelectric immunosensor for the detection of Salmonella paratyphi A by immobilization of antibodies on electropolymerized films. Fresenius J Anal Chem 357:1101-5. Silliker JH, Schmall A, Chiu JY. 1966. The fluorescent antibody technique as a mean of detecting salmonellae in foods. J Food Sci 31:240-4. Skepper JN. 2000. Immunocytochemical strategies for electron microscopy: choice or compromise. J Microscopy 199:1-36. Sveum WH, Hartman PA. 1977. Timed-release capsule method for the detection of salmonellae in foods and foods. Appl Environ Mcrobiol 33:630-4. Swaminathan B, Ayres JC. 1980. A direct immunoenzyme method for the detection of almonella in foods. J Food Sci 45:352-5,361. Swaminathan B, Feng P. 1994. Rapid detection of food-borne pathogenic bacteria. Annu Rev Microbiol 48:401-426. Taitt CR, Shubin YS, Angel R, Ligler FS. 2004. Detection of Salmonella enterica serovar Typhimurium by using a rapid, array-based immuosensor. Appl Environ Microbiol 70:152-8. Tate CR, Miller RG, Mallinson ET, Douglass LW, Johnson RW. 1990. The isolation of salmonella from poultry environmental samples by several enrichment procedures using plating media with and without novobiocin. Poult Sci 69:721-6. 76 Taylor WF. 1965. Isolation of shigellae. I. Xylosse-lysine agars: new media for the isolation of enteric pathogens. Am J Clin Pathol 44:471-5. Todd, E. C. D. 1989. Thomason BM, Cherry WB, Moody MD. 1957. Staining bacterial smears with fluorescent antibody. III. Antigenic analysis of Salmonella typhosa by means of fluorescent antibody and agglutination reaction. J Bacteriol 74:525-532. Thorns, C.J. 1995. Salmonella fimbriae: novel antigens in the detection and control of Salmonella infections. Br Vet J 151:643-650. Thorns CJ, Bell MM, Sojka MG, Nicholas RA. 1996. Development and application of enzyme-linked immunosorbent assay for specific detection of Salmonella enteritidis infections in chichens based on antibodies to SEF14 fimbrial antigen. J Clin Microbiol 34:792-7. Towbin H, Staehelin T, Gordon J. 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci U S A. 76(9): 4350-4. Tsang RSW, Chan KH, Chau PY, Wan KC, Ng MH, Schlecht S. 1987. A murine monoclonal antibody specific for the outer core oligosaccharide of Salmonella lipopolysaccharide. Infect Immun 55:211-6 [USFDA] U.S. Food and Drug Administration. 2003. Bacteriological Analytical Manual online. 8 th edition. U.S. Department of Health and Human Services Food and Drug Administration center for food safety and applied nutrition. Available from:www.cfsan.fda.gov/~ebam/bam-toc.html Van der Zee H. 1994. Conventional methods for the detection and isolation of Salmonella enteritidis. Intl J Food Microbiol. 21:41-6. Van poucke LSG. 1990. Salmonella-Tek, a rapid screening method for Salmonella species in food. Appl Environ Microbiol 56:924-7. van Zijderveld FG, van Ziderveld-van Bemmel AM, Anakotta J. 1992. Comparison of four different enzyme-linked immunosorbent assays for serological diagnosis of Salmonella enteritidis infections in experimentally infected chickens. J Clin Microbiol 30:2560-6. van Vunakis H. 1980. Radioimmunoasssays. An overview. Methods Enzymol 70:201- 210. van Weeman BK, Schuurs HWM. 1971. immunoassay using antigen-enzyme conjugates. FEBS Lett 15:232-5. 77 Vassiliadis P, Pateraki E, Papaiconomou N, Papadakis JA, Trichopoulos D. 1976. Nouveau proc?d? d'enrichissement de Salmonell. Ann Microbiol Inst Pasteur 127B:195- 200. Veling J, van Ziderveld FG, van Ziderveld-van Bemmel AM, Barkema HW, Schukken YH. 2000. Evaluation of three newly developed enzyme-linked immunosorbent assays and two agglutination tests for detecting Salmonella enterica subsp. Enterica serovat Dublin infections in dairy cattle. J Clin Microbiol 38:4402-7. Wang H, Blais BW, Brooks BW, Yamazake H. 1996. Salmonella detection by the polymyxin-cloth enzyme immunoassay using polyclonal and monoclonal detector antibodies. Int J Food Microbiol 29:31-40. Wilson WJ, Blair EMMcV. 1927. Use of glucose bismuth sulphite iron medium for the isolation of Bacillus typhosa and Bacillus proteus. J. Hyg 26:374-391. Wu VCH, Fung DYC. 2004. An improved method for a hydrophobic grid-membrane filter system to detect heat-injured foodborne pathogens in ground beef. J Food Sci 69:85-8. Wyatt GM, Langley MN, Lee HA, Morgan MRA. 1993. Further studies on the feasibility of one-day salmonella detection by enzyme-linked immunosrobent assay. Appl Environ Microbiol 59:1384-1390. Yalow RS, Berson SA. 1959. Assay of plasma insulin in human subjects by immunological methods. Nature 184:1648-9. Yang L, Ruan C, Li Y. 2001. rapid detection of Salmonella typhimurium in food samples using a bienzyme electrochemical biosensor with flow injection. J Rapid Methods Automation Microbiol 9:229-240. Ye Y, Letcher SV, Rand AG. 1997. Piezoelectric biosensor for detection of Salmonella typhimurium. J Food Sci 62:1067-1071,1086.