|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
Detection of Pathogenic Bacteria and Fecal Enterococci in Recreational Water With an Evanescent Wave Fiber Optic Biosensor by Maria Theresa Trindade A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Biology College of Arts and Sciences University of South Florida Major Professor: Daniel V. Lim, Ph.D. Valerie J. Harwood, Ph.D. My Lien Dao, Ph.D. Andrew Cannons, Ph.D. Date of Approval: December 15, 2005 Keywords: seawater, enterococcus, enrichment, marine, microorganisms Copyright 2006, M. Theresa Trindade
ACKNOWLEDGEMENT First and foremost, I would like to thank my family without whom I would not have gotten this far. I would especially like to thank my mom and dad for encouraging me and always telling me that I could do anything. I would like to gi ve a big thank you to my husband Chris for his undying optimism, love, and support. Finally, thank you to my brother, Ismael, and sister, Carmen, for providing support and being an unending source of inspiration. Thanks to Dr. Al Dufour at the U.S. Environmental Protection Agency and George P. Anderson at the Navel Research Laboratories for their donation of strains of Enterococcus faecalis and Enterococcus faecium A special thank you to Paul Stanek with the State of Florida Department of Health for his help in collecting water samples, and especially for his many e-mails and phone calls addressing my endless questions. Thanks also to David Wingfield at the Florida Department of Health Tampa Branch Laboratory for providing me with water samples and data for parts of this study. I would like to give a special thank you to all of the members of my graduate committee: Drs. Valerie J. Harwood, My Lien Dao, and Andrew Cannons who all helped me in various ways and were never too busy to give their guidance and advice. Thanks also to all my labmates in the Lim lab, especially Dr. Marianne Kramer for her support and advice and Crystal for her friendship. Also, thanks to Dr. Joyce Simpson, for her help with experiments and tips on writing. Thank you to Drs. Betty Kearns and Allyson Bissing for showing me all about water work. Finally, I would like to thank my mentor and major professor, Dr. Lim for all his guidance, support, and patience. His generosity, advice, and kindness throughout the past five years has been key to my success in graduate school and I am thankful for the opportunity to have been in his lab. Thank you.
i TABLE OF CONTENTS LIST OF TABLES .............iv LIST OF FIGURES .vi ABSTRACT ... .ix INTRODUCTION ..1 Biosensors . ..1 Background .1 Piezoelectric Biosensors ..4 Surface Plasmon Resonance Biosensors.. 5 Colorimetric Biosensors ... 6 Molecular Recognition Elements Us ed in Biosensor Development 7 Evanescent Wave Fiber Optic Biosensors, Immunoassays, and Their Applications ... 13 The Future of Biosensor Technology 16 Indicators of Fecal Contamination 17 Enterococci-Indicators of Fecal Contamination 23 Biochemical Propertie s and Characterization 23 Reservoirs of Entero cocci; Human and Non-human .24 Group D Antigen and Lipoteichoic Acid ..25 Public Health Significance and Us e as an Indicator of Fecal Pollution 28 Methods for the Detection of Enterococci .31 Emerging and Remergin g Waterborne Pathogens 35 Vibrio cholerae .37 Historical Background ...............................................................................38 V. cholerae O139 ...39 Ecology ...40 Virulence Genes and Pathogenesis 41 Public Health Significance and Epidemiology ..42 Detection methods .43 E. coli O157:H7 .44 Staphylococcus aureus ..46 Summary 46 MATERIALS AND METHODS ...48 Bacterial Strains .48 Media, Culture Conditions and Sample Collection ..51
ii Stock Cultures 51 Enrichment and Recovery Media for Enterococci .52 Water Samples ...53 Enrichment and Recovery Methods for E. coli O157:H7 in Seawater 54 Enrichment and Recovery Methods for V. cholerae in Seawater 55 Characterization of Isolates of V. cholerae : API Identification and PCR Amplification 55 Enrichment and Recovery Methods for S. aureus in Seawater 59 Characterization of S. aureus Isolates 60 Antibodies . 62 Sources ...62 Purification Methods ..63 Labeling .67 Screening 68 Analyte 2000 Biosensor 69 Biosensor Sandwich Immunoassays ..70 Sample Preparation 70 Waveguide Preparation ..72 Sensitivity and Specificity Assays .72 Biosensor Indirect Immunoassays for Enterococci 76 Sample Preparation 76 Waveguide Preparation ..76 Sensitivity and Specificity Assays .76 Data Analysis for Biosensor Assays ..78 Seawater Sample Collecti on and Sample Preparation 80 Unspiked Samples ..80 Spiked Samples ..80 Concentration of Seawater with a Hollow Fiber Filter ..81 Concentration, Enrichment, a nd Biosensor Detection of Fecal Enterococci . 84 SDS PAGE and Western Blot Analysis 84 Immunodot Analysis ..85 Scanning Electron Microscopy Analysis ...86 Growth Curve Experiments for Enterococci ..87 RESULTS .. 88 Development of Biosensor Sandw ich Immunoassay for Enterococci ...88 Antibody Specificity ..88 Crossreactivity Assays 92 Affinity Purification of ARP Antibodies ...95 Varying Capture Antibody Concentration 95 Varying detection Antibody Concentration .97
iii Comparison of Broth Media and Plate Media ...99 Different Antigen Preparations Tested for Assay Development 99 Sonication and Boiling Antigen Preparation on Biosensor Sandwich Immunoassay 101 Immunodot Analysis of Soni cated vs. Unsonicated Cells ..103 SEM Analysis ..104 Lectins as Capture Molecules ..106 Optimization of Biosensor Indirect Immunoassays for Enterococci ...108 Concentration of 20 L iter Volumes of Water ..113 Growth Curve Experiments a nd Enrichment of Enterococci ..114 Concentration, Enrichment, and Biosensor Detection 115 Experiments for Determinin g Criteria for Good Quality Water Samples .117 Experiments for Determining Criteria for Moderate Quality Water Samples 119 Experiments for Determining Criteria for Poor Quality Water Samples 121 Unspiked 20 Liter Volumes of Seawater 124 Isolation of E. coli O157:H7 in Seawater 127 Biosensor Immunoassays for the Detection of E. coli O157:H7 in Seawater 127 Isolation of V. cholerae from Seawater 129 Specificity Studies for V. cholerae Detection .131 Different Sample Preparations for V. cholerae Immunoassay Optimization .. 133 Isolation of S. aureus from Seawater ...135 Specificity Assays for S. aureus Detection ..135 Biosensor Sandwich Immunoassays for S. aureus ..137 DISCUSSION .. 139 REFERENCES 158 ABOUT THE AUTHOR End Page
iv LIST OF TABLES Table 1 Criteria for ideal indicator orga nisms and methods for their detection .22 Table 2 Species included in the genus Enterococcus .24 Table 3 Comparison of common methods to detect bacteria by desirable attributes . 31 Table 4 Emerging and reemerging waterborne pathogens .36 Table 5 Comparison of epidemic and non-epidemic strains of V. cholerae ...38 Table 6 Factors affecting the survival and growth of V. cholerae in the environment ..43 Table 7 Bacteria used in Enterococcus assay development 48 Table 8 Bacteria used in V. cholerae assay development ..51 Table 9 PCR primers for the identification of V. cholerae isolates 58 Table 10 PCR protocol for V. cholerae isolates 59 Table 11 Specificity data for an ti-Strep group D antibody (American Research Products) generated by ELISA ..92 Table 12 Comparison of mean signals from five biosensor experiments ...111 Table 13 Mean signals from four experiments where Ent. faecalis cells were incubated at 36C for 30 minutes 113 Table 14 Recovery of Ent. faecalis cells when concentrated with hollow fiber filter 114 Table 15 Pinellas County Department of Health/EPA water quality criteria .116 Table 16 Biosensor/enrichment crit eria for determining water quality ..117
v Table 17 Experiments for the devel opment of criteria to measure Good quality water 118 Table 18 Experiments for the deve lopment of criteria to measure Moderate quality water .120 Table 19 Experiments in deter mining criteria for Poor Category of water samples ..122 Table 20 Results from environmental samples tested using integrated biosensor method 125 Table 21 Comparison of three me thods for the identification of V. cholerae isolates ..129 Table 22 Organisms tested for specificity with Difco polyclonal V. cholerae antibody through ELISA ..132
vi LIST OF FIGURES Figure 1 Biosensor principles as described by Nakamura et al. (185) ...2 Figure 2a Sandwich immunoassay 14 Figure 2b Indirect immunoassay .14 Figure 3 Structure of ribitol and glycerol teichoic acids .26 Figure 4 Structure of the Enterococcus cell wall .27 Figure 5 Map showing Tampa Bay beaches where seawater samples were collected 53 Figure 6 Identification scheme for V. cholerae isolates ..57 Figure 7 Identification scheme for S. aureus isolates ..61 Figure 8 Standard curve of glucose .66 Figure 9 Analyte 2000 Biosensor (Research International; Monroe WA) ..70 Figure 10 Waveguide and cuvette used for biosensor assay ..73 Figure 11 Sandwich immunoassay format for evanescent wave biosensor ...75 Figure 12 Indirect immunoassay format for evanescent wave biosensor ..78 Figure 13 Concentration apparatus . 82 Figure 14 Processing steps for concentrated seawater samples .. 83 Figure 15 Enterococcus antibodies tested for assay development .. 89 Figure 16 Comparison of high affinity Enterococcus antibodies for use in assay development 90
vii Figure 17 Western blot analysis of two polyclonal antibodies for Enterococcus assay . 91 Figure 18 Crossreactivity between ARP antibody and S. aureus ..94 Figure 19 Affinity purification of ARP polyclonal antibody .95 Figure 20 Various antibody capture concentrations for optimization of Enterococcus sandwich immunoassay ..96 Figure 21 Normalized values for the mean signals shown in Figure 20 97 Figure 22 Determining optimal detection antibody concentration for biosensor sandwich immunoassay .98 Figure 23 Normalized values for the mean signals shown in Figure 22 ... .98 Figure 24 Comparison of biosensor signals when Ent. faecalis cells were grown for 18 hours in Trypticase soy broth (TSB) vs.Trypticase soy agar (TSA) ..99 Figure 25 Comparison of antigen preparations for immunoassay optimization ......100 Figure 26 Comparison of boiled cells and viable cells for biosensor detection using a sandwich immunoassay .. .101 Figure 27 Normalized values for the mean signals shown in Figure 26 .. 102 Figure 28 Sonicated cell preparation used for optimization of biosensor sandwich immunoassay ... 103 Figure 29 Normalized values for the mean signals shown in Figure 28 .. 103 Figure 30 Immunodot analysis of sonicat ed and unsonicated preparations of Ent. faecalis cells in PBS ... . .104 Figure 31 Unsonicated Ent. faecalis cells captured on polystyrene waveguide .. 105 Figure 32 Sonicated Ent. faecalis cells captured on polystyrene waveguide .. 106 Figure 33 WGA as capture molecule for Enterococcus sandwich immunoassay ....107 Figure 34 Normalized values for the mean signals shown in Figure 33 ..108
viii Figure 35 Directly adsorbed Ent. faecalis cells in 0.1 M carbonate buffer (pH 9.3) .109 Figure 36 Directly adsorbed Ent. faecalis cells in seawater .110 Figure 37 Ent. faecalis cells directly adsorbed to polystyrene waveguides after a 30 minute incubation at 36C .. 112 Figure 38 Growth curve of Ent. faecalis in mE enrichment broth .. 115 Figure 39 Detection of E. coli O157:H7 in PBS vs. seawater ..... 128 Figure 40 Normalized values for the mean signals shown in Figure 39 ... ...128 Figure 41 Summary of PCR results for the identification of V. cholerae environmental isolates obtain ed from Tampa Bay Beaches ... .131 Figure 42 Comparison of commercial polyclonal antibodies for V. cholerae detection 132 Figure 43 Detection of viable V. cholerae cells and boiled V. cholerae cells ... ...133 Figure 44 Detection of boiled and sonicated V. cholerae cells 134 Figure 45 Normalized values for the mean signals shown in Figure 44 ..134 Figure 46 Screening of S. aureus commercial antibodies 136 Figure 47 Crossreactivity studies with other Staphylococcus species .137 Figure 48 Detection of S. aureus using biosensor sandwich immunoassay 138 Figure 49 Normalized values for the mean signals shown in Figure 48 ... ...138
ix Detection of Pathogenic Bacteria and Fecal Enterococci in Recreational Water With an Evanescent Wave Fiber Optic Biosensor Maria Theresa Trindade ABSTRACT Development of a rapid method for the detection of fecal enterococci and pathogenic microorganisms in beach water wa s attempted utilizing an evanescent wave fiber optic biosensor. Various assay fo rmats including a sandwich immunoassay were tested in the development of a rapid assay. Fluorophore labeled antibodies were used for specific detection of bacteria captured or adso rbed directly to the surface of a polystyrene fiber optic waveguide. Binding of the fluores cent labeled antibody to its specific target or binding of a fluorescent labeled an ti-IgG within 100-1000 nm of the waveguide surface caused excitation of the fluorescent conjugate resulting in a quantifiable signal. Enterococcus faecalis Staphylococcus aureus Escherichia coli O157:H7, and Vibrio cholerae were used as model organisms for bi osensor detection in phosphate buffered saline and seawater. Seawater samples were selectively enriched for the presence of these model organisms, which were later detected on the biosensor. The sensitivity and specificity of the biosensor was examined by testing various assay formats, sample preparations, and molecules for capture and de tection. Finally, an enrichment protocol combined with filter concentration was utilized to enhance detection of low levels of enterococci.
x The fiber optic biosensor has the potential to be a sensitive and specific system for the detection of fecal enterococci. The lower limit of detection in seawater and phosphate buffered saline was 2.8 x 106 CFU/ml. As few as 6 CF U/100ml (0.06CFU/ml) could be detected in seawater following a 14-24 hour enrichment and c oncentration step. Vibrio alginolyticus was found to grow under the same enrichment conditions as the enterococci. V. alginolyticus crossreacted with the polyclonal anti-Strep group D antibody used in the immunoassay at high cell concentrations. Staphylococcus aureus was the only other organism which showed sign ificant cross-reactivity with this antibody. The biosensor was also able to detect other bacterial pathogens in PBS and seawater. The lower limit for detection of E. coli O157:H7 was 3.6 x 105 CFU/ml. The lower limit for detection of Vibrio cholerae O1 was 1.3 x 108 CFU/ml. The antibodies used in these assays were found to crossreact with other gr am negative microorganisms. The biosensor was not able to detect Staphylococcus aureus.
1 INTRODUCTION Biosensors Background. A biosensor is an analytical detection system that exploits the specific nature of a biological recognition element for its target analyte. This definition is broad and encompasses the most basic systems such as over the counter home pregnancy tests and sophisticated platforms to detect bioterrorist agents in complex matrices. The coupling of a specific biological recognition element with its specific ligand can take many forms in biology including antibodyantigen binding, enzyme-substrate reactions, and receptor-ligand binding. These dynamic inte ractions have all been incorporated into biosensor technology for the speci fic detection of a number of biological agents including bacteria, viruses, and toxin molecules ( 63, 69,159). The technology has also been developed for the monitoring of basic elem ents like oxygen (biological oxygen demand; BOD) or phosphate concentrations (44, 78). Biosensors work on the principle that thro ugh specific receptor-ligand interactions, a signal will be generated that is proportional to the number of interactions between the recognition element and the target (Figure 1). A transducer ultimately receives that signal and produces an electronic signal that can be recorded. The coupling mechanisms between the recognition element and the transducer vary. The biological recognition molecules can be covalently attached to the surface of the transducer, passively adsorbed to the transducer surface through hyrdrophobic interactions, linked to each other through
2 molecules such as avidin and biotin, or a prosthetic group attached to the transducer surface can be arranged to bind an apoenzyme (280). Reaction Surface Transducer Electric Signal Biological Recognition Elements Conversion FIGURE 1. Biosensor principles as described by Nakamura et al. (185). Many different biological recognition elements have been tested for use in biosensor experimentation which exploit the specific binding relationships between a receptor and a ligand. These molecules include low molecu lar weight carbohydrates and peptides, synthetic polymers, enzymes, re ceptors, cells, bacteria, and antibodies. They have been incorporated into a variety of different types of biosensors for the specific capture and detection of several target pathogens (viral or bacteria l), toxins, and chemicals like pesticides or herbicides (280). The biosenso r field has expanded to include an array of instruments incorporating intricate technologie s such as fiber optic biosensors, surface Enzymes Antibodies Receptors Organelles Microbes Animal and Plant Cells Nucleic acids Lectins Chemical Electrode Substance Semiconductor Weight Quartz crystal Light microbalance Sound Photo multiplier Heat Photo diode Sound detector Electrical Signals Thermistor
3 plasmon resonance sensors, piezoelectric immunosensors, a nd colorimetric biosensors (280). Biosensor technology has been in devel opment for over 30 years, and the field has seen many changes during this period. The most common sensors have been those used to detect glucose. This was the first type of sensor developed by Clark and Lyons in 1962 (48). This sensor technology has brought much commercial success to the biosensor field with the invention of por table glucose monitors which measure blood sugar concentration. The success of this t echnology sparked research efforts toward making these instruments diverse, versatile, and widely applicable (260). The use of this technology in other fields of science is expanding. One such application is the monitoring of bacterial conta mination in food and water. De veloping sensitive assays is very difficult for many bacterial pathogens such as E. coli O157:H7 because the infectious dose of this organism is very lo w (23). This is true for immunoassays where high affinity antibody can make the difference between a sensitive assay and an assay which does not work at all. Detection can also be difficult because of the limitations on biomolecules used to detect these pathogen s. Antibodies, for example, can be very crossreactive with other similar targets and this can hinder the development of a specific assay. A complex matrix such as seawater or food can further impair assay performance due to inhibitors and debris that inhibit lig and-target interaction. Many systems are still in development to overcome these obstacles, but biosensors show promise in this field. This is evident as biosensors are now being us ed to detect biological warfare agents in substances such as talc-based pow ders and cornstarch (256).
4 Biosensor instruments have evolved from large bulky machines feasible only for laboratory use to small, portable, automated machines that can fit in ones pocket and give rapid results (135). There are still many obstacles facing the further evolution of biosensor technologies, such as cost re duction and development of systems sensitive enough to detect low levels of agents w ithout enrichment or concentration (135). Piezoelectric biosensors. Piezoelectric biosenso rs (PZ) were first described by Shons and Najarian who reported a sensor based on a quartz crystal microbalance (229). A quartz crystal microbalance (QCM) is made up of a thin quartz disc with electrodes plated on its surface. The QCM is piezoelectric so an oscillating field applied to the balance induces an acoustic wave that is propagated throughout the crystal. The frequency of this wave fluctuates in respons e to material deposited on the surface of the crystal. Changes in the resonant frequency are related to the mass built up on the crystal surface. Antibodies can be immobilized on the surface of the crystal and with increased binding of the target, the resonant frequency decreases. Piezoelectric biosensors have been used to detect several viral agents such as herpes virus and hepatitis C virus as well as bacteria such as E. coli O157:H7 and byproducts of bacteria like staphylococcal enterotoxin B (141, 159, 240, 248). Piezoelect ric biosensors have been used for environmental monitoring by immobilizing pr otein on the crystal surface. One such protein, formaldehyde dehydrogenase, ha s been used to successfully assay for formaldehyde without interferenc e from other air pollutants. Th e bioterrorist agent, ricin and herbicides like atrazine have also been assayed with PZ biosensors (32, 101, 181). There are several commercial piezoelectric bios ensors available from companies such as
5 Elchema (Potsdam, NY) and Maxtek (Torrance, CA) which have been used to test for substances like estrogen and organophosphates (29, 102). While PZ biosensors are promising, several issues still must be addressed for this type of sensor to have a wider applicability. Reproducible immobilization of protein on the crystal surface must be achieved as well as elimination of non-specific binding of other proteins to the crystal surface. The issue of reuse of the quartz crystal needs to be addressed as well (193). Surface plasmon resonance biosensors. Surface plasmon resonance (SPR) biosensors are real-time, label-free, optical detection systems for examining the interactions of soluble analyte with immobilized ligand. They are some of the most successful biosensors on the market. SPR occurs when lig ht is reflected off thin metal films placed between a medium of high refractive index a nd a medium of low refractive index. A fraction of the light energy incident at a sharply defined angle can interact with the delocalized electrons in the metal film and reduces the reflected light intensity (94). The precise angle of incidence at which this occurs is determined by several factors, like the refractive index on the backside of the metal film where target molecules can be immobilized and bind ligands in the mobile pha se. An electromagnetic field called an evanescent wave is formed which extends into the lower refractive media on the other side of the metal. The pr opagation of the evanescent wave depends on the refractive index of the thin layer adjacent to the light interface. The evanescent wave decays exponentially with increased distance from the r eaction interface. If binding of the target occurs, the refractive index changes. The result is a change in SPR angle which can be monitored in real-time by detecting changes in the intensity of the reflected light and will
6 produce a sensorgram. Theoretically, the size of the change in SPR signals is directly proportional to the immobilized mass and this change can be used to calculate the concentration of analyte being assayed. The ra te of change of the SPR signal can also be tracked to rate constants for the association and disassociation of a receptor for its target. SPR has been used in many fields of scienc e including medicine and biotechnology. This technology has given much insight into ki netic rate constants for analyte-ligand relationships as well as aiding in the ch aracterization of the mechanism of certain biomolecular interactions (94). Colorimetric biosensors. Colorimetric biosensors rely on an obvious color change to distinguish a positive from a negative result. Modern colorimetric instruments do not rely on the addition of a chromogenic substr ate to provide that color change. Novel polymers called polydiacetylenes (PDAs) ha ve unique color properties which are dependent on several factors including the ex posure of the material to environmental conditions such as heat, mechanical stre ss, or solvents. Substituted PDAs undergo dramatic color changes from blue to red. Bi osensors have been designed to induce the PDA blue to red conversion by specific binding of the biological target (e.g. antigen) and the molecular recognition element which is directly linked to the transducer (PDA). An example of such a biosensor is a system to de tect influenza A virus by the binding of its viral hemagglutinin to sialic acid residues incorporated into the PDA resulting in a color change to red (42, 43). Another colorimetric biosensor has been developed based on the red color of gold nanoparticles for the dete ction of lead. DNA-functionalized gold nanoparticles are incubated with complementary strands labele d with a specific enzyme (DNAzyme). When complementary sequences come together, they form aggregations
7 that cause a chromogenic change from red to bl ue. This aggregation is reversible. When the aggregates are in the presence of lead, the enzyme on the DNAzyme catalyzes the cleavage of the complementary strand leading to disruption of the nanoparticle assembly and a color change to red (162). Molecular recognition elements used in biosensor development. The list of biological molecules that have been used in biosensor assay development is extensive. It is the recognition element which is the heart of the sy stem and makes a useful biosensor. There are two categories of molecular recognition el ements, catalytic (enzymes or bacteria) and affinity-based (antibodies or receptors). More recently, synthetic recognition elements such as peptide nucleic acids (PNA) have b een developed for biosensor use as well (21). Whole living cells of yeast, bacteria, a nd fungi have been manipulated for several different detection needs, especially in th e environmental and clinical fields (241). Biological oxygen demand (BOD) sensing yeast have been genetically engineered for luminous BOD sensing in environm ental biosensor experiments. Thiobacillus ferrooxidans has been used as a novel microbial sensor for detecting sulfate in rainwater. This bacterium is known to oxidize Fe (II) in the presence of sulfate and the decrease in current at the microbial electrode of the sensor correlates to the sulfate concentration (221). A similar approach has also been done by using recombinant luminescent bacteria. These bacteria were patterned onto a slide and exposed to toxic compounds. When the target compound is present, a cascade is set off within the cell that ultimately results in bioluminescence which can be recorded (147, 196). Enzymes have been used for years to detect metabolites such as lactate, urea and creatinine (268). Enzymatic biosensors have been developed mainly for the detection of
8 pesticides, clinical testing of metabolites such as glucose, and detection of organophosphates in foods a nd environmental samples (268). The most common enzymes used in biosensor detection are acetylcholinesterase (AChE), butyrilcholinesterase, horse ra dish peroxidase, tyrosinase, and urease (168). These enzymes cleave their particular substrate a nd the end result is a product which shows an obvious color change or is fluorescent and can be detected. Once isolated and purified, enzymes can be manipulated genetically fo r specific detection purposes. Pyruvate oxidase from Pediococcus sp. has been isolated and used to detect phosphate in river water (145). The enzyme cyanidase has be en purified and genetically modified to increase its binding affinity for cyanide, wh ich can be a dangerous contaminant of river water (116, 269). Affinity molecules such as antibodies ha ve been very useful for development of detection assays. They are probably the most popular biological recognition elements in use today because of their great specificity and ease of use. Antibodies, both polyclonal and monoclonal, have been a hu ge asset to the field of imm unoassays and rapid detection. They are stable when stored under the proper concentrations and conditions and are easy to use. Antibodies are a group of glycoproteins pres ent in the serum and tissue fluids of all mammals. They are produced by the B cells of the mammalian immune system in response to exposure to foreign antigen. IgG molecules commonly used in immunoassays have an approximate mol ecular weight of 150,000 -160,000 (152). The complex and efficient nature of the immune system causes the body to react to the presence of a foreign antigen with the pr oduction of many different antibody species
9 which are specific for different epitopes on a particular target Several genetic mechanisms are responsible fo r the diverse nature of anti bodies produced by different B cells in response to an antigen, including ge ne recombination and genetic conversion (49, 228). This cocktail of antibodies can be ex tracted as whole serum from whatever host (usually rabbit, goat, or donkey) has been exposed to the foreign material and are referred to as polyclonal antibodies. Polyclonal anti bodies are commonly used in the field of biosensor immunoassays and have been useful in detecting E. coli O157:H7 in ground beef and Salmonella in sprout wash (63, 142). Monoclonal antibody production, introduced by Kohler and Milstein in 1975, has been a valuable tool for achieving a large amount of specific antibody by using hybri doma technology to fuse mouse splenocytes with a B cell myeloma (140, 215). These immortal clones can be maintained indefinitely and can produce unlimited quantities of antibody. While adding enhanced specificity to any immunoassay, this technique could also be limiting by generating a recognition molecule specific only to one epitope on the an tigen. In an assay that is designed to maximize the chance of detecting a pathogen with several antigenic determinants, there are advantages to using an antibody mixture wh ich is specific for more than one target on that cell. There are ways to make these two types of antibodies work synergistically and to the advantage of the assay. A polyclonal antibody can be used to capture the target pathogen, and a monoclonal antibody can be used for detection in a sandwich immunoassay. Another strategy to increase the efficacy of monoclonal antibodies is to make several monoclonal antibodies to many antigens on a partic ular pathogen or target.
10 The monoclonal antibody species could then be mixed together to form a custom polyclonal or an antibody cocktail. In addition to the drawbacks listed above, monoclonal antibodies are very expensive. Furthermore, the immunogenicity (ability to e licit an immune response) of the antigen could limit the use of this strategy. This pr oblem was encountered in this study when monoclonal antibodies were ma de to teichoic acid from Enterococcus cells. The best results obtained were of IgM which has hi gh avidity (combined strength of the interactions between a multivalent antibody and a multivalent antigen), but the affinity (binding strength between an epitope and an antibody paratope) is not very good. Nonetheless, since the advent of the tec hnique, numerous monocl onal antibodies have been developed against an array of different agents including the anthrax lethal toxin, the envelope glycoprotein of the Ebola virus, and botulinum neurotoxin A (103, 114, 167, 203). The last decade has seen the developm ent of recombinant antibodies using phage display, which shows great promise for improvi ng specificity for det ection of biological agents. Phage display uses bacteria and b acteriophage to produce and select synthetic antibodies that have all the target recogniti on attributes of natu ral antibodies (202). These synthetic antibodies are made by the same genes that encode for the variable region (antigen recognition region) of natura l antibodies in mammalian systems (202). In phage display, a library of variant DNA sequences (derived from mammalian B cells) encoding antibody fragments is created and cl oned into phage as a fusion to a viral coat protein (274). This phage library displaying antibody fragments is exposed to the target antigen. Any unbound phage are washed away. Bound phage are eluted and used to
11 infect a bacterial host to amplify the phage population (274). The phage display technique is especially advantageous for th e production of antibodies to antigens which it has previously been very difficult to raise antibodies against (77, 174). Phage display is advantageous because antibody production no longer relies upon the immune response of a mammalian host. Phage libraries can contain up to a billion different antibodies (202). Phage display has been utilized to generate ta rget specific antibodies in many fields of science. The technique has been used in c linical applications to identify epitopes which are produced in the early onset of infection with the hepatitis C virus in an effort to develop tests for early diagnosis (124). Another clinical application has been the development of recombinant antibodies against thyroid stimulating hormone (TSH) for the diagnosis of thyroid diseases (144). Certain ligands have native receptors through which they exert their effects, and receptor-ligand relationships can be effective re cognition strategies for biosensor assays. For example, cholera toxin binds to the ga nglioside receptor GM1 on the surface of human intestinal cells (236). Receptor-ligand binding has also been utilized for hormone detection (241). The estrogen receptor can identify ligands which can act as estrogen-like compounds in humans (60). This research has been conducted to develop drugs which provide the benefits of estrog en and less of the detrimental effects of the hormone (60). Receptor-ligand pairs have the potential to be diversified into detection systems for other molecules like bacterial toxins (236). Some of the most recent advances in recognition elements have been on the molecular level, with the incorporation of nucleic acids as detection molecules. One such group of molecules are peptide nucleic acids (PNA) invented by Peter Nielsen (189).
12 Peptide nucleic acid oligomers are synthe tic DNA analogs with an amide backbone. PNAs have several advantageous features; eg, they are stable under acidic conditions and are resistant to nucleases and proteases (20, 189). Their neutrally charged backbones increase their binding strength when bound to complementary DNA sequences, and also enable the probe sequences to enter whole cells if necessary (20). PNAs are also stable across a wide range of temperatures and pHs; however they are expensive. Nonetheless, PNA is an alternative to less specific and less stable recognition elements available, and has been used to detect E. coli Salmonella and S. aureus (21, 246). Aptamers are another nucleic acid altern ative. They are DNA sequences (less than 100 nucleotides in length) designed to bind pr oteins much as an antibody would bind its target. Their binding affinity is similar to that of antibodies. Anti-lysosyme and antiricin aptamers have already been created for use on detection systems such as the electronic tongue sensor array (134). Aptamers display high affinity and specificity for target molecules, but they can be unstable a nd vulnerable to degredation by nucleases. Lectins are a group of plant proteins with the ability to agglutinate cells and precipitate carbohydrates. They interact with bacterial cell surface structures including teichoic acid, teichuronic acids, peptidogly can, lipopolysaccharide, glycoprotein, group specific polysaccharides, and capsular polysacch arides (70). Lectin-based technologies have numerous applications, including the ability to distinguish bacteria from each other. For example, Bacillus anthracis can be distinguished from other bacilli by growth at 37C and aggregation with the glycine max (soy bean) lectin (50). Lectins can also be used to type blood groups, purify teichoi c acids, and to investigate microbial ultrastructure (70, 277). The most wi dely characterized lectins are wheat germ
13 agglutinin (WGA) and concanavalin A (conA). ConA binds to -D-glucopyranosyl or D-mannopyranosyl groups on the cell surface of gram positive bacteria and WGA binds the teichoic acid of gram positive bacteria (10). Enzyme linked lectinsorbent assay (ELLA) has been used to detect and characterize extracellular polysaccharides and bacteria (99, 151). Biosensors are now bein g developed to exploit the specific binding properties of lectins. A nano-gold plasm on resonance biosensor utilizing gold labeled conA has been developed for the detection of blood glucose levels (11, 172). The current study utilized conA and WGA for the detection of enterococci in a sandwich assay format on a fiber optic biosensor because of their ability to bind carbohydrate groups on the surface of these cells.(27). The applicability of lectins is expanding and there are obvious advantages to their use. Lectin agglu tination assays are rapid, inexpensive, and reproducible. Lectins are also stable, easy to store, readily available from commercial suppliers and reactive at low concentrations (70). Evanescent wave fiber optic biosensors, immunoassays, and their applications. Immunosensors have been developed usin g the principles of well established immunoassay techniques as mode ls (197). The effectiven ess of an immunoassay relies on the following elements: i) the antigen to be detected, ii) the antibody or antiserum used to detect the target, iii) the method to separate bound antigen antibody complexes from free reactants, and iv) the mechanism of signal generation (8). Immunoassays date back as far as 1959 when Yalow and Berson introduced the radioimmunoassay (RIA) for insulin detection (279). The techniques of the immunoassay using labeled molecules, specifically antibodies, to detect antigens or vice versa ar e economical. They can also be sensitive depending on the target molecule. Solid phase assays employing ligands
14 labeled with radioisotopes or enzymes (enzyme-linked immunosorbent assay) are the most widely used immunoassays because of the large sample throughput that can be attained. Radioisotope s and enzyme labels are now bein g replaced by fluorophores and chemiluminescent markers to further increase sensitivity and shelf life of these assays. Two typical immunoassay formats are depicted in Figure 2a and 2 b, a sandwich immunoassay and an indirect immunoassay. a) b) FIGURE 2 a.) Sandwich immunoassa y and b.) Indirect immunoassay. Immunosensors allow for the real time measur ement of interactions between antibodies and antigens, receptors and ligands, and me mbrane reactions between cells. Optical biosensors have made it possible to measure the kinetics of such biomolecular reactions in a rapid and simple fashion (215). Y YYYYY Y Y Y Y Y YYYYY Y Y Y Y Y YYYYY Y Y Y Y YYYYY Y Y Y Y YYYYY Y Y Y Y Y Y Y Y Y Y YY Y Y Y Y Y Y Y Y YY Y Y Y Y Y Y Y Y YY Y Y Y YY Y Y Y Y Y Y Y Y Y Y Y
15 Evanescent wave, fiber optic biosensors combine fiber-optic technology with the many facets of biology and physics. The in strumentation for this type of biosensor is composed of three parts: i) the optics, which includes the platform or waveguide used to collect emitted fluorescence, the light sour ce for excitation, and the detector; ii) the fluidics for delivering solutions; and iii) electr onics and a computer system to convert the signally electronically and record it numerical ly (216). Evanescent wave excitation for use in fluorescent immunoassays was first re ported by Kronick and Little in 1975 (143). Evanescent waves work on the principle that light launched into a waveguide placed into a fluid medium will be reflected internally. Consequently, an electromagnetic field or evanescent wave is generated extending from the waveguide surface into the surrounding medium and any fluorescent molecule with in that evanescent wave is excited. Evanescent wave sensors must satisfy several criteria, beginning with a platform that is transparent to the wavelength used. The platform used must also be as free from impurities as possible so as not to scatter the laser light or cause shifts in the refractive index (216). This type of biosensor has been used to detect viral agents and toxins such botulinum toxin A and staphylococcal enterotoxin B with much success (69, 194, 237, 251). The Analyte 2000 (Research Internationa l; Monroe, WA) biosensor used in this study uses polystyrene fibers as waveguides to collect fluorescence emitted perpendicular to the waveguide (216). The Analyte 2000 biosensor used in this study has been developed to exploit the sensitive and specific nature of antibody-an tigen interactions in a sandwich immunoassay format (157, 230). This biosensor platform is multi-functional and has the unique capability of detecting pollutant s and pesticides in the environment, monitoring food-
16 borne pathogens in the food supply, and det ecting biological warfare agents (158, 260). Evanescent wave fiber-optic biosensors are very well suited for the sandwich immunofluorescence assay. Biotinylated capture antibodies can be affixed on the surface of the waveguide by biotin-avidin bi nding where they can bind target molecules of interest. The sample is incubated with a detection antibody conjugated to a fluorescent molecule. If this fluorophore conjugated antibody is within the evanescent wave, the fluorescent molecule is excited creating a detectable signal (157). This sandwich immunoassay has been used to detect pathog enic bacteria and viruses on the Analyte 2000 biosensor including vaccinia virus, E. coli O157:H7, Listeria monocytogenes and Salmonella (63, 64, 91, 142,157). The sandwich immunoassay is not the only assay format that can be adapted to this instru ment for pathogen detection. An indirect immunoassay format (Figure 2b) has been adap ted to the Analyte 2000 biosensor in this study for the detection of fecal enterococci in beach water. The Analyte 2000 biosensor has the ability to detect pathogens in a complex matrix. This is a desirable characteristic for treating recreational water samples that can be laden with particulate matter and debris. High sensitivity is great importance in waterborne pathogen detection because some bacteria such as E.coli O157:H7 have a low infectious dose (157, 254). The Analyte 2000 biosensor has been used to detect E. coli O157:H7 in a homogenate of ground beef and in apple cide r at low levels. The overall assay time is 20 minutes (63, 64). These results are encouraging for applying this instrument to detection of pathogens in water, especially considering the diverse matrix of beach water. The future of biosensor technology. Biosensors are a cutting edge technology with universal appeal across many fields of science. Over the past decade, there has been
17 increased interest in biosensors for use in environmental, medical, toxicological, and defense applications (197). They are be ing developed as instruments for the rapid detection of possible bioterro rism agents such as ricin, botulinum toxin, and anthrax spores (32, 194, 256). Biosensors like glucose mo nitors have been used as efficient tools for tracking blood sugar level fluctuations observed in diseases like diabetes (18). Research is ongoing to adapt these instrume nts and concepts to other areas of science such as drug discovery and pharmaceuticals. Newly designed biosensors are smaller and more versatile than their predecessors. They are being assembled with miniature components that will enable them to one day be assembled into small multi analyte detection chips (175). A single biosensor coul d theoretically serve as a chip array with multiple detection capabilities in various simple and complex matrices (175). Biosensors such as the fiber optic evanescent wave Anal yte 2000 biosensor have been developed as reusable instruments thereby reducing cost. Continued research and advancements in the fields of biochemistry, fiber optics, engineering, proteomics, and genomics will continue to drive further developments of rapid dete ction methods using biosensor instrumentation (216). Indicators of Fecal Contamination Starting with the first documented case of waterborne illness in 1854, a cholera outbreak in England (243), waterborne diseas e has been a public concern. A major focus of public health programs has been to prevent waterborne disease outbreaks from occurring. This is especially important during times of crisis, such as the events associated with recent hurricane Katrina, which left the city of New Orleans vulnerable to the dissemination of pathogenic bacteria and disease. Documented illness stemming
18 from V. vulnificus infections did indeed occur (34). For over 100 years, U.S. public health officials have relied on indicator orga nisms to assess the mic robiological quality of water (188). Directly assaying for specific pathogens is not easy and can be expensive. Fecal indicator bacteria provide an estimate of the amount of feces and, indirectly, the presence and quantity of fecal pathogens in wa ter (261). Initially, i ndicators were applied to asses the quality of drinking water and, ov er time, indicator measurements combined with chlorination drastically reduced the inci dence of disease associ ated with drinking water like cholera and typhoid (188). Presently, indicator levels are a standard guideline by which the extent of fecal contamination of all types of water are assessed (261, 262). Over the long history of indicators, assays fo r their detection have been standardized and made relatively easy to use for the enumeration of these organisms in water, with the drawback that results still take 24-48 hours (71, 179). The study of waterborne illn ess began with an outbreak of cholera in England in 1854 (243) which was studied by John Snow. A prominent physician, Snow conducted an epidemiological study showing that drinking water from a contaminated well caused the outbreak (243). In 1885, Theodor Escherich showed the frequency of coliform bacteria, specifically E.coli in human fecal matter. The term indicator was coined in the 19th century by Klein and Houston who also repor ted findings on the prevalence of coliform bacteria in sewage contaminated water in (137). Based on the findings of Escherich, Theobold Smith, an employee of the State of New York Department of Health, developed a presumptive test for the presence of E. coli based on its ability to ferment lactose in contaminated samples. This test later became known as the coliform test (242). In 1932, Scott proposed using coliform concentrati ons as a guideline to indicate sewage
19 contamination in coastal beaches in Connectic ut (223). It was later discovered, however, that some of the bacteria that tested positive in the coliform test were not always of fecal origin. Tests based on higher incubation temperatures (44.5 C) and detection of glucoronidase activity in fecal coliforms, a subset of the coliform group, were then developed. This method is still used today to enumerate fecal coliforms in contaminated waters. One such method is the Colilert system (54) developed by Idexx Laboratories (Westbrook, ME). The contamination of recreational bodies of water has become an escalating problem along the coastal United States. There are five commonly-recognized elements involved in the transmission of disease through water: i) the source of the infectious agent, ii) specific mode of water related transmission, iii) characteristics of the organism that allow it to survive and perhaps multiply and move into and within the environment, iv) the infectious dose and virulence of the pathogen, and v) host susceptibility (19). Waterrelated diseases were described by Brad ley and have provided a framework for the classification of waterborne illness (19). In this scheme, water related infections are divided into four groups i) waterborne infec tions ii) water-washed infections, iii) waterbased infections, and iv) infections w ith water related insect vectors. The present study focused on the rapid detec tion of indicator organisms as a means of assessing the extent of fecal contamination in beach water to prevent the incidence of waterborne infectious disease. Along w ith detection of these organisms comes the sometimes difficult issue of pinpointing the sour ce of that contaminant. Point source discharge of waste from wastewater treatment plants into coastal waters has necessitated the implementation of precise environmental regulations for treatment and disposal of
20 such waste (80). A bigger problem than point source discharge such as sewage and industrial waste is nonpoint source contamina tion. Nonpoint source pollution such as urban and agricultural runoff and overflowin g storm drains has generally not been regulated and, therefore, this type of pollution in the past has gone untreated and persists in the environment (41, 93). Human a nd animal feces contribute to nonpoint source contamination and are one of the most ha zardous forms of pollution. Animals and humans can harbor highly pathogenic organisms such as E.coli O157:H7 or hepatitis virus. The result is contaminated coastal beaches, which are unsuitable for the swimming public. Tainted beaches can cause economic impacts due to lagging tourism but, most importantly, the health of the public suffers (90). In response to disease acquired from recr eational water use, many programs have been implemented to monitor coastal water pollution and establish water quality standards. In 1972, Congress passed the Cl ean Water Act and established the U.S. Environmental Protection Agency (EPA). National regulations for water quality were established based on coliform densities, and research began to find other indicators which could be more accurate for a ssessing water contamination. In that same year, the EPA launched a series of epidemiological studies to test the correlation between indicator organisms and recreational water illness. Discriminant comparisons between fecal indicators and disease incidence revealed that elevated levels of enterococci strongly correlated with disease in freshwater and marine waters (25, 26). In 1986, the EPA released Ambient Water Quality Criteria for Bacteria revealing the results of those studies and new guidelines for indicator leve ls in freshwater and marine water beaches (261). E. coli (a member of the fecal coliform gr oup) was found to correlate positively
21 with gastroenteritis in freshwaters. Total coliforms and fecal coliform counts, which were routinely monitored at that time, only weakly correlated with disease (25). It was at this time that well defined criteria for enterococci and coliforms as indicators were developed. In fresh water, levels of E.coli should not exceed 126 CFU/100 ml and levels of enterococci should not exceed a geometric mean of 33 CFU/100ml ( 261). In marine waters, levels of enterococci should not exceed a geometric mean of 35 CFU/100 ml (261). There is still no single indicator which can be used to assess contamination of all types of water. The specific criteria for an ideal indicator organism as well as for the methods required to detect each indicator are listed in Table 1.
22 TABLE 1. Criteria for ideal indicator orga nisms and methods for their detection (90, 188). E.coli is considered to be a good indicator of freshwater fecal contamination. The contamination source is presumed to be wa rm-blooded animals. Fecal coliforms are not good indicators for beach water. The pres ence of fecal coliforms does not directly Criteria for an ideal indicator of fecal contamination Correlates to health risks Found consistently in the feces at higher concentrations than other fecal pathogen s Should not multiply outside the human intestinal tract As resistant to disinfection as other pathogens Should correlate to the presence of fecal pathogens Persists in the environment as long as other fecal pathogens Should be a simple, easy and reliable test to assay for the organism Specific to a fecal source or identifiable as to source of origin Criteria for ideal methods to detect indicator organisms Specific for the target organism Broad applicability Precision Appropriate sensitivity Quick results Allows for enumeration Can measure viability or infectivity Logistically feasible
23 correlate to the increased incidence of gast rointestinal disease among swimmers (261). E. coli has been isolated from freshwater collected from leaf axilae of bromeliads in a tropical rainforest suggesting that E.coli is indigenous to the water environment (213). Clostridium perfringens Staphylococcus aureus phages of Bacteroides fragilis have been mentioned as possible indicators of fecal contamination (90) The saturated sterol coprostanol has also been as an indicator b ecause it is specific to higher mammals (117). The EPA recommends enterococci as the most reliable indicator of fecal contamination in marine waters because it positively co rrelates with human illness (261). Enterococci-Indicators of Fecal Contamination Biochemical properties and characterization. The enterococci are Gram positive organisms that occur singly, in pairs or in short chains. These microorganisms are characterized by their ability to grow at 10C and 45C, in the presence of high sodium chloride (NaCl) concentrations and broad pH ranges. They are capable of survival at 60C for 30 minutes. Enterococci are aeroto lerant fermentative organisms which grow optimally at 35C and hydrolyze esculin in the presence of 40% bile salts. Some species of enterococci can be motile and most hydrolyze pyrrolidonylnaphthylamide (PYR). All strains produce le ucine aminopeptidase. Most strains are homofermentative, producing la ctic acid without gas from glucose. The enterococci are biochemically distinct from streptococci and have been placed in a unique genus, ( Enterococcus ). Enterococci are found in the gastrointestinal tract of warm-blooded mammals. These bacteria can be shed through fecal matter into soil, water, and food. As of 2002, 23 distinct species of Enterococcus were recognized (Table 2).
24 TABLE 2. Species included in the genus Enterococcus (81). Many methods have been developed for the biochemical identification of Enterococcus species. The API 20S (BioMerieux; Hazelwood, MO), the API Rapid ID 32 (BioMerieux; Hazelwood, MO), and the Vitek Gram-Positive Identification Card (BioMerieux; Hazelwood, MO) are examples of miniaturized biochemical test kits. These rapid tests often result in inaccurate identifications, particularly for Ent. gallinarum and Ent. casseliflavus isolates. Molecular techniques ha ve offered a promising tool for the rapid and accurate identification of Ent. faecalis and Ent. faecium Examples are PCR amplification of the Enterococcus protein A genes efaAfs from Ent. faecalis and efaAfm from Ent. faecium and amplification of the Ent. faecalis ace (adhesion for collagen) gene (239). Biochemi cal tests, although more time consuming, are considered by some to be cost effective and user fr iendly than these molecular methods (81). Reservoirs of enterococci; human and non human. As described previously, the enterococci (previously grouped as streptoco cci) are native inhabita nts of the mammalian and avian gastrointestinal tract and comprise approximately 1% of the overall microflora (105 to 107 organims/gram) in the feces (105, 224). They have also been isolated from reptiles and insects on occasion (62). With the emergence of antibiotic resistance within Ent. faecalis Ent. raffinosus Ent. haemoperoxidus Ent. faecium Ent. pseudoavium Ent. moraviensis Ent. avium Ent. cecorum Ent. ratti Ent. casseliflavus Ent. columbae Ent. pallens Ent. durans Ent. saccharolyticus Ent. gilvus Ent. gallinarum Ent. dispar Ent. malodoratus Ent. sulfurous Ent. hirae Ent. asini Ent. mundtii Ent. villorum
25 the genus, there has been increasing interest in identifying the reservoirs of these organisms and their antibiotic resistance gene s. There is a concern that the use of antimicrobial compounds on food animals that has created a large reservoir of antimicrobial resistant enterococci which have the potential to be passed onto human hosts (2, 13). Enterococci have been isolated from plants, soil, and water as fecal contaminants. In water, the predominating species are Ent. faecalis and Ent. faecium Other species have been associated with a particular environmen t or animal host. Examples of host specific species include Ent. assini found in donkeys and Ent. columbae, which is specific to pigeons (61, 67). Ent. faecalis and Ent. faecium have a broad distribution in the environment. Methods have been developed (microbial source tracking) to divide Enterococcus species into host-specific ecovars. These methods are based on genotypic and phenotypic traits that u ltimately discern the possible s ource of contamination. For example, Ent. faecium isolated from poultry retained the ability to ferment raffinose, whereas strains isolated from other origin s did not (66). Pinpointing the source of contamination, especially in beach water, can benefit the public health. Preventative measures to stop contamination at the source are more likely to prevent illness altogether (1). Group D antigen and lipoteichoic acid Teichoic acids were discovered in gram positive bacteria in 1958 by Baddiley and Davi son (12). Most gram positive bacteria produce a cell-wall associated teichoic acid and a membrane associated teichoic acid (lipoteichoic acid or LTA). Lipoteichoic acids are a class of phosphate-containing polymers that are covalently linked to membra ne glycolipids of the plasma membranes of
26 gram positive bacteria (138). They are not synthesized by gram negative bacteria. There are two types of teichoic acids, glycerol phos phates or ribitol phospha tes substituted with various glycosyl and D-alanyl ester groups (Figure 3) (55, 138, 273). In enterococci, the LTA is referred to as the group D antigen (74-76, 272). FIGURE 3. Structure of ribitol a nd glycerol teichoic acids (55). The group D antigen is located below the Enterococcus cell wall and constitutes approximately 1-2% of the b acterial mass (12, 125,126). This location contributes to the difficulty of raising a good immune response against LTA, which is evident in studies dating back to the Lancefield classificati on schemes (146, 170). Wa tson et al. discovered that a good immune response to LTA in gr oup D streptococci could be obtained only when the cells were enzymatically digested with lysozyme to reveal the LTA under the bacterial cell wall (270). Th e limited availability of LTA on the cell surface of intact
27 cells deters the formation of an antibody re sponse against this antigen. Knox and Wicken also noted this difficulty in raising antiserum to LTA using whole cells, but were able to prepare cell fractions which produced a strong antibody response (138). The cell wall structure of a typical Ent. faecalis cell is shown in Figure 4. FIGURE 4. Structure of the Enterococcus cell wall (55). Enterococci have cell wallassociated teichoic acid and membrane-associ ated lipoteichoic acid (group D antigen). The group D antigen is located beneath the cell surface. Several methods have been used to extr act LTA from gram positive bacteria. These methods vary from the preparation of a crude extract using enzymes like DNase and lysozyme to the more complicated hot or co ld phenol extraction method (127). Ideally, for a good immune response against teichoic aci d, LTA must be extracted/separated from protein or extracted with some protein. A single extraction technique does not work for GNAc GNAc GNAc GNAcGNAc GNAcGNAc GNAc GNAc MNAc MNAc MNAc MNAcMNAc MNAc MNAcMNAc MNAcMNAc LipoteichoicAcid (group D Antigen) Teichoicacid E. faecalis species antigen (rhamnopolysaccharide) Peptidoglycan Type specific capsular polysaccharide Cytoplasmic membrane GNAc GNAc GNAc GNAcGNAc GNAcGNAc GNAc GNAc MNAc MNAc MNAc MNAcMNAc MNAc MNAcMNAc MNAcMNAc LipoteichoicAcid (group D Antigen) Teichoicacid E. faecalis species antigen (rhamnopolysaccharide) Peptidoglycan Type specific capsular polysaccharide Cytoplasmic membrane GNAc GNAc GNAc GNAcGNAc GNAcGNAc GNAc GNAc MNAc MNAc MNAc MNAcMNAc MNAc MNAcMNAc MNAcMNAc GNAc GNAc GNAc GNAcGNAc GNAcGNAc GNAc GNAc MNAc MNAc MNAc MNAcMNAc MNAc MNAcMNAc MNAcMNAc GNAc GNAc GNAc GNAcGNAc GNAcGNAc GNAc GNAc MNAc MNAc MNAc MNAcMNAc MNAc MNAcMNAc MNAcMNAc GNAc GNAc GNAc GNAcGNAc GNAcGNAc GNAc GNAc MNAc MNAc MNAc MNAcMNAc MNAc MNAcMNAc MNAcMNAc LipoteichoicAcid (group D Antigen) Teichoicacid E. faecalis species antigen (rhamnopolysaccharide) Peptidoglycan Type specific capsular polysaccharide Cytoplasmic membrane
28 all bacteria especially when bacteria contain LTA and cell wall-associated teichoic acid (138). LTA is a complex and important antigenic structure in gram positive bacteria. Comparisons have been made between the LTA of gram positive bacteria and the LPS of gram negative bacteria. The teichoic acid porti on is thought to be analogous to the core polysaccharide of gram negative cells. While previously thought to be poorly immunogenic, LTAs location within the bact erial cell prevented the generation of antibodies when whole cells were used as the immunogen. To date, however, several monoclonal antibodies have been successfully made against LTA (104, 119). This study will show that disruption of Enterococcus cells leads to a good immune response in rabbits against these bacteria. LTA biosynthesis has been extensively studi ed. LTA is continually synthesized even when bacterial cells are under starvation cond itions. Growth conditions also have a large effect on the extent of LTA synthesis. Sy nthesis of the group D antigen is dependent on the glucose concentration and final pH of the bacterial growth medium. Glucose concentration increases the sugar group substitu tion of the LTA. Antibodies are typically formed against these carbohydrate component s of the LTA. Other studies have also shown that production of LTA is greatest dur ing exponential phase of growth (86). Public health significance of enterococci and use as an indicator of fecal pollution. Fecal enterococci were first discovered in the late 1880s as commensals of the gastrointestinal tract of humans and animals. They have been useful in monitoring the water quality of bathing beaches due to their abundance in human feces. Ent. faecalis was initially named Micrococcus ovalis by Escherich, and later renamed by Andrewes
29 and Horder in 1906 (9). Ent. faecium was first identified in 1899 (195). These two species stand out because they were the main species isolated from sewage-contaminated waters. By 1900, enterococci were recognized as being of fecal origin and were proposed as potentially useful indicators. The genus name Enterococcus was first used by Thiercelin and Jouhaud in 1903 ( 253). Controversy has followe d the classification of the enterococci since the turn of the century. Dible in 1921 proposed that Ent. faecalis be renamed Streptococcus faecalis (68) and Sherman in 1931 proposed that the bacteria known as enterococci be reclassified as Streptococcus (226). Sherman coined the term fecal streptococci instead of fecal enterococci for these indicators. This group of bacteria was characterized as gram positive catalase negative non sporeforming facultative anaerobes which belonged to the Lancefield group D streptococci. In 1984, based on immunological and molecular methods, the fecal Streptococcus group was reclassified as Enterococcus once again. Only a few group D streptococci remained as Streptococcus including Streptococcus bovis and Streptococcus equinus Beginning in the 1940s and 1950s, the development of a method to enumerat e fecal enterococci began. Cabelli at the EPA developed a reliable membrane filtration technique for detecting enterococci in water in 1970 (25). The experiments were conducted as prospective epidemiologicmicrobiological studies to compare rates of gastrointestinal illne ss in swimmers and nonswimmers at freshwater and marine water beach es. As a result, the presence of fecal enterococci in marine water has been consid ered indicative of fecal contamination. The presence of fecal enterococci can also be correlated to the incidence of gastrointestinal illness in swimmers (26, 72, 95, 259, 265, 283). Mean Enterococcus densities were designated as a result of these studies. Enterococci were not to exceed a geometric mean
30 value (over at least 5 samples) of 35 CF U per 100 ml and no one sample should carry greater than 104 CFU of enterococci per 100 ml (261). As previously mentioned, 1986 was marked as the year when strict guidelines for water quality were established and the EPA recommended that Enterococcus counts be used to monitor water quality for marine and coastal areas (261). Coastal states across the U.S. have implemented EPA guidelines for the routine monitoring of recreational bodies of water using coliform or Enterococcus counts or both. The State of Hawaii added a third indicator, Clostridium perfringens that is routinely monitored, (89). Since the 1986 gu idelines were publishe d, several countries have sponsored studies for the further asse ssment of the relationship between indicator organism densities and disease incidence (45, 53, 83). Some results support the 1986 findings of the EPA, that ente rococci should be used to monitor marine waters (261). Noble et al. found that a more integrated a pproach for monitoring these organisms should be used (190) where more than one organism is assayed to monitor marine water in order to obtain additional information on the overall quality of the water (259). The State of Florida monitors fecal coliforms and enteroco cci in marine water (87). Regardless, the enterococci have been found to be reliable and to retain attributes that other indicators have not. Enterococci are not as ubiquitous as coliforms, are always present in the feces of humans and animals, are unable to multiply in contaminated waters, die off less rapidly than coliforms in water, and persist in the water as long as specific pathogens (258). There are disadvantages to using en terococci including their ability to become viable but non culturable, to survive associ ated with zooplankton, and to survive in sediments and sand resulting in resuspension of the organisms at high tide (6). All
31 factors need to be considered when developi ng rapid methods for the detection of fecal enterococci. Methods for the detection of fecal enterococci Improved methods for the detection of enterococci have been developed that are rapid and simple. Table 3 shows a comparison of the most recent methods deve loped to detect bacteria by desirable characteristics. TABLE 3. Comparison of common methods to detect bacteria by desirable attributes (188). The three types of detection met hods were rated from low to high for each criterion listed. Method attributes Culture Immunological Nucleic acid Specificity to desired Low to Moderate High High target Broad applicability High High High Repeatability Moderate Low-High High Sensitivity Moderate/High Low-High High Speed of results Low/Moderate Moderate/Fast Moderate/Fast Quantifiable Moderate Low Low/Moderate Measures viability or Yes No No/ Is possible infectivity Feasibility Training and personnel Lo w/High Moderate Moderate requirements Can be used in the Low/Hi gh Moderate/High Moderate/Hig h field Cost Low/High Moderate Moderate Volume requirements Low/High Low Low
32 The easiest, less labor intensive, most reliable, and most feasible method for the detection of fecal enterococci today are still culture methods. Enrichment culture methods combined with filtration permit bacteria from contaminated samples to be enumerated. Selective plate media and broth cultures can be used to enrich for specific enzymatic activity in enterococci. In both cases, the me dium has been created to target growth of desired bacteria by adding selective agents that will repress the growth and proliferation of the undesired background flora. The medi um also contains components that cause the growth of the target bacteria to appear uni que and distinguishable from non-targets. The most common method to date has been the use of membrane filtrati on of beach water. The filter is then placed on selective and differential agar and enterococci are enumerated. This is the standard method us ed today by local, county, and state public health agencies across the United States (263). Levin et al. described a two-step me mbrane filtration (MF) method for the enumeration of enterococci from beach water (153). This method involved the filtration of a sample of water through a 0.45m filte r. The filter was placed onto primary selective isolation medium (mE) containing nalidixic acid (NA), cycloheximide, and triphenyltetrazolium chloride (TTC). The NA inhibited the presence of gram negative organisms; the TTC permitted for the differentiation of enterococci from other organisms and cyclohexamide inhibited fungi. The medium was then incubated at 41C. Following primary isolation, the filter wa s transferred to EIA media (e sculin-iron media). Pink to red colonies on mE would produce a brownblack precipitate on EIA if they were enterococci due to the hydrolysis of esculin by the enzyme -glucosidase. While this medium was convenient at the time, it took 48 hours to obtain a result.
33 Years later, the two-step MF technique de veloped by Levin was modified by the EPA (263). Building on the foundation of the two step method, the process was made easier and more rapid by addition of the chromogenic substrate indoxyl -D-glucoside (0.75 g/l) to the primary isolation media and reducing the concentration of TTC (from 0.15 g/l to 0.02 g/l). The addition of indoxyl -D-glucoside to the medium resulted in -glucosidase positive colonies of enterococci. The -glucosidase produced an indigo blue complex which diffused into the surrounding agar forming a blue halo around the colony. The modified medium was termed mEI and was referred to by the EPA as Method 1600 (179, 263). All colonies regardless of color were count ed as enterococci if they had a blue halo surrounding them. This method offered severa l advantages including reduction of overall assay time from 48 hours to 24 hour s, and elimination of the need to prepare a secondary medium (EIA) for identification of enterococci, using the existing mE base which was commercially available. Colonies were smaller and more distinct making them more easily counted, and the specificity was better than the two step method (only 6% false positive rate compared to 10% with the two step method) (179). Both the two-step method (using mE and EIA agars) and the modi fied method (using mEI) are listed in the Method 1600 Manual published by the EPA (263). Unfortunately, the indoxyl -Dglucoside substrate is expensive so that the 2-step method is still used by some laboratories. Enterolert developed by Idexx Laborat ories is a semi-automated most probable number (MPN) technique that can enumerate ente rococci. It is intended to be less labor intensive, requires fewer confirmatory tests, less skilled labor, and provides results in 24 hours. The test is based on the activity of the Enterococcus -glucosidase enzyme. The
34 Enterolert medium contains the substrate 4-methylumbelliferone-D-glucoside which is cleaved by -glucosidase to the fluorescent product 4-methylumbelliferylone. 4methylumbelliferylone can be seen when ex posed to long wave UV light of 365 nm. Enterolert is used to test for the presence of enterococci in freshwater and marine waters. This method has been found to be just as effe ctive as membrane filtr ation techniques, less labor intensive requiring shor ter incubation time for results (24, 71, 191, 192). However, because samples need to be diluted, this met hod is incapable of detecting very low levels of enterococci (71). Newer Molecular methods are now availa ble that do not have the limitations of culture methods, specifically, the sensitivity a nd the time it takes for results. Polymerase chain reaction (PCR) to amplify specific DNA se quences from enterococci has been used to overcome the limitations of traditional curr ent methods (57, 220). This method uses DNA polymerase to make thousands of copies of a particular gene sequence from a bacterial genome. The product is then assa yed by common laboratory techniques such as gel electrophoresis. Realtime PCR, which uses fluorescent dyes that can intercalate into double stranded DNA, is a variation of PCR and is quantitative. Several PCR assays have been developed for the detection of en terococci. Danbing et al. developed a novel procedure by amplifying the tuf gene in enterococci which encodes for an elongation factor; however, this procedure has been used for clinical applications only (57). Most procedures are developed to amplify conser ved sequences such as the 16S or the 23S rRNA sequences from enterococci (111). A pr ocedure developed by Santo-Domingo et al. showed the potential for using PCR for detecting enterococci in environmental water samples by amplifying a segment of the 16S rRNA region (220). This procedure was
35 successful for detecting levels of enteroco cci greater than 600 CFU/100ml, but was not able detect lower concentrations directly fro m water. In addition; enterococci could not be detected in samples with heavy fecal contamination because of inhibitory materials. Nonetheless, detection of cl ean samples took only about thr ee hours, and therefore, this procedure has potential for further developm ent (220). A comparative study has been done between PCR, specifically quantitative PCR (QPCR), and the membrane filtration methods and a positive correlation has been observed (110). Whether this technique could potentially replace MF is yet to be seen because epidemiological studies need to be performed. Emerging and Reemerging Waterborne Pathogens Indicator organisms have long been used as a standard in water quality testing. Ideally, an integrated method for the uni versal detection of pathogens, known and reemerging, needs to be developed to preven t waterborne disease, especially in those third world countries where di sease is prevalent due to poverty (247). Cholera in Asia, Africa and now South America remains a conc ern and has been a constant problem in these areas for years. Other waterborne pathogens including E. coli O157:H7, Cryptosporidium parvum and microsporidia are also a health concern (Table 4) (225, 252).
36 TABLE 4. Emerging and reemerging waterborne pathogens (188, 247). Group Pathogen (Source) Diseases Viruses Enteroviruses (polio, Coxsackie) meningitis, paralysis (human feces) rash, fever, respiratory disease, diarrhea Hepatitis A and E infectious Hepatitis, (human feces) jaundice, fatigue Human Caliciviruses (human feces) Norwalk viruses diarrhea, gastroenteritis Rotavirus diarrhea, gastroenteritis Bacteria Salmonella typhoid, gastroenteritis (human and animal feces) Shigella diarrhea (human feces) Campylobacter diarrhea (human and animal feces) Yersinia enterocolitica diarrhea (animal feces and urine) E.coli O157:H7 diarrhea, can lead to (human and animal feces ) hemolytic uremic syndrome in children Legionella pneumophila respiratory pneumonia (freshwater and soil) Protozoa Naegleria meningoencephalitis (freshwater in warm temps. Entamoeba hystolitica amoebic dysentery (human feces) Giardia lamblia chronic diarrhea (human and animal feces) Cryptosporidium parvum acute diarrhea (human and animal feces) Microsporidia diarrhea in (feces) immunocompromised and immunocompetent travelers Cyanobacteria Microcystis diarrhea from toxins (cyanobacterial blooms) produced
37 Vibrio cholerae Vibrio cholerae the etiological agent of cholera, is a facultatively anaerobic, motile, curved or straight, gram negative rod that is 1-3 mm in length. A minimum of 5 mM sodium chloride (NaCl) is required for optimum growth; however, the bacterium can grow without NaCl and in alkaline conditi ons up to pH 10 (204). Recent and ongoing research indicates that this organism is a native member of the aquatic microbial community (129). V. cholerae has caused numerous epidemics in the past and more recently in South America (35). Approximately 200 serovars of V. cholerae have been reported; however, only two serovars, O1 and O139, have been associated with major epidemics. There are two biotypes of V. cholerae O1, classical and El Tor. V. cholerae O1 is classified into two serotypes, In aba and Ogawa, based on agglutination in antiserum (218). A third serotype, Hikojima, has been described, but has rarely been encountered (218). Isolates of V.cholerae O1 and O139 which produce cholera toxin (CT) are considered virulent and capable of causing epidemic cholera. Some isolates of V. cholerae O1 have been found that do not produce CT and do not cause epidemic cholera. These nontoxigenic isolates are impli cated in sporadic diarrheal disease (51). Epidemic and non-epid emic strains of V. cholerae are compared in Table 5.
38 TABLE 5. Comparison of epidemic and non-epidemic strains of V. cholerae (131). a Cholera toxin Historical background. The Ganges Delta in the Indian subcontinent is considered to be the source of cholera disease (14). Pand emics of cholera spread through the world in the nineteenth century. The massive epid emic in Southeast Asia in 1961 was the beginning of the seventh cholera pandemic. The biotype responsible for this pandemic was V. cholerae O1, El Tor, which spread rapidly through south Asia, the Middle East, and southeastern Europe, eventually re aching Africa in 1970 (35, 97, 275). Epidemic cholera reappeared in 1991, but this time in S outh America, specifically, the coastal cities of Peru (35). V. cholerae O1 El Tor serotype Inaba was found to be responsible for this epidemic. It spread swiftly and by 1996, the disease had spread to 21 countries in Latin America. Over one million cases were re ported and close to 12,000 deaths resulted (210). In recent times, an increase in c holera cases has been reported from western Africa and Asia. Cholera in the United Stat es dates back to 1832 in New York (14). Disease spread rapidly and intermittent illness was reported between 1832 and 1875. Cholera then disappeared. No notable cases were reported until 1941 from a New Typing Systems Epidemic-associated Non-epidemic associated Serogroups O1, O139 Non-O1 (>150) Biotypes for V. cholerae O1 Classical and El Tor Biotypes do not apply to (does not apply to non-O1 V. cholerae O139) Serotypes for serogroup O1 Inaba, Ogawa, and Serotypes do not apply Hikojima to non-O1 V. cholerae (does not apply to O139) CTa production Produce cholera toxin Typically not CT producers may produce other toxins
39 Orleans longshoreman and then in 1973 from a shrimp fisherman in Texas (98, 271). Later findings revealed that V. cholerae is indigenous to the U.S. Gulf of Mexico coast and in waters of other countries (129, 200). The majority of isolates from the U.S. Gulf of Mexico coast have been non-O1, but, on occasion, toxigenic V. cholerae has been isolated (65, 155). V. cholerae O139. Prior to the seventh pandemic, out breaks of cholera were traced back to the classical biotype. By 1961, El To r strains had replaced the classical biotype and this strain was determined to be responsible for epide mic cholera which appeared in Peru beginning in 1991 (30). Today, the El Tor biotype is responsible for virtually all of the cholera cases throughout the world. Cl assical isolates are not found outside of Bangladesh. V. cholerae O139 first emerged in 1992 in southeastern India and the Bay of Bengal (209). This serotype appears to have arisen from genetic exchange with V. cholerae O1 El Tor (209). Prior to the seventh pandemic, V. cholerae O139 and nonagglutinating strains of V. cholerae were considered to be a different species of Vibrio. Biochemical and molecular analysis studies ha ve shown that these strains were all of V. cholerae The only distinguishing factor between V. cholerae O1 and V. cholerae O139 is that V. cholerae O139 will not agglutinate with O1 antiserum. O139 antiserum is necessary for identification of this organism Eleven countries reported cases of O139 cholera to the World Health Organization through 1998 (276). While V. cholerae O139 seems to be confined to Asia, imported cases have been reported in the United States as well as other countries (51).
40 Ecology. The ecology of Vibrio cholerae appears to be co mplex and depends on several different environmental factors. This is especially true since V. cholerae is now known to be indigenous in some aquatic environments. Cholera displays a distinct seasonal pattern, especially in those areas wh ere it is endemic (51). The patterns are related to the ecology of V.cholerae Higher cell numbers are seen when water is warmer and when there is an increase in z ooplankton blooms (51, 115). Plankton blooms influence the bacteriums attachment, growth, and multiplication in the aquatic environment, particularly when associated with copepods (121). The physical factors that affect the proliferation of V. cholerae in the environment include warmer temperatures and salinity combined with elevated pH. The incidence of V. cholerae in the environment is thought to follow a predictive pattern that can be monitored in order to predict outbreaks before they occur (165). Those methods include remote sensing to monitor sea surface temperature, sea surf ace height, and plankton blooms (165). Monitoring would at least serve as a pr eventative measure when the potential for outbreaks is high. The mechanisms of survival utilized by endemic V.cholerae are being further characterized. The survival of V. cholerae in the environment is enhanced by their close association with zooplankton and, in particular, copepods whose chitin provides a good source of nutrients for the growth or regrowth of V.cholerae (115). Viable but nonculturable (VBNC) phenomenon in V. cholerae is also a survival mechanism (37, 267). V. cholerae O1 and O139 cells have been found to retain viability and pathogenicity for up to one year in the VBNC state (37).
41 Virulence genes and pathogenesis. Virulence of V. cholerae is due to the production of a heat labile enterotoxin (CT). V. cholerae has two chromosomes, large and a small, and the genes encoding CT are located on the large chromosome (51). Most environmental strains lack th e genes necessary to produce toxin. Virulence genes are distributed among strains of V. cholerae (51). There is the potential of acquiring those genes from these and other pathogenic bact eria, allowing for the emergence of new toxigenic strains of V. cholerae In fact, the V. cholerae O139 serovar acquired novel DNA from the environment (82). Bacterial viru ses also play a role in development of new toxigenic variants and in imparting antibiotic resistance (51, 160). The primary site of infection is the small intestine. Cholera toxin causes an electrolyte imbalance resulting in profuse wa tery diarrhea (154). Pathogenicity of toxigenic V. cholerae depends on several factors, including the presence of a colonization factor (TcpA), production of CT, and the presen ce of associated outer membrane proteins (199, 250). After colonization in the body, pr oduction of CT activates adenylate cyclase elevating the levels of cyclic AMP in the body and ultimately leading to extreme diarrhea. A distinctive symptom of cholera is rice water stool (210). The resultant sideeffects of the massive diarrhea include loss of circulation and bl ood volume, metabolic acidosis, potassium depletion, a nd finally vascular collapse a nd death (210). After onset of disease, death can result from extreme dehydration due to lack of fluids and electrolytes (16, 36). In severe cases, the seve re diarrhea results in the loss of up to 10% of human body weight, leading to hypovole mic shock and death. Infection with V cholerae can vary depending on the amount of exposure. Lower infecting doses may
42 cause, at most, mild illness and discomfort. Only 2% of patients who are infected with V.cholerae see a progression of the disease to a life-threatening condition (36). Research has shown that strains of non-O1 V. cholerae (NT V. cholerae ) are capable of causing disease. Some studies describe pathogenenesis from NT V. cholerae strains that produce a novel toxin that is not CT ( 238). Sporadic cases of disease from non-O1 V. cholerae have been reported to the CDC. Sy mptoms ranged from fever, nausea, and cramps to mild diarrhea (3, 56). None of the cases from NT V. cholerae have been considered life-threatening (17). Public health significance and epidemiology. Transmission of V. cholerae occurs as a result of poor sanitation in highly populated areas where the organism is endemic. The vehicles of transmission are contaminat ed water and undercooked food (180, 207). Water contaminated with free living V. cholerae cells is the main origin of epidemics, followed by the food-borne ingestion of shellfish. The infectious dose of V. cholerae is high and varies between 104 and 1011 CFU, depending on the ma trix (33). Cases of cholera in the U.S. have been sporadic sin ce the epidemic of 1832 in N.Y. (14). More recently, cases of cholera that have emerged in the southern U.S. have been near the Gulf of Mexico, where V. cholerae appears to be endemic (155). These cases have been few and have not resulted in epide mics, and some have been reported to be caused by nonO1 V.cholerae (184). Non-O1 V.cholerae has been found to be native to the Gulf of Mexico. Two of the survival mechanisms in V. cholerae that are being further investigated are the VBNC phenomenon and their close associa tion with copepods in the environment. These mechanisms combined with environmental conditions that allow for the
43 proliferation of this bacterium are factors th at need further study. It is still unknown how an epidemic begins in the absence of any clinical cases of cholera. A summary of the factors influencing the survival of V. cholerae in the environment are listed in Table 6. A sensitive and rapid method is necessary to monitor V.cholerae concentrations in water so that the public can be warned when cell num bers begin to increase. Such a method would be an important tool to prevent the spread of cholera. TABLE 6. Factors affecting the survival and growth of V. cholerae in the environment (133). Detection Methods. Detection methods for Vibrio cholerae in water are largely culture based (132). A sample first has to be enriched in liquid alkaline peptone medium with 1% sodium chloride. Incubation at 37 C is followed by subculture onto selective plating medium. For over twenty years, the medium of choice has been ThiosulfateCitrate-Bile salt-Sucrose Agar (TCBS) (73, 201). TCBS distinguishes sucrose fermenting V. cholerae from other non-sucrose fermenting vibrios. This medium is commonly used because it is commercially available, easy to make, and needs no Factor Effects Temp Refrigeration extends survival; growth occurs above 10C pH Optimum 7.0-8.5; survives at 6-10 Water activity (aw) >.93 NaCl content Best survival .25-3%; optimum 2%; requires Na+ Degree of contamination Longer survival with more contamination Exposure to sunlight Survival is reduced Presence of organic matter Survival is prolonged Osmotic pressure Survival is unfavorable at high pressure Carbohydrate content Survival is reduc ed when high concentrations present Other microflora Survival and growth suppressed by competitors Humidity Survival is longer in high moisture environment
44 autoclaving. Enumeration of V. cholerae in water requires a most probable number method (MPN) for selection because of the large number of vibrio s in environmental water samples. Following isolation on TCBS agar, the bacteria are inoculated into nonselective media for further biochemical and se rological testing. Those tests include the oxidase test, the string test (to differentia te vibrios from non-vibrios), and triple sugar iron test. Serological identification is performed by testing the isolate for agglutination with V. cholerae O1 or O139 antiserum (131, 132). Unfortunately, VBNC V. cholerae cannot be detected by such methods. E. coli O157:H7 E. coli O157:H7 is a foodborne and waterbor ne pathogen, whose main reservoir is cattle; however, the bacterium can be passed by human fecal matter. It was first recognized as a human pathogen following two outbreaks in 1982 in Oregon and Michigan (212). These outbreaks were traced back to ground beef tainted with feces. Transmission of this pathogen occurs from food or water contamin ated with animal and human feces or sewage. E.coli O157:H7 produces a vero-cytotoxin, that causes haemorrhagic colitis (HC) (123). HC manifest s as severe abdominal pain, diarrhea and bloody stools. Progression of the disease to hemolytic uremic syndrome could lead to kidney failure and death (187). Of particular importance to pub lic health officials is the infectious dose which has been reported to be lower than 50 CFU (254). While E.coli O157:H7 typically is transmitte d via undercooked meat or raw milk, waterborne cases have been re ported and are a concern. The first case of waterborne E. coli O157:H7 was reported in 1986 in Philadel phia (176). The bacterium was isolated from a reservoir outside the city and was pr esumed to have a non-human origin. The
45 majority of waterborne infections are from tainted lake or muni cipal water (38). Beaches with higher bather densities and especially those close to sewage treatment plants present a high threat of infection with E.coli O157: H7, especially since this pathogen is capable of entering a viable but non-culturable state in the water prolonging its survival (106). E.coli O157:H7 can also survive in saltwater for up to 15 days (182). Methods for the detection of E. coli O157:H7 include PCR assays to detect the presence of genes encoding shiga-like toxins or by coupling molecular techniques with culture methods. Culture methods involve direct plating on sorbitol MacConkey agar supplemented with cefixime a nd tellurite (CT-SMAC) followe d by serotyping (112). E. coli O157:H7 is unable to ferment sorbitol, re sulting in colorless colonies on CT-SMAC agar. Some laboratories test E. coli O157:H7 strains for the enzyme glucuronidase using a broth or agar medium containing the substrate 4-methylumbelliferyl-Dglucuronide (MUG) (84). Cleavage of MUG pr oduces a fluorescent product is produced. E. coli O157:H7 strains are MUG negative. Immunomagnetic separation (IMS) is also an isolation technique developed to detect E. coli O157:H7 in food and stool samples (130). IMS uses magnetic beads with covalently attached antibodies specific for E. coli O157:H7. E. coli O157:H7 cells in a a sample bind to the antibodies. The antibody coated beads can then be isolated from the rest of the sample using a magnet. This strategy was used in this study for the isolation of E. coli O157:H7 from beach water. E. coli O157:H7 has been successfully detected at low levels in apple juice and ground beef using a sandwich immunoassay format with an evanescent wave fiber optic biosensor (63, 64). The same instrument was used in this study for the detection of E. coli O157:H7 in sea water.
46 Staphylococcus aureus Staphylococcus aureus is a gram positive pathogen which has been considered for use as a parameter for assessing the water qualit y of recreational waters (72, 90). It is a normal inhabitant of the human flora being present in the nose, hands, and respiratory tract. Its overabundance on the body allows it to be easily shed into recreational water. Once in water, this bacterium can infect cuts or abrasions on the skin or mucous membranes of the nose and eyes. Such infecti ons can result in skin rashes, boils and eye infections and other forms of diseas e. The increase in resistance of Staphylococcus aureus to methicillin makes infection even more life-threatening. A rapid method for detection is therefore imperative (39). Summary A rapid method for detection and enumera tion of indicator bacteria and microbial pathogens in water is desirable and nece ssary. In 1997, the EPA launched the Beaches Environmental Assessment Clos ure and Health program (BEACH) to further improve the quality of the nations beaches (262). In 1999, broad goals for the BEACH program were established. These goals included developm ent of methods to detect water pollution within a few hours so that swimmers can be wa rned of a contamination event before they are exposed (262). Public policy and water quality research is now directed towards preventing the incidence of water-borne disease. Biosensor immunoassays that can be highly specific and sensitive are an attractive approach for detecting pathogens and indicator s in mixed samples. If a high affinity antibody is available, a rapid, near real-time assay can be developed for the analysis of contaminated water. The purpose of this st udy is to use the biosensor to detect the
47 presence of these pathogens and/or indicators in beach water samples from Clearwater, FL. and St. Petersburg, FL. and to provide a quantitative assessment of water quality.
48 MATERIALS AND METHODS Bacterial Strains Enterococcus faecalis (ATCC 19433) was used as a model organism in all experiments in seawater and in phosphate-buffered saline pH 7.4 (PBS) to detect fecal enterococci. Additional strains used in this study are listed in Table 7. TABLE 7. Bacteria used in Enterococcus assay development. Microorganism Strain Aerococcus viridans .ATCC 700406 Aeromonas hydrophila ..USF 530 Enterococcus casseliflavus USF 691 Enterococcus casseliflavus USF 707 Enterococcus casseliflavus USF 740 Enterococcus casseliflavus USF 703 Enterococcus durans .USF 694 Enterococcus durans .USF 699 Enterococcus durans USF 732 Enterococcus faecalis .ATCC 19433 Enterococcus faecalis USF 693 Enterococcus faecalis USF 696 Enterococcus faecalis USF 700 Enterococcus faecalis ...USF 708 Enterococcus faecalis USF 709 Enterococcus faecalis USF 710 Enterococcus faecalis ...USF 711 Enterococcus faecalis USF 713 Enterococcus faecalis ...USF 715 Enterococcus faecalis USF 719 Enterococcus faecalis ...USF 720 Enterococcus faecalis ..USF 721 Enterococcus faecalis ...USF 723 Enterococcus faecalis ...USF 724 Enterococcus faecalis ...USF 726
49 Enterococcus faecalis ...USF 727 Enterococcus faecalis ...USF 730 Enterococcus faecalis ...USF 734 Enterococcus faecalis ...USF 741 Enterococcus faecalis ...USF 747 Enterococcus faecalis ...USF 751 Enterococcus faecalis ...USF 764 Enterococcus faecalis ...USF 769 Enterococcus faecalis ..USF 807 Enterococcus faecium ATCC 19434 Enterococcus faecium ..USF 691 Enterococcus faecium ..USF 692 Enterococcus faecium ..USF 725 Enterococcus faecium ..USF 730 Enterococcus faecium ...USF 746 Enterococcus faecium ...USF 753 Enterococcus faecium ..USF 754 Enterococcus faecium ..USF 759 Enterococcus faecium ..USF 763 Enterococcus faecium ..USF 768 Enterococcus gallinarum .USF 767 Escherichia coli ATCC 29417 Escherichia coli O157:H7 USF 640 Group A Streptococcus .8-R291 TGH Group B Streptococcus 1215-RNSJ183V Lactobaccillus arabinosa .USF 541 Lactococcus lactis .ATCC 11454 Leuconcostoc USF 706 Non Protein A Staphylococcus aureus .ATCC 10832 Pediococcus dextrinicus ATCC 33087 Providencia rettgeri .USF 517 Staphylococcus aureus USF 645 Staphylococcus aureus Cowan I ..USF 613 Streptococcus bovis ...ATCC 15351 Streptococcus bovis ..ATCC 49133 Vibrio alginolyticus ..USF 516 Weissella confusa ..ATCC 14434
50 All Enterococcus strains without ATCC numbers were environmental isolates obtained from the following sources: Tumblin Creek, Gainesville, FL; Tampa Palms sewer station (Tampa, FL); sewage from California; a nd beach water from Pinellas County beaches. These isolates were provided by the laborator y of Dr. Valerie J. Harwood (University of South Florida, Tampa, FL) and identified as isolates from various hosts including cows, humans, seagulls and dogs. Another sample of Enterococcus faecalis (ATCC 19433) and one of Enterococcus faecium (ATCC 19434) was obtained fro m Dr. Al Dufour (U.S. Environmental Protection Agency, Cincinnati, OH). The remaining isolates were clinical and environmental isolates obtaine d from USF freezer stocks. The Aeromonas hydrophila strain was isolated from Naples Bay (Naples, FL). The Vibrio alginolyticus strain was a seawater isolate obtained from Courtney Campbell Causeway beach (Clearwater, FL). Escherichia coli O157:H7 (USF 640) isolated from hamburger meat from a taco stand in Massachusetts following an outbreak wa s used to develop biosensor assays for E.coli O157:H7 in seawater. Staphylococcus aureus USF 645 and Vibrio cholerae biotype O1 El Tor serotype Inaba (CDC E5906-1018) were used as model organisms to develop biosensor assays. All other bacteria l strains used in experiments for Vibrio cholerae are listed in Table 8.
51 TABLE 8. Bacteria used in V. cholerae assay development. All Vibrio cholerae isolates were clinical isolates obtained from either the Centers for Disease Control and Preventi on (CDC) (Atlanta, GA) or from Dr. Mark Tamplin (United States Department of Agricu lture; Wyndmor, PA) who was at the University of Florida, Gainesville, FL when these cultures were donated. All other isolates without ATCC numbers are environmen tal isolates. The Aeromonas hydrophila and the Vibrio alginolyticus strain were the same as those used in the Enterococcus experiments. Media, Culture Conditions, and Sample Collection Stock cultures. Cultures of all isolates were obtained from stock cultures maintained at -80 C. The cultures were grown on trypti case soy agar (TSA) (BD, Sparks, MD) or in trypticase soy broth (TSB) (BD) for 18-24 hours at 36C. This working stock was maintained at 4C for up to 1 month. Once gr own, the cells were resuspended in one of three solutions: 0.01M phosphate buffered saline, pH 7.4 (PBS); 0.1M carbonatebicarbonate buffer, pH 9.3; or seawater. This culture procedure was st andard in all assays and viable counts unle ss otherwise indicated. Microorganism Strains Aeromonas hydrophila USF 530 Escherichia coli ATCC 29417 Pseudomonas aeruginosa .ATCC 15442 Pseudomonas aeruginosa USF 620 Salmonella typhimurium ...ATCC 23564 Salmonella typhimurium ..USF 515 Vibrio alginolyticus .USF 516 Vibrio cholerae O1 ..USF 646 Vibrio cholerae biotype O1 El Tor Inaba c holera toxin (+) ..CDC E5906-1018 Vibrio cholerae biotype O1 El Tor Ogawa cholera toxin (-) ..CDC1074-78-1019 Vibrio cholerae O139 ..USF 647 V. cholerae O139 .CDC 2412-93
52 Enrichment and recovery media for enterococci. Azide dextrose broth and ToddHewitt broth were prepared in single strength concentrations according to the recipes in Standard Methods for the Examination of Water and Wastewater for the growth of Enterococcus faecalis (ATCC 19433) in teichoic acid extr action experiments (4). Bile esculin azide broth (4), mE broth, mEI broth (263), brain heart infusion broth (BD), and brain heart infusion broth (BD) with 6.5% Na Cl were evaluated for rapid and selective growth of enterococci by Dr. Allyson Bissing, a postdoctoral fellow formerly associated with our laboratory. Based on her findings, the fastest growth rate for enterococci occurred in bile esculin broth (personal comm unication). Due to its lack of selectivity, this medium was replaced with mE broth. Modified mE broth was prepared in 0.1 M carbonate-bicarbonate buffer, pH 9.3 in stead of deionized water (DI H2O). All enrichments were performed at 37 C in a shaking water bath. mE agar (BD) supplemented with na lidixic acid (0.24 g/l) a nd triphenyltetrazolium chloride (TTC) (0.02 g/l) was used to prepar e the standard mEI medium as suggested by the EPA Method 1600 (263). Indoxyl -D glucoside was added at .75 g/L. mEI agar was used for the recovery of enterococci in all seawater samples analyzed. Environmental water samples were enriched in mE broth at 37C before biosensor analysis. These samples were serially diluted in PBS and plated onto mEI agar and assayed on the biosensor. These plates were incubated in a 36 C incubator and allowed to grow for 1824 hours. Initial counts of e nvironmental water samples we re obtained from filtering 50 and 100 ml samples of seawater through a 0.45 m filter (Milipore, Bedf ord, PA) and this filter was placed on mEI agar to grow at 41C for 18-24 hours. Alternatively, counts
53 were obtained from David Wingfield at the Fl orida Department of Health Tampa Branch Laboratory. Water samples One liter volumes of seawater were obtained biweekly from the Florida Department of Health Tampa Branch Laboratory. These samples were collected according to standard practices in Standard Methods for the Examination of Water and Wastewater (4). Seawater samples were collected from two beaches: North Shore beach (St. Petersburg, FL) and Cour tney Campbell Causeway beach (Clearwater, FL). These two beaches were the sources of seawater samples in all experiments where bacteria were enriched and recovered in this study. Figu re 5 shows the locations of these beaches on the Florida Coast. FIGURE 5. Map showing Tampa Bay beaches where seawater samples were collected. Diamond ( ) Courtney Campbell Be ach; Star ( )North Shore Beach
54 Enrichment and recovery methods for E. coli O157:H7 in seawater Seawater samples were analyzed for E. coli O157:H7 by filtering a volume of seawater (100-300 ml) first through a Whatman #5 filter and th en through a 0.45 m filter using a vacuum pump model number 5KH33DN16AX (General Electric; Ft. Wayne Indiana). The filters were placed onto selective media for growth which included mFC agar (BD) and eosin methylene blue agar (EMB) (BD). The plates were grown at 36C and at 44.5C for 24 hours to evaluate which temperature s howed greater recovery. Any suspect E. coli O157:H7 isolates (green sheen colonies on EMB or blue colonies on m FC agar) were streaked on Sorbitol MacConkey agar (Rem el; Lenexa, KS) supplemented with 0.05 mg/ml Cefixime and 2.5 mg/ml Po tassium tellurite (CT, Dynal; Oslo, Norway) (CTSMAC). Any colorless colonies were tested for agglutination with antiE. coli O157:H7 polyclonal antibody (Kirkegaard and Perry (KPL); Gaithersburg, MD). Later experiments to recover E. coli O157:H7 were performed using immunomagnetic separation with Dynabeads (D ynal Inc.,Oslo, Norway) conjugated to antiE.coli O157:H7 according to manufacturers in structions and as reported by Liu and Li and LeJeune et al. (149, 163). The day before the assay, one liter of seawater was filtered through a 0.45 m filter. This filter was placed in 90 ml of TSB modified with 20 mg/ml novobiocin to inhibit gram positive orga nisms. The sample was incubated at 36C for 6-18 hours. The following day, 20 l of resu spended Dynabeads conjugated to antiE.coli O157:H7 polyclonal antibody was added to 1 ml of the pre-enriched sample. The studies were done in triplicate. These tubes were mounted onto the Dynal magnetic particle concentrator rack. The rack was pla ced on a rotator at 24C and allowed to rock for 10 minutes with continuous agitation to prevent the beads from settling. A strip
55 magnet was inserted into the back of the rack and the rack was placed on a rotator to concentrate the beads into a pellet on the side of the tube. This was done at 24 C for 3 minutes to allow for proper recovery. Th e supernatant fluid was discarded and the magnetic plate was removed. The beads were washed with 1 ml PBST. The strip magnet was reinserted into the rack and the susp ension was rotated again for 3 minutes. The supernatant fluid was discarded and the bead s were washed twice more by the same method. Following the washes, 100 l of PBST was used to resuspend the beads and 50 l of this solution was plated onto CT-SMAC agar. Any col orless colonies on CTSMAC were further tested by slide agglutination with antiE. coli O157:H7 polyclonal antibody and then identified through API (BioMerieux; Hazelwood, MO). Enrichment and recovery methods for V. cholerae in seawater Seawater samples were analyzed in 100 ml volumes. The seawat er was filtered initiall y through a sterile Whatman #5 filter to remove any large debris in the water. Samples were analyzed undiluted and diluted 10-fold and 100-fold. Seawater samples were processed by filtering through a 0.45 m filter using a vac uum pump (General Electric). The filter was removed from the filtration ma nifold with sterile forceps and placed onto Thiosulfate citrate bile salts (TCBS) agar (BD) (244). Plates were gr own at 36 C and at a more selective temperature, 42C for 18-24 h. A ll samples were filtere d in duplicate. The following day any yellow colonies that were 2-3 mm in diameter were removed from the filter with a sterile toothpick and inoculated on TSA and allowed to grow for 18-24 h at 36C. These isolates were further characterized. Characterization of V. cholerae isolat es: API identification and PCR amplification. All isolates that appeared yellow on TCBS agar and were 2-3 mm in diameter were
56 identified. A gram stain and an oxidase test were performed on all suspect isolates. If the oxidase test was positive and the cells were gram negative rods, the arginine dihydrolase test and iron esculin hydrolysis te st were performed to presumptively identify any suspect isolates according to the two-test method (46). If the arginine dihydrolase test and the iron esculin hydrolysis were bot h negative, these isolates were identified using the API 20 E biochemical test strip (BioMerieux). Any putative V. cholerae isolates were tested for O1 somatic antigen (Inaba or Ogawa) by agglutination with polyclonal antiserum to V.cholerae O1 (Difco Laboratories; Fr anklin Lakes, NJ). All isolates identified as Vibrio cholerae through API were further identified using polymerase chain reaction (PCR) (47). Figure 6 outlines the identification process followed.
57 FIGURE 6. Identification scheme for V. cholerae isolates. Positive Negative Positive Negative Positive Negative Positive Negative Yellow colonies 2-3 mm in diamete r Gram Stain Oxidase Test Arginine Dihydrolase Iron Esculin Hydrolysis API, PCR, slide agglutination
58 Primers (IDT; Coralville,IA) ) targeted to a 300 base pair region of the 16S-23S rRNA intergenic spacer (ITS) region of Vibrio cholerae (47, 161) were used for PCR. Both the forward and reverse primers were diluted to a 20 M stock concentration. For each PCR, 2-4 l of whole Vibrio cholerae cells suspended in sterile H2O was added to a 96-98 l master mix prepared for each sample by using a TaKaRa TaqTM PCR kit from TaKaRa Biomedicals (Shiga, Japan). All reaction mix tures included 10 l of 10X PCR buffer, 2.5 mM (each) deoxynucleoside triphospha tes (dNTP), 0.2 M (each) forward and reverse primers, TaKaRa Taq DNA polymerase (5 units/ l) (TaKaRa Biomedicals), and sterile H2O to raise the total volume to 100 l. The primer sequences are listed in Table 9. All PCR reactions were pe rformed on the Bio-Rad I Cycler (Bio-Rad Laboratories; Hercules, CA). The PCR protocol to identify isolates of V. cholerae is outlined in Table 10. TABLE 9. PCR primers for the identification of V. cholerae isolates (161). Target Sequence Length Vibrio cholerae ITS F 5TTA AGC STT TTC RCT GAG AAT G 3 ~ 300 bp R 5AGT CAC TTA ACC ATA CAA CCC G 3
59 TABLE 10. PCR protocol for V. cholerae isolatesa. a PCR was performed using 100 l volumes in the I Cycler thermocycler (Bio-Rad Laboratories). Primers to the 16S-23S rRNA ITS region of V.cholerae were amplified to identify environmental isolates. Four microliters of the PCR pr oduct was combined with 1 l of the gel loading dye (BioRad Laboratories). Five microliter samples were loaded into a 1% agarose gel and run in TBE electrophoresis buffer (1X working solu tion: 89 mM Tris base, 89 mM boric acid, 2 mM EDTA) at 100 volts. The marker used was a 1 KB benchtop PCR marker (Promega; Madison, WI). Enrichment and recovery methods for S. aureus. Vogel Johnson agar (BD) and Baird Parker agar (BD) were tested for recovery efficiency. These culture media were found to be successful in isolation of S. aureus in swimming pool water (136). Baird Parker agar was chosen as the recovery medium because of its better recovery of S. aureus from seawater when compared to Vogel Johnson ag ar in preliminary experiments. Seawater samples were analyzed in 100 ml volumes. The seawater was pre-filtered through a sterile Whatman #5 filter to remove any large debris in the water. Samples were analyzed undiluted and diluted 10-fold and 100-fold. Seawater samples were processed by filtering through a 0.45 m filter using a vacuum pump (General Electric). The filter was removed from the filtration manifold with sterile forceps and placed onto Baird Segment Number of Cycles Temperature Duration 1 1 95 C 2 min s 2 30 95 C 1 mins 58 C 1 m ins 72 C 2 m ins 3 1 72 C 10 mins Holding mode 4C
60 Parker agar. These plates were placed in a 36C incubator and allowed to grow for 48 hours. Characterization of S. aureus isolates. Any black colonies with surrounding clear halo area were removed from the culture plat e with a sterile toothpick and restreaked on TSA. These isolates were grown for 18-24 hour s in a 36C incubator. A gram stain and a catalase test were performed on all suspect isolates. Any catalase positive isolates were streaked on Mannitol Salt Agar (MSA) (Difco Laboratories). Any isolates which tested positive for mannitol fermentation were tested for coagulase and DNase activity. Citrated rabbit plasma (Remel; Lenexa KS) was rehydrated according to manufacturers instructions. A heavy inoculation of the susp ect isolate was made in 0.5 ml fractions of citrated rabbit plasma. These tubes were incubated for up to 4 hours in a 36C water bath. Visible clotting of the plasma was considered a positive result. DNase test agar (Difco Laboratories) was made according to manufacturers instructions. Suspect isolates and a positive control were streaked onto the agar. After 18-24 h of growth, the agar was flooded with hydrochloric acid. Any visible clearing of the agar around the growth was considered a positive result. Any isolates which were positive for one or both coagulase and DNase were frozen down in tubes of TSB and glycerol. They were stored in a -80C freezer. The identification scheme for S. aureus is outlined in Figure 7.
61 FIGURE 7. Identification scheme for S. aureus isolates. Positive Negative Positive Negative Positive Negative Positive Negative Yellow colonies 2-3 mm in diamete r Gram Stain Catalase Test F e rm e n tat i o n o f m a nni to l Coagulase & DNase test Staphylococcus aureus
62 Antibodies Sources Antibodies were purchased from the following companies: rabbit polyclonal anti-Strep Group D antibody, American Resear ch Products (ARP) (Belmont, MA); rabbit polyclonal antiEnterococcus antibody, Virostat (Portland, ME); rabbit polyclonal Group D antiserum, Lee Laboratories (Franklin Lakes, NJ); affinity purified polyclonal antiE.coli O157:H7, Kirkegaard and Perry (KPL) (Gaithersburg, MD); rabbit polyclonal antiVibrio cholerae O1 serotypes Inaba, Ogawa, Hikojima, Difco Laboratories (Detroit MI); rabbit polyclonal antiVibrio cholerae O1, Murex (Lenexa, KS); rabbit polyclonal antiVibrio cholerae serotypes Ogawa, Inaba, Denka Seiken (Tokyo, Japan); monoclonal antiVibrio cholerae O1, AUSTRAL Biologicals (San Ramon, CA); monoclonal antiVibrio cholerae O1, New Horizons Diagnostics Inc. (Columbia, MD); polyclonal rabbitantiS. aureus antibody Maine Biotechnology Services Inc. (Portland, ME); goat antirabbit IgG labeled with Cy5 (Jackson Immunoresearch; West Grove, PA)and monoclonal anti-peptidoglyca n, Chemicon International (Temecula, CA). Antibodies used in development of assays for S. aureus were screened through ELISA by Dr. Marianne Kramer, University of South Flor ida, Tampa, FL. All commercial antibodies were screened by enzyme linked immunosorbent assay (ELISA). Custom prepared antibodies were also te sted. A rabbit polyclonal antibody was made by Strategic Biosolutions (SBS) (Newark, DE). Two rabbits were injected with 2.0 x 109 CFU/ml of heat-killed Enterococccus faecalis ATCC 19433. Cells were serially diluted in PBS and a viable count was performed. All cell preparations were heat killed by boiling for 20 minutes before being sent for injection. Test bleeds and final exsanguination samples were screened thro ugh ELISA. Any antibodies which showed
63 high binding affinity for the target antige n were purified and used to develop the biosensor immunoassay. A monoclonal antibody to the immunodeterminants of Ent. faecalis and Ent. faecium was made by the Hybridoma Core Laboratory at the University of Florida, Gainesville, FL. Six mice were injected with heat-killed cell preparations of Ent. faecalis and Ent. faecium Two of the mice were injected with 3.5 x 108 CFU/ml of Ent. faecium (ATCC 19434) cells, two were injected with 1.2 x 109 CFU/ml of Ent. faecalis (ATCC 19433) cells and two were injected with a mixture of both types of bacteria. The cells were heatkilled by boiling for 20 minutes before injec tion. IgM from Clone HL 1869 was screened through ELISA for assay development. A second set of six mi ce was immunized after the first monoclonal antibody project was fi nished. Three mice were immunized with 5.0 x 108 CFU/ml of Ent. faecalis cells (ATCC 19433) cells. Two mice were immunized with 1.0 x 108 CFU/ml of Ent. faecium (ATCC 19434) cells. One mouse was immunized with 25 g of a crude extract of teichoic acid prepared from Ent. faecalis (ATCC 19433) cells. IgG from clone 1H8-3E8 was screened through ELISA for assay development. Purification methods. Polyclonal antiE.coli O157:H7 (KPL) was affinity purified by the manufacturer. The antibody was rehydrated according to manufacturers instructions to a final concentration of 1 mg/ml. IgG purification using a protein A column kit (Amersham Biosciences; Piscataway, NJ) was utilized to purify whole serum used in this study. The purification was done accordin g to manufacturers instructions. A 1 ml aliquot of antibody was passed through the IgG column. The column was washed with 10 ml of 0.1 M sodium phosphate buffer (pH 7.0) to remove any unbound antibody. The IgG was eluted with 6 ml of 0.1 M sodium citrate (pH 3.0) and collected in 0.5 ml
64 fractions in microcentrifuge tubes cont aining 0.5 ml of 0.1M carbonate-bicarbonate buffer (pH 9.3) to neutralize the eluted sa mples. A DU-64 spectrophotometer was used to analyse fractions for protein content. The samples with highest A280 nm readings were combined for labeling. A buffer exchange was performed using 0.1 M carbonatebicarbonate buffer (pH 8.5) for labeling purpos es and to prevent denaturation of the antibody. The exchange was done in an Amic on Ultra-4 centrifugal filter (Milipore) with a 30,000 molecular weight cutoff and by centrifu gation in an Allegra 6R centrifuge at 4,000 x g for 30 minutes. This exchange was performed three times. In preliminary experiments, ARP polyclonal antibody was affinity purified through a procedure developed by Dr. My Lien Dao, University of South Florida, Tampa, FL. A high concentration of cells (~108 CFU/ml) was passively adsorbed to nitrocellulose paper by incubating the membrane with the bacterial cells at 24C for 1 h with continuous agitation. The membrane was washed w ith 2 ml of 0.01 M phosphate-buffered saline 0.1% Tween 20 (PBST). Two hundred microliters of antibody was incubated with the nitrocellulose membrane at 24C for 1 h with continuous agitation. The antibody was aspirated and the filter was washed with 2 ml of PBST. The antibody was eluted with 0.1 M glycine for 10 minutes. Protein concen tration was calculated by measurement of A280 on a DU-64 spectrophotometer. A buffer exchange was performed as described previously. A Carbo-Link affinity column kit fro m Pierce Biotechnology (Rockford, IL) was tested for its effectiveness in affinity purif ication of the ARP antibody. A sample of lipoteichoic acid was prepared for conjugation to the gel column. Two methods were attempted to obtain a crude extract of teichoic acid for attachement to the Carbo-Link
65 affinity column. The first method was one us ed for serological identification of group D Streptococci in Standard Methods for the Examination of Water and Wastewater (4). Cells of Ent. faecalis were grown in Todd-Hewitt broth and concentrated by centrifugation in an Allegra 6R centrifug e at 3,000 x g for 5 minutes. Cells were resuspended in 0.5 ml saline and autoclaved for 15 minutes at 121 C. The bacteria were centrifuged again as described above and th e supernatant fluid was screened by ELISA for teichoic acid. Another method of extractin g teichoic acid was performed by taking a heavy suspension of Ent. faecalis cells in 10 ml PBS and adding a pinch of lysozyme, DNase and RNase to the suspension and allo wing the sample to sit at 24C for 24-48 hours. Ethanol was added at two and a ha lf times greater volume than the original volume. The entire sample was then centrifu ged in an Allegra 6R centrifuge at 4,000 x g for 1 minute. The supernatant fluid was de canted and 1 ml of water was added to the pellet. The sample was centrifuged once mo re at 1,500 x g for 30 seconds to remove all insoluble debris (76). The concentration of the sample was calculated by extrapolation from a calibration curve (Figure 8). This calibration curve was created by analyzing known glucose concentrations on a DU-64 spec trophotometer after reacting the glucose with phenol and sulf uric acid (58).
66 FIGURE 8. Standard curve of glucose. The phenol and sulfuric acid reaction with different glucose concentrations was mon itored by DU-64 spectrophotometer at 488 nm and the readings were used to create a stan dard curve (58). A concentration between 0.510 mg of carbohydrate was needed to attach to the affinity column. The crude extract was conjugated to the carbo-link column according to manufacturers instructions. The teichoic ac id sample was oxidized using sodium-metaperiodate provided in the kit. The sample was allowed to react at 24C for 30 minutes. The oxidized sample was then applied to a desalting column and three 2 ml fractions were collected. These fractions were analyzed on a spectrophotometer DU-64 spectrophotometer at 488 nm. The fraction with the highes t absorbance contained the oxidized carbohydrate. A carbo-link gel column was equilibrated with 6 ml of PBS. A porous polyethylene disc was used to pack th e column and the Carbo-link gel slurry was added to the column (~4 ml slurry containing 2 ml gel). The liquid from the column was drained and the oxidized sample was applied to the gel matrix. The column was gently rotated end-over-end for 6-24 h at 24C. Following incubation, the liquid was drained from the column. The fractions were anal yzed by a DU-64 spectrophotometer at 488 nm Standard Curve: Glucose0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 010255075 Glucose (ug/ml) Glucose Absorbance ( 488 nm )
67 and based on those reading, the coupling effi ciency was calculated. After conjugation, the column was stored in PBS 0.02% sodium azide. Antibody was purified according to manufacturers instructions. The column was equilibrated with 6 ml of column buffer (0.01M PBS, pH 7.4). One a nd a half milliliters of the antiserum was applied to the column and allowed to enter the gel bed and react with the affinity column for 1 h at 24C. After incubation, the column was wa shed with 12 ml of column buffer. The bound antibody was eluted with 100 mM glyc ine buffer (pH 3.0). One millilite r fractions were collected in 1.5 ml microcentrifuge t ubes containing 0.1 M carbonate-bicarbonate buffer (pH 9.3) to neutralize the fractions. The fractions were analyzed using a DU-64 spectrophotometer at 280 nm. The fractions of interest were pooled and a buffer exchange was performed as described previously. All antibody samples were screened initially as whole serum using ELISA. Those which showed high affinity to the target cells were purified by one of the methods mentioned above. Antibodies we re then labeled with biotin and or cyanine 5 (Cy5) as needed. Only those which showed high bindi ng affinity to the target of interest on ELISA were eventually tested on the fiber optic biosensor. Labeling A labeling kit (FluoroLink Cy5 Reactive 5-pack; Amersham Biosciences, Piscataway, NewJersey) was used for labeling antibodies with cyanine 5 (Cy5) dye molecules. The labeling was done according to manufacturers instructions. A concentration of at least 1 mg/ml of antibody was necessary to efficiently label. Following purification, 0.5 ml of antibod y diluted with 0.1 M carbonate-bicarbonate buffer (pH 9.3) was added to a Cy5 reactive dye pack. The dye vial was incubated in the dark at 24C for at least 1 h or alternativel y, the vial was incubated at 4C for 18-24 h.
68 After incubation, free dye was separated fro m labeled antibody by separation on a BioGel P-10 (Bio-Rad Laboratories, Hercules,C A) column with an exclusion limit of 1,50020,000 molecular weight. The column was equ ilibrated with PBS0.02% sodium azide. Labeled antibodies were collected in 0.5 ml fractions and analyzed on a DU-64 spectrophotometer for protein c ontent. Molar concentrations of dye and protein were calculated and the final dye to prot ein ratio was also calculated. At least 1 mg of antibody was necessary for efficient biotin labeling. Biotin labeling was performed according to manufacturers instructions using EZ-Link NHS-LC-LCBiotin (succinimidyl-6-(biotinamido) hexanoate) (Pierce; Rockford IL). One mg of EZLink NHS-LC-LC-Biotin was dissolved in 1 ml of N,N -dimethylformamide (DMF). Seventy five microliters of this solution wa s mixed with 425 l of antibody (~1mg/ml). This solution was incubated at 4 C for 1824 h. Unattached biotin was separated from the labeled antibody by using gel filtration on a Bio-Gel P10 column equilibrated with PBS with .02% sodium azide. Labeled anti bodies were collected in 0.5 ml fractions and analyzed on a DU-64 spectrophotometer (Bec kman Coulter) for protein content. Antibody samples whose final concentration was less than 1 mg/ml were stabilized by adding bovine serum albumin (BSA) to raise the concentration to 1 mg/ml. All labeled antibodies were stored at 4C until neede d. Whole serum samples that were unpurified and unlabeled were stored in a -20C freezer indefinitely. Screening. All antibodies, commercial and cust om made, were screened using ELISA. Enterococcus faecalis (ATCC 19433) and the other organisms listed in Tables 7 and 8 were serially diluted in PBS and placed in a 96-well ELISA plate (Nunc, Rochester, NY). The plate was incubated fo r 18-24 h at 4C. The following day, the
69 plate was washed with PBST. The plate was blocked with PBS containing 2mg/ml Casein and 2 mg/ml bovine seru m albumin (BSA) for 1 h and then washed again with PBST. Primary antibody specific to the target antigen was diluted in PBS to a concentration of 10 g/ml. The antibody was incubated on the plate at 24C for 1 h and then the plate was washed again with PBST. A 1 mg/ml solution of anti-rabbit IgG (Jackson ImmunoResearch, West Grove, PA) co njugated to horse radish peroxidase (HRP) was diluted to 2 g/ml in PBS and incubated on the plate at 24C for 1 h. The plate was washed again with PBST. A Quanta Blue kit (Pierce; Rockford, IL) provided the substrate to react with the HRP enzyme. The substrate was mixed together with the stabilizing solution provided in the kit according to manufacturers instructions. The mixture was a combined 9 parts substrate solu tion to 1 part stable solution. One hundred microliters per well of the substrate was incubated on the plate for 25 minutes. After the incubation period, 100 l per we ll of stop solution was added to each well and the plate was read on a Spectramax Gemini XS microplate reader (Molecular Devices; Sunnyvale, CA). Analyte 2000 Biosensor The Analyte 2000 biosensor (Research International; Monroe, WA) was used to perform all biosensor experiments in this study (Figure 9).
70 FIGURE 9. Analyte 2000 biosensor (Resea rch International; Monroe WA). This instrument has four channels allowing fo r four samples to be processed at once. In this instrument, optical fibers act as wa veguides carrying electromagnetic waves. An evanescent wave sensing scheme is utilized where only fluorophores close to the surface of the fiber (~100 nm) will generate a signal (157). This is achieved when a 635 nm diode laser sends light throug h the proximal end of a polys tyrene waveguide propagating electromagnetic waves within the optical fiber which are reflected off its exposed surface. An electric field is generated and causes th e excitation of fluorescent molecules at a 1001000 nm distance of the wave guide. The e mission energy is recoupled through the waveguide. The emitted light is collected and the photodiode allows for the quantitation of the emission at wavelengths above 650 nm The binding event is recorded as a numerical signal on a computer. Biosensor Sandwich Immunoassays Sample preparation. All samples assayed using a sandwich immunoassay format on the fiber optic biosensor were prepared by ta king a sample of the test organism from the
71 freezer stock culture and growing the culture on TSA or TSB for 18-24 h at 36C. After this growth period, some of the culture was taken directly from the plate using either a sterile polyester fiber tipped plastic swab or a loop and resuspended in PBS to approximately 108 CFU /ml. This starter sample was serially diluted 10-fold in PBS and plated onto TSA to obtain a vi able count of cells. Aliquot s from this dilution set were assayed directly on the biosensor from lowest to highest concentration. If TSB was used, the culture was centrifuged at 4,000 x g in an Allegra 6R centrifuge to pellet the cells. The supernatant fluid was decanted and the cells were resuspended in PBS. These samples were assayed directly after serially diluting and plating on TSA for a viable count. Preliminary experiments were performed on the biosensor using viable cells. Different sample preparations were used in an attempt to improve assay sensitivity. In all experiments, a starter culture tube was made in PBS to 108 CFU/ml and then serially diluted in PBS. A viable c ount was performed by plating ont o TSA. In heat killed cell experiments, the tubes to be analyzed were boiled for 20 minutes. After cooling, the solutions were assayed directly on the biosen sor. Cell samples were also treated by freeze thawing. Dry ice was combined with denatured ethanol in a Styrofoam cooler. The samples were then placed in the dry ice solution for 1 minute and then allowed to thaw in a 36C waterbath. This procedure was repeated three times to disrupt cells as much as possible. Sonication was also performe d as an alternative cell prep to disrupt the bacterial cells. A 4 ml volume of bacteria resuspended in PBS was placed in a 50 ml plastic conical tube. The tube was placed in a styrofoam cooler with ice to ensure that the tube did not become warm and sonicated at setting number 2 on the Dismembrator 100
72 (Fisher Scientific) for different time intervals. These different sample preparations were compared through ELISA and those which gave the highest signals when reacted with the target specific antibody were assayed on the biosensor. Waveguide preparation. Polystyrene waveguides (Research International), approximately 4 cm in length, were cl eaned by sonication for 30 seconds in 2isopropanol using an ultrasonic cleaner. The wa veguides were rinsed in deionized water and then black ink was placed on the tip of each waveguide to prevent the escape of laser light. The waveguides were allowed to air dry completely and were then placed in a glass capillary tube approximately 4.5 cm in length which served as the reaction chamber. The capillary was cut to hold appr oximately 100 l of solution. A solution of streptavidin was prepared in PBS (pH 7.4) at a concentration of 100 g/ml. One hundred microliters of solution was used to fill the reaction chamber and the waveguides were incubated for 18-24 hours at 4C with this solution. Sensitivity and spec ificity assays. Following incubation with streptavidin, the waveguides were placed into the cuvettes made specifically to fit the instruments laser diode manifold. Figure 10 shows the waveguide and the Analyte 2000 cuvette which holds the fiber during the assay. Waveguides were secured with a cap that fits and secures the waveguide into the cuvette. Two ports coming out of the cuvette were connected to lines of tygon tubing. Wash solution and or antibody solutions were introduced into the cuvette through the proximal inlet. The distal outlet at the top of the cuvette was connected to a waste line.
73 FIGURE 10. Waveguide and cuvette used for biosensor assay. The polystyrene waveguide used in all biosensor assays wa s housed in the black cuvette above. This cuvette was made specifically for th e Analyte 2000 biosensor instrument. The waveguide was then washed with 3 ml of PBST injected into the cuvette using a 1 ml syringe to rinse out any unbound streptavidin. Different capture molecules were tested throughout the course of the study. Concanavalin A (ConA) (Nordic Immunology ; Tilburg, The Netherlands) and Wheat Germ Agglutinnin (WGA) (Vector Laboratories, Burlingham, CA) lectins were tested as potential capture molecules for the sandwi ch immunoassay based on their ability to bind surface carbohydrate groups on Enterococcus cells (27). They were screened using ELISA as described above. After screen ing the two lectins, it was observed that extensive cross-reactivity existed between ConA and the polyclonal anti-Strep Group D antibody (ARP) so this was not pursued. WGA crossreacted to a lesser degree with the antibody and was tested on the biosensor. Biotinylated WGA was diluted to10 g/ml in PBS and then incubated on the streptavidin coated waveguide at 4C for 18-24 h. Polyclonal antibody labeled with biotin specific to the target being tested was prepared in PBS to a concentration of 100 g/ml. Other concentrations of capture antibody were also tested. An initial reading was taken to observe the inherent Polystyrene Waveguide Cuvette Inlet Port for wasteline
74 fluorescence of the waveguide. In later experiments to reduce overall assay time, all antibody preparations were made in blocking buffer (2 mg/ml casein and bovine serum albumin in 0.01 M PBS (pH 7.4)). One hundred microliters of capture antibody was allowed to react in the cuvette for 1 h. Fo llowing incubation, the waveguide was rinsed with 1 ml of PBST. In early experiments, a blocking step was included in the assay to block non-specific binding on the waveguide. The waveguide was incubated with 100 l of blocking buffer for 20 minutes. In later experiments this step was removed and all antibody solutions were prepared in blocking buffer. Background fluorescence caused by nonspeci fic binding of Cy5 labeled antibodies was determined by incubating the Cy5 la beled antibodies (10-20 g/ml) at 5 minute intervals on the waveguides 4 times. After each incubation, the waveguide was rinsed with 2 ml of PBST. The signal recorded after the PBST rinse was considered the background signals. These readings were anal yzed later to deter mine limits of detection (LOD). Cells were serially diluted in PBS for use in the assay. One milliliter of each test sample was allowed to react with the cap ture antibody for 10 minutes. Uncaptured bacteria were washed out with 1 ml PBST. Cy5 labeled detection antibody specific for the target being tested was allowed to react with captured cells for 5 minutes. Any unbound detection antibody was washed out of th e cuvette with 2 ml PBST and a reading was taken. The assay sche me is illustrated in Figure 11.
75 laser on FIGURE 11. Sandwich immunoassay format fo r evanescent wave biosensor. The standard assay time is 20 minutes excluding the attachment of capture antibody on the waveguide (1 h) and sample preparati on (10-30 minutes depending on preparation method). Biotin labeled antibody attached to waveguide: 1 hour Dirty sample is applied to waveguide: 10 minutes Target specific antibodies bind bacterial cells Cy5 labeled antibody applied to captured cells on waveguide-laser excites fluorophore: 5 minutes YYYYY YYYYY YYYYY YYYYY YYYYY YYYYY YYYYY YYYYY YYYYY Y Y Y Y Y YYYYY Y Y Y Y Y YYYYY YYYYY YYYYY Y Y Y Y Y
76 Biosensor Indirect Immunoassays for Enterococci Sample Preparation. The test organism was grown on TSA at 36C for 18-24 h. A starter tube was made by taking a sample from the TSA culture plate with a sterile polyester fiber tipped plastic swab or loop and resuspending the cells in 0.1M carbonatebicarbonate buffer (pH 9.3) to approximately 108 CFU /ml. A 10-fold serial dilution of this sample was made in bicarbonate-carbonate buffer (pH 9.3). Samples from the dilution set were plated onto TSA to obtain a viable count of bacteria. Waveguide preparation. Polystyrene waveguide fibers were cleaned and prepared as described previously with one exception, after the waveguides were dried from cleaning, they were placed in a glass capillary tube, but were not incubated with streptavidin. Bacterial cells resuspended in PBS or 0.1 M carbonate-bicarbonate buffer (pH 9.3) were used to fill the glass capillary tube. The cells were allowed to passively adsorb to the waveguide at 36C for a minimum of 30 minutes Experiments were performed to test the time requirements for adsorption of cells to the waveguide so at times, cells were incubated up to 18-24 hours. A negative c ontrol consisting of either 0.1M carbonatebicarbonate buffer or uninoculated enrichment broth was always included in each assay. Sensitivity and sp ecificity assays Following incubation with the bacterial cell suspension, the waveguides were placed in the Analyte 2000 cuvettes as described previously. The waveguides were rinsed with 2 ml of PBST to wash out any unadsorbed sample. An initial reading was taken to determine the inherent fluorescence of the waveguide and of the dirty sample. Back ground fluorescence of the Cy5 labeled antirabbit IgG was determined by first, allowing a polyclonal antibody non-specific to the target antigen (polyclonal antiV. cholerae O1, Difco Laboratories) to react with the
77 sample on the waveguide for 5 minutes. The wa veguide was rinsed with 2 mls of PBST. Then a Cy5-labeled anti-IgG was incubated on the waveguide for 5 minutes. Any unbound antibody was rinsed with 2 ml PBST and then a reading was taken. This procedure was repeated four times and a limit of detection was calculated based on these readings. Unlabeled polyclonal anti-Strep group D antibody (ARP) (American Research Products, Belmont, MA) was diluted as descri bed previously to a concentration of 10 g/ml. One hundred microliters of antibody was allowed to react with the attached cells for 5 minutes. Any unbound antibody was rinsed away using 2 ml of PBST. An anti-rabbit IgG labeled with Cy5 was diluted to1.5 g/ml in PBS or blocking buffer. One hundred microliters of the anti-IgG was allowe d to react with the primary IgG already bound to the bacteria on the waveguide for 5 minutes. Any unbound anti-IgG was rinsed away with 2 ml PBST and a reading was taken. The assay scheme is illustrated in Figure 12.
78 FIGURE 12. Indirect immunoassay format for evanescent wave biosensor. The standard assay time for the indirect immunoassa y for the evanescent wave biosensor is 10 minutes excluding the sample preparation step (10 minutes for a pure culture) and adsorption of the sample to the polystyre ne waveguide (at least 30 minutes). Data Analysis for Biosensor Assays The data for the waveguides and bacterial concentrations tested were analyzed differently for the sandwich immunoassays a nd for the indirect immunoassays. Data from each waveguide for the sandwich immunoassays were normalized to allow comparison of independent waveguides (266). Values were normalized by setting the signals obtained at the highest concentration tested to 100. The normalized values were Sample containing bacterial cells incubated with waveguide at least 30 minutes at 36C Polyclonal anti-Strep group D antibody incubated with cells for 5 minutes Anti-rabbit IgG labeled with Cy5 dye incubated with cells and primary antibody for 5 minutes. Y Y Y YY Y Y Y Y YY Y Y Y Y YY Y Y YYYY Y Y Y Y YY Y Y YYYY Y Y Y Y YY Y Y Y Y YY Y Y YYYY Y
79 treated as ratios of 100%. An average normaliz ed signal was calculated for each bacterial concentration tested and the standard deviation between the normalized signals for a particular concentration was determined. The limit of detection (LOD) for each waveguide was determined by calculating the standard deviation of the backgrounds and multiplying by three. This was done after th ree background readings were taken. In later experiments the amount of backgrounds performed was increased to four and LODs were calculated differently. The LOD was calculated by averaging the differences between each consecutive background reading. Then th e standard deviation of the mean was calculated and multiplied by three. Those two values were added together and were considered the LOD for each waveguide. Any signal over the LOD was considered a positive. The data for the indirect immunoassays fo r enterococci could not be expressed as normalized signals because different waveguides were used to test various concentrations of bacteria. The limit of detection for each wa veguide was calculated in exactly the same fashion as described above in later exper iments when four background readings were performed. A negative control was always tested along with test samples to assess the extent of non-specific binding of the antig en specific antibody (anti-Strep group D polyclonal antibody). A sample was considered positive only if the signal generated was greater than the limit of detection and also greater than the signal generated by the negative control.
80 Seawater Sample Collecti on and Sample Preparations Unspiked samples. Twenty liter volumes of seawater were collected using a sterile 20 liter carboy by the EPA Method 1600, American P ublic Health Standard Methods for the Examination of Water and Wastewater, a nd the methods described by Simmons and Francy (4, 88, 234, 263). Water was colle cted from Courtney Campbell Causeway beach (Clearwater,FL). Samples were collected three feet from shore and at a depth of approximately 6-12 inches below the water su rface. Water samples were kept on ice for 30 minutes during transit to the laboratory. All samples were processed immediately upon arrival or within 6 hours of collection. On Tuesdays when samples from this site were also collected by the Pinellas County Department of Healt h, counts were obtained from the Florida Department of Health, Tamp a Branch Laboratory, which processed the samples collected by Pinellas County officials. On all other days when samples were collected, initial Enterococcus counts were obtained by using standard EPA methods (263). Fifty and 100 ml samples were filtered through a 0.45 m filter. After filtration, the filter and the unit was rinsed with bu ffered water [1.25 ml stock phosphate solution (34.0 g KH2PO4 in 500 ml reagent grade H2O) + 5.0 ml magnesium chloride solution (81.1 g MgCl2 6H2O/L)/L]. The filter was then placed with sterile forceps on mEI media and incubated at 41C for 24 h. The following day, any colonies with blue halos surrounding them indicating the presence of -D-glucosidase positive enterococci and were counted. Final concentrati ons were reported as CFU/100ml. Spiked Samples. Twenty liter volumes of water were sterilized and used to fill a sterile 20 liter carboy. Test organisms were grown on TSA for 12-24 h at 36C. A sample was taken from the culture plate a nd resuspended in PBS to approximately
81 108 CFU/ml. A serial dilution of the starter tube was made and viable counts were performed to determine the number of live bacteria in each tube. Spiked assay samples were prepared by inoculating the carboy with 3 mls of a particular d ilution of bacteria to obtain the desired CFU/100ml. All assays performed in sterile water were also performed in Instant Ocean (Aquarium Syte ms Inc.; Mentor OH) (3.5% in DI water) following the same procedure described above. Concentration of Seawater with a Hollow Fiber Filter Twenty liter samples of spiked sterile water and unspiked sea water samples were concentrated by using an 80,000 molecular we ight cutoff hollow fiber filter (Fresenius USA; Lexington, ME) (88, 234). Initial expe riments were carried out by the same method but in 20 liters of st erile deionized water. The day before the assay was performed, the filter was bloc ked with 0.1% Laureth-12 (PA LL Corporation; East Hills, NY) elution buffer (blocking buffer) made in 0.01 M PBS (pH 7.4). The filter was incubated at 4C for 18 h. Concentrati on was performed using a Masterflex 4S motorized pump (Model # 77201-62, Cole Parmer Instrument Company; Barrington IL). The filter was connected to the carboy using 36 gauge tygon tubing connected by plastic adaptors and secured with metal clamps. The apparatus used to concentrate water is illustrated in Figure 13.
82 FIGURE 13. Concentration apparatus. Twen ty liter volumes of spiked and unspiked sterile water and seawater were concentrated using an 80 K MWCO hollow fiber filter. Water was continuosly recirculated through the system releasing partial volumes as permeate until the whole sample was filtered. Bacteria were eluted from the filter using 1% Laureth 12 solution. Following concentration, the filtrate was el uted with 250 ml of 1% Laureth-12 elution buffer made in PBS. The elution was perform ed for 1 minute with a visual reduction in buffer volume. A sample from the elution was e ither plated directly or diluted and then 100 l was plated on TSA, mEI and marine agar to obtain a viable count from the concentrated sample and to calculate a percent recovery. This was also performed on the three types of media listed above to obser ve any background flora which may also be growing along with the enterococci. The eluted sample was then filtered through a 0.45 m filter (Millipore). This filter was rinsed with buffere d water and then cut out of the filtration apparatus. Using sterile forceps, the filter was placed into a beaker containing mE enrichment broth prewarmed to 36C. One hundred microliters of this Orbital shaker Pump Pressure Gauge Carboy Clamp Bacteria in Filte r
83 sample were also plated on TSA, mEI and mari ne agar. The beaker was then placed into a 36C shaking water bath to enrich for 8-24 hours. Figure 14 shows the procedure used to process and enrich concentrated seawater samples. Following enrichment and concentration, a sample wa s removed and directly assayed on the biosensor. FIGURE 14. Processing steps for concentrated seawater samples. Eluted sample 250 ml 250 ml filtration unit 400 ml beaker 400 ml beaker Shaking water bath Eluted sample is filtered through a 0.45 m filte r The filter is cut out and placed in 50 mls of mE Enrichment broth with filter placed in 36C shakin g water Biosensor detection
84 Concentration, Enrichment and Biosenso r Detection of Fecal Enterococci Concentration with a hollow fiber filter was combined with biosensor detection to create a sensitive and specific method to de tect fecal enterococci in sea water. Concentration was carried out as described above. During the enrichment period, a 100 l sample was plated at strategic time pe riods, usually at the 8 h, 10 h, 12 h, 14 h and 24 h marks. Samples were plated on mEI, TSA, and marine agar to obtain a viable count at that time period. An Enterococcus count of at least 106 CFU/ml was necessary to obtain positive detection on the fiber optic biosensor. At those key time periods, a 1 ml sample was removed and incubated with the polystyrene waveguide as described previously. The glass capillary reaction chamber was filled with the enriched sample and cells were allowed to adsorb to the waveguide at 36C for a minimum of 30 minutes. The biosensor assay was performed according to the protocol described for the indirect immunoassay. SDS PAGE and Western Blot Analysis Enterococcus faecalis cells were grown on TSA agar for 18-24 h at 36C. A sample of the cell culture was taken w ith a sterile loop or swab a nd the sample was resuspended in sterile H2O. Samples were prepared by combining 13 l of sample (whole cell samples of Ent. faecalis untreated or treated by sonica tion for 20 minutes) with 5 l NuPAGE LDS sample buffer (4x) (Inv itrogen; Carlsbad,CA) and 2 l H2O and heated at 95C for 5 minutes. A 4-12% Bis-Tris polya ccrylamide gel (Invitrogen; Carlsbad, CA) was loaded with 10 l of the prepared sample Running buffer used to run the gel was 1x SDS-MES running buffer (Invitrogen; Carlsb ad,CA). The gel was run at 200 volts for 30 minutes. A protein color marker (Sigma; St. Louis, MO) with a molecular weight
85 range between 6,500-205,000 was used to r un the gels. The cells samples were transferred to a nitrocellulo se membrane (Sigma) in Tobin buffer (pH 8.3) for up to 48 h at 20 volts. The following day, the membranes were removed from the transfer cell and blocked with 40 ml of blocking solution (5% non-fat dry milk made in PBST) for 30 minutes with continuous agitation. The me mbrane was washed with 40 ml of PBST. Target specific antibody (anti-Strep group D antibody) was added to the membrane at a 1:1000 dilution in PBS and incubated for 30 minutes with continuous agitation. The membrane was rinsed with 40 ml of PBST three times. Anti-rabbit IgG conjugated to alkaline phosphatase was dilu ted 1:5000 and was incubate d on the membrane for 30 minutes with continuous agitation. The memb rane was washed with 10mM EDTA for 30 seconds with agitation. The membrane was then washed with 40 ml of PBST three times. Forty milliliters of substr ate (50 ml borate buffer (pH 9.7) -13 mg o-dianisidine, 13 mg naphthyl acid phosphate) was a dded to the membrane and th e membrane was allowed to develop until bands appeared (59). Immunodot Analysis Nitrocellulose membrane was cut into a 4 inch square. Wells were embossed using a 96-well ELISA plate (Nunc, Fish er Scientific). Cells of Enterococcus faecalis were grown on TSA at 36C for 18-24 h. The cells were resuspended in PBS and a 10-fold serial dilution was made and plated on TSA to obtain a viable count. On some occasions, sonicated cells extracts (sonicated for 20-30 minutes) were analyzed. Ten microliters of each dilution were dotted in the embossed wells and allowed to dry. The membrane was incubated at 4C for 10 minutes. The rest of the immunodot procedure was as described above for Western blot analysis.
86 Scanning Electron Microscopy Analysis Waveguides were prepared as described above for the biosensor assays. After cleaning, incubation with streptavidin a nd biotin-labeled anti Strep group D capture antibody (ARP), the waveguides were incubated with 100 l of an Ent. faecalis cell suspension (3.8 x 108 CFU/ml in PBS). Two of the waveguides were incubated with untreated cells and the other two were incu bated with cells which were sonicated for twenty minutes on setting number 2 using the Di smembrator 100 (Fisher Scientific). The waveguides were washed with 2 ml of PBST and then the waveguides were taken to Betty Loramm, USF Department of Biology to prepare the waveguides for SEM analysis. Waveguides were prepared according to her instructions. The waveguides were fixed in Trump fixative (Electron Microscopy Sciences; Ft. Washington, Pennsylvania). They were rinsed three times in 0.1M sodi um cacodylate buffer (pH 7.4) for 10 minutes each rinse. The waveguides were then placed into 2% osmium tetroxide in 0.1M sodium cacodylate buffer (pH 7.4) for 1 h. Each waveguide was then rinsed three times again in 0.1M sodium cacodylate buffer (pH 7.4) for 1 h. The samples were then dehydrated according to the following sequence: 30% ethyl alcohol-10 minutes 60% ethyl alcohol-10 minutes 90% ethyl alcohol-10 minutes 100% ethyl alcohol-15 minutes 100% ethyl alcohol-15 minutes The waveguides were dried chemically with hexamethyldislazane (HMDS). The whole procedure was repeated and th en the waveguides were critical point dried with liquid CO2. Double stick adhesive was used to mount dried waveguides onto SEM stubs
87 Hitachi S800 (Ernest F. Fullam Inc., Latham, NY). The waveguides were then sputter coated and mounted with Gold-paladium targ et using a Pelco Model 3 sputter coater. Once sputter coated, the waveguides were ready for SEM analysis. Growth Curve Experiments for Enterococci Cells of Ent. faecalis were grown in TSB broth at 36C for 18-24 h. A one percent inoculation was performed in the enrichment br oth being tested (mE broth or bile esculin azide broth). One hundred microliters of th e enrichment culture was plated on either TSA or mEI or both every 30-60 minutes. Ab sorbance (OD 540 nm) readings were also taken in growth curve experiments usin g a DU-64 spectrophotometer. The growth experiments were performed over a 6 hour period.
88 RESULTS Development of Biosensor Sandwich Immunoassay for Enterococci All initial experiments were performed on the Analyte 2000 biosensor using a sandwich immunoassay format. Whole cells of Enterococcus faecalis were resuspended in PBS and serially diluted by 10-fold dilu tions. The samples were assayed from lowest to highest cell concentration. Antibody specificity Several commercial a nd custom made antibodies were tested for use in the development of a sensitive and specific assay for the detection of fecal enterococci in seawater. Seawater of high quality contains a very low concentration of enterococci in seawater (0.01 CFU/ml 1 CFU/ml) and necessita tes a method sensitive enough to enumaerate such a low concentration. A high quality antibody was a very important part in developing such an assa y using the Analyte 2000 biosensor. Figure 15 shows the results of an ELISA screening of all the antibodies tested. Antibodies were diluted to 10 g/ml except the monoclonal anti bodies tested which were diluted 1:10 in PBS.
89 0 10000 20000 30000 40000 50000 60000 70000 10^7Enterococcus faecalis concentration (CFU/ml) HL 1869 IgG Mab 1H8-3E8 IgM Mab Virostat polyclonal IgG American Research Products IgG Lee Labs polyclonal IgG Stategic Biosolution polyclonal IgG Relative fluorescence FIGURE. 15. Enterococcus antibodies tested for biosensor assay optimization. All antibodies were tested against Ent. faecalis cells at 10 g/ml except the monoclonal antibodies which were diluted 1:10 in PBS. The commercial polyclonal antibodies from Am erican Research Products (ARP), Lee Laboratories, and the custom made Strategi c Biosolutions (SBS) polyclonal antibodies showed the greatest affinity for the Enteroccoccus faecalis cells. These antibodies were tested to evaluate their efficiency at detecting low levels of Ent. faecalis on the biosensor. Figure 16 shows the results of an ELISA used to determine the binding affinity of the selected antibodies to low concentrations of enterococci. The custom made SBS antibody and the commercial ARP gave the highest signals when tested against low concentrations of Ent. faecalis
90 40000 45000 50000 55000 60000 65000 10^810^710^610^5 Enterococcus faecalis concentration (CFU/ml) Lee labs polyclonal IgG American Research Products polyclonal IgG Startegic Biosolutions polyclonal IgG Relative fluorescence FIGURE 16. Comparison of high affinity Enterococcus antibodies for use in assay development. Polyclonal antibodies from Lee Laboratories, American Research Products (ARP) and Strategic Biosolutions (SBS) were tested against Ent. faecalis cells at 10 g/ml. Both the ARP and the SBS had a higher affinity for Ent. faecalis than the polyclonal antibodies from Lee Laboratories did. ARP and SBS antibodies were assayed using We stern blot analysis, as described in the Materials and Methods, to see which antibody sh owed the greatest affinity for different epitopes of Ent. faecalis cells. The cells were tested sonicated and unsonicated with SDS-PAGE. Part of each sample was centri fuged to assay the pellet and supernatant fluid to determine which frac tion contained a greater amount of antigen to bind with the ARP antibody. Subsequent Western blot anal ysis revealed that the SBS antibody showed greater affinity for a low molecular weight molecule, but the ARP antibody showed a higher affinity for more than one epitope on the Ent. faecalis cells (Figure 17) Also, it was observed that after sonication, more sol uble antigen was present in the supernatant fluid than in the pellet fraction. The AR P antibody was chosen for further assay development on the Analyte 2000 biosensor.
91 Strategic Biosolutions Antibody 1 2 3 4 5 6 7 8 9 10 American Research Products Antibody 1 2 3 4 5 6 7 8 9 10 FIGURE 17. Western blot analysis of polyclonal ARP and SBS antibodies for Enterococcus biosensor assay. Lane 1: Marker Lane 2: Unsonicated Ent. faecalis ATCC 19433 cells Lane 3: Sonicated Ent. faecalis ATCC 19433 cells Lane 4: Supernatant from s onicated cells after centrifugation Lane 5: Pellet from sonicated cells after centrifugation Lane 6: Supernatant unsonicat ed cells after centrifugation Lane 7: Pellet unsonicated cells after centrifugation Lane 8: N/A Lane 9: 100 g/ml lipoteichoic acid from Ent. faecalis (Sigma) Lane 10: 50 g/ml lipoteichoic acid from Ent. faecalis (Sigma) 205 kDa 116 kDa 66 kDa 20 kDa 45 kDa 29 kDa 14.2 kDa 6.5 kDa 205 kDa 116 kDa 66 kDa 20 kDa 45 kDa 29 kDa 14.2 kDa 6.5 kDa
92 Of note is the lack of a distinct ba d for unsonicated cell preparations of Ent. faecalis This is probably due to the in accessibl e location of the group D antigen (LTA) of enterococci. The LTA is located under the cell surface where antibodies do not have binding access. Crossreactivity assays. The ARP polyclonal antibody was assayed for specificity against environmental isolates of enteroco cci and also other potentially cross-reactive species. Positive samples were recorded as any fluorescence readings that gave a signal to noise ratio of 2 or greater. Those micr oorganisms tested are listed in Table 11 along with the signal to noise ratio that was obtai ned at the concentration required to give a positive. Several bacteria were shown to crossreact with the ARP antibody at the same concentrations as the enterococci tested. Those species included Lactobacillus arabinosa Leuconostoc Pediococcous dextrinicus Staphylococcus aureus TABLE 11. Specificity data for anti-St rep group D antibody (American Research Products) generated by ELISA. Microorganism Signal to noise ratioa Concentration required for positive signal (CFU/ml) Aerococcus viridans N/Ab N/Ab (ATCC 700406) Ent. casseliflavus (USF 691) 2.6 3.2 x 107 Ent. casseliflavus (USF 703) 3.6 1.0 x 108 Ent. casseliflavus (USF 707) 3.1 2.3 x 107 Ent. casseliflavus (USF 740) 2.0 2.7 x 107 E. coli (USF 640) N/Ab N/Ab Ent. faecalis (ATCC 19433) 4.6 7.0 x 106 Ent. faecalis (USF 693) 2.7 3.4 x 105 Ent. faecalis (USF 696) 6.0 3.2 x 106 Ent. faecalis (USF 709) 2.2 5.8 x 105 Ent. faecalis (USF 710) 2.2 4.5 x 105 Ent. faecalis (USF 711) 2.4 3.2 x 105 Ent. faecalis (USF 713) 4.8 4.2 x 106 Ent. faecalis (USF 715) 6.1 3.5 x 106 Ent. faecalis (USF 719) 3.7 7.8 x 106
93 a A signal to noise ratio over 2.0 was requi red for a value to be considered positive. bN/A, no positives were observed S. aureus consistently showed crossreactivity with the ARP antibody. While Ent. faecalis and S. aureus both contain a lipoteichoic acid antigen, S. aureus has a Ent. faecalis (USF 720) 6.5 7.6 x 106 Ent. faecalis (USF 721) 2.0 1.0 x 106 Ent. faecalis (USF 723) 4.1 1.3 x 107 Ent. faecalis (USF 724) 2.1 1.3 x 106 Ent. faecalis (USF 726) 5.4 1.2 x 107 Ent. faecalis (USF 727) 4.9 7.9 x 106 Ent. faecalis (USF 734) 2.1 1.0 x 106 Ent. faecalis (USF 747) 3.2 1.4 x 106 Ent. faecalis (USF 751) 2.4 4.3 x 105 Ent. faecium (USF 692) 2.8 6.1 x 106 Ent. faecium (USF 725) 3.9 7.0 x 106 Ent. faecium (USF 730) 2.4 6.0 x 106 Ent. faecium (USF 753) 3.6 1.6 x 107 Ent. faecium (USF 754) 5.1 5.6 x 107 Ent. faecium (USF 759) 3.9 6.0 x 106 Ent. durans (USF 694) 2.4 1.8 x 106 Ent. durans (USF 732) 2.0 1.9 x 107 Ent. gallinarum (USF 767) 2.0 1.5 x 107 Group A Streptococcus 4.1 1.8 x 108 (8-R291 TGH) Group B Streptococcus N/Ab N/Ab (1215-RNSJ183V) Lactobaccillus arabinosa 2.0 2.0 x 106 (USF 541) Lactococcus lactis (USF 690) N/Ab N/Ab Lactococcus lactis (USF 776) 3.0 6.2 x 107 Leuconostoc (USF 706) 2.6 8.3 x 106 Pediococcus dextrinicus 2.8 1.2 x 106 (ATCC 33087) S. aureus (USF 645) 2.0 2.6 x 107 Streptococcus bovis 4.5 2.7 x 108 (ATCC15351) Streptococcus bovis 2.2 8.7 x 108 (ATCC 49133) Vibrio alginolyticus (USF 516) 2.5 8.4 x 108 Weissella confusa 3.2 2.9 x 107 (ATCC 14434)
94 ribitol teichoic acid and Ent. faecalis has a glycerol teichoic acid. The nature of this crossreactivity, however, was found to be pr edominantly due to the protein A molecules on the surface of S. aureus as this phenomenon has been reported in previous studies (100). This characteristic was tested on ELISA by using an S. aureus strain mutated to not produce protein A (Figure 18). This strain was tested against a Cowen I strain that had abundant protein A molecules on its surface In previous studies, it was found that protein A bound the Fc portion of antibody mol ecules with great affinity. This was found to be the case with the ARP antibody as well. There was still some degree of crossreactivity, however, even with the non-protein A S. aureus 0 5 10 15 20 25 30 10^910^810^710^610^510^410^3Staphylcoccus aureus concentration (CFU/ml) Non-Protein A S. aureus Cowen I S. aureus Signal to noise rati o FIGURE 18. Crossreactivity between ARP antibody and S. aureus. Crossreactivity was assessed through ELISA using ARP at 10 g/ml. Non-protein A S. aureus ATCC 10832 and USF 613 Cowen I S. aureus were tested to see if ARP antibody specifically bound epitopes on the cell surface of S. aureus or if crossreactivity was due to nonspecific binding of the Fc portion of the antibody to protein A on the cell surface.
95 Affinity purification of ARP antibodies The ARP antibody was affinity purified to attempt to improve the sensitivity and specificity of the test assays. The antibody was purified using an affinity column consisting of a gel bed that had oxidized teichoic acid from Ent. faecalis cells attached to it. The column was made according to manufacturers instructions and the antibody was purified also according to manufacturers instructions. Figure 19 shows the mean signal to noise ratio obtained on ELISA at various Ent. faecalis concentrations. Affinity purification did not improve the binding affinity of the ARP antibody to Ent. faecalis cells. 0 2 4 6 8 10 12 14 16 18 10^810^710^610^510^410^3Ent. faecalis concentration (CFU/ml) Affinity purified ARP IgG purified ARP Signal to noise ratio FIGURE 19. Affinity purification of ARP pol yclonal antibody. An affinity column kit from Pierce was used to create an affinity column conjugated with teichoic acid extracted from Ent. faecalis cells. This column was used to affinity purify polyclonal ARP antibody. A protein A column (Amersham Biosciences) was used to IgG purify the ARP antibody as well. Varying capture antibody concentration. Various conditions were tested to develop a sensitive sandwich immunoassay on the Analyte 2000 biosensor. Two different ARP
96 capture antibody concentrations of, 100 g/ml and 200 g/ml, were tested to determine which concentration resulted in optimum capture on the polystyrene waveguides. The mean LOD for the 200 g/ml assay was 5.1. The mean LOD for the 100 g/ml assay was 16.3. The mean fluorescence over background of those waveguides incubated with 200 g/ml of capture antibody went up with in creased cell concentra tions; however, the signals did not increase logarithmically as was expected in this type of immunoassay (Figure 20). In fact, the signals were very low, even at 108 CFU/ml. There also was no discernible pattern when normalized si gnals were graphed (Figure 21). 0 5 10 15 20 25 30 35 40 10^410^510^8 Ent. faecalis concentration (CFU/ml) 200 ug/ml capture antibody 100 ug/ml capture antibody Mean fluorescence over background FIGURE 20. Various antibody capture concentrations for optimization of Enterococcus sandwich immunoassay. Biotinylated ARP antibody was tested at 200 g/ml and 100 g/ml to optimize the sandwich im munoassay. The mean LODs were 5.1 and 16.3, respectively.
97 0 20 40 60 80 100 120 140 160 10^410^510^8Ent. faecalis concentration (CFU/ml) 200 ug/ml capture antibody 100 ug/ml capture antibody Mean normalized signal (%) FIGURE 21. Normalized values for the mean signals shown in Figure 20. Results in Figures 20 and 21 are data obtained from two experiments. Varying detection antibody concentration Two different concentrations of ARP detection antibody, 20 g/ml and 40 g/ml, were tested to try and optimize the biosensor immunoassay as well. The 40 g/ml detection antibody concentration showed improved detection of Ent. faecalis but only at 108 CFU/ml (Figures 22 and 23). This still, did not improve the sandwich assay sufficiently to develop the assay in this format for screening environmental sea water samples. The mean LOD of the assays tested with 20 ug/ml of detection antibody was 26.1. The mean LOD for the assays tested with 40 ug/ml detection antibody was 93.
98 -50 0 50 100 150 200 250 02468 Log Ent. faecalis concentration (CFU/ml) 20 ug/ml 40 ug/ml Mean signal over background (p Am p s ) FIGURE 22. Determining op timal detection antibody concentration for biosensor sandwich immunoassay. Two c oncentrations of ARP Cy5 labeled polyclonal antibody were tested: 20 g/ml and 40 g/ml. Biotinylated capture antibody was used at 100 g/ml. Mean fluorescence signals over the limit of detection for each concentration tested are represented. The mean LODs were 26.1 and 93, respectively. 0 20 40 60 80 100 120 02468Log Ent. faecalis concentration (CFU/ml) 20 ug/ml 40 ug/ml Normalized signal (%) FIGURE 23. Normalized values for the mean signals shown in Figure 22. Figures 22 and 23 are summarized data from two different experiments.
99 Comparison of broth media and plate media. Ent. faecalis cells were grown on plate media or in broth media and assayed direct ly on the biosensor to see if the culture medium would make a difference in the amount of antigen expressed by the cells that the antibody could bind. The results suggested that the broth culture made a difference in the binding of the antibodies to antigenic dete rminants on the cells, but not significantly enough to improve the sensitivity of the sandwich immunoassay (Figure 24). The cell signals were very low. The mean LOD for the assay was 23.9. 0 10 20 30 40 50 60 70 80 10^310^410^510^8 Enterococcus faecalis concentration (cfu/ml) TSB culture TSA culture Mean signal over background (pAmps) FIGURE 24. Comparison of biosensor signals when Ent. faecalis cells were grown for 18 hours in trypticase soy broth (TSB) vs.t rypticase soy agar (TSA). These results were obtained from two replicate experiments. Mean fluorescence signals over the limit of detection for each concentration tested are represented. This assay was performed using a sandwich immunoassay and a polyclona l antiStrep group D antibody (American Research Products) to capt ure and detect cells of Enterococcus faecalis The mean LOD was 23.9. Different antigen preparations tested for assay development. Various antigen preps were attempted to try to optimize th e immunoassay and obtain detection of Ent. faecalis
100 within the limits required by the EPA for beach water quality. Antigen preparations attempted included sonication, boiling and fr eeze-thawing. The methods for these antigen preparations were described in the Materials and Methods. Boiling the cells was the least efficient preparation for improving the immunoassay. This was most likely due to the denaturation of some of the antigenic proteins by the heat (Figure 25). Freezethawing also showed an improvement, however, sonication showed a drastic improvement in sensitivity when screened with ELISA. The ELISA, however, was not a sandwich assay. The sandwich immunoassay form at needed to be tested on the Analyte 2000 biosensor. 0 1 2 3 4 5 6 7 8 9 10 10^810^710^610^510^410^310^2Ent. faecalis concentration (CFU/ml) Untreated Ent. faecalis Boiled Ent. faecalis Freeze-thawed Ent. faecalis Sonicated Ent. faecalis Signal to noise rati o FIGURE 25. Comparison of antigen prep arations for immunoassay optimization. An ELISA was performed to compare three different sample preparations: repeated freeze thawing with dry ice and denatured ethanol; boiling for 20 minutes; and sonication for 10-30 minutes. Samples were tested in duplicate. Anti-Strep group D antibody (10 g/ml) (American Research Products) was used to test the effect of each sample preparation on assay sensitivity. Sonication of Ent. faecalis cells proved to be the most efficient sample preparation.
101 Sonication and boiling antigen preparation on biosensor sandwich immunoassay. Viable and dead cells were assayed on the bios ensor to verify results obtained by ELISA. Cells were grown on TSA overnight as desc ribed in the Materials and Methods and resuspended in PBS. After a 10-fold serial dilution was made, the cells were boiled and allowed to cool and then assayed on the biosensor instrument. Mean signals and normalized signals are shown in Figures 26 and 27. The mean LOD was 48.4. There were minor increases in mean signals ove r background and the normalized signals, but not logarithmic as would be expected with the addition of higher cell concentrations. 0 10 20 30 40 50 60 70 80 90 10^310^410^510^7 Ent. faecalis concentration (CFU/ml) Live cells Boiled cells Mean signal over background (pAmps) FIGURE 26. Comparison of boiled cells a nd viable cells for biosensor detection using a sandwich immunoassay. Mean fl uorescence signals over the limit of detection for each concentration tested are represente d. Biotinylated polyclonal anti-Strep group D antibody (ARP) was used for capture (100 g/ml) and Cy5 labeled ARP was used for detection (20 g/ml) of Ent. faecalis cells in PBS. The mean LOD was 48.4.
102 0 20 40 60 80 100 120 10^310^410^510^7Ent. faecalis concentration (CFU/ml) Live cells Dead cells Normalized signal (%) FIGURE 27. Normalized values for the mean signals shown in Figure 26. Figures 26 and 27 are summarized data from four different experiments. Sonication of Ent. faecalis cells showed improvements in assay sensitivity on ELISA and was tested on the biosensor using the sa ndwich immunoassay. Sonication showed a marked increase in mean fluorescence over the previous signal (Figures 28 and 29). The mean LOD was 15.1. The sensitivity of the sandwich immunoassay on the biosensor was improved by one log concentrations to 107 CFU/ml. This sensitivity was still not low enough for the development of an assay to comply with EPA water quality standards using the biosensor alone. Figures 28 and 29 are summarized data from two different experiments.
103 -100 0 100 200 300 400 500 10^610^710^8 Ent. faecalis concentration (CFU/ml) Ent. faecalis cells unsonicated Ent. faecalis cells sonicated Mean signal over previous signal (p Am p s ) FIGURE 28. Sonicated cell preparation us ed for optimization of biosensor sandwich immunoassay. Mean fluorescence values over the limit of detection for each concentration tested are represented. Bi otinylated polyclonal anti-Strep group D antibody (ARP) was used for capture (100 g/ml) and Cy5 labeled ARP was used for detection (20 g/ml) of Ent. faecalis cells in PBS. The mean LOD was 15.1. 0 20 40 60 80 100 120 10^610^710^8Ent. faecalis concentration (CFU/ml) Ent. faecalis unsonicated Ent. faecalis sonicated Normalized signal (% ) FIGURE 29. Normalized values for the mean signals shown in Figure 28. Figures 28 and 29 are summarized data from three different experiments. Immunodot analysis of sonicated vs. unsonicated cells. An immunodot was performed to verify results obtained from bi osensor assays testing the efficacy of
104 sonicating Ent. faecalis cells. The immunodot showed a di stinct difference in antibody binding to cells that were sonicated and not sonicated (Figure 30). The ARP antibody binds with greater affinity to sonicated cells, suggesting that the antibody has better access to those epitopes such as the group D antigen which is located beneath the cell surface. FIGURE 30. Immunodot analysis of sonicated and unsonicated preparations of Ent. faecalis cells in PBS. Ent. faecalis cells were sonicated for 30 minutes at setting #2 on the Dismembrator (Fisher Scientific). Polyclonal ARP antibody was diluted 1:1000 in PBS was used for detection. SEM analysis All SEM analysis was performed by Jay Bieber at the USF College of Engineering. Waveguides were prepared as described in Materials and Methods. The waveguides were prepared for SEM by Bett y Loramm, USF Department of Biology as described in Materials and Methods. The co mparison of sonicated captured cells and unsonicated captured cells is obvious. There ar e scattered debris and salt deposits on the waveguide analyzed with cells that were unsoni cated (Figures 31 and 32). This debris is 108 107 106 105 104 103 Sonicated Unsonicated
105 more evident on the waveguide exposed to unsonicated cells, than on the other waveguide analyzed. More efficient capture is obtained on the waveguide allowed to react with sonicated cells and further ex plains the improved assay sensitivity when sonication is used for sample preparation FIGURE 31. Unsonicated Ent. faecalis cells captured on polystyrene waveguide. Biotinylated ARP antibody was used to capture the cells.
106 FIGURE 32. Sonicated Ent. faecalis cells captured on polystyrene waveguide. Biotinylated ARP antibody was used to capture the cells. Lectins as capture molecules. Lectins were tested as capture molecules for enterococci on the fiber optic waveguide surface because of their known abilities to agglutinate carbohydrate groups on the surface of gram positive cells. ConA and WGA have been observed to agglutinate enteroco cci and group D streptococci so they were chosen as test molecules. Initial screening with ELISA revealed that ConA cross-reacted extensively with the polyclonal anti-stre p group D antibody used as the detection molecule in the assay. Antibodies are gl ycoproteins so this crossreactivity was not surprising. WGA crossreacted to a lesser degree with the ARP antibody and so WGA was pursued for use on the biosensor sandw ich immunoassay. Because some cross reactivity was also observed between WGA and the Enterococcus antibody, backgrounds were performed by running negative samples (s terile PBS) and then the labeled detection molecules. Cy5 labeled primary anti body did not directly detect captured Ent. faecalis
107 cells, so an anti-IgG labeled with Cy5 was used to amplify the signal of antibodies bound to cells. This antibody was diluted 1:1000 in 0.1M PBS. LODs were calculated as described in Materials and Methods. Using WGA as a capture molecule was successful, but not sensitive enough, to improve the effi ciency of the sandwich assay. Sensitivity was 2.1 x 108 CFU/ml. Figures 33 and 34 show the results of four experiments testing this WGA as an alternative capture in the sandwich assay. The mean LOD was 12.5. 0 50 100 150 200 250 10^610^710^8Ent. faecalis concentration (CFU/ml) WGA as capture, polyclonal anti-Strep Group D antibody as detection Mean signal over background (pAmps) FIGURE 33. WGA as capture molecule for Enterococcus sandwich immunoassay. Biotinylated WGA diluted 1:50 was used to capture Ent. faecalis cells. Mean LOD was 12.5.
108 -20 0 20 40 60 80 100 120 10^610^710^8Ent. faecalis concentration (CFU/ml) WGA as capture, polyclonal anti-Strep Group D antibod y as detection Normalized signal (%) FIGURE 34. Normalized values for the mean signals shown in Figure 33. Figures 33 and 34 are summarized data from four experiments. Optimization of Biosensor Indir ect Immunoassays for Enterococci EPA standards for recreational water quality require a very sensitive assay for proper water quality assessment. The sandwich immunoassay was optimized, but the sensitivity was not adequate to be a feasible test for e numerating indicator organism density in beach water. An indirect assay format (illustra ted in Materials and Methods) was optimized as an alternative to increase assay sensitivity. In preliminary experiments, Ent. faecalis cells suspended in 0.1M carbonate-b icarbonate buffer (pH 9.3) were assayed to ascertain the sensitivity of this type of assay (Figure 35). This format of immunoassay was used to screen antibodies using ELISA and the sens itivity of the ARP antibody was found to be 106 CFU/ml. Using the same assay principles, this format was tested on a polystyrene waveguide on the Analyte 2000 biosensor. Tw o polyclonal antibodies were used in the assay, one specific to group D streptococ ci (ARP), and one used for determining
109 background signals which was specific for V. cholerae O1 (Difco Laboratories). Antirabbit IgG labeled with Cy5 was used to generate a specific signal when excited at 635 nm. Figure 35 was generated with data from five experiments and the mean LOD was 12.2. The sensitivity of th is indirect immunoassay was 105-106 CFU/ml. -1000 0 1000 2000 3000 4000 5000 6000 Neg. control 10^810^710^610^510^410^310^2 Ent. faecalis concentration (CFU/ml) Ent. faecalis cells in carbonate buffe r (pH 9.3) Mean signal ove r background (pAmps) FIGURE 35. Directly adsorbed Ent. faecalis cells in 0.1 M carbonate buffer (pH 9.3). Polyclonal ARP antibody (10 g/ml) was used to detect Ent. faecalis cells in carbonatebicarbonate buffer. Anti-IgG labeled with Cy5 was used to detect ARP antibody bound to the target antigen. Mean LOD was 12.2. Ent. faecalis cells were also tested in filter-sterilized seawater (Figure 36). A negative control was tested to assure that there was no inherent or background fluorescence from the seawater. Figure 36 is summarized da ta from three different experiments and the mean LOD was 13.6. The sensitivity in seawater, detection of 106 CFU/ml, was comparable to that in sterile carbonate-bicarbonate buffer (pH 9.3).
110 -500 1500 3500 5500 7500 9500 11500 Neg. control 10^810^710^610^510^410^3Ent. faecalis concentration (CFU/ml) Ent. faecalis in seawater (pH 6.91) Signal over background (pAmps) FIGURE 36. Directly adsorbed Ent. faecalis cells in seawater. Polyclonal ARP antibody (10 g/ml) was used to detect Ent. faecalis cells in carbonate-bicarbonate buffer. Anti-IgG labeled with Cy5 was used to detect ARP antibody bound to the target antigen. Mean LOD was 13.6. Table 12 shows the mean signal s from five experiments where Ent. faecalis cells in seawater were directly adsorbed to the wave guide. In four out of four experiments, positives signals were seen when negative samples were assayed on the biosensor. This was probably due to nonspecific binding of the detection antibody to the waveuide. These false positives were very close to signals generated by the 105 concentrations tested so positive signals were not significant until 106 CFU/ml.
111 TABLE 12. Comparison of mean signals from five biosensor experiments. aStandard deviation Incubation times were also tested to try and shorten the time required for cells in carbonate-bicarbonate buffer (pH 9.3) to adsorb to the polystyrene waveguides. Figure 37 shows summarized data obtained from four experiments and the mean LOD was 33.6. High concentrations still gave strong positive signals with as short an incubation time as 30 minutes (Figure 37). Table 13 shows th e mean signals from those 4 experiments. False positive were seen in three out of four control samples and because the mean signals at 105 were so close to the false positives, the sensitivity of this assay was 106 CFU/ml. This assay improved the overall sensitivity of the Enterococcus assay by one log concentration, however, the assay was still not sensitive enough to detect enterococci from the environment directly. The sensitivity, detection of 106 CFU/ml, in both matrices was comparable to that on ELISA. Avg. CFU/ml Mean signal SDa Number of (pAmps) positive samples/ total num ber of sam ples 0 56.6 50.5 4/4 3.1 x 105 23.8 22.1 0/5 3.1 x 106 162.4 73.3 2/2 3.0 x 107 395.2 131.1 2/2 4.7 x 108 7053.9 4121.4 5/5
112 0 1000 2000 3000 4000 5000 6000 7000 Neg. control10^710^610^5Ent. faecalis concentration (CFU/ml) Ent. faecalis in carbonate buffer (pH 9.3) Mean signal over background (pAmps) FIGURE 37. Ent. faecalis cells directly adsorbed to a polystyrene waveguide after a 30 minute incubation at 36C. Polyclonal ARP antibody (10 g/ml) was used to detect Ent. faecalis cells in carbonate-bicarbonate buffer. Anti-IgG labeled with Cy5 was used to detect ARP antibody bound to the target antigen. Table 13 shows the mean signal s from four experiments where Ent. faecalis cells in carbonatebicarbonate buffer (pH 9.3) were in cubated at 36C for 30 minutes. The same results were observed in these assays that we re observed in the seawater assays. Signals at 105 CFU/ml were very close to the false positives which were observed. Positive detection was obtained at 106 CFU/ml.
113 TABLE 13. Mean signals from four experiments where Ent. faecalis cells were incubated at 36C for 30 minutes. a Standard deviation Concentration of 20 Liter Volumes of Water Twenty liter water samples were concentrat ed according to the method of Simmons et al (234). Hollow fiber filter concentration was used to enhance the sensitivity of the bioassay. Preliminary experiments were perfor med in 20 liter volumes of sterile water to assess the recovery efficiency of the filter and also to see if samples could be concentrated sufficiently to detect entero cocci directly on the biosensor following filtration (Table 14). Avg. CFU/ml Mean signal SDa Number of positive (pAmps) samples/ total number of samples 0 74.5 59.5 3/4 5.6 x 105 128.6 75.7 4/4 5.6 x 106 994.0 358.6 4/4 5.6 x 107 3695.7 2257.7 4/4
114 TABLE 14. Recovery of Ent. faecalis cells when concentrated with hollow fiber filter. a Experiment was performed with an envi ronmental seawater sa mple. Counts were obtained from the Florida Department of Health Tampa Branch Laboratory. The mean recovery using the Fresensius hollow fiber filter was 89.6%. The lower the cell concentration, the more erratic the recove ry became. This was especially the case when a concentration of ~ 1 CFU/ml was filtered. Overall, this process concentrated water samples 100-fold. This improved assay sensitivity, but an enrichment step was needed to further enhance assay sensitivity. Growth Curve Experiments and Enrichment of Enterococci To enhance the sensitivity of the biosensor assay, an enrichment step was added to the method along with concentration. Initial experiments for enrichment of Ent. faecalis were performed in bile esculin azide broth. This selective enrichment broth, however, did not prove sufficiently selective for use with actual beach water samples. mE broth Concentration Concentration Eluted volume % Recovery in 20 L (CFU/ml) after filtration (mls) (CFU/ml) 2.6 x 104 2.6 x 106 225 109 3.6 x 104 3.0 x 106 185 76.4 3.3 x 104 3.3 x 106 210 105 3.0 x 104 2.0 x 106 230 76.7 2.3 x 104 1.3 x 106 208 60 2.7 x 104 2.7 x 106 268 133 255 3.0 x 104 198 116 225 1.7 x 104 230 86.9 210 2.0 x 104 182 86.7 375 2.3 x 104 105 32 269 1.6 x 104 230 72.5 480 6.0 x 104 170 106 2.1 32 180 13 1.8 200 226 125 1 200 226 214 4.8 100 244 25 .08 10 138 86a
115 was used to perform the rest of the enrichment experiments. The mean generation time of Ent. faecalis in mE broth was 40.5 minutes. The growth curve was performed as described in the Materials and Methods and the results appear in Figure 38. 0 0.5 1 1.5 2 2.5 3 3.5 4 0100200300400Time (minutes) mE broth OD 540 n m FIGURE 38. Growth curve of Ent. faecalis in mE enrichment broth. Based on this growth curve, the mean generation time for Ent. faecalis in mE broth was 40.5 minutes. Concentration, Enrichment and Biosensor Detection Seawater samples were processed in 20 liter volumes on Tuesdays when the Pinellas County Department of Health (DOH) also test ed water from this same beach (Courtney Campbell Causeway beach, Clearwater, FL). Initial counts were obtained from David Wingfield at the Florida Department of Hea lth Tampa Branch Laboratory. A guide was developed similar to what is used in Pine llas County for gauging the quality of beach water based on established EPA guidelines. The Pinellas county guidelines are listed in Table 15.
116 TABLE 15. Pinellas County Health De partment/EPA wate r quality criteria. a These criteria were derived from standard set by the U.S. Environmental Protection Agency. A similar guideline to that described above was created to gauge water quality based on enrichment times and biosensor detection. Seawater samples were processed as described in the Materials and Methods. Fo llowing concentration with a hollow fiber filter, samples were enriched in 50 ml of mE broth. Samples were taken at key time intervals for detection on the biosensor and also to perform a viable count. Based on these data a guide was developed to deter mine the water quality of unspiked samples. The guide consisted of three parts, Good qualit y water, Moderate qua lity water, and Poor quality water, similar to the Pinellas County me thod. These categories were established based on the time it took for a sample to enrich to 106 CFU/ml (the limit of detection for the indirect immunoassay on the biosensor) wher e there should also be positive detection on the biosensor. That general guideline is listed in Table 16 for this integrated method for determining water quality. Pinnellas County Water Quality Guidelines for Enterococcia Good 0-34 CFU/100 ml Moderate 35-104 CFU/100ml Poor > 105 CFU / 100 ml
117 TABLE 16. Biosensor/ enrichment cr iteria for determining water quality. a In two out of three experiments, positive detection was observed after 14 h of enrichment in mE broth. In on e of the replicate exper iments, detection was observed at 14 h. b In two out of three experiments, positive detection was observed between 12-14 h of enrichment in mE broth. One of the rep licate experiments, detection was not observed until14 h of enrichment. c In three out of three experiments, positive detection was observed between 8-10 h of enrichment in mE broth. All experiments to create these guidelines were performed in triplicate in sterile water and in simulated sea water (Instant Ocean, 3.5% in DI H2O). Results were the same in both matrices. Experiments for determ ining criteria for Good qua lity water samples. Sterile water was spiked with a concentration of Ent. faecalis cells that would give an approximate count equal to what the Pinellas County Hea lth Department consid ers good quality water (0-34 CFU/100 ml of enterococci). The proce ss was performed as de scribed in Materials and Methods. Three experiments were perfor med and the results of those experiments are listed in Table 17. Category Viable counts Enrichment time needed for positive biosensor detection Good 106 CFU/ml over 14 hoursa Moderate 106 CFU/ml 12-14 hoursb Poor 106 CFU/ml 8-10 hoursc
118 TABLE 17. Experiments for the developm ent of criteria to measure Good quality water. a Limit of Detection b N/A= samples were not assayed on the biosensor. c Negative control consisted of uninoculated mE broth which was allowed to react with the polystyrene waveguides at 36C fo r the same amount of time as the samples. d False positives were seen for negative controls. Positives had to be higher than the LOD and higher than the negative control. Time points were judged ba sed on viable counts and if positive detection was observed when counts reached the biosensors limit of detection (106 CFU/ml). Those points are boldfaced in the Table 17. To be consider ed a positive result, not only did the signal have to be higher than the LOD, but it also had to be higher than the false positive observed for the negative control. The mean LOD was 20.8 + 15.2. The mean recovery was 72%. Two times out of thr ee, it required over 14 hours to enrich a concentration this low (~0.01 CFU/ml) to obtain the minimum CFU/ml (106 CFU/ml) required for positive Experiment Time point Viable count LODa Biosensor signal (hrs) (CFU/ml) (pAmps) (pAmps) 1 8 2.3 x 103 N/Ab N/Ab ( 5 CFU/ 10 7.8 x 103 12.6 8.0 100 ml) 12 2.4 x 104 N/Ab N/Ab 14 2.0 x 105 16.4 68.3 72 1.6 x 108 21.4 3579.9 Neg. contro lc 53.1 179.3d 2 10 2.8 x 103 4.2 24.7 (3 CFU/ 12 5.0 x 103 N/Ab N/Ab 100 ml) 14 6.1 x 104 18.7 23.2 24 1.7 x 107 13.6 573.3 Neg. Controlc 5.5 48.7 d 3 10 7.2 x 104 12.4 35.8 (3 CFU/ 12 1.6 x 106 N/Ab N/Ab 100 ml) 14 9.6 x 106 14.3 379.4 24 2.6 x 109 32.8 5512.8 Neg. controlc 44.9 139.1 d
119 detection on the Analyte 2000 biosensor. The growth rate for Ent. faecalis was calculated to observe any discrepancies in growth due to differences in initial inoculum between the three categories. The mean growth rate for these experiments was calculated to be 43.6 + 2.6 minutes. All of these data were used to establish criteria for the Good category of water quality de veloped in this method. Experiments for determining criteria for Moderate quality water samples. Sterile water was spiked with a concentration of Ent. faecalis cells that would give an approximate count equa l to what Pinellas County Health Department considers moderate quality water (35-104 CFU/100 ml of enteroco cci). The process was performed as described in Materials and Methods. Three experiments were performed and the results of those experiments are listed in Table 18.
120 TABLE 18. Experiments for the development of criteria to measure Moderate quality water. a Limit of Detection b Negative control consisted of uninoculated mE broth which was allowed to react with the polystyrene waveguides at 36C for the same amount of time as the samples. c False positives were seen for negative controls. Positives had to be higher than the LOD and higher than the negative control. d N/A= samples were not assayed on the biosensor. e Results fit into Good category, but this ex periment was done as part of moderate quality experiments. When an initial inoculum of ~0.1 CFU /ml wa s tested, two out of three experiments showed that 12-14 hours of enrichment was required to obtain the minimum concentration of Ent. faecalis (106 CFU/ml) to get positive detection on the Analyte 2000 biosensor. Those time points are boldfaced in Table 18. Again, to be considered a positive result, the signal had to be higher than the LOD and higher than the false positive observed for the negative control. The mean LOD was 30.9 + 16.6. The mean recovery was over 100%. The mean growth rate of Ent. faecalis in this set of three experiments Experiment Time point Viable count LODa Biosensor signal (hrs) (CFU/ml) (pAmps) (pAmps) 1 10 5.1 x 105 9.8 33.9 ( 71 CFU/ 12 2.9 x 106 42.5 308.6 100 ml) 14 1.6 x 107 24.9 781.9 24 2.5 x 109 31.2 5280.2 Neg. controlb 50.2 135.9c 2 10 1.1 x 105 8.9 24.9 (48 CFU/ 12 4.9 x 105 N/Ad N/Ad 100 ml) 14 4.4 x 106 7.0 167.5 36 2.0 x 107 34.6 5595.2 Neg. controlb 47.3 136.2c 3 10 2.0 x 103 13.0 11.5 (32 CFU/ 12 6.4 x 104 N/Ad N/Ad 100 ml)e 14 1.1 x 105 33.9 32.1 36 1.7 x 108 52.2 3950.9 Neg. controlb 46.1 158.7 c
121 was calculated as 46.9 + 3.2 minutes. This data was used to establish criteria for samples of moderate quality using concentrati on, enrichment and biosensor detection. Experiments for determ ining criteria for Poor quality water samples. Sterile water was spiked with a concentration of Ent. faecalis cells that would give an approximate count equal to what Pinellas County Health Department consider s poor quality water (> 105 CFU/100 ml of enterococci). The process was performed as desc ribed in Materials and Methods. Three experiments were perfor med and the results of those experiments are listed in Table 19.
122 TABLE 19. Experiments for the developm ent of criteria to measure Poor quality water. a Limit of Detection b N/A= samples were not assayed on the biosensor. c Negative control consisted of uninoculated mE broth which was allo wed to react with the polystyrene waveguides at 36C for the same amount of time as the samples. d False positives were seen for negative controls. Positives had to be higher than the LOD and higher than the negative control. When an initial inoculum of ~1 CFU/ml was tested, three out of three experiments showed that 8-10 hours of enrichment was required to obtain the minimum concentration of Ent. faecalis (106 CFU/ml) to get positive detection on the Analyte 2000 biosensor. Those time points are boldfaced in Table 19. To be considered a positive result, the signal had to be higher than the LOD and higher than the false positive observed for the Experiment Time point Viable count LODa Biosensor signal (hrs) (CFU/ml) (pAmps) (pAmps) 1 5 1.4 x 104 N/Ab N/Ab ( 180 CFU/ 6 4.5 x 104 N/Ab N/Ab 100 ml) 7 1.1 x 105 N/Ab N/Ab 8 3.7 x 105 N/Ab N/Ab 9 6.1 x 105 N/Ab N/Ab 10 1.7 x 106 14.1 268.1 24 3.9 x 108 39.3 4185.5 Neg. control c 36.3 107.4 d 2 8 2.2 x 106 41.6 318.8 (600 CFU/ 10 2.6 x 107 40.8 2993.8 100 ml) 24 1.1 x 109 46.3 4923.3 Neg. control c 20.3 104.6 d 3 8 1.0 x 106 13.1 66.7 (330 CFU/ 10 5.4 x 106 11.2 191.6 100 ml) 36 3.0 x 107 74.6 4698 Neg. control c 50.1 97.9 d
123 negative control. The mean recovery was 148% and the mean LOD was 35.3 + 19.3. The mean growth rate of Ent. faecalis in this set of three experiments was calculated as 43.3 + 8.4 minutes. These data were used to es tablish criteria for sa mples of poor quality using concentration, enrichment and biosen sor detection. In all the experiments performed, the growth rates were similar. Of note is the false positives seen with the negative control waveguides. The backgr ounds were taken using a primary antibody other than the anti-Strep group D polyclonal antibody. A Difco antiV. cholerae O1 polyclonal antibody was used to ascertain all background noise wh ich the Cy5 labeled anti-IgG would give due to non-specific binding. While the antiV. cholerae antibody did not seem to show affinity to any of the media components or non-specifically adhere to the polystyrene waveguide, the anti-Strep group D antibody did show some nonspecific activity leading to higher background signals and false positive when allowed to react with the secondary antibody. While this water quality guide is only semi-quantitative, it definitively distinguishes between Good and Poor quality of water very well. A draw back to the method is that the geometric mean, which is routinely calculat ed by the Pinellas County Department of Health, cannot be calculated ba sed solely on this method becau se it is semi-quantitative. Initial counts for all these ex periments were obtained from viable counts on TSA before spiking 20 L of sterile water in a carboy, but in blind samples, the water would just be processed to observe in what category the wa ter would fall into (23). This category as explained previously would be an estimate of the counts that are truly present in the water. Criteria for closing a beach not only rely on the three categories mentioned
124 previously, but also on the calculated geomet ric mean which takes into account specific counts from the previous five sampling week s. If the geometric mean exceeds 35 CFU/100 ml, than the beach is posted and warn ings are sent out to the news media and posted on the county website and at the be ach. This method would be a very good preliminary screening tool and re sults would be obtained faster than the standard methods which can require biochemical confirmation to assert the presence of fecal enterococci. Unspiked 20 liter volumes of seawater. Samples of unspiked seawater were assayed as described in Materials and Methods. Table 20 shows the results of three environmental seawater samples which were processed using the integrated biosensor/concentrati on/enrichment method.
125 TABLE 20. Results from envi ronmental samples tested using integrated biosensor method. a Limit of Detection b N/A= samples were not assayed on the biosensor. c Negative control consisted of uninoculated mE broth which was allowed to react with the polystyrene waveguid es at 36C for the same amount of time as the samples. d False positives were seen for negative controls. Positives had to be higher than the LOD and higher than the negative control. e Water sample falls into good category acco rding to counts and biosensor detection. Also matches good category in Pinellas county guidelines. f Water sample falls into good category acco rding to Pinellas county guidelines. According to counts and biosensor data, it falls into the m oderate water quality category. g Water sample falls in mode rate category for both Pine llas county guidelines and biosensor detection. h Counts obtained from the Department of Health i Filtration counts on mEI performed in the lab. Experiment Time point Viable count LODa Biosensor signal (hrs) (CFU/ml) (pAmps) (pAmps) 1e 8 N/Ab N/Ab N/Ab ( 6 CFU/ 10 6.0 x 103 N/Ab N/Ab 100 ml)h 12 3.0 x 104 35.0 -2.2 (9 CFU/ 14 4.8 x 105 35.4 -3.3 100 ml)i 16 2.3 x 106 N/A N/A 24 6.1 x 108 58.0 777.5 2f 8 N/Ab N/Ab N/Ab (4 CFU/ 10 2.4 x 105 N/Ab N/Ab 100 ml)h 12 1.2 x 106 75.0 37.7 (69 CFU/ 14 1.0 x 107 218.9 403.3 100 ml)i 16 2.6 x 108 -39.1 4028 Neg. controlc 13.7 104 d 3g 10 2.6 x 105 -26.5 -28.7 (60 CFU/ 12 6.4 x 106 422.9 282.2 100 ml)h 14 7.0 x 106 79.0 138.5 (43 CFU/ 16 3.9 x 197 N/Ab N/Ab 100 ml)i 18 1.4 x 108 N/Ab N/Ab Neg. control c 15.5 29.7 d
126 Two out of three times, the method correctly co rrelated the growth of fecal enterococci in mE enrichment broth to the actual density of the organisms in beach water. The mean recovery was 71.5% and the mean LOD was 80.7 + 132.5. The mean growth rate was 46.3 + 1.1. The increased growth rate is ex pected as the cells from environmental samples are stressed and then undergo furthe r stress through filtration with the hollow fiber filter. In all unspiked samples, representative colonies with blue halos on mEI were confirmed by inocula ting into brain heart infusion broth with 6.5% NaCl and streaking onto bile esculin azide agar. These isolates were identified using API Strep and a biochemical key that was put together by Manero (171). The predominating species identified were Ent. faecium and Ent. faecalis. The yellow bacterium Ent. casseliflavus was found only occasionally. Two red coloni es which resulted on mEI in one of the experiments were not able to be identified through API Strep and these results remained inconclusive. In one of the experiments, one red colony was found to grow which had no blue halo on mEI. This isolate was identified as Aerococcus viridans In experiment 3, Vibrio alginolyticus thrived in the mE enrichment brot h along with the enterococci. This was only found to happen in this experiment. V. alginolyticus was assayed on the biosensor and with ELISA to check for crossreactivity with the anti-Strep group D antibody. This organism was not found to crossreact significantly. The fact that this organism grew in the mE broth which contai ned nalidixic acid was surprising considering that this organism is very susceptible to this antibiotic. An ex periment performed on isolates of V. alginolyticus isolated from marine water from Indonesia showed that 1.3 % of the strains recovered were re sistant to nalidixic acid (183). The high pH (9.3) of the medium did not present an obstacle as vibrios are known to thrive in alkaline
127 environments. The recovery from the filtration of the water sample was erratic between experiments including those used to create th is water quality guideline. While unusual, preliminary experiments where low concentrations of enterococci were concentrated also showed these fluctuations. At higher concentrations, the loss is less evident and may account for the greater and more consistent re covery. This unusual recovery was seen in all concentrations with a low st arting bacterial concentration. Isolation of E. coli O157:H7 in seawater Seawater samples were analyzed and enriched for E. coli O157:H7 as described in Materials and Methods. Sample s were analyzed from two beaches and no isolates were found to be E. coli O157:H7. Any suspect isolates of E. coli O157:H7 which were sucrose negative on CT-SMAC were identified through API. These isolates were found to be either nonO157:H7 E. coli or Klebsiella spp. Biosensor Immunoassays for the Detecti on of E. coli O157:H7 in Seawater. Previous work by DeMarco and Lim showed that the KPL affinity purified polyclonal antibody against E. coli O157:H7 was only mildly crossre active with one other organism, Escherichia hermanii (63, 64). Assays were performed as described in Materials and Methods using the sandwich immunoassay. Th e data in Figures 39 and 40 are results from seven different experiments. The mean LOD was 69.5. Detection in seawater was comparable to detection in 0.01M PBS (pH 7.4). Sensitivity in PB S and seawater was 3.6 x 105 CFU/100 ml.
128 -200 0 200 400 600 800 1000 1200 1400 1600 Neg control 10^610^510^410^310^2 E. coli O157:H7 concentration (CFU/ml) E.coli O157:H7 in PBS E. coli O157:H7 in seawater Mean signals over backgroun d (pAmps) FIGURE 39. Detection of E. coli O157:H7 in PBS vs. seawater. A sandwich immunoassay format was used to detect E. coli O157:H7 in seawater. Biotinylated KPL polyclonal antibody (100 g/ml) was used to capture. Cy5 labeled KPL (10 g/ml). -20 0 20 40 60 80 100 120 140 10^610^510^410^310^2 E. coli O157:H7 concentrartion (CFU/ml) E.coli O157:H7 in PBS E.coli O157:H7 in seawater Normalized signal (%) FIGURE 40. Normalized values for the mean signals shown in Figure 39.
129 Isolation of V.cholerae from Seawater A total of 128 isolates were recovered fro m seawater by isolation on TCBS agar as described in Materials and Methods. Any organisms that were yellow and 2-3 mm in diameter on TCBS agar were re streaked onto TSA and further analyzed. After gram stain and oxidase test, the organisms were examin ed using the two test method developed by Chopun et al. (46). The organisms were in oculated into iron esculin medium and arginine dihydrolase broth. One drawback to the two test method was the difficulty in interpreting the iron esculin hydrolysis test as described by Chopun et al. The black precipitate which formed from esculin hydrol ysis could also be the result of melanin production. This was sidetste pped by checking for fluorescence of esculin under UV light which resulted from hydrolysis. While this made interpretation easier, it was still difficult to interpret tubes which were completely black. Negative results in both of these test lead to further identification through the API test. Table 21 compares the three methods used to identify V. cholerae in this study. TABLE 21. Comparison of three methods for the identification of V. cholerae isolates. a Using the 2-test method, 121 out of 127 is olates obtained from e nvironmental samples were identified as V. cholerae. b 87 out of 127 isolates obtained from e nvironmental samples were identified as V. cholerae through API. c 67 out of 127 isolates obtained from e nvironmental samples were identified as V. cholera. Two-test method a API ID b PCR c ______________________________________________________________________ Number of Number other/ Number Number other/ Number of V. cholerae / total number V. cholerae / total number V. cholerae / total number of isolates number of of isolates number of of isolates total isolates total isolates 121/127 6/127 87/127 40/127 67/127
130 Some of the isolates identified wh ich were sucrose negative included V. vulnificus and V. parahaemolytcus These isolates were included in the final count of isolates and were API identified, but were not assayed through PCR or the two-test method. All sucrose positive isolates were identified through AP I even if the two-test method results suggested a microorganism other than V. cholerae Many of the isolates which may have been considered V. cholerae from using just the two-method test were, in fact, identified as Vibrio alginolyticus when assayed using an API test strip. V. alginolyticus is sucrose positive and also negative for arginine dihydrolase and iron esculin hydrolysis. Another sucrose positive organism isolated was identified as Aeromonas hydrophila through API. This isolate, however, gave the expected re sults for arginine dihydrolase (+) and iron esculin hydrolysis (+) in the two test method. In some instances the API results suggested A. hydrophila Group II and V. alginolyticus as alternatives if the isolate assayed was not V. cholerae In these cases, the isolate was identified through PCR along with other isolates which were identified as V. cholerae at greater confidence levels (95% and greater) through API. When assayed through PCR by amplifying the ITS region of the 16S-23S rRNA sequence of V. cholerae a total of 67 isolates from the 127 isolated from the two beaches tested resulted as being V. cholerae (Table 21). This result was confirmed by the presence of the 300 bp fr agment upon gel electrophoresis. Figure 41 shows the typical results of five isolates a ssayed by PCR and gel electrophoresis. Four out of the five isolates assayed were confirmed as V. cholerae All other isolates were assayed the same way. All confirmed V. cholerae isolates were non-O1 strains of V.
131 cholerae as confirmed through slide agglutination with Difco polyclonal antiV. cholerae O1 polyclonal serum. 1 2 3 4 5 6 7 8 FIGURE 41. Summary of PCR results for the identification of V. cholerae environmental isolates obtained from two Tampa Bay beaches. Specificity Studies for V.cholerae Detection Five different antibodies were tested for purposes of biosensor assay development. One monoclonal and four polyclonal were tested against V. cholerae O1 whole cells (Figure 42). The Austral Biologicals monoc lonal and the Difco polyclonal outperformed all the other antibodies tested. The Difco pol yclonal was chosen for use on the biosensor Lane 11 Kb Marker (Bench Top PCR marker) Lane 2 Negative control, sterile H2O Lane 3 Positive V. cholerae O1 (CDC E5906-1018) Lane 4 Isolate USF 877 from N. Shore beach Lane 5 Isolate USF 878 from N. Shore beach Lane 6 Isolate USF 879 from N. Shore beach Lane 7 Isolate USF 880 from N. Shore beach Lane 8 Isolate USF 883 from Courtney Campbell beach 1 Kb 300 b p 300 b p 500 b p 750 b p 150 b p 50 b p
132 due to its higher affinity for lower concentrations of V. cholerae O1 cells and the lower cost. 0 1 2 3 4 5 6 7 8 10^810^7'10^610^5 V. cholerae O1 concentration (CFU/ml) Remel polyclonal antibody Difco polyclonal antibody Austral monoclonal antibody Denka Seiken polyclonal antibody New Horizons polyclonal antibody Signal to noise ratio FIGURE 42. Comparison of commercial polyclonal antibodies for V. cholerae detection on ELISA. All antibodies were assa yed as whole serum diluted 1:500 in PBS. A signal to noise ratio above 2.0 was necessa ry for a value to be considered positive. Several organisms were assayed for specifi city against the Difco polyclonal antibody against V. cholerae O1 using ELISA (Table 22). TABLE 22. Organisms tested for crossreactivity with Difco polyclonal V. cholerae antibody through ELISA. a A signal to noise ratio above 2.0 was require d for the ELISA to be considered positive. Microorganism Signal to noise ratioa Concentration needed for (+) signal (CFU/ml) Aeromonas hydrophila (USF 530) 2.0 2.2 x 107 E.coli ATCC 29417 2.8 2.6 x 108 Salmonella typhimurium (USF 515) 4.8 4.3 x 108 V. cholerae O1 (CDC E5906-1018) 2.0 2.2 x 105 Pseudomonas aeruginosa (USF 620) N/A N/A Vibrio alginolyticus (USF 516) 3.5 8.4 x 106 V. cholerae O139 (CDC 2412-93) 3.3 3.7 x 109
133 Several of the organisms tested did crossreact with the Difco polyclonal. The majority of the organisms only crossreacted at high concentrations. Vibrio alginolyticus crossreacted to a greater degree. This crossreactivity c ould be problematic cons idering how prevalent V. alginolyticus is in marine waters. Different Sample Preparations for V. cholerae Immunoassay Optimization Different sample preparations were also tested for assay optimization. Cells of V.cholerae were tested untreated, boiled for 20 mi nutes, or sonicated for 10-20 minutes. Boiled cells suspensions were compared to untreated cell suspensions using ELISA (Figure 43). 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 10^910^810^710^610^5 V. cholerae O1 concentration (CFU/ml) Untreated V. cholerae O1 cells Boiled V. cholerae O1 cells Signal to noise ratio FIGURE 43. Detection of viable V. cholerae cells and boiled V. cholerae cells. Cells were boiled for 20 minutes and allowed to c ool before assaying through ELISA. Difco polyclonal antiV. cholerae O1 antibody (10 g/ml) was used to test the different cell preparations. Boiled cells showed much stronger fluorescence readings than untreated cells. Sonication was compared to boiling on the biosensor to see if sensitivity could further be enhanced (Figure 44 and 45). These data are results from 3 experiments. Boiling still appeared to be the best sa mple preparation method for V. cholerae O1 assay
134 development. The mean LOD was 30.8. The a ssay sensitivity for this assay was 1.6 x 109 CFU/ml. 0 50 100 150 200 250 300 350 10^710^810^9 V.cholerae O1 concentration (CFU/ml) Sonicated V. cholerae O1 Boiled V. cholera e O1 Mean Signal (pAmps) FIGURE 44. Detection of boiled and sonicated V. cholerae cells. A sandwich immunoassay format was used to detect V. cholerae O1 cells on the biosensor. Mean fluorescence over the limit of detection for each concentration tested is represented. Biotinylated Difco polyclonal antibody (100 g/ml) was used to capture cells. Cy5 labeled Difco polyclonal antibody (10 g/ml) was used to detect the captured cells. Mean LOD was 30.8. 0 20 40 60 80 100 120 10^710^810^9 V. cholerae O1 concentration (CFU/ml) Sonicated V. cholerae O1 Boiled V. cholerae O1 Normalized Signal (%) FIGURE 45. Normalized values for the mean signals shown in figure 44.
135 Isolation of S. aureus from Seawater Samples of seawater from North Shor e beach and Courtney Campbell Causeway beach were collected, filtered and enriched for the presence of S. aureus as described in Materials and Methods. Brie fly, 100 ml samples (diluted a nd undiluted) were filtered through a 0.45 m filter and the filter was pl aced onto Baird Parker agar. Any black colonies were gram stained and then further tests were performed to identify any potential S. aureus isolates. Only 12 isolates were identified as S. aureus and stored in glycerol and TSB at -80 C. Specificity Assays for S. aureus Detection Two antibodies were tested for purposes of assay development. A polyclonal rabbitantiS. aureus antibody from Maine Biotechnology Services Inc and a monoclonal antibody against peptidoglycan from Chemicon In ternational, Temecula CA were tested against viable and dead cells of S. aureus (Figure 46). These tests were performed by Dr. Marianne Kramer, University of South Flor ida, Tampa, FL. The monoclonal antibody did not show good affinity for S. aureus Based on these results, further specificity studies were performed using the Main e Biotechnology polyclonal antibody.
136 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 10^810^710^610^510^410^310^2 S. aureus concentration (CFU/ml) Maine polyclonal antibody Live S.aureus cells Maine polyclonal anitbodyBoiled S. aureus cells Chemicon antipeptidoglycan mab-Live S.aureus cells Chemicon antipeptidoglycan mabBoiled S.aureus cells Si g nal to noise FIGURE 46. Screening of S. aureus commercial antibodies. Cells were boiled for 20 minutes and allowed to cool be fore assaying with ELISA. All antibodies were assayed at 10 g/ml. The Maine Biotechnology polyclonal antibody wa s tested for crossreactivity with other species of Staphylocococcus (Figure 47) using ELISA. A signal to noise ratio over 2 was considered a positive result. All the organi sms tested were crossreactive with this polyclonal antibody; however, S. aureus gave much stronger signals. This antibody was used to develop the biosensor assay for S. aureus
137 0 1 2 3 4 5 6 7 10^810^710^610^510^410^3 Cell concentration (CFU/ml) S. aureus S. simulans S. saprophyticus S. epidermidis Signal to noise rati o FIGURE 47. Crossreactivity studies with other Staphylococcus species. The Maine Biotechnology polyclonal aantibody was tested at 10 g/ml on ELISA. A signal to noise ratio over 2.0 was required for a value to be considered positive. Biosensor Sandwich Immunoassays for S. aureus. Waveguides were prepared as describe d in Materials and Methods. The Maine Biotechnology polyclonal antibody used in the biosensor assay for S. aureus failed to detect S. aureus to any degree. Figure 48 and 49 is data from three different experiments. The mean signals did not increase in proporti on to the increase in cell concentrations assayed and no detection of S. aureus was observed at any concentration (Figure 48). The mean LOD was 68.2. Normalized signals failed to show any distinctive pattern of detection either (Figure 49). This assay was not pursued any further due to the lack of a high affinity antibody.
138 -4 -2 0 2 4 6 8 10 12 10^910^610^5 S. aureus concentration (CFU/ml) Maine Biotechnology polyclonal antibody Mean signal over background (pAmps) FIGURE 48. Detection of S. aureus using biosensor sandwich immunoassay. Mean fluorescence (pAmps) over the limit of detec tion for each concentration tested are represented. -60 -40 -20 0 20 40 60 80 100 120 10^510^610^9S. aureus concentration (CFU/ml) Maine Biotechnology polyclonal antibody Mean normalized signal (%) FIGURE 49. Normalized values for the mean signals shown in figure 48.
139 DISCUSSION The U.S. Environmental Protection Agency ha s set forth strict guidelines with respect to recreational water quality and the degree of contamination which is permissible. Fecal indicator bacteria provide an estimation of the amount of feces in the water, and indirectly, the presence and quantity of fecal pathogens in the water. Enterococci have been recommended as one of the most accura te indicators of seawater quality. In accordance with EPA regulations, a single sample containing > 105 CFU/100ml (~1 CFU/ml) of enterococci is indicative of a high degree of fecal contamination and, therefore, could pose a dangerous threat to th e health of the swimming public (261). This danger is not due to the presence of enteroco cci, which can be pathogenic themselves, but more so because the presence of this particular bacterium indicates the possibility that other, more pathogenic bacteria and viruses of fecal origin could be present in the water as well. Indicator organisms, particularly E. coli have been assayed for over a century to assess water quality because these organisms are consistently found in the feces of warmblooded animals. In the infancy of the sc ience of waterborne diseases, a test for specific pathogens such as V. cholerae did not exist. Since Escherichs discovery of the abundance of E. coli in fecal matter, contaminated water has been analyzed by assessing concentrations of fecal indicator bacteria (79) Presently, assays for indicator organisms are routinely performed and preferred to a ssays for pathogenic microorganisms because
140 they are easy, inexpensive, do not requi re skilled personnel and depending on the indicator, and indicator bacteria densities have been found to correlate with the incidence of bather illness. There are several emerging and reemerging waterborne pathogens that are becoming of increasing public health concer n. Many of those bacteria, viruses, and protozoans were listed in Ta ble 4. While assays for indicator organisms are convenient and easy, they are a preventative measure and tests for pathogenic microorganisms are also necessary, especially when a waterborne disease outbreak occurs. Such was the case in New Orleans during recent Hurricane Katrina where cases of Vibrio vulnificus infections were reported to the CDC (34). Conventional methods for the detection a nd enumeration of enterococci and other pathogenic bacteria have largely been culture based. Methods for detecting enterococci include membrane filtration and enrichment on mEI agar and the Enterolert most probable number system. As explained previously, these methods are simple, user friendly, do not require highly skilled pers onnel and give results within 24-48 hours. However, colonies obtained from the membrane filtration technique are routinely confirmed through biochemical tests which add an extra day or longer for final results. The Enterolert system is unable to detect very low levels of enterococci since all samples need to be diluted 10-fold before testing (71) Special equipment is also needed to run this test. In spite of the advantages and di sadvantages of culture methods, bacteria have evolved in such a way that culture based met hods may not accurately depict the extent of contamination by these bacteria in the environment. Research is now revealing that any method developed for the detection of bacteria in the environment needs to be designed to take into account conditions in the envi ronment and how these conditions affect the
141 organism. The viable but non-culturable (VBNC) phenomenon, for example, has been observed in many bacteria including the enterococci, E. coli O157:H7, and Vibrio cholerae (37, 106, 233). The VBNC phenomenon results from a drastic change in environmental conditions such as changes in temperature or nutrients which stresses the bacteria and renders them unculturable. Conventional culture t echniques cannot account for these VBNC bacteria. Furthermore, these bacteria while unculturable still can be capable of causing disease if they find the right host. Due to the limitations of culture methods molecular and immunological methods are being increasingly researched. Immunomagne tic separation (IMS) has been used to rapidly detect low levels of E. coli O157:H7 and Salmonella typhimurium in food, water and feces (15, 40, 257, 282). This method is s imple and requires little expertise. Rapid immunodiagnostic test kits using monoclonal antibodies to V. cholerae O1 such as the Cholera-Screen coagglutination test ( 52) and Cholera SMART from New Horizons Diagnostics, a colloidal gold based colorimetr ic test (109), have proved highly specific and sensitive when compared to culture methods. A direct fluorescent antibody technique using monoclonal anti bodies labeled with fluorescein isothiocyanate is able to detect 1.5 x 102 CFU/ml of V. cholerae O139 in water when combined with filtration (108). Molecular methods such as P CR have been used to detect V. cholerae and Enterococcus faecalis in seawater and simulated seaw ater (111, 169, 220). PCR has been used to detect V. cholerae cells and toxin genes in raw oy sters and in synthetic seawater with much success (6 to 8 CFU per gram in oysters and 10 CFU per gram in seawater) (169, 211). Some studies have even shown th e ability to detect 10 CFU per gram of
142 oyster homogenate (139). PCR has obvious a dvantages such as sensitivity, specificity and rapidity of results however, there are limitations to the technique. The presence of inhibitors in the environment, which can interf ere with the assay, is a large problem when rapid results from a dirty sample are desire d. Additional steps to clean up and purify a sample add extra time to the assay. More importantly, the fact that PCR cannot distinguish between live and dead cells is one of the biggest disadvantages of this technique. Nonpathogenic dead ce lls would be detected along with viable cells and could lead to an inaccurate assessment of a partic ular sample. If those bacteria are nonviable cells from a previous contamination event, results could lead to unsubstantiated beach closings. Other nucleic acid techniques are fi nding a place in the field of rapid detection. Fluorescent in situ hybridization (FISH) is a method used to identify bacteria with a high degree of specificity. FISH uses fluorophor e-labeled nucleic acid probes complementary to a specific rRNA sequence in a particular bacterium to visibly identify the organism. This method has been used in combination w ith membrane filtration in a technique called microcolony filter hybridization to more rapidly identify indicator bacteria such as E. coli S. aureus and enterococci (156, 177, 178, 198). In this method, a volume of water is filtered through a 0.45 m filter and pre-enriched on selective media. Cells are then fixed with formaldehyde and nucleic acid probes are introduced to hybridize with a target sequence. The membranes are then visua lized using a fluorescent microscope. An advantage to using probes targeted to rRNA sequences is the flexibility of probe specificity. By targeting regions of greater or lesser conservation, probes can be designed to be specific at a kingdom, genus, or species specific level (156). While novel, this technique differs in efficiency between gram positive and gram negative cells. Due to the
143 thicker cell wall of gram positive bacteria, added steps are required to allow the probe to enter the cell in these organisms. This adds time to the overall procedure. The sensitivity of this assay is high and 102-105 cells/membrane are usually ne cessary (245). FISH using peptide nucleic acid probes has been used to detect 104 CFU of E. coli with high specificity (245). This technique also offers the added benefit that if multiple probes are used with different conjugated fluorophores, multiplexing can be achieved (198). While molecular and immunological met hods have proven to be co mparable in sensitivity and specificity to culture methods, they often require preenrichment of up to 24 hours for positive detection. The method develope d in this study was no exception. Very low levels of enterococci in water can be indicative of larger amounts of fecal contamination in beachwater. With the many epidemiological studies and reviews which have been published correlating the incidence of disease in bathers to the levels of enterococci in seawater, the development of rapid and specific me thods to detect low levels of enterococci in environmental samples has become more important and necessary (25, 72, 95, 259, 265). Few if any immunoassays have been described for the detection of enterococci in seawater. The primary goal of this study was to detect the low levels of enterococci required by EPA gu idelines using an immunoassay format and an evanescent wave fiber optice bi osensor. This could not be achieved using the biosensor alone and, therefore, sample concentr ation and enrichment were required. An evanescent wave fiber optic biosenso r, the Analyte 2000, was used to detect enterococci in unspiked samples of seawater. The limit of detection of the sandwich assay was 1.5 x 107 CFU/ml when Enterococcus cells were sonicated for a minimum of 10 minutes. An indirect assay format was developed where by Enterococcus cells were
144 adsorbed directly onto the polystyrene wavegu ide and detected using a specific antibody and a Cy5 labeled anti-IgG. Using this method, 1 x 106 CFU/ ml of enterococci could be consistently detected in PBS (pH 7.4) and in environmental seawater samples. Neither method alone was successful in detecting the low levels of enterococci required by the U.S. Environmental Protection Agency. Samp le concentration methods and enrichment were used to improve the assay sensitivity. A guideline was developed to assess environmental samp les for water quality based on enrichment times in mE broth and biosen sor detection. In accordance with this guideline, environmental water samples were assayed through concentration, enrichment and biosensor detection. Envi ronmental water samples were collected in 20 liter volumes from two Gulf Coast beaches known to consiste ntly yield samples of poor quality. When concentrated with a hollow fiber filter and subsequent enriched in mE broth, water samples containing as few as 6 CFU/100ml of enterococci could be detected with the Analyte 2000 biosensor in 16-24 hours and determin ed to be water of good quality. This result was found to correlate with the criteri a set forth in the guideline developed in simulated water quality experiments. The biosensor assay alone, excluding enrichment and sample preparation, took only 10 minutes total. The anti-Strep group D antibody (American Research Products) was found to crossreact significantly with only S. aureus cells. The nature of this crossreactivity wa s attributed mainly to nonspecific binding of the primary antibody used to the protein A molecules found on the surface of S. aureus This crossreativity has been noted in other immunoassays developed to specifically detect S. aureus (100). Crossreactivity could be eliminated to an extent by blocking the protein A binding sites on S. aureus cells with another species of Ig G. In this study, horse serum
145 was used as a blocking agent. Crossreactivity could not be completely eliminated using this method. Studies were also conducted to test the effi ciency of the hollow fiber filtration step of the assay. This process consistently conc entrated water 100-fold and recovery was found to vary between 13% and greater than 100% depending on the concentration of the initial inoculum. Lower recoveries were observed fo r lower concentrations of cells where the loss was more evident due precisely to the fact that lower cells are present and loss is more evident in this realm. Only one non-target bacterium, Vibrio alginolyticus was found to enrich along with the enterococci in mE enrichment broth. This organism was assayed on ELISA with the anti-Strep group D antibody to check for cross-reactivity and no significant cross reaction was found. While this integrated method was successful at detecting low levels of enterococci directly from seawater, this method could not be used alone without confirmatory biochemical tests. This is because of the cross-reactivity that was found to exist between the anti-Strep group D antibody and other organisms such as S. aureus Five percent of the colonies obtained from viable counts on mE I agar were inoculated on bile esculin agar and into Brain Heart Infusion broth with 6.5 % Nacl to confirm that they were enterococci. This type of confirmation must be done in all immunoassays not just the one developed in this study because of antibody crossreactivity. A secondary goal of this project was to analyze water samples collected from two of the most highly contaminated beache s around Tampa Bay for the presence of Vibrio cholerae E. coli O157:H7, and Staphylococcus aureus Furthermore, biosensor assays were developed to detect these pathogens wh ich could be potentially shed in the water
146 simultaneously with enterococci. Standard a ssay methods for total coliforms in seawater could not be used to isolate E. coli O157:H7 from seawater as the common temperature to isolate these organisms is 44.5C. This te mperature was too high for the growth of E. coli O157:H7. Maximum growth temperature for E. coli O157:H7 is 41.542.5 C (96, 208). Ultimately, methods were modified to optimize chances of isolating E. coli O157:H7. Immunomagnetic separation has been used to recover E. coli O157:H7 with great success from food, fecal samples, and wa ter, so this method was employed to try and recover the organism from seawater (15, 40, 257, 282). No isolates of E. coli O157:H7 were found in the waters from Cour tney Campbell Causeway beach and North Shore Beach. The Analyte 2000, evanescent wave fiber op tic biosensor was used to detect the pathogenic E. coli O157:H7, Vibrio cholerae O1 and S. aureus in PBS and seawater. A sandwich immunoassay format was used to detect E. coli O157:H7 in seawater and was able to detect the organism at 105 CFU/ml. The zero tolerance policy of the FDA for E. coli O157:H7 in food matrices necessitates the improvement of assay sensitivity for rapid detection in food. Simila r low detection limits are requi red for water contaminated with E. coli O157:H7 because the infectious dose of E. coli O157:H7 is thought to be very low (50 CFU>) (254). Isolates of V. cholerae and S. aureus were obtained from environmental water samples taken from Courtney Campbell Causeway beach in Clearwater, FL and North Port beach in St. Petersburg FL. Sixty seven isolates were confirmed as V. cholerae through PCR. None of the isolates were V. cholerae O1 strains as determined by slide aggluntination with V. cholerae O1 antiserum. Twelve isolates of
147 S. aureus were also recovered from the two Gulf beaches using culture methods. A suggested direction for this part of the pr oject would be to further characterize the presence or absence of toxin genes in the 67 Vibrio cholerae isolates and to test the S. aureus isolates for methicillin resistance. The presence of these pathogens in marine waters further substantiated the developmen t of rapid assays for the detection and monitoring of these organisms in seawater. The sensitivity of the biosensor sandwich immunoassay for V. cholerae O1 was 1.6 x 109 CFU/ml. The evanescent wave fiber optic biosensor was evaluated for detection of S. aureus in seawater as well. This attempt was not successful in PBS and was not pursued further because a high quality antibody was not available. Many limitations to the developed biosensor assays were observed in this study. For instance, inherent variability exists between polystyrene waveguides made by the manufacturer for the Analyte 2000 biosensor. This however, was not the root of the sensitivity issues for the biosensor assay for enterococci and the other organisms tested. One factor was the evanescent wave biosensor itself. Any assay based on evanescent wave principles is limited by the distance of that evanescent wave. The evanescent wave of the Analyte 2000 biosensor extended ~100 nm from the surface of the waveguide so anything outside of that evanescent wave was not detected. Bacteria are large in size (enterococci are approximately 1.5 m in length) and any antibodies attached to the cells that are outside the evanescent wave could not be detected (216). In an assay format where capture and detection antibodies al ready compete for epitopes on the target antigen, the potential area for antibody binding is reduced greatly by the limitations of the evanescent wave. This handicap was overcome to some extent by using the indirect assay
148 format with signal amplification to detect enterococci. Signal amplification could bypass the effects of the evanescent wave phe nomenon by increasing the signal of those antibodies which are bound to the target with in the evanescent wave. This was done in this study by using a fluorophore labeled anti-IgG. The biggest limiting factor in the Enterococcus Vibrio cholerae O1, and S. aureus assays was the lack of high quality antibodies As with all immunoassays, the test is only as good as the antibody and antigen. If either one is less than optimal, than the success of the assay is compromised. The antibody used in assays for V. cholerae O1 was found to be very crossreactive with seve ral other organisms, especially Vibrio alginolyticus, and, in previous studies, it has been found that the KPL polyclonal antiE. coli O157:H7 antibody minimally crossreacted with E. hermanii (63). A high quality antibody to S. aureus cells could not be found and, theref ore, the assay was unsuccessful. Compounding this problem is that any antibody used in an immunoassay for S.aureus would be problematic due to non-specific bindi ng of the Fc portion of the antibody to protein A on the surface of S. aureus cells. Specificity in any immunoassay for S. aureus would be compromised and methods to overcome this would need to be developed. This would include finding a way to block t hose protein A binding sites. As for the enterococci, two antibodies we re commissioned for use in optimizing the biosensor assay for enterococci. They were made by immunizing mice and rabbits with boiled cell preparations of Ent. faecalis and Ent. faecium The first antibody created was a monoclonal antibody made by the Hybridoma Core Laboratory at the University of Florida, Gainesville, FL. This process failed to yield a high affinity IgG for use on the biosensor assay. IgM produced from this pr ocess was tested on the biosensor using the
149 sandwich immunoassay format a nd detection was not improve d. The other antibody was a polyclonal antibody made by a commercial source (Strategic Biosolutions) which showed high affinity for cells of Ent. faecalis when screened through ELISA. On further analysis through Western blot, it was found that this antibody was specific for one primary target on the Ent. faecalis cells. When compared to the American Research Products (ARP), polyclonal anti-Strep group D antibody, the custom-made antibody was found to be inferior. The ARP antibody showed high affinity for several targets on the Enterococcus cells. In theory, this antibody would allow for the maximum capture of cells and would minimize competition between capture and detection antibodies for the same epitopes (Figure 13). One future direction for this part of the Enterococcus project would be to further research and charact erize all the epitopes for which the ARP antibodies showed highest affinity. Those proteins or carbohydrates could then be isolated and used to create several monocl onal antibodies to be combined later on to make a custom prepared polyclonal antibody. The success of this procedure could yield a more specific and higher affinity antibody which could improve the sensitivity of the Enterococcus assay. Within the past decade, research has revealed many new characteristics about the enterococci. First, the enterococci have been found to subsist embe dded in sediments and sand which would make counts obtained from water underestima tes of the actual counts that are in the environment (92). Second, en terococci have been known to enter a viable but non culturable state and attach to copepods much like V. cholerae does to enhance their survival under nutrient limiting conditions (164, 173, 231-233) These findings drive the need to move away from or at least add to culture based methods of detection which
150 probably underestimate Enterococcus concentrations in the water. This of course is a double edged sword with detection of indicators. While it is desirable to be able to completely enumerate enterococci in seawater at any given time, if VBNC enterococci are present in the environment, their detection could compromise risk assessment. The inability to distinguish the presence of enterococi due to current contamination event from those presnt from a previous contamin ation event could lead to unnecessary beach closings. These are all factors which must be considered when developing methods for the detection of indicators of fecal pollution. Viable but nonculturable cells present a unique problem for the biosensor and any immunoassay geared toward the detection of ba cteria. Antibodies for these assays are made to specific epitopes and antigens on the surface of a bacterial cell. Many bacteria undergo physiological changes when in the vi able but non culturable state which can lead to changes in antigenicity and cell surface structure. However, an immunoassay format for detection of enterococci in seawater is f easible even when cells are in a VBNC state. The group D antigen is consistently produced by Enterococcus cells, even when the cells go VBNC. In fact, it has been observed th at teichoic acid production in these cells increases when they are in th e VBNC state (86, 233). This would certainly work to the benefit of an immunoassay as antibodies c ould bind cells even in times of stress. E. coli O157:H7, has been found to undergo structural changes that could interfere with an immunoassay. The antibody used in assay deve lopment is directed against the O antigen of the cells. This antigen has been found to be shed and not produced when these cells are in the VBNC state, however, verotoxin can still be produced (106). This structural
151 change presents a problem because the immunoass ay would not be able to detect the cells that are still virulent. VBNC V. cholerae is less of a concern in immunoa ssay development. As described previously, this state has been shown to affect the production of O antigen in E. coli O157:H7 affecting the ability of an antibody to bind to the target cell. The antiserum used in this study was created against the O antigen of V. cholerae calling into question whether these VBNC cells may not be de tected when assayed by immunological methods. Research has shown that the loss of O antigen is not a problem with V.cholerae in the VBNC state as these cells do not lose O antigenicity when they are under these stressed conditions (37). However, a limitation of the biosensor assay for V. cholerae would be the inability to detect nonagglutinable V. cholerae which does not react with O1 antiserum. While these vibrios are not as highly pathogenic as O1 V. cholerae they have been found to cause disease. This is a limitation for any immunoassay for V. cholerae not just the biosensor assay. Other supplem ental methods such as culture or molecular methods would be needed to detect these vibrios. There are several ways that the fiber optic biosensor assay could be improved in order to make this method more feasible fo r field use. The most important factor would be to improve the sensitivity of the assays by several log concentrations. One way to address this problem would be to search fo r alternative capture and detection molecules besides antibodies. One such alternative w ould be to model the fluorescent in situ hybridization technique (FISH) described above and use DNA or PNA probes specific for the 16S rRNA region of the enterococci labe led with a fluorescent molecule (120, 177, 178, 255). Advandx Inc. (Woburn, MA) has deve loped such a system for detecting
152 Ent. faecalis using a fluorescein isothiocyanate (F ITC) labeled PNA probe. The system is intended for use in blood culture however, it could be tested for use in other matrices. Although this system has been developed for use on a microscope slide, it could be adapted for use in cell suspension and ultimately, for use on the biosensor. Aptamers which are unique antibody-like nucleic acids are also promising nucleic acid alternatives which could be used to improve biosensor assays. Several molecules could be suggested as specific altern ative capture and detection molecules for enterococci. For example, ente rococci have been found in clinical studies to bind biological molecules (extracellular matrix proteins ) found in the human body such as collagen, fibronectin, fibrinogen, and vitronectin (186, 217, 278). Bacterial surface adhesions called MSCRA MMs (microbial surface components recognizing adhesive matrix molecules) have b een found in isolates of Ent. faecalis and Ent. faecium One such molecule, Ace, in Ent. faecalis binds collagen (206). One of these molecules could potentially be used as a capture molecule much like lectins were used in this study. A specific antibody could then be used for the detection step. Also, human mannose binding protein has been found to bind the liptoeichoic acid of gram positive bacteria including enterococci, so this molecule is also an alternative for capture and detection of enterococci (205). Lectins could also be test ed as alternative capture molecules in the assays for V. cholerae E. coli O157:H7 and S. aureus since they were not done so in this study. Another way to improve assay sensitivity would be to add the sensitivity of PCR to the immunoassay. One way to accomplish this would be to employ the use of a more novel technique, immuno polymerase chain reaction (immuno-PCR). This method
153 developed by Sano et al. combines the speci ficity of both nucleic acids and antibodies into a versatile ELISA (219). In immuno-PCR, antibodies are linked to specific nucleic acid sequences. Using a standard sandwic h immunoassay format, target antigen is captured by an antibody. An antibody tagged with a specific nucleic acid sequence is used to detect the captured target. Then nucleotides, primer, and polymerase are added to the mix and the specific DNA sequence attached to the detection antibody is amplified and used to generate a fluorescent signal using a real time PCR machine. Immuno-PCR has been used to detect protein A of S. aureus with much success (113). This methodology could be applied to the biosen sor assay described in this study as a confirmatory method and a met hod of signal amplification. It has been shown in our laboratory by Dr. Joyce Simpson, a postdocto ral fellow, that PCR can be performed on captured cells on the waveguide when the waveguide is cut into pieces (235). This method could be used to try immunoPCR on captured cells as well. Novel methods are being studied and de veloped everyday for the detection and recovery of enterococci in water and thes e methods could potentially be adapted for biosensor detection. Chemiluminescence has b een discovered in enterococci when the bacteria are grown in the presence of a ferm entable sugar and this property is now being developed as a method to try and detect fecal contamination in water (7). Chemiluminescence is based on the detection of photon emission produced by a chemical such a luminol in response to oxygen metabo lites. In the case of the enterococci, it was found that when the respiratory chain is imp aired, extracellular superoxide is generated and can react with luminol to generate a chemiluminescent signal (7). This property
154 could be used in an array type biosensor format using a camera which could detect the chemiluminescent signal. Emerging biosensor technologies offer th e opportunity to further expand detection assays. DNA microarrays have been develope d as a way to detect multiple targets at once. The principle behind microarray technology is that DNA probes can be synthesized and patterned onto a glass slid e for the detection of a specific cDNA sequence. The patterned slide is exposed to the test sample labeled with either a radioactive label or a fluorescent tag and a signal is generated or not based on the hybridization event. This technique has been used for years to monitor and record gene expression in different types of clinical research and now is being applied to pathogen detection in environmental sa mples (28, 197). This technology has already been used to detect disease agents, but instead of DNA, antibodies were patterned onto a slide (216, 249). Multiplex methods such as a micro array would allow for the simultaneous monitoring of all suggested indicator organi sms not just enterococci, and specific agents of waterborne disease such as V. cholerae O1, E. coli O157:H7, and S. aureus It has already been done to detect Staphylococcus enterotoxin B, ricin, Francisella tularensis and Bacillus globigii using a portable version of the bi osensor used in this study, the RAPTOR (5). An approach like this would also be very helpful for epidemiological studies because they are always being conduct ed to try and correlate the incidence of disease caused by specific pathogens to the dens ities of indicator organisms in water. There are many future directions for this project. The first would be to attempt a study to determine the origin of fecal contamination at the Courtney Campbell beach in Clearwater. Scott etal. recently reported the use of a host associated molecular marker
155 (esp; enterococcal surface protein) in Ent. faecium which could be used as an index of human fecal pollution (222). This could be a way in which to ascertain if indicator bacteria are present in the water due to human fecal contamination. Another recent study of Pensacola beach, addressed not only source-t racking of contamination at the beach, but also the survival of enterococci in the swash zone interstitial waters (92). The swash zone is described as the area on shore where the waves continuously wash up on the sand. This swash zone is known to harbor genera lly higher densities of organisms than the actual water (85, 122, 227). This becomes a problem because people spend more time on the beach rather than in the actual water. Analyzing the swash zone for higher densities of enterococci at Courtney Campbell beach woul d be a worthy direction to pursue, as this beach is lacking in restrooms and pets are a llowed on the beach at all times. While it has been shown in this study that V. cholerae and S. aureus can be isolated from the seawater in Tampa Bay, these data should be followe d up with a study to track the abundance of these organisms in the water. A rapid assay for the detection of S. aureus and E. col i O157:H7 in seawater is both desirable and necessary for the prevention of illness. The swimming public is exposed to numerous potential health hazards that could end up in serious infection unless a rapid method is developed to preempt the danger. S. aureus has been frequently isolated from recreational waters using cultu re based methods (4). The need for rapid methods of detection of S. aureus is necessary, especially now with the emergence of methicillin resistant Staphylcoccocus aureus (MRSA) and their isolation from marine waters (39). There is a real need not only to develop methods to rapidly detect low levels of S. aureus but also to assess the prevalence of S. aureus in seawater. A study addressing the
156 abundance of S. aureus in Tampa Bay seawater could add to research that this organism is and indicator of contamination that shoul d be routinely monitored for in seawater (281). The majority of water related outbreaks of E. coli O157:H7 have been due to contaminated freshwater and drinking water. At least one study in the past has suggested that toxigenic E. coli is not a health risk associated with coastal waters (148). While not a prevalent health risk, enterotoxigenic E. coli has caused beachwater related illness in at least two instances. The first beach water outbreak of E. coli O157:H7 took place in the United Kingdom in 1999 (107). Since then, ther e has also been an outbreak reported at a beach in Montreal in 2001 (22). These out breaks necessitate the need for a rapid assay for E. coli O157:H7 in seawater especially in those instances of disease outbreak. Currently, outbreaks of cholera are ongoing in West Africa and are constantly being reported to the World Health Organization (WHO). Some of those African countries include Benin, Mauritania, and Mali. In Guinea alone, 1956 cases were reported from mid-July to September of this year. Of thos e cases, 72 resulted in death (276). Kaper et al. found in 1982 that there was a resident toxigenic strain of V. cholerae in the Gulf coast (128). It is for this reason that there is a need for screening and detection methods to monitor this bacterium in seawater. Th ere is a requirement for rapid, specific, and sensitive methods for detecting V. cholerae in shellfish as well as in water, particularly in the third world. Methods such as these ar e especially necessary for the complete assessment of seawater due to the presence of VBNC cells. The infectious dose of Vibrio cholerae in food is between 104 and 107 CFU, but in seawater the infectious dose is 1010 CFU because of the absence of food which can absorb stomach acid and improve the
157 chances of survival of the cells in the body (33). The biosensor immunoassay could be used as a screening tool for consistent monitoring of beach water so that when temperatures started to increase and optimum conditions for the proliferation of V. cholerae in the environment arose, this shift could be detected and illness could be prevented. The biosensor has the potential to specifically detect low levels of V. cholerae in seawater and seafood as well if a high quality antibody can be found. Biosensors are currently being used in ma rine science to detect various pollutants including algal toxins (brevetoxins, saxitoxi n) (31), pesticides (264), phosphates (78), pathogens (150), and trace me tals (lead, zinc, cadmium) (118, 166, 214) directly from sea water which can many times be a very complex matrix. Biosensor immunoassays are ideal for the detection of small molecules like toxins or pesticides; however, detection sensitivity can be affected with large target analytes such as bacteria. This problem can be resolved to some extent through signal amplification technique s and the use of novel capture and detection molecules; however, the task of detecting very low concentrations of microorganisms in a comple x matrix without the use of selective enrichment will remain difficult. The low numbers of ente rococci required by the EPA to designate a beach uncontaminated or minimally conta minated necessitates a sensitive method for detection. In the method devel oped in this study, such low levels of detection could be obtained using the Analyte 2000 biosensor in co mbination with sample concentration and selective enrichment in mE broth for 8-24 hours. The biosensor assay alone took only 10 minutes after the samples was attached to th e waveguide and results could be obtained in 24 hours or less. This method could serve as an efficient preliminary screening tool for the assessment of recreational water quality.
158 REFERENCES 1. Aarestrup, F., P. Butaye, and W. Witte. 2002. Nonhuman reservoirs of enterococci, p. 55-99. In M. S. Gilmore (ed.), The Enterococci: Pathogenesis, Molecular Biology, and Antibiotic Resi stance. ASM Press, Washington, D.C. 2. Aarestrup, F. M., F. Bager, and J.S. Andersen. 2000. Association between the use of avilamycin for growth promotion and the occurrence of resistance among Enterococcus faecium from broilers: epidemiological study and changes over time. Microb. Drug. Resis. 6: 71-75. 3. Aldova, E., K. Laznickova, E. Stepankova, and J. Lietava. 1968. Isolation of nonagglutinable Vibrios from an enteritis outbreak in Czechoslovakia. J. Infect. Dis. 118: 25-31. 4. Amercan Public Health Association. 1995. Standard methods for the examination of water and wastewater 19 ed. American Public Health Association, Washington D.C. 5. Anderson, G. P., K.D. King, K.L. Gafney, and L.H. Johnson. 2000. Multianalyte interrogation using the fiber op tic biosensor. Biosens. Bioelectron. 14: 771-777. 6. Anderson, K. L., J.E. Wh itlock, and V.J. Harwood. 2005. Persistence and diffferential survival of fecal indicator bacteria in subtropical waters and sediments. Appl. Environ. Microbiol. 71: 3041-3048. 7. Andre, P., C. Metzger, S. Petey, D. Muller, and D.J-M. Vidon. 2005. Chemiluminescence of enterococci isolates from freshwater. FEMS Microbiol. Lett. 245: 123-129. 8. Andreotti, P. E., G.V. Ludwig, A.H. Peru ski, J.J. Tuite, S.S. Morse, and L.F. Peruski Jr. 2003. Immunoassay of infectious agents. Biotech. 35: 850-859. 9. Andrewes, F. W., and T.J. Horder. 1906. A study of the streptococci pathogenic for man. Lancet 2: 708-713. 10. Archibald, A. R., and H. E. Coapes. 1971. The interaction of Concanavalin A with teichoic acids and bact erial walls. J. Biochem. 123: 665-667.
159 11. Aslan, K., J.R. Lakowicz, and C.D. Geddes. 2004. Nanogold-plasmonresonance-based glucose sensing. Anal. Biochem. 330: 145-155. 12. Baddiley, J., and A. L. Davison. 1961. The occurrence and location of teichoic acids in Lactobacilli. J. Gen. Microbiol. 24: 295-299. 13. Bager, F., M. Madsen, J. Christensen, and F.M. Aarestrup. 1997. Avoparcin used as a growth promoter is associat ed with the occurrence of vancomycinresistant Enterococcus faecium on Danish poultry and pig farms. Prev. Vet. Med. 31: 95-112. 14. Barua, D. 1992. History of cholera, p. 1-36. In D Barua and W.B. Greenough III (ed.), Cholera. Plenum Me dical Book Co., New York. 15. Bennett, A. R., S. MacPhee, and R.P. Betts. 1996. The isolation and detection of Escherichia coli O157:H7 by use of immunomagnetic separation and immunoassay. Lett. Appl. Microbiol. 22: 237-243. 16. Bennish, M. L. 1994. Cholera: pathophysiology, c linical features, and treatment, p. 229-255. In P. A. Blake, I.K. Wachsmuth, and O. Olsvik (ed.), Vibrio cholerae and Cholera: Molecular to Global Pers pectives. ASM Press, Washington D.C. 17. Blake, P. A. 1994. Endemic cholera in Austra lia and the United States, p. 309319. In I.K. Wachsmuth, P.A. Blake, and O. Olsvik (ed.), Vibrio cholerae and Cholera: Molecular to Global Persp ectives ASM Press, Washington, D.C. 18. Boutelle, M. G., H. Allen, O. Hill, M. Berners, R. John, P.D. Dobson, and P. Leigh 1996. New technologies for amperometric biosensors. J. Mol. Recognit. 9: 664-71. 19. Bradley, D. 1977. Health aspects of water suppl ies in tropical countries. p. 3-17. Water, Wastes, and Health in Hot Countries. John, Wiley, and Sons Ltd., London. 20. Brandt, O., J. Feldner, A. Stephan, M. Schroder, M. Schnolzer, H. F. Arlinghaus, J.D. Hoheisel, and A. Jacob. 2003. PNA microarrays for hybridisation of unlabelled DNA sa mples. Nucleic Acids Research. 31: 1-9. 21. Brandt, O., and J. D. Hoheisel. 2004. Peptide nucleic ac ids on microarrays and other biosensors. Trends in Biotechnology 22: 617-622. 22. Bruneau, A. 2004. Outbreak of E. coli O157:H7 associated with bathing at a public beach in the Montreal-Centre region. Can. Commun. Dis. Rep. 30: 133136.
160 23. Buchanan, R. L., and M. P. Doyle. 1997. Foodborne disease significance of Escherichia coli O157:H7 and other enterohemorrhagic E.coli Foodtechnol. 51: 69-76. 24. Budnick, G. E., R.T. Howard, and D.R. Mayo. 1996. Evaluation of Enterolert for enumeration of enterococci in recr eational waters. Appl. Environ. Microbiol. 62: 3881-3884. 25. Cabelli, V. J. 1983. Health effects criteria for ma rine recreational waters. U.S. Environmental Protection Agency publi cation no. EPA-600/1-80-031. Office of Research and Development, U.S. Envi ronmental Protection Agency, Washington, D.C. 26. Cabelli, V. J., A.P. Dufour, L.J. McCabe, and M.A. Levin. 1982. Swimmingassociated gastroenteritis and water quality. Am. J. Epidemiol. 115: 606-616. 27. Calamita, H., B.P. Thuan, M. Kemper, G. Buck, A.M. Calderon de la Barca, and R.J. Doyle. 2000. Drift in lectin reactivity of group D Streptococci between the years 1983 and 1998 In D. Rogers (ed.), Lectins, Biology, Biochemistry, Clinical Biochemistry, vol. 14, Hellerup. 28. Call D.R., M. K. Borucki., and F.J. Loge. 2003. Detection of bacterial pathogens in environmenta l samples using DNA microarrays. J. Microbiol. Met. 53: 235-243. 29. Carmon, K. S., R.E. Baltus, and L.A. Luck. 2005. A biosensor for estrogenic substances using the quartz crysta l microbalance. Anal. Biochem. 345: 277-283. 30. Carrillo, C. 1991. Difference between toxigenic Vibrio cholerae O1 from South America and U.S. gulf coast. The Lancet 337: 1097-1098. 31. Carter, R. M. 1993. Immunoelectrochemical bi osensors for detection of saxitoxin and brevetoxin. Immunometh. 3: 128-133. 32. Carter, R. M., M.B. Jacobs, and G.J.Lubrano 1995. Piezoelectric detection of ricin and affinity purified goat anti-ricin antibody. Anal. Lett. 28: 1379-1386. 33. Cash, R. A., S.I. Music, J.P. Libonati, M.J. Snyder, R.P. Wenzel, and R.B. Hornick. 1974. Response of man to infection with Vibrio cholerae I. Clinical serologic, and bacteriologic responses to a known inoculum. J. Infect. Dis. 129: 45-52. 34. Centers for Disease Control and Prevention. 2005. Vibrio illnesses after hurricane Katrina-multiple states, Au gust-September 2005. Morbid. Mortal. Weekly Rep. 54: 1-4.
161 35. Centers for Disease Control and Prevention. 1991. Cholera-Peru. Morbid. Mortal. Weekly Rep. 40: 108-109. 36. Centers for Disease Control and Prevention. 1999. Laboratory methods for the diagnosis of epidemic dysentery and chol era. p. 41-54. Division of Bacterial and Mycotic Diseases, Atlanta. 37. Chaiyanan, S., A. Huq, T. Maugel, and R.R. Colwell 2001. Viability of the nonculturable Vibrio cholerae O1 and O139. Syst. Appl. Microbiol. 24: 331-341. 38. Chalmers, R. M., H. Aird, and F.J. Bolton. 2000. Waterborne Escherichia coli O157:H7. J. Appl. Microbiol. 88: 124-132. 39. Chambers, H. F. 2001. The changing epidemiology of Staphylococcus aureus ? Emerg. Infec. Dis. 7: 178-182. 40. Chapman, P. A., A.T. Cerdan Malo, C.A. Siddons, and M. Harkin 1997. Use of commercial enzyme immunoassays and immunomagnetic separation systems for detecting Escherichia coli O157 in bovine fecal sa mples. Appl. Environ. Microbiol. 63: 2549-2553. 41. Charles, K., N. Ashbolt, C. Ferguson, D. Roser, R. McGuinness, and D. Deere. 2003. Centralised versus decentralised sewage systems: a comparison of pathogen and nutrient loads released into Sydney's drinking water catchments. Water. Sci. Technol. 48: 53-60. 42. Charych, D. 1999. Biochromic polydiacetylene synthetic membranes, p. 1-24. In V. C. Yang (ed.), Biosensors and Thei r Applications. Kluwer Academic/Plenum Publishers, New York. 43. Charych D.H., J. O. Nagy, W. Spevak, and M.D. Bednarski. 1993. Direct colorimetric detection of a receptor-lig and interaction by a polymerized bilayer assembly. Science. 261: 585-588. 44. Chen, D., Y. Cao, B. Liu, and J. Kong. 2002. A BOD biosensor based on a microorganism immobilized on an Al203 sol-gel matrix. Anal. Bioanal. Chem. 372: 737-739. 45. Cheung, W. H. S., K.C.K. Chang, and R.P.S. Hung. 1989. Health effects of beach water pollution in Hong Kong. Epidemiol. Infect. 105: 139-162. 46. Choopun, N., V. Louis, A. Huq, and R.R. Colwell. 2002. Simple procedure for rapid identification of Vibrio cholerae from the aquatic environment. Appl. Environ. Microbiol. 68: 995-998.
162 47. Chun, J., A. Huq, and R.R. Colwell. 1999. Analysis of 16S-23S rRNA intergenic spacer regions of Vibrio cholerae and Vibrio mimicus Appl. Environ. Microbiol. 65: 2202-2208. 48. Clark, L. C., and C. Lyons. 1962. Electrode systems for continuous monitoring in vascular surgery. Ann. NY Acad. Sci. 102: 29-45. 49. Clarke, S. H., and R. Rudikoff. 1984. Evidence for gene conversion among immunoglobulin heavy chain variable region genes. J. Exp. Med. 159: 773-782. 50. Cole, H. B., J.W. Ezzell, K.F. Keller, and R.J. Doyle. 1984. Differentiation of Bacillus anthracis and other Bacillus species by lectins. J. Clin. Microbiol. 19: 48-53. 51. Colwell, R. R. 2004. Infectious disease and envi ronment: cholera as a paradigm for waterborne disease. International Microbiol. 7: 285-289. 52. Colwell, R. R., J.A.K. Hasan, A. Huq, L. Loomis, R.J. Siebeling, M. Torres, S. Galvez, S. Islam, M. Tamplin, and D. Bernstein. 1992. Development and evaluation of a rapid, simple, se nsitive, monoclonal antibody-based coagglutination test for direct detection of Vibrio cholerae O1. FEMS Microbiol. Lett. 97: 215-220. 53. Corbett, S. J., G.L. Rubin, G.K. Curry, and D.G, Kleinbaum. 1993. The health effects of swimming at Sydney beaches. Am. J. Pub. Health 83: 17011706. 54. Cowburn, J. K., T. Goodall, E.J. Fricker, K.S.Walter, and C.R. Fricker. 1994. A preliminary study of the use of co lilert for water quality monitoring. Lett. Appl. Microbiol. 19: 50-52. 55. Coyette, J., and L.E. Hancock. 2002. Enterococcal cell wall, p. 177-218. In M. S. Gilmore (ed.), The Enterococci: Pa thogenesis, Molecular Biology, and Antibiotic Resistance. ASM Press, Washington, D.C. 56. Dakin, W. P. H., D.J. Howell, R.G.A. Sutton, M.F. O'Keefe, and P. Thomas. 1974. Gastroenteritis due to nonagglutinable (non-cholera) Vibrios Med. J. Aust. 2: 487-490. 57. Danbing, K., F.J. Picard, F. Martineau, C. Menard, P.H. Roy, M. Ouellette, and M. G. Bergeron. 1999. Development of a PCR assay for the rapid detection of enterococci. J. Clin. Microbiol. 37: 3497-3503.
163 58. Daniels, L., R.S. Hanson, and J.A. Phillips. 1994. Chemical analysis, p. 518519. In R. G. E. Murray, W.A. Wood, and N. R. Krieg (ed.), Methods for general and molecular bacteriology. ASM Press, Washington D.C. 59. Dao, M. 1985. An improved method of antigen de tection on nitrocellulose: in situ staining of alkaline phosphatase conj ugated antibody. J Immunol Methods 82: 225-231. 60. De, S., I.G. Macara, and D.A. Lannigan. 2005. Novel biosensors for the detection of estrogen rece ptor ligands. The J. Ster. Biochem. and Mol. Biol. 96: 235-244. 61. De Vaux, A., G. Laguerre, C. Divies, and H. Prevost. 1998. Enterococcus assini sp. nov. isolated from the caecum of donkeys. Int. J. Syst. Bacteriol 48: 383-387. 62. Deibel, R. H. 1964. The group D Streptococci. Bacteriol. Rev. 3: 330-366. 63. DeMarco, D. R., and. D. V. Lim. 2002. Detection of Escherichia coli O157:H7 in 10and 25gram ground beef samples w ith an evanescent wave biosensor with silica and polystyrene waveguides. J. Food. Prot. 65: 596-602. 64. DeMarco, D. R., and D. V. Lim. 2001. Direct detection of Escherichia coli O157:H7 in unpasteurized apple juice with an evanescent wave biosensor. J Rap. Meth. Auto. Microbiol. 9: 241-257. 65. DePaola, A. 1992. Isolation of Latin American epidemic strain of Vibrio cholerae O1 from the U.S. Gulf Coast. Lancet 339: 624-625. 66. Devriese, L. A., B. Pot, and M.D. Collins. 1995. Phenotypic identification of the genus Enterococcus and differentiation of phylogene tically distinct enterococcal species groups. J. Appl. Bacteriol. 75: 399-408. 67. Devriese, L. A., K. Ceyssens, U.M. Rodrigues, and M.D. Collins. 1990. Enterococcus columbae, a species from pigeon intestines. FEMS Microbiol. Lett. 71: 247-252. 68. Dible, H. 1921. The Enterococcus and the faecal streptococci: their properties and relations. J. of Pathol. Bacteriol. 24: 3-12. 69. Donaldson, K. A., M.F. Kramer, and D.V. Lim. 2004. A rapid detection method for Vaccinia virus, the surrogate fo r smallpox virus. Biosens. Bioelectron. 20: 322-327.
164 70. Doyle, R. J., and M. Slifkin. 1989. Applications of lec tins in microbiology. ASM News 55: 655-658. 71. Eckner, K. F. 1998. Comparison of membrane filtration and multiple tube fermentation by the Colilert and Enterolert methods for the detection of waterborne coliform bacteria, Escherichia coli and enterococci used in drinking and bathing water quality monitoring in southern Sweden. Appl. Environ. Microbiol. 64: 3079-3083. 72. Efstratiou, M. A., A. Mavridou, S. C. Richardson, and J.A. Papadakis. 1998. Correlation of bacterial i ndicator organisms with Salmonella spp., Staphylococcus aureus and Candida albicans in sea water. Lett. Appl. Microbiol. 26: 342-346. 73. Elliot, E. L., C. A. Kaysner, L. Jackson, and M.L. Tamplin. 1995. Bacteriological analytical manual 8th ed. AOAC International. Gaithersburg, MD. 74. Elliot, S. D. 1959. Group and type specific polysaccharides of group D Streptococci. Nature 184: 1342. 75. Elliot, S. D. 1962. Teichoic acid and the group antigen of group D streptococci. Nature. 193: 1105-1106. 76. Elliot, S. D. 1960. Type and group polysaccharides of group D Streptococci. J. Exp. Med. 111: 621. 77. Emanuel, P. A., J. Dang, J.S. Gebhardt, J. Aldrich, E.A. Garber, H. Kulaga, P. Stopa, J.J. Valdes, and A. Dion-Schultz. 2000. Recombinant antibodies: a new reagent for biological agent de tection. Biosens. and Bioelectron. 14: 751759. 78. Engblom, S. O. 1998. The phosphate sensor. Biosens. Bioelectron. 13: 981-994. 79. Escherich, T. 1887. Ueber darmbakterien im al lgemeinen and diejenige der sauglings im besonderen, sowie der bezie hungen der letzteren zur aetiologiebder darmerkrankungen. Centralbla tt. fur. Bakteriologie. 1: 705-713. 80. Exall, K., J. Marsalek, and K. Schaefer 2004. A review of water reuse and recycling, with reference to Canadian practice and potential: 1. incentives and implementation. Water. Qual. Res. J. Can. 39: 1-12. 81. Facklam, R. R., M. da Gloria, S. Carvalho, and L.M. Teixeira. 2002. History, taxonomy, biochemical characteristics, and antibiotic susceptibility testing of enterococci, p. 1-54. In M. S. Gilmore (ed.), The Enterococci; Pathogenesis, Molecular Biology, and Antibiotic Re sistance ASM Press, Washington, D.C.
165 82. Faruque, S. M., and G.B. Nair. 2002. Molecular ecology of toxigenic Vibrio cholerae Microbiol. Immunol. 46: 59-66. 83. Fattal, B., E. Peleg-Olvesky, T. Agurshy, and H.I. Suval. 1987. The association between sea water pollution as measured by bacterial indicators and morbidity of bathers at Mediterranean beaches in Israel. Chemosphere 16: 565-570. 84. Feng, P., and P. A. Hartman. 1982. Fluoregenic assays for immediate confirmation of Escherichia coli Appl. Environ. Microbiol. 43: 1320-1329. 85. Ferguson, D. M., D.F. Moore, M.A. Getrich, and M.H. Zhowandi. 2005. Enumeration and speciation of enteroco cci found in marine and intertidal sediments and coastal water in southe rn California. J. Appl. Microbiol. 99: 598608. 86. Fischer, W. 1988. Physiology of lipoteichoic aci ds in bacteria. Adv. Microbial Phys. 29: 233-296. 87. Florida Department of Environmenta l Protection, Bureau of Watershed Management. 2004. Integrated water quality assessment for Florida: 2004 305 (b) Report and 303 (d) list update. Tallahassee, FL. 88. Francy, D. S., O.D. Simmons III, M .W. Ware, E.J. Granger, M.D. Sobsey, and F.W. Schaefer III. 2004. Effects of seeding procedures and water quality on recovery of Cryptosporidium oocysts from stream water by using U.S. environmental protection agency method 1623. Appl. Environ. Microbiol. 70: 4118-4128. 89. Fujioka, R., B. Roll, and M. Byappanahalli. 1997. Appropriate recreational water quality standards for Hawaii a nd other tropical regions based on concentrations of Clostridium perfringens In Proceedings of WEF-TEC 97, 70th Annual Conference and Exposition, Water Environment Federation, Alexandria, VA 4: 405-411. 90. Fujioka, R. S. 2002. Microbial indicators of marine recreational water quality, p. 234-243. In R. L. Crawford, C.J. Hurst, G.R. Knudsen, M.J. McInerney, and L.D. Stetzenbach (ed.), Manual of Environm ental Microbiology, 2nd ed. ASM Press, Wasington, D.C. 91. Geng, T., M.T. Morgan, and A.K. Bhunia. 2004. Detection of low levels of Listeria monocytogenes cells by using a fiber-optic immunosensor. Appl. Environ. Microbiol. 70: 6138-6146.
166 92. Genthner, F. J., J.B. James, D.F. Yates, and S. D. Friedman. 2005. Use of composite data sets for source-tracking enterococci in the water column and shoreline intersitial waters on Pensacola Beach, FL. Marine Poll. Bull. 50: 724732. 93. Gessel, P. D., N.C. Hanson, S.M. Goyal, L.J. Johnston, and J. Webb. 2004 Persistence of zoonotic pathogens in surfa ce soil treated with different rates of liquid pig manure. Appl. Soil. Ecol. 25: 237-243. 94. Glaser, R. W. 1999. Surface plasmon resonance biosensors, p. 195-212. In V. C. Yang (ed.), Biosensors and Their A pplications. Kluwer Academic/ Plenum Publishers, New York. 95. Godfree, A. F., D. Kay, and M.D. Wyer. 1997. Faecal streptococ ci as indicators of faecal contamination in water. J. Appl. Microbiol. 83: 110-119. 96. Gonthier, A., Guerin-Faublee, B. Tilly, and M.L Delignette-Muller. 2001. Optimal growth temperature of O157 and non-O157 Escherichia coli strains. Lett. Appl. Microbiol. 33: 352-356. 97. Goodgame R.W., and. W. B. Greenough III. 1975. Cholera in Africa: a message from the west Ann. Intern. Med. 82: 101-106. 98. Gordon, A. M. 1970. Cholera in the new world. J. Ky. Med. Assoc. 68: 657. 99. Graham, K., K. Keller, J. Ezzell, and R. Doyle 1984. Enzyme-linked lectinosorbent assay (ELLA) for detecting Bacillus anthracis Eur. J. Clin. Microbiol. 3: 210-212. 100. Guidry, A. J., K.E. Squiggins, W.F. Vann, and D.C. Westhoff. 1991. Prevention of nonspecific binding of immunoglobulin to Staphylococcus aureus protein A in ELISA assays. J. Immunol. Meth. 143: 159-165. 101. Guilbault, G. G. 1983. Determination of formaldehy de with an enzyme coated piezoelectric crystal. Anal. Chem 55: 1682-1684. 102. Halamek, J., J. Pribyl, A. Makower, P. Skladal, and F.W. Scheller. 2005. Sensitive detection of organophosphates in ri ver water by means of a piezoelectric biosensor. Anal. Bioanal. Chem. 382: 1904-1911. 103. Halverson, K., R.G. Panchal, T.L. Nguyen, R. Gussio, S.F. Little, M. Misakian, S. Bavari, and J.J. Kasianowicz. 2005. Anthrax biosensor, protective antigen ion channel asymmetric blockade. J. Biol. Chem. 280: 34056-34062.
167 104. Hammonds, S. J., T.J. Easterbrook, and M.B. McIllmurray. 1991. Differentiation of enterococci from othe r group D streptococci by means of a specific monoclonal antibody. FEMS Microbiol. Lett. 82: 91-94. 105. Hancock, L. E., and M. S. Gilmore. 2000. Pathogenicity of enterococci, p. 251258. In V. A. Fischetti (ed.), Gram Positive Pathogens. ASM Press, Washington D.C. 106. Hara-Kudo, Y., M. Miyahara, and S. Kumagai. 2000. Loss of O157 O antigenicity of verotoxin-producing Escherichia coli O157:H7 surviving under starvation conditions. Appl. Environ. Microbiol. 66: 5540-5543. 107. Harrison, S., and S. Kinra. 2004. Outbreak of Escherichia coli O157 associated with a busy bathing beach. Commun. Dis. Pub. Health 7: 47-50. 108. Hasan, J. A. K., A. Huq, B. Nair, S. Garg, A.K. Mukhopadhyay, L. Loomis, D. Bernstein, and R. R. Colwell. 1995. Development and te sting of monoclonals antibody-based rapid immunodiagnostic test kits for direct detection of Vibrio cholerae O139 synonym Bengal. J. Clin. Microbiol. 33: 2935-2939. 109. Hasan, J. A. K., A. Huq, M.L. Tamp lin, R.J. Siebeling, and R. R. Colwell. 1994. A novel kit for rapid detection of Vibrio cholerae O1. J Clin. Microbiol. 32: 249-252. 110. Haugland, R. A., S.C., Siefring, L.J. Wym er, K.P. Brenner, and A. P. Dufour. 2005. Comparison of Enterococcus measurements in freshwater at two recreational beaches by quantitative polymerase chain reaction and membrane filter culture analysis. Water Res. 39: 559-568. 111. He, J. W., and S. Jiang. 2005. Quantification of enterococci and human adenoviruses in environmental sample s by real-time PCR. Appl. Environ. Microbiol. 71: 2250-2255. 112. Hitchins, A. D., P.Feng, W.D. Watkins S.R. Rippey, and L.A. Chandler. 1995. Escherichia coli and the coliform bacteria. AOAC. Bacteriological analytical manual, 8th ed. Gaithersburg, MD. 113. Huang, S. H., and T. C. Chang. 2004. Detection of Staphylocococcus aureus by a sensitive immuno-PCR assay. Clin. Chem. 50: 1673-1674. 114. Huber, M., U. Vor Dem Esche, R. Grunow, and W.G. Bessler. 2005. Generation of mouse polyclonal and human monoclonal antibodies against Bacillus anthracis toxin. Drugs Exp. Clin. Res. 31: 35-43.
168 115. Huq, A., E.B. Small, P.A. West, M.I. Huq, R. Rahman, and R.R. Colwell. 1983. Ecological relationships between Vibrio cholerae and planktonic crustacean copepods. Appl. Environ. Microbiol. 45: 275-283. 116. Ikebukuro K., A. Miyata, S.J. Cho, Y. Nomura, S.M. Chang, Y. Yamauchi, Y. Hasebe, S. Uchiyama, and I. Karube. 1996. Microbial cyanide sensor for monitoring river water. J. Biotechnol. 48: 73-80. 117. Isobe, K. O., M. Tarao, N. H. Chiem, L.Y. Minh, and H. Takada. 2004. Effect of environmental factors on the rela tionship between concentrations of coprostanol and fecal indicator bacteria in tropical (Mekong Delta) and temperate (Tokyo) freshwaters. Appl. Environ. Microbiol. 70: 814-821. 118. Ivask, A. 2002. Construction and use of specific luminescent recombinant bacterial sensors for the assessment of bioavailable fracti on of cadmium zinc, mercury, and chromium in soil. Soil. Biol. Biochem. 34: 1439-1447. 119. Jackson, D. E., W. Wong, M.T. Largen, and G.D. Shockman. 1984. Monoclonal antibodies to immunodeterminan ts of lipoteichoic acids. Infec. Immun. 43: 800-803. 120. Jansen, G.J., M. Mooibroek, J. Idema, H.J. M., Harmsen, G.W. Welling, and J.E. Degener. 2000. Rapid identification of bacteria in blood cultures by using fluorescently labeled oligonucleotid e probes. J. Clin. Microbiol. 38: 814-817. 121. Jiang, S.C., and W. Fu. 2001. Seasonal abundance and distribution of Vibrio cholerae in coastal waters quantified by a 16S-23S intergenic spacer probe. Microb. Ecol. 42: 540-548. 122. Jin, G., A.J. Englande, H. Bradford, and H.W. Jeng. 2004. Comparison of E.coli ,enterococci, and fecal coliform as indicators for brackish water quality assessment. Water Environ. Res. 76: 245-255. 123. Johnson, W. M., H. Lior, and G.S. Bezanson. 1983. Cytotoxic Escherichia coli O157: H7 associated with haemo rrhagic colitis in Canada. Lancet i: 76. 124. Jolivet-Reynaud C., A. A., S. Michel, G. Deleage, G. Paranhos-Baccala, V. Gonin, N. Battail-Poirot, X. Lacoux and D. Rolland. 2004. Characterization of mimotopes mimicking an immunodominan t conformational epitope on the hepatitis C virus NS3 helicase. J. Med. Virol. 72: 385-395. 125. Jones, D. and P. M. Shattock. 1960. The location of the group D antigen of Streptococcus J. Gen. Microbiol. 23: 335-343.
169 126. Joseph, R., and G. D. Shockman. 1975. Cellular localization of lipoteichoic acid in Streptococcus faecalis J. Bacteriol. 122: 1375-1386. 127. Josephson, S. L., M.W. Stinson, S.J. Millar, and R.E. Cohen. 1986. Purification of lipoteichoi c acid by chromatography in water-organic solvent systems. Infec. Immun. 51: 378-384. 128. Kaper, J. B., H.B. Bradford, N. C. Roberts, and S. Falkow. 1982. Molecular epidemiology of Vibrio cholerae in the U.S. Gulf Coast. J. Clin. Microbiol. 16: 129-134. 129. Kaper, J. B., J.P. Nataro, N.C. Robert s, R.J. Siebeling, and H.B. Bradford. 1986. Molecular epidemiology of non-O1 Vibrio cholerae and Vibrio mimicus in the U.S. Gulf coast region. J. Clin. Microbiol. 23: 652-654. 130. Karch, H., C. Janetzki-Mittman, A. Aleksic, and M. Datz. 1996. Isolation of enterohemorrhagic Escherichia coli O157:H7 strains from patients with hemolytic uremic syndrome by using immunmagnetic separation, DNA-based methods, and direct culture. J. Clin. Microbiol. 34: 516-519. 131. Kay, B. A., C.A. Bopp, and J.G. Wells. 1994. Isolation and identification of Vibrio cholerae O1 from fecal specimens, p. 3-25. In P. A. Blake, I.K. Wachsmuth, O. Olsvik (ed.), Vibrio cholerae and cholera: Molecular to Global Perspectives. ASM Press, Washington, D.C. 132. Kaysner, C. A., and A. DePaola Jr. 2004. V. cholerae V. parahaemolyticus V. vulnificus and other Vibrio spp. AOAC. Bacteriological analytical manual, 8th ed. Gaithersburg, MD. 133. Kaysner, C. A., and W.E. Hill. 1994. Toxigenic Vibrio cholerae O1 in food and water, p. 27-39. In P. A. Blake, I.K. Wachsmuth, and O. Olsvik (ed.), Vibrio cholerae and Cholera: Molecular to Global Pe rspectives. ASM Press, Washington D.C. 134. Kirby, R., E.J. Cho, B. Gehrke, T. Bay er, Y.S. Park, D.P. Neikirk, J.T. McDevitt and A.D. Ellington. 2004. Aptamer-based sensor arrays for the detection and quantitation of proteins. Anal. Chem 76: 4066-4075. 135. Kissinger, P. T. 2005. Biosensor-a perspective. Biosens. Bioelectron. 20: 25122516. 136. Klapes, N. A. 1983. Comparison of Vogel-Johnson and Baird-Parker Media for membrane filtration recove ry of Staphylococci in swimming pool water. Appl. Environ. Microbiol. 46: 1318-1322.
170 137. Klein, E., and A.C. Houston. 1899. Twenty-eighth annual report of the local government board containing the report of the medical officer. Her majesty's stationary office. 138. Knox, K. W., and A. J. Wicken. 1973. Immunological proper ties of teichoic acid Bacteriol. Rev. 37: 215-257. 139. Koch, W. H., W. L. Payne, B. A. Wentz, and T.A. Cebula. 1993. Rapid polymerase chain reaction method for detection of Vibrio cholerae in foods. Appl. Environ. Microbiol. 59: 556-560. 140. Kohler, G., and C. Milstein. 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256: 495-497. 141. Konig, B., and M. Gratzel. 1994. A novel immunosensor for herpes viruses. Anal. Chem. 66: 341-344. 142. Kramer, M. F., and D.V. Lim. 2004. A rapid and automated fiber optic based biosensor assay for the detection of Salmonella in spent irrigation water used inthe sprouting of sprout seeds. J. Food. Prot. 67: 46-52. 143. Kronick, M. N., and W. A. Little. 1975. New immunoassay based on fluorescence excitation by internal refl ection spectroscopy. J. Immunol. Meth. 8: 235-240. 144. Kubitschko, S., J. Spinke, T. Bruckner, S. Pohl, and N. Oranth. 1997. Sensitivity enhancement of optical imm unosensors with nanoparticles. Anal. Biochem. 253: 112-122. 145. Kubo, I., M. Inagawa, T. Sugawara, Y. Arikawa, and I. Karube. 1991. Phosphate sensor composed from immobilized pyruvate oxidase and an oxygen electrode. Anal. Lett. 24: 1711-1727. 146. Lancefield, R. C. 1940. Specific relationship of ce ll composition to biological activity of hemolytic streptococci. Harvey Lect. 36: 251. 147. Lee, J. H., R.J. Mitchell, B.C. Kim, D.C. Cullen, and M.B. Gu. 2005. A cell array biosensor for environmental toxi city analysis. Biosens. Bioelectron. 21: 500-507. 148. Lee Lang, A., Y. Tsai, C.L. Mayer, K.C. Patton, and C.J. Palmer. 1994. Multiplex PCR for detection of the heat-labile toxin gene and shiga-like toxin I and II genes in Escherichia coli isolated from natural waters. Appl. Environ. Microbiol. 60: 3145-3149.
171 149. LeJeune, J. T., T.E. Besser, D. H. Rice, and D.D. Hancock. 2001. Methods for the isolation of water-borne Escherichia coli Lett. Appl. Microbiol. 32: 316-320. 150. Leonard, P. 2003. Advances in biosensors for detection of pathogens in food and water. Enzyme Microb. Technol. 32: 3-13. 151. Leriche, V. P. S., and B. Carpentier 2000. Use of enzyme-linked lectinsorbent assay to monitor the shift in polysacharide composition in bacterial biofilms. Appl. Environ. Microbiol. 66: 1851-1856. 152. Leslie, R. G. Q., and S. Cohen. 1970. Chemical proper ties of guinea-pig immunoglobulins G1, G2, and M. Biochem. J. 120: 787-795. 153. Levin, M. A., J.R. Fischer, and V.J. Cabelli. 1975. Membrane filter technique for enumeration of enterococci in marine water. Appl. Microbiol. 30: 66-71. 154. Levin, M. M., J.P. Kaper, D. Herrin gton, J. Ketley, G. Losonsky, C.O. Tacket, B. Tall, and S. Cryz. 1983. New knowledge on pathogenesis of bacterial enteric infections as applied to vaccine development. Microbiol. Rev. 47: 510550. 155. Levine, W. C., and P. M. Griffin. 1993. Vibrio infections on the Gulf Coast: results of first year of regional surveillance. J. Infect. Dis. 167: 479-483. 156. Lewis, G. D., S. A. Anderson, and S.J. Turner. 2002. Detection of Enterococci in freshwater and seawater (16S and 23S rRNA Enterococcus oligonucleotide probes). Methods in Mol. Biol. 179: 159-169. 157. Lim, D. V. 2003. Detection of microorganisms and toxins with evanescent wave fiber-optic biosensors. Proceedings of the IEEE 91: 902-907. 158. Lim, D. V., J.M. Simpson, E.A. Kearns, and M.F. Kramer. 2005. Current and developing technologies for monitoring ag ents of bioterrorism and biowarfare. Clin. Microbiol. Rev. 18: 583-607. 159. Lin, H. C., and W. C. Tsai. 2003. Piezoelectric crystal immunosensor for the detection of staphylococcal ente rotoxin B. Biosens. Bioelectron. 18: 1479-1483. 160. Lipp, E. K., A. Huq, and R.R. Colwell. 2002. Effects of global climate on infectious disease: the choler a model. Clin. Microbiol. Rev. 15: 757-770. 161. Lipp, E. K., I.N.G., Rivera, A.I. Gil, E.M., Espeland, N. Choopun, V.R. Louis, E. Russek-Cohen, A. Huq, and R.R. Colwell. 2003. Direct detection of Vibrio cholerae and ctx A in Peruvian coastal wate r and plankton by PCR. Appl. Environ. Microbiol. 69: 3676-3680.
172 162. Liu, J., and Y. Lu. 2004. Colorimetric biosenso rs based on DNAzyme-assembled gold nanoparticles. J. Fluoresc. 14: 343-354. 163. Liu, Y., and Y. Li. 2002. Detection of Escherichia coli O157:H7 using immunomagnetic separation and absorbance measurement. J. Microbiol. Met. 51: 369-377. 164. Lleo, M. M., B. Bonato, M.C. Tafi, C. Signoretto, M. Boaretti, and P. Canepari. 2001. Resuscitation rate in different en terococcal species in the viable but non-culturable state. J. Appl. Microbiol. 91: 1095-1102. 165. Lobitz, B., L. Beck, A. Huq, B. Wood, G.Fuchs, A.S.G. Faruque, and R. Colwell. 2000. Climate and infectious disease: use of remote sensing for detection of V. cholerae by indirect measurement. PNAS 97: 1438-1443. 166. Lu, Y. 2003. New highly sensitive and sel ective catalytic catalytic DNA biosensors for metal ions. Biosens. Bioelectron. 18: 529-540. 167. Lucht, A., R. Grunow, C. Otterbein, P. Moller, H. Feldmann, and A. Becker. 2004. Production of monoclonal antibodies and development of an antigen capture ELISA directed against the envel ope glycoprotein GP of Ebola virus. Med. Microbiol. Immunol. 193: 181-187. 168. Luque de Castro, M. D., and M.C. Herrera. 2003. Enzyme Inhibition based biosensors and biosensing systems: que stionable devices. Biosens. Bioelectron. 18: 279-294. 169. Lyon, W. J. 2001. TaqMan PCR for detection of Vibrio cholerae O1, O139, NonO1, and Non-O139 in pure cultures, raw oysters, and synthetic seawater. Appl. Environ. Microbiol. 67: 4685-4693. 170. Maekawa, S. M., M. Yoshioka, and Y. Kumamoto. 1992. Proposal of a new scheme for the serological typing of Enterococcu faecalis strains. Microbiol. Immunol. 36: 671-681. 171. Manero, A., and A. R. Blanch. 1999. Identification of Enterococcus spp. with a biochemical key. Appl. Environ. Microbiol. 65: 4425-4430. 172. Masarova, J., E.S. Dey, and B. Danielsson. 2004. Optical lectin based biosensor as tool for bacteria identification. Pol. J. Microbiol. 53: 23-27. 173. Maugeri, T. L., M. Carbone, M.T. Fera, G.P. Irrera, and C. Gugliandolo. 2004. Distribution of potentially pathogenic bacteria as free living and plankton associated in a marine coas tal zone. J. Appl. Microbiol. 97: 354-361.
173 174. McCafferty, J., A. D. Griffiths, G. Winter, and D.J. Chiswell. 1990. Phage antibodies: filamentous phage displaying antibody variable domains. Nature. 348: 552-554. 175. McFadden, P. 2002. Broadband biodetection: holmes on a chip. Science. 297: 2075-2076. 176. Mcgowan, K., E. Wickersham, and N.A., Strockbine. 1989. Escherichia coli O157:H7 from water. The Lancet 1: 967-968. 177. Meier, H. 1998. Detection and identification of indicator bacteria using rRNA targeted probes. Presented at the OECD Workshop Molecular Methods for Safe Drinking. 178. Meier, H., C. Koob, W. Ludwig, R. Amann, E. Frahm, S. Hoffman, U. Obst, and K.H. Schleifer. 1997. Detection of enterococci with rRNA targeted DNA probes and their use for hygienic drinking water control. Wat. Sci. Tech 35: 437444. 179. Messer, J. W., and A. P. Dufour. 1998. A rapid, specific membrane filtration procedure for enumeration of enterococc i in recreational water. Appl. Environ. Microbiol. 64: 678-680. 180. Mintz, E. D., T. Popovic, and P.A. Blake. 1994. Transmission of Vibrio cholerae O1, p. 345-356. In P. A. Blake, I.K. Wachsmuth, and O. Olsvik (ed.), Vibrio cholerae and Cholera: Molecular to Gl obal Perspectives. ASM Press, Washington, D.C. 181. Minunni, M. G. G. Guilbault., and B. Hock. 1995. The quartz crystal microbalance as a biosensor. Anal. Lett. 28: 411-419. 182. Miyagi, K., K. Omura, A. Ogawa, M. Hanafusa, Y. Nakano, S. Morimatsu, and K. Sano. 2001. Survival of Shiga-toxin producing Escherichia coli O157:H7 in marine water and frequent detection of the shiga to xin gene in marine water samples from an estuary port. Epidemiol. Infect. 126: 129-133. 183. Molitoris, E., S.W., Joseph, M.I., Krichevsky, W. Sindhuhardja, and R.R. Colwell. 1985. Characterization and Distribution of Vibrio alginolyticus and Vibrio parahaemolyticus isolated in Indonesia. Appl. Environ. Microbiol. 50: 1388-1394. 184. Morris, G. J., and R.E. Black. 1985. Non-O group 1 Vibrio cholerae : a look at the epidemiology of an occasional pathogen. Epidemiol. Rev. 12: 179-191.
174 185. Nakamura, H., and I. Karube. 2003. Current research activity in biosensors. Anal. Bioanal. Chem. 377: 446-468. 186. Nallapareddy, S. R., G.M. Weinstock, and B.E.Murray. 2003. Clinical isolates of Enterococcus faecium exhibit strain-specific co llagen binding mediated by Acm, a new member of the MSCRAMM family. Mol. Microbiol. 47: 1733-1747. 187. Nataro, J. P., and J. B. Kaper. 1998. Diarrheagenic Escherichia coli Clin. Microbiol. Rev. 11: 142-201. 188. National Research Council of the National Academies. 2004. Indicators for waterborne pathogens. The National Academies Press, Washington D.C. 189. Nielsen, P. E., M. Egholm, R.H. Berg, and O. Buchardt 1991. Sequenceselective recognition of DNA by strand di splacement with a thymine-substituted polyamide. Science 254: 1497-1500. 190. Noble, R. T., D.F. Morre, M.K. Leeca ster, C.D. McGee, and S.B. Weisberg. 2003. Comparison of total coliform, fecal coliform, and enterococcus bacterial indicator response for ocean recreational water quality testing. Water Res. 37: 1637-1643. 191. Noble, R. T., M.K. Leecaster, C.D. McGee, S.B. Weisberg, and K. Ritter. 2004. Comparison of bacterial indicator anal ysis methods in stormwater affected coastal waters. Water Res. 38: 1183-1188. 192. Noble, R. T., S.B. Weisberg, M.K. L eecaster, C.D. Mcgee, K. Ritter, K.O. Walker, and P.M. Vainik. 2003. Comparison of beach bacterial water quality indicator measurement met hods. Environ. Mon. Assess. 81: 301-312. 193. O'Sullivan, C. K., and G.G. Guilbault. 1999. Piezoelectric immunosensors: theory and applications, p. 159-173. In V. C. Yang (ed.), Biosensors and Their Applications. Kluwer Academic /Plenum Publishers, New York. 194. Ogert, R. A., J.E. Brown, B.R. Singh, L.C. Shriver-Lake, and F.S. Ligler. 1992. Detection of Clostridium botulinum toxin A using a fiber optic based biosensor. Anal. Biochem. 205: 306-312. 195. Orla-Jensen, S. 1919. The lactic acid bacteria. Me m. Acad. Roy. Sci. Danemark. Sec. Sci. Ser. 5: 81-197. 196. Paitan, Y., D. Biran, I. Biran, N. Shechter, R. Babai, J. Rishpon, and E.Z. Ron. 2003. On-line and in-situ biosensors for monitoring environmental pollution. Biotech. Adv. 22: 27-33.
175 197. Pancrazio, J. J., J.P. Whelan, D.A. Bo rkholder, W. Ma, and D. A. Stenger. 1999. Development and applica tion of cell-based biosensors. Ann.Biomed. Eng. 27: 697-711. 198. Perry-O'Keefe, H., S. Rigby, K. Oliveira, D. Sorensen, H. Stender, J. Coull, and J.J. Hyldig-Nielsen. 2001. Identification of indi cator microorganisms using a standardized PNA FISH me thod. J. Microbiol. Methods 47: 281-292. 199. Peterson, K. M., and J.J. Mekalanos. 1988. Characterization of the Vibrio cholerae Tox R regulon: identification of novel genes involved in intestinal colonization. Infec. Immun. 56: 2822-2829. 200. Peterson, M. E., B.J. Yokel, and D.V. Lim. 1984. Recovery of Selected Pathogens from Naples Bay, Florida, and Associated Waterways. Estuaries 7: 133-138. 201. Pfeffer, C., and J. D. Oliver. 2003. A comparison of thio silphate-citrate-bile salts-sucrose (TCBS) agar and thiosulpha te-chloride-iodide (TCI) agar for the isolation of Vibrio species from estuarine environements. Lett. Appl. Microbiol. 36: 150-151. 202. Pini, A., and L. Bracci. 2000. Phage display of antibody fragments. Curr. Protein. Pept. Sci. 1: 155-169. 203. Pless, D. D., E.R. Torres, E.K. Reinke, S. Bavari. 2001. High-affinity, protective antibodies to the binding domai n of botulinum neurotoxin type A. Infec. Immun. 69: 570-574. 204. Pollitzer, R. 1959. Cholera, Monograph no. 43. World Health Organization. 205. Polotsky, V. Y., W. Fischer, R. Al an, B. Ezekowitz, and K.A. Joiner. 1996. Interactions of human mannose-binding protein with lipoteichoic acid. Infec. Immun. 64: 380-383. 206. Ponnuraj, K., Y. Xu, D. Moore, C.S.Champion, L. Boque, M. Hook, S. V.L. Narayana. 2002. Crystallization and preliminary X-ray crystallographic analysis of Ace: a collagen binding MSCRAMM from Enterococcus faecalis Biochimica. Biophysica. Acta. 1596: 173-176. 207. Rabbani, G. H., and W.B. Greenough. 1999. Food as a vehicle of transmission of cholera. J. Diarrhoeal Dis. Res 17: 1-9. 208. Raghubeer, E. V., and J. R. Matches. 1990. Temperature range for growth of Escherichia coli serotype O157:H7 and selected coliforms in E.coli medium. J. Clin. Microbiol. 28: 803-805.
176 209. Ramamurthy, T., S. Garg, R. Sharma, S.K. Bhattacharya, G.B. Nair, T. Shimada, T. Takeda, T. Karasawa, H. Kurazano, A. Pal, and Y. Takeda. 1993. Emergence of novel strain of Vibrio cholerae with epidemic potential in southern and eastern India. Lancet 341: 703-704. 210. Reidl, J., and K. E. Klose. 2002. Vibrio cholerae and cholera:out of the water and into the host. FEMS Microbiol Rev 26: 125-139. 211. Ren, S., Y. Qin, X. Zhang, D. Lin, B. Yan, and W. Wang. 1996. Rapid detection of cholera toxin gene of Vibrio cholerae O1 by polymerase chain reaction. Chin. Med. J. 109: 494-496. 212. Riley, L. W., R.S. Helgerson, S.D. McGee, H.B. Wells, J.G. Davis, B.R. Hebert, R.J. Olcott, H.M., Johnson, L. M. Hargrett, N.T. BLake, and P.A. Cohen 1983. Hemorrhagic colitis associated with a rare Escherichia coli serotype. N. Engl. J. Med. 308: 681-685. 213. Rivera, S. C., T.C. Hazen, and G.A. Toranzos. 1988. Isoaltion of fecal coliforms from pristine sites in a tropical rainforest. Appl. Environ. Microbiol. 54: 513-517. 214. Rodriguez, B. B. 2004. Development of urease and glutamic dehydrogenase amperometric assay for heavy metals and screening in polluted samples. Biosens. Bioelectron. 19: 1157-1167. 215. Roitt, I., J. Brostoff, and D.Male. 2001. Immunology. Harcourt Publishers, London. 216. Rowe Taitt, C., G.P. Anderson, and F.S. Ligler. 2005. Evanescent wave fluorescence biosensors. Biosens. and Bioelectron. 20: 2470-2487. 217. Rozdzinski, E., R. Marre, M. Susa, R. Wirth, and A. Muscholl-Silberhorn. 2001. Aggregation substance-mediated adherence of Enterococcus faecalis to immobilized extracellular matrix proteins. Microbial. Patho. 30: 211-220. 218. Sakazaki, R., and K. Tamura. 1971. Somatic antigen variation in Vibrio cholerae Jpn. J. Med. Sci. Biol 24: 93-100. 219. Sano, T. 1992. Immuno-PCR: very sensitive antigen detection by means of specific antibody-DNA conjugates. Science. 258: 120-122. 220. Santo Domingo, J. W., S.C. Siefring, and R.A. Haugland. 2003. Real-time PCR method to detect Enterococcus faecalis in water. Biotech. Lett 25: 261-265.
177 221. Sasaki, S., K. Yokoyama, E. Tamiya, I. Karube, C. Hayashi, Y. Arikawa, and M. Numata. 1997. Sulfate sensor using Thiobacillus ferrooxidans Anal. Chim. Acta. 347: 275-280. 222. Scott, T. M., T.M. Jenkins J. Lukasik, and J.B. Rose. 2005. Potential use of a host associated molecular marker in Enterococcus faecium as an index of human fecal pollution. Environ. Sci. Technol. 39: 283-287. 223. Scott, W. J. 1932. Survey of Connecticut's shor e bathing waters. Public Health Eng. 19: 316. 224. Sghir, A., G. Gramet, A. Suau, V. Rochet, P. Pochart, and J.Dore. 2000. Quantification of bacterial groups within human fecal flora by oligonucleotide probe hybridization. A ppl. Environ. Microbiol. 66: 2263-2266. 225. Sharma, S., P. Sachdeva, and J.S. Virdi. 2003. Emerging water-borne pathogens. Appl. Microbiol. Biotechnol. 61: 424-428. 226. Sherman, J. M. 1937. The streptococci. Bacteriol. Rev. 1: 3-97. 227. Shibata, T., H.M. Solo-Gabriele, L.E. Fleming, and S. Elmir. 2004. Monitoring marine recreational water quality using multiple microbial indicators in an urban tropical environment. Water Res. 38: 3119-3131. 228. Shmuel, A., and B. Sasson. 1979. Immunoglobulin differentiation is dictated by repeated recombination sequences within the V region prototype gene: A hypothesis. Proc. Natl. Acad. Sci. USA 76: 4598-4602. 229. Shons, A. F. D., and J. Najarian. 1972. Piezoelectric quartz immunosensor. J. Biomed. Mater. Res. 6: 565-575. 230. Shriver-Lake, L. C., G.P.Anderson, J. P. Golden, R.A.Ogert, D. Wijesuria, L.Cao, J.M. Mauro, and F. Ligler. 1994. Fiber-optic sens or detection of biological agents. Directors Series on Proliferation 4: 47-54. 231. Signoretto, C., G. Bulacchini, C. Pruzzo, and P. Canepari. 2005. Persistance of Enterococcus faecalis in aquatic environments vi a surface interaction with copepods. Appl. Environ. Microbiol. 71: 2756-2761. 232. Signoretto, C., G. Bulacchini, M. del Mar Lleo, C. Pruzzo, M. Zampini, L. Pane, G. Franzini, and P. Canepari. 2004. Adehsion of Enterococcus faecalis in the nonculturable state to plankton is the main mechanism responsible for persistance of this bacterium in both la ke and seawater. Appl. Environ. Microbiol. 70: 6892-6896.
178 233. Signoretto, C., del Mar, L., Tafi, M.C., and P. Canepari. 2000. Cell wall chemical composition of Enterococcus faecalis in the viable but nonculturable state. Appl. Environ. Microbiol. 66: 1953-1959. 234. Simmons, O. D., M.D. Sobsey, C.D. Heaney, F.W. Schaefer, and D.S. Francy. 2001. Concentration and detection of Cryp tosporidium oocysts in surface water samples by method 1622 using ultrafiltr ation and capsule filtration. Appl. Environ. Microbiol. 67: 1123-1127. 235. Simpson, J. M., and D.V. Lim. 2005. Rapid PCR c onfirmation of E.coli O157:H7 after evanescent wave fiber optic biosensor detection. Biosens. Bioelectron. 21: 881-887. 236. Singh, B. R., S.H. Harrison, and J.S. Schoeniger. 2000. Gangliosides as receptors for biological toxins: deve lopment of sensitive fluoroimmunoassays using ganglioside-bearing liposomes. Anal. Chem. 72: 6019-6024. 237. Singh, B. R., and M. A. Silvia. 1996. Detection of botulinum neurotoxins using optical fiber-based biosensor. Adv. Exp. Med. Biol. 391: 499-508. 238. Singh, D. V., M.H. Matte, G.R. Matte, S. Jiang, F. Sabeena, B.N. Shukla, S.C. Sanyal, A. Huq, and R.R. Colwell. 2001. Molecular analysis of V. cholerae O1, O139, non-O1, and non-O139 strains: clonal relationships between clinical and environmental isolates. Appl. Environ. Microbiol. 67: 910-921. 239. Singh, K. V., T.M. Coque, G.M. Weinstock, and B.E. Murray. 1998. In vivo testing of an Enterococcus faecalis efaA mutant and use of efaA homologs for species identification. FEMS Immunol. Med. Microbiol. 21: 323-331. 240. Skladal, P., C. dos Santos Riccardi, H. Yamanaka, and P.I. da Costa. 2004. Piezoelectric biosensors for real-time monitoring of hybridization and detection of hepatitis C virus. J. Virol. Methods 117: 145-151. 241. Skretas, G., and D. W. Wood. 2005. A bacterial biosensor of endocrine modulators. J. Mol. Biol 349: 464-474. 242. Smith, T. 1891. A new method for determinin g quantitatively the pollution of water by fecal bacteria. 13th Annual Re port of the New York State Board of Health. Albany, NY. 243. Snow, J. 1954. On cholera, L. Clendening, e d. ed. Paul B, Hoeber, Inc., New York. 244. Sood, S., B.K. Das, A. Kapil, and A. Rattan. 1996. TCBS for the isolation of Vibrio cholerae. Trop. Gastroenterol. 17: 177.
179 245. Stender, H., K. Oliveira, S. Rigby, F. Bargoot, and J. Coull. 2001. Rapid detection, identificati on, and enumeration of Escherichia coli by fluorescence in situ hybridization using an array scanner. J. Microbiol. Methods 45: 31-39. 246. Stender, H., M. Fiandaca, J.J.Hylig-Nielsen, and J. Coull. 2002. PNA for rapid microbiology. J. Microbiol. Met. 48: 1-17. 247. Straub, T., and D. P. Chandler. 2003. Towards a unified system for detecting waterborne pathogens. J. Microbiol. Met. 53: 185-197. 248. Su, X. L., and Y. Li. 2004. A self-assembled monol ayer-based piezoelectric immunosensor for rapid detection of Escherichia coli O157:H7. Biosens. Bioelectron. 15: 563-574. 249. Taitt, C. R., J.P. Golden, Y.S. Shubin, L.C. Shriver-Lake, K.E. Sapsford, A. Rasooly, and F.S. Ligler. 2004. A portable array biosensor for detecting multiple analytes in complex samples. Microb. Ecol. 47: 175-185. 250. Taylor, R. K., V.L.Miller, D. B. Furlong, and J.J. Mekalanos. 1987. Use of phoA gene fusionsto identify a pilus colonization factor coordinately regulated with cholera toxin. Proc Natl. Acad. Sci. USA 84: 2833-2837. 251. Tempelman, L. A., K.D. King, G.P. Anderson, and F.S.Ligler. 1996. Quantitating staphylococcal enterotoxin B in diverse media using a portable fiberoptic biosensor. Anal. Biochem. 233: 50-57. 252. Theron, J., and T.E. Cloete. 2002. Emerging water borne infections: contributing factors, agents, and de tection tools. Crit. Rev. Microbiol 28: 1-26. 253. Thiercelin, M. E., and L. Johaud. 1903. Reproduction de l'enterocoque;taches sentrales; graduations periferiques et microblastes. CRS Soc. Biol. Fil. 55: 686688. 254. Tilden, J. J., W. Young, A.M. McNama ra, C. Custer, B. Boesel, M.A. Lambert-Fair, J. Majkowski, D. Vugia, S.B. Werner, J. Hollingsworth, and J.G. Morris Jr. 1996. A new route of transmission for Escherichia coli : infection from dry fermented salami. Am. J. Public Health 86: 1142-1145. 255. Timm, E. A., and C. Stewart. 1992. Fluorecent in situ hybridization en suspension (FISHES) using digoxigeninlabeled probes and flow cytometry. Bio.Techniques 12: 362-367. 256. Tims, T. B., and D. V. Lim. 2004. Rapid detection of Bacillus anthracis spores directly from powders with an evanes cent wave fiber optic biosensor. J. Microbiol. Meth. 59: 127-130.
180 257. Tomoyasu, T. 1998. Improvement of the immuno magnetic sepa ration method selective for Escherichia coli O157:H7 strains. Appl. Environ. Microbiol. 64: 376-382. 258. Toranzos, G. A., G.A. McFeters, and J.J. Borrego. 2002. Detection of microorganisms in environmental freshwa ters and drinking waters, p. 205-219. In R. L. C. C.J. Hurst, G.R. Knudsen, M.J. McInerney, L.D. Stetzenbach (ed.), Manual of Environmental Microbiolog y. ASM Press, Washington, D.C. 259. Turbow, D. J., N.D. Osgood, and S.C. Jiang. 2003. Evaluation of recreational health risk in coastal waters based on enterococcus densities and bathing patterns. Environ. Med. 111: 598-603. 260. Turner, A. P. F. 2000. Biosensors---sense and sensitivity. Science. 290: 13151317. 261. U.S. Environmental Protection Agency. 1986. Ambient water qualtiy criteria for bacteria--1986. U.S. Environmental Protection Agency publication no. EPA 440-5-84-002. Office of Research and Development, U.S. EPA. Washington, D.C. 262. U.S. Environmental Protection Agency. 1999. EPA action plan for beaches and recreational waters. U.S. Environmen tal Protection Agency publication no. EPA/600/R-998/079. Office of Research and Development, U.S. EPA. Washington, D.C. 263. U.S. Environmental Protection Agency. 2005. Method 1600: Membrane filter test method for enterococci in water. U.S. Environmental Protection Agency, Washington D.C. 264. Vedrine, C. 2003. Optical whole-cell biosensor using Chlorella vulgaris designed for monitoring herbicides. Biosens. Bioelectron. 18: 457-463. 265. Wade, T. J., N. Pai, J.N.S. Eisenberg, and J.M. Colford. 2003. Do U.S. Environmental Protection Agency water qu ality guidelines for recreational waters prevent gastrointestinal illnes? Environ. Health. Persp. 111: 1102-1109. 266. Wadkins, R. M., J.P. Golden, and F.S. Ligler. 1995. Calibration of biosensor response using simultaneous evanescent wave excitation of cyanine-labeled capture antibodies and antigens. Anal. Biochem. 232: 73-78. 267. Wai, S. N., Y. Mizunoe, and S. Yoshida. 1999. How Vibrio cholerae survive during starvation. FEMS Microbiol. Lett. 180: 123-131. 268. Wang, J. 1999. Electroanalysis and biosensors. Anal. Chem 71: 329-332.
181 269. Watanabe, A., K. Yano, K. Ikebukuro, and I. Karube. 1998. Investigation of the potential active site of a cyanide dihydratase using site-directed mutagenesis. Bichim. Biophys. Acta. 1382: 1-4. 270. Watson, B. K., R. C. Moelleing, and L.J. Kunz. 1976. Effects of penicillin and lysozymes on the immunofluorescent a nd precipitin reactivity of group D streptococci. Am. J. Clin. Pathol. 66: 73-79. 271. Weissman, J. B., W.E. DeWitt, J. Thomps on, C.N. Muchnick, B.L. Portnoy, J.C. Feeley, and E.J. Gangarosa. 1975. A case of cholera in Texas, 1973. Am. J. Epidemiol. 100: 487-498. 272. Wicken, A. J., S.D. Elliott, and J. Baddiley. 1963. The identity of Streptococcal group D antigen with teichoic acid. J. Gen. Microbiol. 31: 231-239. 273. Wicken, A. J., and K. W. Knox. 1975. Lipoteichoic acids: A new class of bacterial antigen. Science 187: 1161-1167. 274. Willats, W. G. T. 2002. Phage display: practicalities and prospects. Plant Mol. Biol. 50: 837-854. 275. Wilson, A. 1971. The spread of cholera to and within Nigeria 1970-71. J Clin Pathol 24: 768. 276. World Health Organization. 2005. Cholera in West Africa. Wkly Epidemiol Rec. 80: 297-298. 277. Wu, A. M., J.H. Wu, W.M. Watkins, C.P. Chen, S.C. Song, and Y.Y. Chen. 1998. Differential binding of human bl ood group Sd (a+) abd Sd (a-) TammHorsfall glycoproteins with Dolichos biflorus and Vicia villosa -B4 agglutinins. FEBS Lett. 429: 323-326. 278. Xiao, J., M. Hook, G.M. Weinstock, B.E. Murray. 1998. Conditional adherence of Enterococcus faecalis to extracellular matrix pr oteins. FEMS Immunol. Med. Microbiol. 21: 287-295. 279. Yalow, R., and S. A. Berson. 1959. Assay of plasma insulin in human subjects by immunological methods. Nature 184: 1684-1649. 280. Yang, V. C., and T. T. Ngo. 2000. p. ix-x. In V. C. Yang (ed.), Biosensors and Their Applications. Kluwer Acade mic/Plenum Publishers, New York. 281. Yoshpe-Purer, Y., and S. Golderman. 1987. Occurrence of Staphylococcus aureus and Pseudomonas aeruginosa in Israeli coastal water. Appl. Environ. Microbiol. 53: 1138-1141.
182 282. Yu, H., and J. G. Bruno. 1996. Immunomagnetic electrochemiluminescent detection of Escherichia coli O157:H7 and Salmonella typhimurium in foods and environmental water samples. Appl. Environ. Microbiol. 62: 587-592. 283. Zmirou, D. 2003. Risks assoicated with the microbiological quality of bodies of fresh and marine water used for recreational purposes: summary estimates based on published epidemiological st udies. Arch. Environ. Health 58: 703-711.
ABOUT THE AUTHOR Theresa Trindade received a Bachelors Degr ee in Biology with a minor in Chemistry from Florida Southern College in 1999. She en tered the Ph.D. program in Biology at the University of South Florida in 2000. While in the Ph.D. program at the University of South Florida, Ms. Trindade was active in the American Society for Microbiology (ASM). She made several poster and oral pr esentations at various regional and national ASM meetings and was awarded travel grants to attend two of those meetings.
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001788257
007 cr mnu|||uuuuu
008 070521s2005 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001443
Trindade, Maria Theresa.
Detection of pathogenic bacteria and fecal enterococci in recreational water with an evanescent wave fiber optic biosensor
h [electronic resource] /
by Maria Theresa Trindade.
[Tampa, Fla] :
b University of South Florida,
ABSTRACT: Development of a rapid method for the detection of fecal enterococci and pathogenic microorganisms in beach water was attempted utilizing an evanescent wave fiber optic biosensor. Various assay formats including a sandwich immunoassay were tested in the development of a rapid assay. Fluorophore labeled antibodies were used for specific detection of bacteria captured or adsorbed directly to the surface of a polystyrene fiber optic waveguide. Binding of the fluorescent labeled antibody to its specific target or binding of a fluorescent labeled anti-IgG within 100-1000 nm of the waveguide surface caused excitation of the fluorescent conjugate resulting in a quantifiable signal. Enterococcus faecalis, Staphylococcus aureus, Escherichia coli O157:H7, and Vibrio cholerae were used as model organisms for biosensor detection in phosphate buffered saline and seawater. Seawater samples were selectively enriched for the presence of these model organisms, which were later dete cted on the biosensor. The sensitivity and specificity of the biosensor was examined by testing various assay formats, sample preparations, and molecules for capture and detection. Finally, an enrichment protocol combined with filter concentration was utilized to enhance detection of low levels of enterococci. The fiber optic biosensor has the potential to be a sensitive and specific system for the detection of fecal enterococci. The lower limit of detection in seawater and phosphate buffered saline was 2.8 x 106 CFU/ml. As few as 6 CFU/100ml (0.06CFU/ml) could be detected in seawater following a 14-24 hour enrichment and concentration step. Vibrio alginolyticus was found to grow under the same enrichment conditions as the enterococci. V. alginolyticus crossreacted with the polyclonal anti-Strep group D antibody used in the immunoassay at high cell concentrations. Staphylococcus aureus was the only other organism which showed significant cross-reactivity with this antibody.^^^^^ The biosensor was also able to detect other bacterial pathogens in PBS and seawater. The lower limit for detection of E. coli O157:H7 was 3.6 x 105 CFU/ml. The lower limit for detection of Vibrio cholerae O1 was 1.3 x 108 CFU/ml. The antibodies used in these assays were found to crossreact with other gram negative microorganisms. The biosensor was not able to detect Staphylococcus aureus.
Dissertation (Ph.D.)--University of South Florida, 2006.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 182 pages.
Adviser: Daniel V. Lim, Ph.D.
t USF Electronic Theses and Dissertations.