Phage Display Single-Chain Variable Fragments for Detection of Listeria monocytogenes

Material Information

Phage Display Single-Chain Variable Fragments for Detection of Listeria monocytogenes
MESSER JR, HARALD G.P. ( Author, Primary )
Copyright Date:


Subjects / Keywords:
Antibodies ( jstor )
Antigens ( jstor )
Bacteriophages ( jstor )
Enzyme linked immunosorbent assay ( jstor )
Flagella ( jstor )
Listeria ( jstor )
Listeria monocytogenes ( jstor )
Reactivity ( jstor )
Reagents ( jstor )
Signals ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Harald G.P. Messer, Jr. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:


This item is only available as the following downloads:

Full Text




Copyright 2005 by Harald G.P. Messer, Jr.


This document is dedicated to my parents.


iv ACKNOWLEDGMENTS It is my sincerest wish to thank first a nd foremost my family, Renee and Kathy, for their unending belief in me. I especially want to thank my father, Harald Sr. for his constant encouragement and my mother, Lourdes, for her love and to whom I dedicate my work in life. I want to thank my mentor, Paul Gulig, for challenging and nurturing my growth as a scientist. He has been a great source of help and unders tanding. I would also like to give special thanks to my committee memb ers, Donna Duckworth and Jeannine Brady, for their insight, encouragement, and motivati on. I want to thank my co-workers in the Gulig lab for their support and assistance . Finally, I want to give my most heartfelt thanks to Scott, for never giving up on me.


v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT....................................................................................................................... ..x CHAPTER 1 INTRODUCTION........................................................................................................1 Listeria monocytogenes ................................................................................................1 Antigenic Structure of Listeria.....................................................................................5 Bioterrorism..................................................................................................................9 Public Food Safety......................................................................................................10 Methods for Bacteriological Detection and Analysis.................................................12 The Biosensor System................................................................................................14 Monoclonal Antibody Production..............................................................................15 Recombinant Phage Display.......................................................................................19 M13 Phage Biology.............................................................................................20 Tomlinson I + J Human Synthetic VH + VL Phagemid Libraries........................23 2 MATERIALS AND METHODS...............................................................................26 Bacterial Strains, Phage Stra ins, and Growth Conditions..........................................26 Enzyme Linked ImmunoSorbent Assays....................................................................28 BioPanning of Phage Display Libraries.....................................................................30 Panning on Immunotubes....................................................................................30 Panning in Suspension.........................................................................................31 Titering the Phage................................................................................................32 Amplification of the Selected Phage...................................................................32 Panning without Amplification...........................................................................33 Production of Soluble Antibody Fr agments (scFv antibodies)...........................33 DNA Manipulations....................................................................................................34 Plasmid Extractions.............................................................................................34 Genomic Extraction.............................................................................................34 Restriction Enzyme Manipulations.....................................................................35


vi Agarose Gel Electrophoresis...............................................................................35 Polymerase Chain Reaction (PCR).....................................................................35 Construction of lm-Auto-pET19b Plasmid Expression Vector (pGTR1499).....36 Construction of scFv-Avitag Plasmid Vectors....................................................37 Protein and Flagella Manipulations............................................................................38 Extraction of Flagella..........................................................................................38 Recombinant Protein Expre ssion of Auto Protein...............................................39 Bacterial Cell Lysis Extraction of Recombinant Auto Protein...........................40 Nickel-column Affinity Purifi cation of His-tagged Proteins..............................41 Chemical Biotinylation of scFv...........................................................................41 Determination of Protein Concentration.............................................................42 Sodium Dodecyl Sulfate-Polyacrylam ide Gel Electrophoresis (SDS-PAGE)....42 Coomassie Blue Staining for Proteins.................................................................43 Tsai-Frasch Silver Staining.................................................................................43 Immunoblotting...................................................................................................44 3 RESULTS...................................................................................................................45 Rationale for Study.....................................................................................................45 Specific Aim 1: Isolation of Phage Display Antibody Fragments to Listeria ............45 scFv-Phage Antibodies to Whole L. monocytogenes Cells Demonstrate Variable Reactivities....................................................................................46 scFv-Phage Antibodies to L. monocytogenes Flagella Recognize Both Glycosylated and Ungl ycosylated Epitopes.................................................62 Recombinant lmo1076 (Auto) from L. monocytogenes is Suitable as Antigen for Phage Display Panning.............................................................76 scFv-Phage Antibodies to rAuto Discriminate L. monocytogenes from L. innocua .........................................................................................................81 Conclusion of Specific Aim 1......................................................................83 Specific Aim 2: Optimization of Immunological Reagents.......................................86 scFv Antibodies Can be Genetically Fused to a Biotin-receiving Peptide (Avitag)..87 Improved Detection of L. monocytogenes Cells Through Chemical Biotinylation of scFv.....................................................................................93 Anti-Listeria Auto Streptabody Has Improved Detection of rAuto.............96 4 DISCUSSION...........................................................................................................102 Specific Aim 1: Isolation of Phage Display Antibody Fragments to Listeria ...103 Specific Aim 2: Optimization of Immunological Reagents..............................123 LIST OF REFERENCES.................................................................................................131 BIOGRAPHICAL SKETCH...........................................................................................144


vii LIST OF TABLES Table page 2-1 Bacterial strains used................................................................................................27 3-1 Detection of L. monocytogenes serovar 1/2a with whole cell phages by ELISA following periodate and protei nase treatment of cells..............................................56 3-2 Detection of listeria fl agella and whole cells by ELISA using Avitag scFvs..........90 3-3 Detection of listeria flagella and cells by ELISA without primary antibody...........91 3-4 Detection of listeria flagella and cells by ELISA without primary antibody...........92 3-5 Detection of listeria flagella by ELI SA using chemically-biotinylated antiflagella scFv.............................................................................................................95


viii LIST OF FIGURES Figure page 2-1 Vector map of pIT2 phagemid vector. The vector map of pIT2 phagemid vector used in constructing th e Tomlinson libraries...........................................................38 2-2 pET-19b cloning/expression region.........................................................................40 3-1 ELISA analysis of phages isolat ed following whole cell pannings to L. monocytogenes serovar 1/2a cells in suspension. ....................................................48 3-2 ELISA cross-reactivity of five whole cell Listeria phage clones. ..........................51 3-3 Competition effect of mixing phages and cells in an inhibition ELISA. ...............53 3-4 Comparison of phage and scFv reactivities to L. monocytogenes cells by ELISA. ............................................................................................................................... ...59 3-5 Expression and detection of clone 1-3 and BSA scFv in E. coli HB2151 supernatant and cell fractions. ................................................................................61 3-6 Electron micrograph of ura nyl acetate negative stain of L. monocytogenes serotype 1/2a flagella preparation. .........................................................................63 3-7 Analysis of L. monocytogenes serovar 1/2a flagella by SDS-PAGE. ....................64 3-8 ELISA of anti-Listeria flagella phages following panning on immunotubes. .......65 3-9 Comparison of ELISA signal to noise rati os of anti-listeria flagella phages to flagellated and non-flagellated Listeria . .................................................................67 3-10 Reactivity of anti-listeria flagella scFv phages to flagella or whole L. monocytogenes serovar 1/2a cells by Western blot. ..........................................68 3-11 SDS-PAGE and Western blot analysis of listeria flagel la, whole flagellated cells, and non-flagelleted cells with anti-f lagella phages..................................................70 3-12 SDS-PAGE and Western Blot analysis of periodate-treat ed listeria flagella using anti-listeria phages. .................................................................................................71 3-13 Anti-flagella scFv reactivities to listeria flagella by ELISA. .................................73


ix 3-14 Effect of concentrating an ti-flagella clone 4 scFv for the detection of flagella by ELISA. ....................................................................................................................74 3-15 Effect of concentrati ng anti-flagella clone 4 sc Fv for the detection of L. monocytogenes serovar 1/2a cells by ELISA. ........................................................75 3-16 SDS-PAGE and Western Blot analysis of recombinant Auto protein, postaffinity column purification.....................................................................................78 3-17 SDS-PAGE analysis of CellLytic extracted recombinant Auto protein from E. coli DE3 Tuner cells. ..............................................................................................79 3-18 SDS-PAGE analysis of nickel-affi nity-column purified recombinant Auto protein following Cell Lytic extraction. .............................................................80 3-19 ELISA of anti-Auto phages following panning to recombinant Auto and whole cells. ....................................................................................................................... .82 3-20 ELISA of anti-Auto scFv to recombinant Auto and L. monocytogenes EGDe cells. ....................................................................................................................... .84 3-21 Western blot analysis of anti-Auto scFv clones to both recombinant Auto and L. monocytogenes EGDe cells. ...................................................................................85 3-22 SDS-PAGE and Western blot analysis of biotinylated Avitag-scFvs. ...................88 3-23 Detection of recombinant Auto prot ein using using anti-Auto phages, scFv, biotinylated scFv (scFv-B), and st reptabody (stAb) via an ELISA. ......................98 3-24 Detection of L. monocytogenes EGDe and L. innocua cells via ELISA using phages, scFvs, scFv-B, and stAbs. .......................................................................100


x Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science PHAGE DISPLAY SINGLE-CHAIN VARI ABLE FRAGMENTS FOR DETECTION OF Listeria monocytogenes By Harald G.P. Messer, Jr. December 2005 Chair: Paul A. Gulig Major Department: Molecular Genetics and Microbiology Listeria monocytogenes is a psychrophilic gram-positive bacterium responsible for over 2,500 cases of listeriosis re sulting in 500 deaths annually in the United States. Immunocompromised individuals and pregnant women are at highest risk of infection. This organism is well adapted to both cold temperatures and high salt environments; Listeria may therefore be transmitted in ready-toeat foods that have been kept properly refrigerated. In addition to the health impact , sporadic outbreaks of listeriosis can cause great economic losses to the food industry, espe cially the dairy and meat industries. Therefore, detection of Listeria as well as other pathogens is important. The quality of an immunological-based detection system, esp ecially a fiber-optic biosensor system, ultimately relies on the quality of the captu re and detection antibodies. Recombinant antibody phage display libraries offer a pow erful, economical, and rapid method of screening libraries of large complexities of recombinant antibody molecules, on the order of 109 variants.


xi We used the Tomlinson I Human Syntheti c scFv phagemid libraries to isolate antibodies that recognize surface epitopes of L. monocytogenes . Panning was done under various conditions using whole bacterial cells , flagella, or a recombinant Auto surface protein as antigens. Initial attempts at is olating scFv-expressing pha ge to whole bacterial cells using a suspension technique did not yield useful reagents. Unfortunately, antibodies were not specific for L. monocytogenes , but reacted with other unrelated bacterial species, as well. However, pa nning experiments to both flagella and Auto yielded several promising scFv-encoding pha ges that recognized th ese antigens by both ELISA and Western Blot. Anti-flagella phages recognized listeria flagellin, both purified and on the surface of whole L. monocytogenes cells. Initial results also demonstrated that the reactivity of some scFv phages may have been specific to glycosidic residues on the listeria flagellin protein because they reac ted with nonflagellated strains. This was confirmed by periodate-treating flagella and losing reactivity of some phages. Phages that recognized a recombinant form of Auto were unreactive to nonpathogenic Listeria innocua and reacted with whole cells of L. monocytogenes . These phages demonstrated specificity to all serovars of L. monocytogenes , excluding serovar 4b, as expected. We used anti-flagella and anti-Auto phagemids to obtain soluble antibody molecules (scFv) instead of fusion proteins displayed on the phage surface. The scFvs were genetically and chemically modified with biotin tags to increase their usefulness in the biosensor system. Initial attempts we re also made to form streptabodies by tetramerizing the biotinylated scFvs to streptavidin. These studies will aid in the continuing development of immunological tools for use with the real time fiber optic biosensor system as well as other detection systems for L. monocytogenes .


1 CHAPTER 1 INTRODUCTION Listeria monocytogenes In 1924, E.G.D Murray isolated gram-positive rods from the blood of rabbits, and not recognizing the organism for clear classification, he named it Bacterium monocytogenes (1,2). Later in 1940, J.H.H. Pirie renamed the organism to Listeria , recognizing it as a catalase-positive, gram-pos itive rod that was being isolated not only from rabbits, but also from humans, the environm ent, other animals, and food (3). Still, general awareness of the potential pathogeni city of this organism was not recognized until years later when an epidemic of 85 cases of infant infections, at the time called “granulomatosis infantiseptica”, occurred in Germany in 1949 (4) . Histological examination of both the livers and brains of these patients mistakenly led to being identified as Corynebacterium . A scientist from the Univer sity of Bonn, H.P.R. Seeliger, examining samples from skin lesions, correctly identified the infections as Listeria and not Corynebacterium , based on the presence of flagellar motility. It has been Seeliger who has contributed the most significant amount of information through his constant and consistent research on Listeria . His publications brought Listeria into recognition as a pathogen (5). L. monocytogenes is a ubiquitous, gram-positive, psychrophilic, environmental bacterium that has adapted to relative extr emes in living conditions, existing comfortably in high salt and cold temperatures, not only in soil and animals, but also on kitchen and food-processing machinery. For decades following its discovery, Listeria was considered


2 a single organism. Using specific antibodies to lipoteichoic acids found on the outer cell wall and to H-antigens of flagella, Listeria has been since characterized into seven serovars. A Frenchman, Rocourt, was the firs t to separate the se rovars by biochemical and genetic comparisons (6). All L. monocytogenes are pathogenic with the majority of human isolates in serovars 1 and 4. Other species including Listeria innocua , Listeria seeligeri , Listeria welshimeri , and Listeria ivanovii are nearly non-pathogenic (7). Suprisingly, bona fide acceptance of L. monocytogenes as a serious potential pathogen acquired through the consumption of contamin ated food did not arise until outbreaks in the 1980’s in Canada, the United Stat es, and Switzerland, which connected Listeria to the consumption of coleslaw and cheese (8). Listeriosis is the disease caused by L. monocytogenes , and is acquired by the ingestion of contaminated foods. Between 1980 and 2000, fifteen food borne outbreaks of listeriosis were reported to the Centers for Disease Control and Prevention (CDC; Atlanta, GA); ten of these outbreaks have occurred since 1998, and three of these involved eleven states. Raw unpastuerized milk has been a common source of contamination and in 2000 led to an outbreak in North Carolina i nvolving contaminated Mexican soft cheese (9,10). This particular outbreak resulted in five still births, three premature deliveries, and two infected ne wborns (CDC, 2001). Although unpasteurized dairy products are a common source, ready-to-e at meats can also serve as sources. In 1998, 40 listeriosis cases were linked to contamin ated hotdogs and resulted in four deaths (CDC, 1998). Since 1996, the FDA has ma intained a zero-to lerance policy for L. monocytogenes in ready-to-eat foods. Most, if no t all, contaminations have been attributable to just three se rotypes: 1/2a, 1/2b, and 4b.


3 Through an active surveillance in 1997 by Food Net and the CDC, there were a reported 2,500 cases of listeriosis annually in this country and a reported 500 deaths (11,12). A high mortality rate is in part due to immunocompr omised patients, who are in fact some of the most at risk for listerio sis, succumbing to inf ection. Clinically its diagnosis can be frequently overlooked as a possible cause of illness due to its unique growth capabilities. Although Listeria can be confused with common less harmful bacterial contaminants and disregarded, an im portant delineation exists regarding growth conditions. While most bact eria grow poorly at lower temperatures (below 25C), Listeria survives in temperatures from below freezing (0C) to body temperature (37C) with robust growth at low temperatures, includi ng the temperature range that is normally used for refrigeration (2-8 C). As a result, Listeria may be transmitted in ready-to-eat foods that have been kept properly refrigerat ed. This is cause of great concern and as such the impact of antimicrobial ingredie nts on ready-to-eat fo ods has been studied extensively in the food-industr y (13,14). The ability of Listeria to grow in such diverse environments is just one of the many challe nges presented by this dangerous bacterium. Certain groups of individuals are at greater risk for listeri osis. These are recipients of organ transplants who are placed on i mmunosuppressive drugs, pregnant women (and their unborn children), and immunocompromised persons. Persons with AIDS suffer listeriosis 65 to 145 times more frequently than the general popul ation, and persons who take glucocorticosteroid medi cations (also called cortisone ) that depre ss the immune system are also at increased risk. The elde rly and certain debilitated patients (such as those on dialysis or alcoholics) are at minor in creased risk for listeriosis. In infants, listeriosis occurs when the infection is tr ansmitted from the mother, either through the


4 placenta or during the birthing process. Ther efore there are severa l host factors, along with the amount of bacteria ingested and the virulence of the strai n, which determine the risk of disease. Although the infectious dose is not known, ingestion of L. monocytogenes bacteria can result in illness. There is normally a period of incubation that follows ingestion of bacterial cells that can last from 3 to 70 days, but usually lasts 4 to 21 days. This variable incubation time also complicates self-diagnosis. Five days to three weeks after ingestion, Listeria can become systemic and able to spre ad to all body areas including the central nervous system, heart, eyes, or other locat ions depending on host predisposed health. Fetuses of pregnant women are particularly vulnerable to Listeria (15,16). Symptoms of listeriosis are usually by themselves general: fever, muscle aches, and gastrointestinal symptoms such as nausea or diarrhea. If infection spreads to the nervous system, symptoms include headache, stiff neck, loss of balance, confusion, obtundation or convulsions. With brain involvement, listeriosis may mimic a stroke. Infected pregnant women experience only a mild, flu-like illnes s; however, infecti on during pregnancy can lead to miscarriage, infection of the new born, or even stillbirth. The perinatal and neonatal mortality rate is 80%. Human cases of listeriosis are, for the most part, sporadic and treatable. Nonetheless, listeriosis remains an important threat to public hea lth, especially among those most susceptible to this disease. With the increase of the numbers of immunocompromised people, the risk increases. The fact that liste riosis is a disease easily transmitted from mother to fetus th rough the placenta is worrisome, especially since pregnant women themselves would not necessarily show obvious signs of infection.


5 For unknown reasons, in i mmune-deficient hosts, Listeria invades and grows best in the central nervous system, especially in the brain parenchyma causing encephalitis or brain stem causing meningitis (17-19). Studies ut ilizing sheep diagnosed with listerial encephalitis demonstrated the trigeminal ga nglion and axons as routes of entry and infection (18,20). In pregnant women, the fe tus is most heavily infected, leading to spontaneous abortion, stillbirths, or sepsis in infancy most likely due to the lowered immune states of both mother and fetus during pregnancy. These statis tics indicate true misfortunes, as listeriosis is a preventable condition. A positive diagnosis of listeriosis is confirmed by sampling patient blood, cerebra l spinal fluid (CSF), or focal lesions yielding -hemolytic, gram-positive rods on blood agar plates. Treatment is normally with the antibiotic gentamicin combined with ampicillin, penicillin G, or trimethoprim / sulfamethoxazole. Antigenic Structure of Listeria The pathogenicity of Listeria combines both common and unique features of an intracellular pathogen that targets macrophage s (21-23), as well as showing a high degree of tropism for non-phagocytic liver hepatocy tes due to an uncommon bacterial-induced phagocytosis (16,24). Antibodies have little role in protective immunity, and acquired immunity is generally regarded as being dependent on cytotoxic CD8+ T-cells (25,26). Nave peritoneal resident macrophages and polymorphonuclear cells of mice have been shown to kill intracellular L. monocytogenes in hepatocytes (24,27) , and therefore these cytotoxic macrophages may also play a role in early defense ag ainst intracellular infections (28,29). Listeria can invade the body through a norm al intact gastrointestinal tract and once in the body can travel through the bloodstr eam exposing the brain or placenta to infection; however the bacteria are most often found inside cells (28,30).


6 Several interactions must occur between surface proteins on the bacterium and with the host cell for infection to occur (31). In the case of L. monocytogenes , there have been many surface proteins identifie d with critical roles in pathogenicity (31,32) including binding, entry, and persistence within host cells. There ha ve been many studies on the biochemistry and homology of gram-positive bacterial proteins that underscore their importance between Listeria (33) and other gram-positives such as highly studied B subunit of the RNA polymerase holoenzyme. The B -subunit modulates stress response in the food borne pathogens Bacillus cereus and Stapyhlococus aureus , and is responsible for conferring resistance to high osmolar ity, high and low pH, oxidizing agents, and biofilm formation. This pr otein is also present in Listeria, having a role in its survival at low temperatures (34,35). L. monocytogenes is a non-encapsulated rod and is phenotypically identical to other gram-positive bacteria, containing a cytosol, a single cytoplasmic membrane, and a thick multilayered outer wall of peptidoglycan cove red with various surface proteins and two groups of anionic carbohydrate polymers: (i) te ichoic acids anchored covalently into the peptidoglycan and (ii) lipoteichoic ac ids (LTA) which are polyphosphoglycerol substituted with a D-alanyl (D-Ala) ester or a glycosyl residue and anchored in the membrane by their glycolipid moiety (36). Cell wall teichoic acids are uniformly distributed over the entire peptidoglycan outer wall. Although the function and biosynthesis of LTA is not entirely underst ood, it has been proposed that anionic teichoic acids function to capture divalent cationic s ubstances, serve as binding sites for enzymes that cleave the peptidoglycan such as LytA amidase (33), or to prevent diffusion of substances by providing a physic al barrier. Many murein hy drolases such as Auto and


7 other surface proteins such as internalin B (InlB) are believed to be closely associated with LTA and are bound to the bacterial surface by interaction of their tail G-W modules (37-39). The teichoic acids found on Listeria and other gram-positive organisms may indeed be species-specific. Complete genome sequencing of L. monocytogenes indicates that 133 of its 2,853 genes encode surface proteins (37) . The main differences between L. monocytogenes and L. innocua are, in fact, the make-up of their su rface proteins. Uptake by one or more nonprofessional phagocytic cells is mediated by the co-ordination of a cadre of bacterial surface proteins, namely proteins of the internalin family and possibly murein hydrolases. Not surprisingly, transcriptional activation and regulation of many listeria virulence factors such as internalin families are coordi nately regulated by a transcriptional activator protein, PrfA (40). Interestingly, Listeria cells can initiate phagosomal fusion through cell surface proteins, and cell line invasion studies demonstrate that both murein hydrolases and internalin family of proteins are sufficient for intern alization into cells (41,42). To mediate internalization, the majo r virulence factor Internalin A (InlA) is required to bind to the host cell receptor, Ecadherin (40,43), and InlB is required to bind to a Met receptor tyrosine-kinase thus act ing via a PI3-kinase (39,44). Internalized bacteria are compartmentalized in a vacuole and eventually escape into the cytosol. Studies on intracellular survival of Listeria have attributed a single surface protein, the hly -encoded pore-forming listeria-lysin O (L LO), for escape from the vacuole (4547). Homologous related lysins have been id entified in other gram-positive organisms including streptolysin, perfringol ysin, and pneumolysin produced by Streptococcus pyogenes , Clostridium perfringens , and Streptococcus pneumoniae, respectively. The


8 mechanism for escape from the vacuole is not entirely clear, but invol ves acidification of the vacuole which then signals transcripti onal activation of LLO (21,46) along with two other listeria secretable phospholipases C (PLCs). LLO and PLC puncture the vacuole membrane from within, dissipate the internal pH gradient, and halt the maturation of the phagosome (48,49). This subsequently and ac tively forms a route of escape through the puncture. LLO-expressing Bacillus subtilis and Escherichia coli are capable of phagosomal escape in cell culture (50). Th e bacterial phosholipases and possibly host vacuolar components then channel through the phagosome and lead to its dissolution. Surprisingly, there is apparent host cell re pair of lesions, and the host cell survives; however necrosis is common (28). Once in the mammalian cytosol, Listeria induces an unusual pol ymerization of host actin filaments and uses the force generated to move first within the cytoplasm ultimately to spread from cell to cell in what is termed a listeriopod. This phenomenon is attributable to a single prot ein, ActA, which is also regul ated by PrfA (40,51). The Nterminus of ActA polymerizes actin filaments, utilizing the activity of the host cell Arp2/3 complex and forming what microsc opically appears to be a cloud around the bacterium (52,53). This binding mechanism mi mics cellular Wisco tt–Aldrich syndrome proteins (WASP) (54,55) by recruiting hos t-cell enabled vas odilator-stimulated phosphoprotein (Ena/VASP). This mechanis m is normally important in cell motility, platelet shape, axon-guidance, and T-cell activation (56). Homologous proline-rich regions are found on both ActA and Ena/VASP proteins and attract cellular profilin through binding to Ena/VASP homology domai n 1 (EVH1) (56-58). In summary, ActA mimics WASP-family proteins and recrui ts the Arp2/3 complex to induce actin


9 nucleation and filament elongation and branch ing. By spreading from cell to cell, Listeria avoids host cell immune responses (e.g., cyto toxic T-cells and phagocytes) (59). Listeria can also move by means of flagel la-based motility in extracellular environments. Flagella expression in L. monocytogenes is temperature-dependent, and transcription of flaA , the gene encoding for the major fl agellin protein, is downregulated at physiological temperature (37 C) and above (59-61). Tran scriptional regulation is controlled through the motility gene re pressor, MogR, which binds to the flaA promoter region and is required for full virulence ( 62). The 29-kDa flagellin protein, FlaA, activates TOLL-like receptor 5 in mammalian i mmune systems (63). Therefore the lack of flagellated Listeria in host cells may indeed repres ent a unique adaptation of this intracellular pathogen that has modified it s movement as a means to avoid host cell detection. Bioterrorism The use of biological and chemical wea pons has a history that parallels all prominent human history, spanning many cultures and countries (64). It gained further modern prominence during the First World Wa r in the use of mustard gas, which was studied by universities and implemented thr ough governments. This essentially marked several countries’ foray into strategic and sanctioned use of biological weapons, and their overwhelming effect resulted in high human cas ualties. Following these events and the end of the Cold War, many countries looked to the prevention of bioterrorism, and in 1972 the Biologicals Weapons Convention pr ohibited the development, production, stockpiling, and transfer of biol ogical agents for use as weap ons (64). It was ratified by more than 100 nations, yet experts believe that several signatory countries may be


10 violating the convention's terms and deve loping offensive biological weapons (64,65). Tragically, as the recent anthra x attacks in the U.S. have demonstrated or the uncovering of six Algerians plot to use ricin in London in 2003, terrorists will not hesitate to use such state-derived biologicals or crude versions of them as weapons to further their goals. However, if high casualties are the intended use, these are not the most effective means. Chemical-biological agents are more useful in creating fear and anxiety. Although there are many practical obstacles to terrorists in using these agents, terrorism exists as a psychological weapon intended for mo stly political effect (66,67). Although there are many reasons, both hypot hetical and real, it is difficult to ascertain the motivations behind terrorists’ use of chemical and biological weapons (66). Dispersal of biological stockpiles (68) and th e availability of free information through a variety of sources such as the Internet, have allowed individual seekers access to technical sources and information once regard ed only accessible through governments. Recently there has been growing at tention too by many countries through government agencies in agriculture, water, public health, and environmental offices to address the threat of biological agents as w eapons. Following recent terror attacks, notably the mailing of letters containing anthrax spores ( Bacillus anthracis ) resulting in five deaths, the U.S. responded by passing stricter federal laws, such as the Public Health Security and Bioterrorism Preparedness and Response Act of 2002 (the Bioterrorism Act) regarding both enhanced contro l of biological agents and toxins including but not limited too the protection / safety of food, drug, and water supplies. Public Food Safety Food transmission is a common route fo r an estimated 200 diseases, caused by viruses, bacteria, toxins, and prions. The Foodborne Diseases Active Surveillance


11 Network (FoodNet) operated under the Cent ers for Disease Control and Prevention (CDC, Atlanta, GA) estimates that there are over 76 million cases of illnesses, 352,000 hospitalizations, and 5,000 deaths attributable to food borne diseases in the United States annually (12). Although these statistics may se em high, this is actually a downward trend in food borne related illnesse s since 1996. The ease at wh ich opportunistic pathogens can infiltrate public food supplies was demonstr ated recently in February 2002, when Abbott Choice brand cheese products were recalled by the Abbott Cheese Company of Cowichan Bay, British Columbia (9). An epidemiol ogical investigation linked the company’s cheeses to at least 21 L. monocytogenes infections among people who had consumed the cheeses throughout Canada on Vancouver Island and through lower British Columbia. A woman who became ill with a listeria infecti on after consuming the contaminated cheese subsequently had a miscarriage. It is the ubiquitous nature a nd the relative ease to isolate some of these organisms from the environment that mark them as important targets for food safety. There then exists a need to easily det ect pathogens in both environmental and food supplies. It should also not be overlooked what th e huge economic effect s a bioterrorist attack would have on public food and water su pplies, particularly on the food processing industry as a whole. Although there are regul ations and agencies such as the United Stated Department of Agriculture (USDA) and the Food and Drug Administration (FDA) who monitor food and drug supplies within th is country, improved standardized methods would increase security and safety. The goal of such te sting involves a high volume and high number of samples and would ideally be simple, reliable, and fast, while exhibiting a high degree of sensitivity and specificity.


12 Beyond this, systems used for detecting c ontamination with microbes and toxins should be portable, less time consuming, invol ve minimal handling and processing of the samples, and obviate the need for costly a nd bulky equipment. The need for skilled technicians should be minimal. Methods for Bacteriological Detection and Analysis There are several reliable bacterial de tection methods applied clinically and environmentally. However, these tests are often labor intensive, time consuming, and technically challenging. Ma ny traditional methods rely on determining microbial contamination involving patient sampling and empirical enumeration. Estimating contamination empirically is not very specific and not sensitive. The presence of bacteria in water, foods, and biological or envir onmental samples can also be monitored by measuring physiochemical changes that occu r from bacterial metabolism and growth, such as biochemical assays that detect adenosine tri-phosphate (ATP). These biochemical methods identify the microbes in a very sensitive manne r; however, they are time consuming and not very specific. Gene rally, many tests in us e today for microbial detection involve growing the pathogens in a select pre-enrichment step followed by growth in a selective medium; the entire process can take more than 24 hours to complete. Several biochemical tests are needed for characterizing the microbes. Beyond this though, many laboratories have addressed this issue by incorporating modern molecular genetic techniques for rapid and sensitive detection of microbes which involve amplification of specific nucleic acid seque nces by polymerase chain reaction (PCR), immunomagnetic separation, enzyme tests wi th synthetic chromogenic or fluorogenic substrates (Direct Epifluorescent Filter Tec hnique, DEFT), real-time PCR, and flow


13 cytometry. Many commercially available k its including API Rapid 20E (bioMerieux, Los Angeles, CA), GN Microplate (Biolog, San Jose, CA), Directigen (BD Diagnostic Systems, Franklin Lakes, NJ) are available for detection of microbes and can save time but, for example, detect only gram-negative species. Imporvements that integrate engineering and biology include high-end, techinical equipment such as optically based biosenso rs. The surface plasmon resonance (69) and the IAsys (Affinity Sensor, Cambridge, UK) ar e optically based sensors that have been used to detect Staphylococcal enterotoxin A in foods down to 10 to 100 ng/g, but no research publications exist to support detection of whole b acterial cells (70). Several other biosensors such as pha ge-coupled piezoelectric biosen sors, electrical biosensors, electrical impedance biosensors, poten tiometric immunosensors, and amperometric immunosensors have been evaluated and work well for detection of various other food borne pathogens or their toxins (71-73). Many of the above-mentioned techniques may identify microbes sensitively and specifical ly, but require costly equipment, skilled personnel, and sterile conditions to perform the analysis. Additionally, most of these systems are not portable and cannot be used in field conditions. For immunochemical based detection systems to work and be successful, the antibodies utilized must demonstrate extrem e specificity. Murine monoclonal antibodies have been used for many years in research for detecting the antigens and separating the particles from a complex mixt ure (74). Some of the above -mentioned detection systems employ antibodies in conjunction with some ot her technique to identify the pathogen. Examples for such techniques us ing immunological reagents include


14 immunofluorescence, fluorescence auto mated cell sorting, immunoprecipitation, immunohistochemistry, and immunomagnetic separation. The Biosensor System Frances Ligler, Senior Scientist for Bios ensors and Biomaterials at the Naval Research Laboratory's Center for Bio/Molecu lar Science and Engin eering is attributed with development of the first commercia lly available antibody-based fiber optic biosensor for the detection of pathogens and their toxins, the RAPTOR (Research International, Woodinville, WA )(75). The principle of the RAPTOR biosensor is based on a sandwich fluoroimmunoassay. Capture anti bodies are coated on the surface of four separate polystyrene fiber-optic probes by pa ssive absorption. Following an overnight incubation at 4 C, the unbound antibody is rinsed off w ith deionized-water and incubated with immunoassay stabilizer so lution (ABI, Columbia, MD). The probes are then dried and mounted into an assay housing cartri dge referred to as a coupon. For sample analysis, the assay coupon is inserted in to the RAPTOR coupon compartment which aligns the four fiber optics into their optical paths and also forms the fluidic connections which are required for the fully-automated featur es used in sample analysis. A pneumatic pump then moves buffer at a flow rate of approximately 9 mL/minute, contained in an on-board reservoir over the probes to rehydrate the bound antibody. The serpentine channels in the coupon are used in all step s of the sandwich assa y for reagent addition over the probes. A five-minute baseline prot ocol is automatically initiated prior to sample analysis by incubating the rehydrated probe with the fluorescent tracer antibody used for detection. This esta blishes an important background signal that determines the rate of non-specific binding in the absence of a positive antigen. When sample is flowed


15 over the probes, it binds to the fiber optic co ated with capture an tibody specific for that antigen; unbound material is th en washed away with a phos phate-buffered saline solution (PBS) containing a detergent, normally Tween or TritonX-100. Sample size is normally 3 mL. Fluorescently Cy5-conjugated antibody or Alexafluor633 (Invitrogen, Carlsbad, CA)-conjugated antibody (tracer antibody) is then introdu ced and binds to the antigenantibody complex on the probe. The Cy5-conjugat ed antibody is recycl ed and stored in the RAPTOR reagent compartment and maintained at 2-8 C. Following a second rinse with PBS containing Tween to wash unbound c onjugated antibody, an excitation light is emitted from four 5mW Sanyo laser diodes at 635 nm and is focused through the fiber optics through a light-gathering short focal le ngth ball lens fused to the probe’s end. A 665 nm filter rejects reflected laser light. The fiber optic’s distal end contains an aluminum coated mirrored end that returns the evanescent wave back up the fiber optic probe to a photodiode sensor. On-board soft ware discriminates signal generated from the bound fluorophores within 1000 nm of the probe . Signal exponentially decreases as distance increases away from the fiber optic. Data analysis is completely automated, and results are displayed on a LCD screen. Total analysis time is approximately 20 minutes. The biosensor is capable of multiple analyte detection (Table1-1) (70). Monoclonal Antibody Production Over one hundred years ago Emil von Behri ng and Shibasaburo K itasato searching for the mechanism that induces immune prot ection found that vaccinated individuals’ serum contained certain substances that bound to relevant pathogens. A few years later, Paul Erhlich first described his side-cha in theory on the interactions of these “immunebodies” with their antig ens, their formation, and the idea of antibodies as “magic


16 bullets”. In 1975, seventy-five years after Ehrlich, George s Kohler and Cesar Milstein described an in vitro method of manipulating an immort alized cell line by fusing murine myelomas and splenocytes to secrete mono-reactive antibodies of a single known specificity (76). They called these cells hyb ridomas and their antibodies as monoclonal. Previously, the use of the antisera from immunized mammals, though useful in many applications, has some limited biological significance, since the mammalian immune system will produce a heterogenous populati on of antibodies that can possess crossreactive patterns to other antigens regardless if the antisera were raised in genetically identical animals immunized with identic al immunogen. Since then, the exquisite specificity and affinity exhibited by the t housands of monoclonal antibodies that have been created using Kohler and Milstein ’s hybridoma technology has had far-reaching implications in the treatment of disease (77). The generation of monoclona l antibodies (mAbs) has not varied much from the original Kohler and Milstein hybridoma techni que and can be separated into four steps; immunization, cell fusion, cell selection, and expansion. Animals are injected with antigen that must represent the target substa nce of the antibody wanted to be generated by the immune system. This elic its B-lymphocyte priming, and th e animal is subsequently boosted with a second injection to i nduce B-cell clonal expansion. Following immunization, B-cells are taken from the sp leen and fused to immortalized myeloma cells using low-centrifuagtion in the presence of polyethylene-glycol (PEG), a polywax substance that promotes membrane fusion and exchange of nuclei. The original method utilized BALB/c mouse SP-O or SP-1 mous e myeloma cells for fusion, although murine non-immunoglobulin secreting ce ll line P3-X63-Ag8.653 or rat myeloma cells have also


17 been used (78). Care must be taken to use myeloma cells that are non-antibody secreting since this could potentially “contaminate” th e pool of mAbs orig inating from the Blymphocytes. Unfused splenoctyes have a li mited life span in culture, and myelomas lack the purine nucleotide sa lvage pathway that utili zes the enzyme hypoxanthineguanine phospho-ribosyltransferase needed fo r survival in the presence of a purine inhibitor, aminopterin, used in the growth medium hypoxanthine aminopterin thymidine. Therefore, this is the selective medium us ed for the growth of the hybridomas because they posses both the immortality of the myel oma cells and the purine nucleotide salvage pathway derived from the splenocytes. Ot her methods such as electroporation and transfection have been utilized for hybridoma production but ar e not as efficient as PEG (79,80) for unknown reasons. The quality of the monocl onal antibody and the stability of the hybridomas are crucial to obtaining a homoge nous cell line secreting mAb. Th is is ensured by rigorously testing for mAb specificity and cloning the cultures by limiting dilution until a near 100% clone efficiency is achieved representing a pure single clonal cell line. The hybridomas thusly formed can be stored in liquid nitr ogen for years making these clones virtually immortal. Maintenance in high-density gr owth chambers of mA b-secreting hybridomas has been well documented. In cell culture ro ller bottles, yields are approximately 10-100 g/mL of mAb (81). Yields can be increas ed to approximately 50 mg using mediumsized fermentators (81) such as the Mini PERM bioreactor (Sartorius AG; Goettingen, Germany) or to even 1 g of antibody utilizi ng large-scale novel hol low-fiber cell culture biorectors that utilize reci rculation pumps and oxygenation units to add nutrients and remove waste (82).


18 There have been described many syst ems for the culturing of monoclonal antibodies and antibody fragments including expression in bacteria and yeast as well as in plant, insect, and mammalian cells (83,84). These systems, in part, attempt to address some obstacles encountered with classic hybridoma technology. Instability of hybridoma cell lines is not uncommon, as well as, possess ing less than desirable growth, cloning, or poor mAb production. Additionally, over 50% of the mAbs generated by hybridomas are also in the form of pentameric IgM, which is not always practical for use and is not readily purified. IgM production is especia lly of concern if th e antigen used for immunization is heavily glycosyl ated and thus unable to invo ke an anamnestic “memory” B-cell response. These B-cells never mature beyond production of IgM even in hybridomas. More common problems encountered with hybrdioma technology are inherent in the use of murine cell lines in the genera tion of mAbs. Use of murine monoclonal antibodies therapeutically in humans can le ad to a human anti-mouse antibody immune response in which the murine mAb is recogni zed as foreign. To address this problem, monoclonals can be “humanized” by vari ous methods. Human mAbs have been generated with some success ut ilizing mouse / human heteromy elomas as fusion partners (85,86) or immortalizing antibody secreting human B-lymphocytes through infection with Epstein-Barr virus; alt hough this cloning has proven to be difficult and infected cells typically have low mAb yield (78). Alternately, transgenic mice have been created by replacement of murine IgH and Ig loci corresponding to antibody heavy and light chain genes respectively by homologous recombina tion in embryonic stem (ES) cells. Once crossbred, these mice inherit a double knock-out JH -/and C -/in their germline DNA.


19 The ES cells from these double-knockouts can be fused with yeast spheroplasts containing human antibody DNA, a so-called ye ast artificial chromosome. The progeny mice obtained secrete only human IgM and IgG antibodies (87). The use of transgenic animal or insect cell lines, how ever, requires a great deal of highly technical knowledge. Plants, on the other hand are ea sier to manipulate, and have also been utilized for the generation of human antibodies with supr isingly high yield (83,88). However, mAbs generated in plants can be heavily glycolsyla ted and the time frame in which to harvest the plants to test antibody specifi city can be upwards of months. Recombinant Phage Display For years now, following the work of G.P. Smith and G. Winter on genotypic and phenotypic linkage of polypeptides to f ilamentous bacteriophage M13 (89-91) recombinant display of randomized pep tides, antibodies, and antibody fragments on phage has been an invaluable research tool in the field of monoclonal antibody technology. The insertion of an tibody genes into phage geno mes has been developed for phage T7, phage lambda, and the Ff cla ss of filamentous phages f1, fd, and M13 (89,90,92). In the Ff class of pha ge, replication of DNA and a ssembly of the phage is not constrained by the size of the incorporated DNA. Additionally, assembly and production of Ff phage occurs without ki lling the host bacterial cell. Therefore members of the Ff class of phages have been the choice for not only molecular cloning and recombination but also antibody phage display. Smaller antibody fragments are commonly used for display, mostly due to E. coli’s inability to properly fold and generate an entire human i mmunoglobulin (Ig) molecule.


20 Both fragment antigen-binding (Fabs) and si ngle-chain Fvs (scFvs) are useful in phage display. M13 Phage Biology A number of filamentous bacteriophage have been identified that are able to infect gram-negative bacteria for their survival. Of these, the best characterized are phage that infect E. coli and because of their dependence on the presence of the F conjugative plasmid, they are referred to as the Ff class of phages. This class includes f1, fd, and M13. Their genomes have completely been sequenced and share 98% homology (93). M13, like other Ff phage, has a single-str anded covalently closed genome of DNA surrounded by a slightly distorte d left-handed cylindrical tube of outer coat protein (9395) measuring approximately 7 nm wide and 900 to 2000 nm in length. The genome encodes phage proteins I thr ough XI that are important in phage infection, replication, and assembly and are numerically arranged by function according to stages of the phage life cycle (Table 1-2). Infection of E. coli by M13 is normally divided into a two-step process: binding then integration. The binding of Ff phage to E. coli is mediated by the interaction of the distal gene-3-protein (g3p or pIII) end of the phage with the tip of the F-conjugative pilus. The horseshoe-shaped bilobal pIII protein consists of two amino-terminal domains, N2 and N1, and a third carboxy-terminal domain, CT, embedded in the phage coat that non-covalently inte racts with pVI. All domain s are separated by flexible glycine-rich hinge regions, not ably important in the N2-N1 di-domains, which work as a proline-switch (96,97). The bindi ng of N2 to the F-pilus init iates a kinetically favorable trans to cis isomerization of the Gln212-Pro213 hinge region of N2-N1, exposing the N1 domain and allowing for binding to the TolA cell co-receptor (96,98) . Retraction of the


21 pilus occurs, mediated presumably by depolymerization of the pilin subunits into the inner membrane of the bacterium, which then brings the tip of the phage in contact with the membrane surface. It is not known wh ether depolymerization is triggered by the binding of the phage or is actually an i ndependent natural func tion of polymerizationdepolymerization (93,99). The second step is integration of the pha ge through incorpora tion into the outer membrane of not only the five copies of pIII that reside on the end of the phage but also the approximate 2,700 copies of pVIII that surround the genome. Penetration of the phage and release of the viral DNA is depende nt on bacterial TolQRA proteins (98,100). The Tol proteins are important in membrane st ability and binding of bactericidal group A colicins (100) as well as mediating phage g3p binding. TolA is anchored in the inner membrane through its N-terminal end (TolAI) which interacts with both TolQ and TolR. The -helical TolAII region spans the periplasmi c space to the outer membrane to the Cterminal TolAIII region, where the N1 domain of g3p interacts. Upon retraction of the Fpilus, cooperating TolA proteins assume a more compact structure that bring the outer and inner membrane closer drawing in the pha ge g3p. Phage g3p is then inserted into the inner membrane and the phage head begins to open, releasing the phage’s genomic DNA into the bacterial cell. In the cytoplasm, the infective single-s tranded genome is converted by bacterial enzymes into a supercoiled double-stranded re plicative form (RF) molecule. This molecule serves as a template for transcription and translation from which all phage proteins are synthesized. Fo llowing early translation of pha ge proteins, pII binds and nicks the RF molecule, forming a free 3’ end th at serves as a primer for synthesis of a


22 new (+) strand. Rolling circle replication con tinues and is circularized by pII. The production of phage proteins continues to increa se with an increase in the number of RF molecules. When the single-stranded DNA binding protein, pV, reaches proper concentration of approximately 800 dimers in the cytoplasm, it begins to package the phage single-stranded DNA into a left-handed helix with a small hairpin stretch of DNA referred to as the packaging signal (PS) left exposed for initiation of phage assembly (93,101). Products of the genes, III, VI, VII, VIII, and IX constitute capsid proteins of the phage particle. All of these pr oteins contain signal sequences that direct their transport to the inner membrane via the Sec system, where they accumulate perpendicularly to the surface until necessary for phage assembly. Binding of the DNA PS with pI initiates assembly (101,102), and the assembly site may be a result of the intera ctions of the minor capsid proteins, pI, pXI, and pIV, al though there is no dir ect evidence (103,104). Speculative models demonstrate twelve to thirteen -sheet monomers of pIV form a gated channel through which the assembled phage eventually passes through (105,106). Additionally, there is no eviden ce of the order of assembly of the capsid ends, pVII and pIX, or even the major coat protein, pV III, during initiation (103,105). Hydrolysis of ATP by the cytoplasmic portion of pI through it s interaction with the reduced form of host cell thioredoxin (TrxA) catalyzes remova l of the pV dimers from the phage DNA and replaces them with membrane-embedde d pVIII as the DNA is extruded through the pIV/pXI channel (106). The 2,700 copies of pVIII reside as pentameters at a 23 angle out from the phage axis, with positively ch arged residues facing inward towards the DNA and acidic “double-hook” residues facing out wards (105,107). When the end of the DNA


23 is reached, assembly is terminated by the addition of pVI and pIII and release of the phage particle. It is not known whether the assembly site can be used for the synthesis of more than one phage particle. Ff phage repl ication and assembly is tolerated well by the bacterial cell. Infected cells continue to grow but at a lowered rate of division. The advantages of this membrane associat ed assembly of proteins are as follows: Capsid formation around the DNA as it is extruded from the bacterial membrane allows for packaging of large pieces of DNA limited in theory only by shear DNA size. Foreign peptides can be fused to the periplasmic portions of pIII and pVIII and not influence assembly of the phage particle as long as the protein can be translocated acro ss the inner membrane efficiently. Tomlinson I + J Human Synthetic VH + VL Phagemid Libraries An alternative for the generation of hu man monoclonal antibodi es involves the construction of combinatorial libraries of VH and VL antibody genes and their expression on filamentous phage. The variable domain (F v) of the antibody has inherent instability since the VH and VL domains are not covalently linked. To address this, a flexible GlySer linker between the carboxy terminal of one domain is linked to the amino terminal end of the second domain to form an single-chain Fv, or scFv. scFv library construction involves the use of reverse transcription and polymerase chain reaction to amplify rearranged anti body genes from B-lymphocytes found in a donor’s blood. Universal degenerate primers are first used to anneal to the 5’ end of the exons encoding the antibody V-gene, which is conserved in humans. The mRNA extracted from the B-lymphocyt es is used to create cDNA th at represents rearranged VH


24 and VL genes. The cDNA is PCR assembled using overlap extension technique and contains restriction enzyme sites for subcl oning into a pIT2 vect or containing an M13 origin of replication, an ampicillin resistance gene ( bla ), and both His6 and myc tags. For displaying the scFv sequence on the surf ace of the phage as fu sion protein, the scFv gene was cloned in frame with the gIII gene coding for the pIII protein in the phagemid vectors. When E. coli harboring these phagemids are supe rinfected with a helper phage, phage particles are produced which display the scFv-pII I fusion protein. The phage particle expresses the antibody fragment on its surface and contains the antibody-encoded genes. Tomlinson I + J libraries are “nave li braries” constructed from non-immunized donors who do not have ongoing immune responses to a specific antigen of interest. Nave libraries offer higher di versification of antibody repert oires, and the Tomlinson I+J libraries were additionally constructed to incl ude side-chain diversif ication as well as a shortened CDR3 binding site on the heavy chain. This thesis describes the efforts to isol ate recombinant phage display antibodies to L. monocytogenes and the attempts at modifying the antibodies for their potential use in the biosensor system. The specific aims for this study are: 1. Isolate recombinant phage display antibodies to either whole L. monocytogenes cells or to critical unique surface prot eins, such as those mentioned above. Obtaining specific antibodies to whole bacterial cells improves the advantage of being able to quickly pre-select for anti-whole cell clones. More precisely, specificity through selection of unique Listeria surface antigens such as flagella or


25 murein hydrolases that are pref erably lacking in nonpathogenic L. innocua , increases the utility of such antibodies in their detection of L. monocytogenes . 2. To manipulate and optimize these imm unological reagents for ultimate use in the biosensor system. The usefulness of the biosensor system is greatly dependent on the quality of the immunologica l reagents utilized. Therefore, those antibodies selected that showcase utility in Listeria detection can be chemically or genetically enhanced to improve overall a ssay quality and that of the biosensor.


26 CHAPTER 2 MATERIALS AND METHODS Bacterial Strains, Phage Strains, and Growth Conditions The bacterial strains used and their genot ypes are listed in 2-1. Tomlinson I + J Human Synthetic VH + VL phagemid libraries constructe d by the Medical Research Council, Cambridge, U.K. (108) were obtaine d from the Interdisciplinary Center for Biotechnology Research Hybridoma Core, Univer sity of Florida. Hyperphage (109,110) was obtained from Progen Biot echnik (Heidelberg, Germany). E. coli TG1 cells were grown in 2xTY broth (16 g tryptone, 10 g y east extract, 5 g NaCl, and 3 mL of 1 M NaOH in 1 L water) or on 2xTY agar pl ates containing 1.5% (w/v) agar. E. coli TG1 cells infected with phage from the Tomlins on library or with the phage obtained after elution and neutralizat ion after a round of panning were grown on 2xTY agar plates with 100 g/mL ampicillin and 1% (w/v) glucose (2xT Y AG plates). When the bacteria were superinfected with Hyperphage to produce pha ge particles, they were grown in 2xTY broth with 100 g/mL ampicillin and 50 g/mL kanamycin. All Listeria species and strains were grown in brain-heart infusi on broth (BHI, (Difco)) or on BHI plates containing 1.5% (w/v) agar. All of the othe r bacteria were grown in modified LuriaBertani broth (LB-N: 10 g tryptone, 5 g y east extract, 8.5 g NaCl, and 3 mL of 1 M NaOH in 1 L water) or on LB-N plates containing 1.5% (w/v) agar. All of the bacteria were grown as overnight standi ng cultures in either BHI, LB-N, or 2xTY at 37 C. The bacteria from the starter culture were dilu ted 1:40 into fresh medium and grown with aeration at 37 C until the optical density at 600 nm (OD600) reached approximately 0.4.


27 Table 2-1. Bacterial strains used. Strain Genotype / Antigenic Formula Source / Reference Fmc r A ( mrr-hsd RMSmcr BC) 80 lac Z M15 lac X74 rec A1 ara D139 ( araleu ) 7697 gal U gal K rps L (StrR) end A1 nup G L. monocytogenes 1/2a unspecified Martin Wiedman L. monocytogenes 1/2b unspecified Martin Wiedman L. monocytogenes 1/2c unspecified Martin Wiedman L. monocytogenes 3a unspecified Martin Wiedman L. monocytogenes 3b unspecified Martin Wiedman L. monocytogenes 3c unspecified Martin Wiedman L. monocytogenes 4a unspecified Martin Wiedman L. monocytogenes 4b unspecified Martin Wiedman L. monocytogenes 4c unspecified Martin Wiedman L. monocytogenes 4d unspecified Martin Wiedman L. innocua unspecified Martin Wiedman L. seeligeri unspecified Martin Wiedman L. ivanovii unspecified Martin Wiedman L. welshimeri unspecified Martin Wiedman L. monocytogenes 104030s unspecified Fred Southwick L. monocytogenes 104030s unspecified Helene Marquis L. monocytogenes HEL-304 L. monocytogenes EGDe genomic strain Alan Pavinski Bitar E. coli BL21 DE3 Tuner Fgal LDE3 ( lacUV-5 T7 gene1 imm21 and nin5 ) hsdS F.W. Studier Ref:J Mol Biol 188 (1986) flaA 104030s isogenic Helene Marquis E. coli HB2151 ara, ( lac-pro ), thi/F' proA+B+, lacIqZ M15 Linda Green, ICBR E. coli TOP10Invitrogen E. coli TG1 K12 (lac proAB ) supE thi hsdD 5/F’ traD 36 proA+B lacIq lacZ M 15 MRC, Cambridge, UK E. coli AVB100 C06 aa 39 (aa leu )7696 ( lac )l74 galU galK hsdR 2(rKmK+) mcrB1 rpsL(Strr) Avidity, LLC, Denver, CO


28 Enzyme Linked ImmunoSorbent Assays Enzyme Linked ImmunoSorbent Assays (ELISA) were used in the screening and characterization of both polyclonal and monoclonal antibodies, antibody fragments (scFvs), and phage particles recovered from the panning experiments. Briefly, standard 96-well polystyrene, flat-botto med microtiter plates (BectonDickinson, Franklin Lakes, NJ) were coated with antigen (cells, cellular extracts, or recombinan t proteins) diluted in either phosphate buffered saline (PBS : 137 mM NaCl, 2.7 mM KCl, 4.3 mM Na2HPO4, 1.4 mM KH2PO4, pH 7.3) or carbonate buffer (13 mM Na2CO3, 87 mM NaHCO3, pH 9.2), covered and incubated overnight at 4 C. The following day the plates were allowed to equilibrate to room temperature for one hour. Wells were washed once with 300 L of PBS and aspirated in an ELX 800 Strip Washer (BioTek, Winooski, VT). Wells were then blocked with the addition of 200 L casein blocker for one hour at room temperature and aspirated prior to loading pre-diluted phage or antibody. For consistency polyclonal and monoclonal antibodies were generally diluted to a working range of 4 to 10 g/mL in casein blocking buffer. Polyethylene-glycol (PEG, Sigma-Aldrich, St. Louis, MO)) -precipita ted phage particles were diluted to 109 TU/mL. If untitered phage or scFv was used in the ELISA from culture supernatant, reagents were diluted 1:1 in casein blocking buffer prior to loading. One hundred microliters of prediluted phage, antibody, or scFv was then added to blocked wells and allowed to incubate one hour at room temperature with no agitati on. The microwell plate was then washed a minimum of three times with 300 L of PBS containing 0.05% v/v Tween-20 (Fisher) (PBS-T (0.05)) before loading secondary antibody.


29 All secondary antibodies and detecting reagents were conjugated to horseradish peroxidase (HRP): either monoclonal anti -M13:HRP for phage detection, protein-L:HRP for scFv detection, or goat-a nti-rabbit:HRP for rabbit polyc lonal. Secondary reagents were pre-diluted to a working range in casein blocking buffer. One hundred microliters of diluted HRP-conjugated secondary immunoreagent was added to washed wells and allowed to incubate for 30 minutes at room temperature. The microwells were then washed again as above in PBS-T (0.05). Substrate for development reaction for HRP-conjugated secondary enzymes was prepared by dissolving one capsule of phospha te citrate buffer with sodium perborate (Sigma) in 100 mL of water (0.05 M phos phate-citrate buffer pH 5.0, 0.03% (w/v) sodium perborate). A 10 mg t of 3, 3’, 5, 5’-tetramethylbe nzidine substrate (Sigma) was added to 10 mL of the buffer to give a fi nal concentration of 1 mg/mL. Two hundred microliters of the substrate solution was added to each well, and the plate was read in an ELx 800 UV plate reader (BioTek, VT) at 630 nm for HRP conjugates following approximately 20 to 30 minutes of substrate de velopment. The data were analyzed with KC Junior (BioTek) and Microsoft Excel software. Epitope Characterization The epitope recognized by the scFv-phage s was tested for its sensitivity to proteinase K treatment and periodate oxidati on by ELISA. Coated plates were washed three times with PBS, and antigens (cells or protei ns) were exposed to 250 g/ml of proteinase K in PBS and then incubated at 24C for 60 min. Wells were then washed five times with 300 L/well PBS, and the ELISA procedure was continued using the prepared phages or antibodies.


30 Sodium meta-periodate oxidation was used with the ELISA to determine if the scFv-phages were specific for carbohydrate. Microwell plates were prepared for the ELISA by coating 100 L of L. monocytogenes cells at 107 to 109 CFU/mL in PBS. Plates were washed once with 300 L/well of 10 mM HEPES buffer. Wells were then incubated with 200 L per well of 10 mM sodium me ta-periodate (Sigma) in 50 mM sodium acetate buffer, pH 4.5 for 1 hour at 24 C in the dark. Wells were then rinsed three times with 300 L sodium-sulfite (2 mg/mL). The ELISA procedure was then continued with the addition of pr epared antibodies or phages. For the treatement of listeria flagella, 30 g of the L. monocytogenes flagellar prepartion (at approximately 0.6 mg/mL in PB S) was aliquoted into a microcentrifuge tube and incubated with 100 L of 4 mM sodium meta-periodate / 10 mM sodium acetate buffer, pH 4.5 for 30 minutes at 24C in the dark. Twenty-five microliters of sodium sulfite (2 mg/mL) was added prior to dilu tion in Laemmli sample buffer for SDS-PAGE. BioPanning of Phage Display Libraries Panning on Immunotubes Flagella or recombinant antigen was diluted to 10 g/mL in PBS and added to a polystyrene (Nunc) immunotube to comple te fullness. The immunotube was then allowed to incubate overnight (> 16 hours) with the antigen suspension at 4 C. Following overnight incubation the suspension was emptied from the tube by inversion, aspirated, and rinsed three times with PBS. The rinsed tube was then filled with casein blocker to coat uncoated areas in the i mmunotube and reduce nonspecific binding. The immunotube with blocker sat wi th static incubation for a minimum of 1 hour at room temperature (25 C), was then emptied by inversion, and drained over paper towels.


31 Hyperphage-amplified Tomlinson I + J Libr ary was diluted in casein blocker to a minimum of 1011 TU/mL. The immunotube was ca pped, and the phage library was incubated with end-over-end turning for one hour at room temperature. The tube was then incubated an additional one hour at r oom temperature with no agitation. The diluted unbound phage library was emptied by inversion a nd aspirated, and the tube was rinsed a minimum three times with PBST (0.1% tween-2 0). Incubating the rinsed tube end-overend with 1 mL of 1 M triethanolam ine for 30 minutes eluted bound phage. Triethanolamine-eluted phages were transfer red into a clean microcentrifuge tube and neutralized by adding 500 L of 1 M Tris-HCl, pH 7.4. Neut ralized phages were stored at 4 C until ready to use for further rounds of panning or for amplification of phage clones. Panning in Suspension Bacterial cells were diluted to 109 CFU/mL in PBS in a 1.5 mL microcentrifuge tube, and 1011 TU of Tomlinson I + J Hyperphage amplified library was added directly into the bacterial suspension. The panning suspension was incubated end-over-end for a minimum of four hours at room temperatur e. Cells and bound phage were pelleted by centrifuging the microcentrifuge tube 10 minut es at 10,621 x g at room temperature. The unbound phages were then removed by aspirating and collecting the liquid layer into a separate microcentrifuge tube. The bacter ial pellet containing bound phage was washed by repeating the PBS rinse, resuspension, a nd centrifugation once to minimize loss of pellet. For elution of bound phage, the resu spended pellet was tr ansferred to a fresh microcentrifuge tube, and 500 L of 0.5 M glycine, pH 2.8 was added. The tube was incubated end-over-end at room temperature for 10 minutes and quickly centrifuged at


32 10,621 x g for 10 minutes to pellet bacterial cel ls. The eluted phages contained in the supernatant were collected and transferred to a new microcentrifuge tube and neutralized with 50 L of 1 M Tris-HCl, pH 8.0. El uted phages were stored at 4 C until use. Titering the Phage Culture supernatant containi ng untitered phage were seri ally diluted in PBS, and 100 L of the10-2, 10-4, 10-6, 10-7, and 10-8 was added to 1 mL of exponentially growing E. coli TG1 (OD600 of 0.2 to 0.4). Infected TG1 was th en incubated for 20 minutes in a 37 C water bath. One hundred microliters of ea ch dilution of infected cells was spread onto 2xTY AG plates. Plates we re incubated overnight at 37 C, and colonies were counted and recorded corresponding to each phage dilution. Amplification of the Selected Phage Following elution of phage from panning, 0.5 mL of neutralized eluted phage was used to infect 1.5 mL of exponentially growing E. coli TG1 cells. The TG1 cells were incubated for 30 minutes at 37 C, and the full 2.0 mL of culture was spread onto five 2xTY AG plates using 0.4 mL of culture per pl ate. Plates were incubated overnight at 37 C. To grow phage, the lawn of bacterial cells was scraped off the plate the following day by first adding 1 mL of 2xTY AG medium to the plate and pipetting the lawn suspension into a microcentrifuge tube. The lawn scraping was repeated for all plates and cell suspensions were pooled and vortexed to mix thoroughly. Fi fty microliters of the concentrated suspension of cells was added to 100 mL of 2XTY AG medium, and an initial OD600 reading was recorded. The cells were grown shaking with aeration to late exponential phase (OD600 of 0.4 to 0.6). Ten mililiters of this culture was aliquoted to a


33 separate container and superinfected with H yperphage to a calculated multiplicity of infection (MOI) of 20. Cells were incubated in a 37 C water bath for 30 minutes and spun down in a centrifuge at 10,621 x g for 5 minutes at room temperature to pellet bacteria. The supernatant was discarded, a nd the cells were resuspended in 100 mL of 2xTY containing 1% (w/v) glucose, 100 g/mL ampicillin, and 50 g/mL kanamycin (2xTY AK). This resuspension was incuba ted shaking with aera tion overnight at 30 C. The following day the culture was spun down at 10,621 x g for 10 minutes at 4 C, and the supernatant containing phage was collec ted. Phages were then precipitated by slowly adding 25 mL 20% (w/v) of polye thylene glycol (PEG)/2.5 M NaCl with continuous swirling. The mixt ure incubated on ice at 4 C for a minimum of four hours. The precipitation mix was then spun down at 5,400 x g for 20 minutes at 4 C. The small whitish pellet was resuspended in 20 mL of PBS, and the spin was repeated to remove excess PEG. Following this pellet rinse, the pha ge pellet was suspended in 1 mL of PBS, aliquoted, and saved at 4 C until titered and used for the next round of panning. The phages were titered and diluted up to 4 mL in casein blocking buffer and used in either a pre-coated immunotube or to cells in suspension as described above. Panning without Amplification Phages recovered from a prior round of pa nning were used for further rounds of panning without amplification by diluting the eluted phage in up to 4 mL of casein blocking buffer and added to a pre-coated im munotube or to cells in suspension. Production of Soluble Antibody Fragments (scFv antibodies) Phage possessing positive binding to a se lected antigen via ELISA and Western blot were used to infect nonsuppressor E. coli HB2151, which does not translates the


34 TAG amber stop codon between the scFv-pIII fu sion as a stop in the phagemid vector. The scFv gene is preceded by a pelB leader sequence, and thus, the scFv protein is secreted to the periplasmic space and to the culture media. One mililiter of e xponentially growing E. coli HB2151 was infected with 100 l of phage and incubated for 30 minutes at 37 C. One-hundred microliters of infected cells were plated on 2xTY AG plates to select for transformed colonies and allowed to incubate overnight at 37 C. The following day, randomly se lected colonies were picked, grown overnight in 2xTY AG (0.1% glucose) at 37C, then were diluted 1:20 in 2xTY AG (0.1% glucose). Cells grew to an OD600 of 0.6 to 0.8, then induced with the addition of IPTG to a final 1 mM concentration. Th e induced culture conti nued to grow shaking overnight at 30 C. The supernatant was collected the following day by centrifuging cells out at 4 C for 10 minutes at 13,800 x g. DNA Manipulations Plasmid Extractions Plasmid extractions were done using QIAp rep Spin Miniprep kit (Qiagen) for cultures up to 10 mL and Plasmid Midi kit (Q iagen) for culture volumes of 50 to 100 mL. All extractions were performed according to the manufacturer instructions. Genomic Extraction Genomic extraction of L. monocytogenes sequenced strain EGDe was performed using DNeasy (Qiagen) Genomic Extraction kit. A 3 mL culture was pelleted and suspended into 180 L of enzymatic lysis buffer (20 mM Tris-HCl, pH 8.0, 2 mM EDTA, 1.2% Triton-X 100 (Fisher), 20 mg/mL lysozyme (Sigma)).


35 Restriction Enzyme Manipulations Restriction enzyme digestions were done with enzymes purchased from Invitrogen, Rockville, MD; New England BioLabs, Beve rly, MA; and Promega, Madison, WI, and were used according to ma nufacturer instructions. Agarose Gel Electrophoresis DNA was resolved on 0.7 to 1% (w/v) agar ose gels using Tris -borate-EDTA buffer (89 mM Tris, 89 mM boric acid, 2 mM EDTA) containing 10 g/mL ethidium bromide. Gel electrophoresis was done at 100 V, a nd the DNA bands were visualized on a UV transilluminator. Polymerase Chain Reaction (PCR) Polymerase chain reaction was used to amplify the listeria Auto protein ( lmo1076 ) from genomic DNA extracted from EGDe ce lls for subcloning into pET19b protein expression vector. The recombinant protein was subsequently used for panning. Primers were designed to exclude the 26 amino acid signal sequence and to anneal and amplify the N-terminal autolysin domain between amino acids 75 and 243. Primers used were lm-Auto-5’ and lm-Auto-3’containing XhoI and BamHI restriction sites respectively (Sigma GENOSYS). lm-Auto-5’: CCTCGAG GCTGAAACAACTAATGGAGTAG lm-Auto-3’: GGGATCC TTAATCATATTGAGTCAAATTATATGAAG The PCR conditions were 30 cycles of denature at 95 C, 1 minute; anneal at 57 C (first 10 cycles) and 65 C (last 20 cycles), 1 mi nute; extension of 72 C, 1 minute. The final PCR product has a calculate d size of 668 base pairs.


36 Construction of lm-Auto-pET19b Pl asmid Expression Vector (pGTR1499) The lm-aut-PCR product was resolved on a 0.7% agarose gel, and a single band corresponding to the 668-bp lm-aut amplif ied product was excised. Since polymerase chain reaction was performed with high fide lity Pfu polymerase (Stratagene), the PCR product was reacted additionally with Taq polymerase to incorporate overhang dATPs necessary for subcloning into pCR 2.1-TOPO linear vector (Invitr ogen). TOPO was selected for ease of capture of PCR product. Following cloning of the lm-aut into pCR 2.1-TOPO, a mixture of the Taq treated-fragme nt and vector was el ectroporated at 1.25 kV into E. coli TOP10 cells (Invitrogen ) and incubated with 1 mL of LBN medium at 37 C for 30 minutes. The culture was pl ated on LBN agar spread with 40 L of 40 mg/mL 5-Bromo-4-chloro-3-indolyl -D-galactopyranoside and supplemented with 40 g/ml kanamycin to select for transfor med colonies. Separately, pET19b was electroporated into E. coli BL21 DE3 Tuner cells and plated on a LBN-Amp plate. The plates were incubated overnight at 37 C. The following day both TOP10 and Tuner wh ite colonies were selected, passaged on LBN-Kan or LBN-Amp plates, and plasmid was extracted. Plasmids were miniprep extracted (Qiagen) and resolved on a 0.7% agar ose gel to confirm plasmid size. The pCR 2.1-TOPO-lm-aut plasmid was restriction enzyme digested with Xho I and Bam HI to obtain the fragment corresponding to lm-aut prior to ligation. pET19b was also Xho I / Bam HI digested to linearize the plasmid. Digested plasmid and insert fragments were resovled on a 0.7% agarose gel. The bands corresponding to lm-aut and to linearized digested pET19b were gel excised / extracted (Qiagen), li gated, and electroporated into E. coli TOP10. Cells were


37 plated on LBN-Amp plates, and random colonies were selected for plasmid extraction. Extracted plasmids were resolved on a 0.7% agaorse gel to confirm sizes. The correct fragment orientation within the plasmid was further conf irmed by restriction enzyme digestion with Xho I / Bam HI and resolved on a 0.7% agarose gel. Finally, selected plasmids meeting these criteria were then electroporated into E. coli Tuner, plated on LBN-Amp for transformation selection, and co lonies were selected the next day. Plasmids were extracted and resolved on a 0.7% agarose gel for verification. Construction of scFv-Avitag Plasmid Vectors To improve the usefulness of the scFv anti bodies, they were biotinylated by cloning the scFv gene into pAC Avitag vectors and expressing them in E. coli AVB100 cells (Avidity). Avitag is a 15-peptide sequence sp ecifically biotinylated by the BirA enzyme of E. coli. Plasmid vectors pAC4, pAC5, and pAC6 contain the sequence coding for the Avitag peptide distal to a multiple cloning site. The E. coli AVB100 strain has birA stably integrated into the chromo some under the cont rol of AraC. The genes coding for anti-liste ria whole cell scFv, anti-fla gella scFv, and anti-BSA scFv were cloned into the pAC5 Avitag vect or. Anti-BSA antibodies were biotinylated and tested as a model for the scFv antibodies. Plasmid DNA was extracted from E. coli TG1 cells containing plasmids encoding the anti-whole cell, an ti-flagella, and anti-BSA scFv an tibodies. Restriction digestions were done with Hind III and Not I to obtain the fragment c ontaining the ribosome binding site, the pelB leader sequence, and the entire scFv gene (Figure 1-1). The 5’ overhangs were filled in with Klenow fragment of DNA polymerase I (Invitrogen). The digestion mixture was resolved on a 0.7% (w/v) agarose gel, and the fragment corresponding to Hin dIII/ Not I band was excised. The DNA was extracted from the


38 agarose gel using a Gel Extracti on kit (Qiagen). Plasmid vect or pAC5 was digested with SmaI to linearize the plasmid, and the Hind III/ Not I fragment was blunt end ligated into the vector. With anti-BSA scFv, another plasmid was constructed without the pelB leader sequence by digesting the plasmid with Nco I and Not I to obtain only the scFv gene. After digestion with Not I, the 5’ overhang was filled in with Klenow fragment of DNA polymerase I and then digested with Nco I, which cuts at the 5’ end of the scFv gene. The digestion mixture was resolved on a 0.7% (w/v) agarose gel, and the fragment corresponding to Nco I/NotI band was excised. pAC5 Av itag vector was digested with Nco I and Sma I, the gel-purified fragment was ligated into the vector. Figure 2-1: Vector map of pIT2 phagemid vector. The vector map of pIT2 phagemid vector used in constructing the To mlinson libraries. RBS-Ribosome binding site. pelB-leader peptide sequence. VH and VL-genes coding for heavy and light chains, respectively. Linker(Gl y4Ser)3 linker sequence. Amber stop codon is at the junction of c-myc tag and gIII gene. Not shown in the figure are E. coli ColE1 ori , M13 ori , and bla . Protein and Flagella Manipulations Extraction of Flagella L. monocytogenes cells were enriched for flagella by using a sterile needle to vertically puncture soft agar motility pl ates (LB with 0.5% w/v agar) from an exponentially growing listeria culture. The plate was incubated at room temperature lac promoter Hind III RBS S f i I / Nco I pelB leader VH Linker VL Xho I Sal I Not I Amber codon c-myc tag g III


39 (25 C) to induce flagellar expression. Followi ng 48 hours, a sterile loop was used to carefully excise bacterial cells from the outer perimeter of the growth ring on the motility plate. This region should be enriched fo r bacterial cells possessing the flagellated phenotype, which were therefore used to inoc ulate a 1L shaking culture in BHI following visual motility confirmation under light micros cope. Cells were shaken with aeration overnight at 25 C, and the following day the entir e volume was spun down at 15,000 x g for 10 minutes at room temperature. The pelle ts were serially resuspended in 100 mL of PBS, and the cell suspension was transferred to an Osterizer blender. Cells were homogenized at maximum speed for 2 minutes to mechanically shear off the flagella. The homogenized mixture was spun at 15,000 x g for 10 minutes at 4 C to pellet cellular debris and cells. The liquid portion was re moved from the pellets, and the spin was repeated. The flagella-containing supernatant was ultracentrifuged in a swinging-bucket rotor at for 3 hours at 4 C at 180,000 x g. A small whitish pellet was observed, and the liquid layer was carefully decanted off. The pellet was then suspended in 10 mL of PBS containing 0.02% (w/v) sodium azi de (Sigma) and stored at 4 C. Recombinant Protein Expression of Auto Protein Target genes cloned into pET19b are under control of a strong bacteriophage T7 promoter, and therefore target gene expr ession is accomplished by providing a source of T7 polymerase. The T7 polymerase transcripts on the E. coli Tuner genome is controlled by an upstream lac promoter which is bound by the lac repressor (LacI). Target gene transcription begins by the addition of IPTG which removes the lac repressor and allows for transcription and transl ation of T7 RNA poylmerase. Once a source of T7 RNA polymerase is provided within the cell, the target gene is also transcribed and translated at


40 a high rate since IPTG is non-metabolizable. Th erefore, when this vector is used, the lac repressor functions as both a promoter in the host chromosome to repress T7 RNA polymerase and to also function at the vector to block any transcription of the target gene by any T7 polymerase that is made. Fig 2-2: pET-19b cloning/expression region. Following verification of the correct subc loned lm-auto fragment into pET19b in E. coli Tuner cells, a 10 mL standing overnight culture of E. coli Tuner (pGTR1499) was diluted into 1 L of 2xTY containing no glucose and supplemented with 100 g/mL ampicillin. The culture was grow n shaking with aeration at 37 C until an OD600 of approximately 0.7 was reached. One mililiter of pre-induced cells was collected for SDS-PAGE quantitation purposes. Ten mililiter s of 100 mM IPTG was then added to the 1 L of cells for a final concen tration of 1 mM IPTG. The culture was shaken for 4 hours at 37 C with one mililiter aliquots removed every hour. The culture was then spun down at 6,000 x g for 10 minutes at 4 C. The supernatant was saved, and the pellets were frozen overnight at -20 C. Bacterial Cell Lysis Extraction of Recombinant Auto Protein Extraction of the recombinant Auto protein from E. coli Tuner host expression cells was performed using CellLytic B II Cell Lysis Extraction Reagent (Sigma-Aldrich) according to manufacturer’s instructions. For large scale purification, protein was T7 promoter lac operator RBS HIS tag T7terminator Xba I Nco I Nde I Xho I BamH I


41 extracted using 35 mL of CellLytic B II from approximately 7 grams of a frozen cell paste. Nickel-column Affinity Purification of His-tagged Proteins Proteins with exposed histidine residues, especially (His)6 can be selectively retained on affinity columns pre-charged with Ni2+ ions. Tagging the scFv or lm-Auto with (His)6 allows for strong binding of these protei ns to the charged column, His-tagged proteins can be eluted with buffers contai ning imidazole, thus allowing for purification from E. coli culture supernatant. HiTrap HP 1 mL columns, (Amersham Bi oScience, Piscataway, NJ) samples, and buffers were incubated at room temperatur e prior to proceeding with purification of proteins. After column was rinsed with fi ve column volumes (5 mL) of binding buffer (20 mM imidazole, 1X PBS), 25 to 50 ml of 20 mM imidazole-equilibrated sample was loaded onto the column with a syringe with an approximate flow rate of 1 mL per minute. The column was rinsed with another five column volumes of binding buffer, and bound sample was eluted by utilizing a step -gradient of 100 mM, 300 mM, and 500 mM imidazole in 1X PBS in 5 mL column loads. All column flow-through, rinse, and elutions were collected and examined by SDSPAGE to verify purification of protein. Chemical Biotinylation of scFv Alternately, simple chemical biotinylati on of scFvs and polyclonal antibodies was utilized as a quick method to assist in detection enhancement. Following nickel column purification of His-tagged scFv, imidazole was removed by buffer exchange using AmiconUltra (Millipore) 10 KDa molecular-w eight cut-off spin concentrators and suspending in 0.5 M sodium-carbonate bicarbona te buffer, pH 9.2. Pr otein concentration


42 was determined by DC Protein assay (BioRa d) and the scFv was diluted to a final concentration of > 1.0 mg/mL in bicarbonate buffer. For biotinylation, a 1.0 mg/mL solution of (+)-Biotin N -hydroxysuccinimide ester (NHS-biotin, (Sigma-Aldrich)) was made by dissolving 3 to 5 mg in an equivalent volume ratio (i.e. 3 to 5 mL) of dimethyl sulfoxide (DMSO, (Sigma-Aldrich)). Onemilligram of antibody was aliquoted into a glass vial, and 100 L of the NHSbiotin/DMSO solution was added dropwise with constant stirring. The biotin-antibody solution was incubated for three hours at r oom temperature in the dark. To halt the biotinylation reaction, the biotinylated antibody was dialyzed overnight at 4 C in 1 L of PBS, an additional three hours at room temperature in fresh 1 L of PBS, and aliquoted/stored at 4 C until ready for use. Alternativel y, excess unreacted biotin can be buffer exchanged using filter spin concentrators. Antibody concentration is assumed to still be approximately 1.0 mg/mL. Determination of Protein Concentration Quanitation of protein was performed by use of DC Protein Assay (BioRad) read at 630nm with an ELx 800 UV plate reader (BioTek, VT). Proteins suspended in buffers containing imidazole were first buffer exch anged by centrifuging protein in Amicon Ultra 10 KDa molecular-weight cut-off spin con centrators (Millipore) at 4,000 x g for 10 minutes at room temperature and suspending in either bicarbonate or PBS to remove interfering substances prior to determining protein concentration. Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE) ReadyGel Tris-glycine 10% (w/v) PAGE gels (Bio-Rad) were used with the MiniProtean Electrophoresis system (Bio-Rad ) for SDS-PAGE of protein samples. The


43 samples were diluted 1:2 in Laemmli sample buffer (Bio-Rad), and dilutions were made as necessary. When whole bacterial cells were used as antigens, 108 CFU were suspended in Laemmli sample buffer and boiled for 10 min. Electrode buffer was 25 mM Tris, 0.17 M glycine, and 0.1% (w/v) SDS. The samples were electrophoresed for 1 hour at 100 V. Coomassie Blue Staining for Proteins Proteins in SDS-PAGE were fixed in 50% (v/v) methanol, 10% (v/v) glacial acetic acid in water for 30 min. The gel was staine d with 0.1% (w/v) Coomassie Blue R, 50% (v/v) methanol, 7% (v/v) glacial acetic acid and agitated gently for 30 min. The gel was destained with 5% (v/v) methanol, 7% (v/v) glaci al acetic acid in water. When the protein bands appeared the gel was dr ied, and a photograph was taken. Tsai-Frasch Silver Staining Proteins in SDS-PAGE were fixed in 40% (v/v) ethanol, 5% (v/v) glacial acetic acid in water overnight. Carbohydrates were oxidized by incubating the fixed gel in 0.7% (w/v) periodic acid, 40% et hanol, 5% glacial acetic acid for five minutes. The gel was then washed three times in ddH2O using 15 minutes per wash. Fresh stain is prepared by titrating Reagent A (27.7 mL ddH2O, 280 L 10M NaOH, 2 mL NH4OH) with dropwise additions of Reagent B (20% w/v AgNO3, 5 mL ddH2O). The titrated reagent was added to 115 mL of ddH2O and stirred briefly. St aining solution was quickly transferred to a dish and the gel was added. Staining was allowed to proceed no longer than 10 minutes. The gel was then devel oped with 1X Developer (5X Tsai-Frasch developer: 50 mg citric acid, 0.5 mL 37% (v/v) formaldehyde, 200 mL ddH2O) for 3 to 10 minutes with constant gentle shaking. The developed gel was then placed in dH2O to finish.


44 Immunoblotting The antigens in the SDS-PAGE gels were transferred onto a nitrocellulose membrane for reaction with monoclonal antibod ies, phage, or scFv antibodies according to the procedure of Towbin et al. A Mini TransBlotting cell (Bio-Rad) was used for the wet blotting technique. The transfer buffer used was 25 mM Tris, 192 mM glycine, and 20% (v/v) methanol. A voltage of 100 V was a pplied for 1 hour at room temperature. Protein transfers were confirmed by staining the membrane with 0.3% Ponceau S dissolved in 3% trichloroa cetic acid (TCA) and rinsing with dH2O or 0.3% TCA. The nonspecific binding sites on the me mbrane were blocked with casein blocking buffer. Primary antibody diluted in casein bl ocking buffer was added and incubated for 1 h with continuous rotation. The membrane wa s washed with 100 mL Tris buffered saline with 0.1% (w/v) Tween 20 with gentle agit ation. The membrane was incubated with secondary antibody diluted in blocking buffer for 30 minutes with continuous rotation followed by another three washes as describe d above. The developm ent of the reaction was done by placing the membrane in 100 mL of substrate solution and observing the appearance of bands. The substrate used fo r secondary antibodies conjugated with horseradish peroxidase was 4-chloro-1-napht hol (4CN) (Fisher) prepared according to manufacturer instructions. The developm ent reaction was stopped by removing the membrane from the substrate and placing in water. The membrane was scanned using an UMax photo scanner.


45 CHAPTER 3 RESULTS Rationale for Study There needs to exist a rapid, specific, and efficient dete ction system of food borne pathogens to address the ease with which pa thogens can infiltrate public food supplies, the growing threat of bioterrorism, and the increased interest in biodefense. Such a system would be of great utility not only for the protection of food, water, and drug supplies for the general public (111). Th e fiber-optic waveguide biosensor uses antibody-based reagents that optimize critical specific ity to pathogenic agents, incorporating both portability and ease of use. This thesis describes the efforts made to isolate recombinant antibody fragments specific for the detection of path ogenic L. monocytogenes , utilizing phage display. The specific aims of this study were: 1. To isolate recombinant phage display antibody fragment s to L. monocytogenes, both whole cells and surface proteins. 2. To optimize and manipulate these imm unological reagents to improve their usefulness in the biosensor system. Specific Aim 1: Isolation of Phage Display Antibody Fragments to Listeria A nave recombinant human antibody library wa s used for the selection of monoclonal scFv antibodies to L. monocytogenes cells in suspension. For panning, 109 listeria cells were suspended in casein blocking buffer, and 1011 TU of Tomlinson I Human Synthetic VH + VL Library was added. Following one hour, four hours, or overnight incubations


46 (see Materials and Methods), th e phage antibodies were elut ed with a low pH glycine buffer. Phages recovered from the elutions were utilized for any subsequent panning rounds or to infect E. coli TG1 for amplification of phages. To grow phages, infected TG1 cells were superinfected with Hyperphage overnight. The following day the phagecontaining culture supernatant was used fo r immunological analysis or was used for concentrating phages by PEG precipitation. An ELISA either with the phage produced from all the pooled clones (polyclonal phage ELISA) or with the phage produced from each individual clone (monoclonal phage ELISA) tested the amplified phages thus obtained. scFv-Phage Antibodies to Whole L. monocytogenes Cells Demonstrate Variable Reactivities L. monocytogenes serovar 1/2a cells were chosen for all panning experiments. Panning was performed using 109 cells rinsed in PBS and su spended in casein blocking buffer. The effects of the addition of 0.1% (w /v) Tween-20 detergent in the wash steps of a panning round were considered in the isola tion of high affinity scFvs. Because the library is comprised of a mixed pool of pha ges displaying scFvs possessing low to high affinities for their target antigen and because we were unsure of the nature of the cell surface target, our initia l goal was to obtain a high quantity of phages irrespective of their affinities for L. monocytogenes, although a high affinity phage was desirable. A high number of wash steps that included dete rgent would remove those phages that bound with low affinity to cell surface epitopes, although this would simultaneously increase the probability of losing potentially useful pha ges (phages with a lower affinity but possessing high specificity). Therefore, we performed the initial panning attempt to


47 whole cells in two parallel experiments: wash steps with detergent or wash steps using PBS alone. For expediency, panning to whole cells wa s performed without th e amplification of eluted phages. Phages that were eluted from the cells in the first round were neutralized then immediately used for the subsequent tw o further rounds of panning to whole cells in suspension. From the original 1011 phages used in the first panning round, no phages were recovered by the end of the third round when using detergent. Forty phages were recovered from panning without using detergent. After the final round of panning, the eluted phages were used to infect E. coli TG1 cells. Because we did not recover phages from the third round using detergent, we used phages eluted from the second round, which had a titer of 700 TU/mL. Fifteen col onies were randomly selected and screened from the two parallel panning procedures. Seven clones, 1-1 through 1-7, were isolated from the panning to whole cells utilizing detergent in the wash steps, and eight clones, 21 through 2-8, were isolated from the panni ng to whole cells using PBS alone. The reactivities of these clones were examined in a whole cell ELISA. S:N ratios were calculated for all phages screened in the ELI SA using the mean absorbance signal of wells that contained antigen di vided by the mean absorbance signal of wells coated with buffer alone (blank wells). An arbitrary S:N cut-off value of 2 was used to initially discriminate positive clones from negative. The ELISA results of the first panning attempt without amplification are summarized in Fig. 3-1A. In our first screening, a single clone, 13, had a S:N twice that of the blank, uncoated wells (S:N=2, n=2). Repeated optim ization of the assay (excluding day-to-day variability) later increased the S:N to 4. Th ere was a significantly greater quantity of


48 1-3* 0 1 2 3 4 5 6 7 8 clone signal to noise 24* 17* 0 1 2 3 4 5 6 7 8 clonesignal to noise B Figure 3-1. ELISA analysis of phages isolated following whole cell pannings to L. monocytogenes serovar 1/2a cells in suspensio n. For the assay, listeria cells were diluted in PBS, coated into mi crowells, and probed with phages (see Materials and Methods). Bound phages we re detected with a mouse anti-M13 secondary antibody conjugated to HRP. Absorbances measured at 630 nm, and used to calculate signal to no ise ratios (mean antigen-coated well absorbance / mean blank well absorbance) for each phage clone. Whole cell panning isolated 160 clones performed in three attempts: (A) Attempt #1, (B) Attempts #2 and #3. (n=2 wells) A


49 phages that eluted when we di d not include detergent in the wash steps, and we therefore reasoned that adding 0.1% Tween required phages to bind with higher affinity. Thus, all later panning attempts included the a ddition of Tween in the wash steps. Subsequent panning experiments to whole cel ls in suspension also included an amplification step as described in the Mate rials and Methods to obtain a higher quantity of phage clones after elution. Two additiona l panning attempts were made to cells in suspension. Three rounds of panning were us ed for each experiment. Phages obtained following each round of panning were titered prior to beginning any subsequent panning round. Each panning experiment to whole cells began with a library of approximately 1011 phages. Following the first round of panning to whole cells with amplification, the concentration of phages el uted was approximately 104 TU/mL. Following three rounds of panning, the number of phages eluted in creased to a final concentration of 107 TU/mL. Listeria specific clones were being enriched. One hundred sixty three of 107 phages were screened to whole L. monocytogenes serovar 1/2a cells by ELISA. Two panning experiments were performed with amplificatio n, and their ELISA results are summarized in Figures 3-1A and 3-1B. Ten of the 163 clones were positive in their detection of listeria, demonstrating S:N ratios between 2 to 6. From the three panning experiments, we selected nine phage clones that demonstrated the highest S:N ratios to characterize their binding reactivities further. Phage clones can be substituted for monocl onal antibodies in most typical antibody applications. The nine phage clones we select ed were used in Western blots as primary detecting antibodies. Listeria cells were boiled in 2% (w/v ) SDS, and the proteins were


50 resolved by SDS-PAGE. These antigens were then transferred to a nitrocellulose membrane and were probed with each of th e nine phages obtained from the whole cell pannings or with a rabbit antiL. monocytogenes polyclonal antibody (PAb). Except for the PAb, there were no visible bands using th e whole cell clones (d ata not shown). The ability of the phages to distinguish L. monocytogenes serovar 1/2a from 10 different serovars of L. monocytogenes as well as L. innocua , L. welshimeri , L. seeligeri , and L. ivanvoii was tested by ELISA. The results in Fig. 3-2 show that reactivity was highly variable, and none of the clones de monstrated outstanding specificity to any Listeria species or serovar. Of the five clones in this figure, all demonstrated a relatively high degree of reactivity to se rovars 1/2a, 1/2b, 1/2c, 3a, 3b, a nd 3c. This reactivity (S:N between 3 to 7) demonstrated S:N ratios twice that of serovars 4a though 4d (S:N of 1 to 2). These reactivity values s lightly rose in detection of L. ivanovii and L. seeligeri (S:N of 2 to 4) and dropped for L. innocua (S:N of 1 to 2). The multispecific PAb, as expected, had relatively high reactiv ities (S:N of 4 to 13) to all Listeria spp . except L. monocytogenes serovar 1/2c (S:N of 2). However, in repeated attempts, the phages tested here reacted equally well with L. monocytogenes serovar 1/2a (the original panning serovar) as with non-pathogenic L. innocua . Repeated attempts did not reproduce any noticeable reactive patterns to specific serova rs or species outside day-to-day variability for either the phages or the PAb.


51 0 2 4 6 8 10 12 14 1-32453831PAbphage clone signal to noise 1/2a 1/2b 1/2c 3a 3b 3c 4a 4b 4c 4d L. ivanovii L. seeligeri L. innocua L. welshimeri 10403S, wt Figure 3-2. ELISA cross-reac tivity of five whole cell Listeria phage clones. Listeria cells were diluted in PBS, coated into microwells, and probed with phages or a rabbit anti-Listeria poly clonal antibody using proce dures described in the Materials and Methods. -M13-HRP or goat anti-ra bbit Ig-HRP antibodies were used as secondary reagents. Key indicates L. monocytogenes serovars 1/2a through 4d and L. monocytogenes 10403S. (n=2 wells)


52 A competitive inhibition ELISA was also performed to demonstrate the binding affinity of the phages to whole L. monocytogenes . The inhibition ELISA allowed for observations of specific or non-sp ecific interactions of the phage clone s with the native states of surface-exposed antigens on the listeria cells. This also mimicked the conditions used in the panning experiment. A quantified amount of phages (109 TU/mL) was first incubated with L. monocytogenes or Enterococcus faecalis cells, then loaded into listeriacoated microwells. The ELISA was then pe rformed using the standard procedures as stated in Material and Methods. Phages c ould either bind to cells within the pre-mix suspension (pre-absorbed phages) or to the cell s coated in the microwells. Pre-absorbed phages were washed away with the cells th ey were bound to, leaving a smaller quantity of phages bound to cells in the microwells. Th erefore, the decreases in absorbance signal values were directly relate d to the quantity of phages removed by pre-absorption. We represented these absorbance values (or inhi bition effect) as a percent decrease. The inhibition effect was calculated using the A630 generated in microwells that received preabsorbed phage to the A630 of microwells that received unabsorbed phage . Inhibition to BSA-coated wells was also measured us ing a phage specific for BSA, not for Listeria , in the competition ELISA. Additionally, E. faecalis was substituted for L. monocytogenes in a parallel competition reaction. The fo rmat of our ELISA did not allow for 109 CFU to be competed with the anti-BSA phage. The results of the inhibition ELISA are represented in Fig. 3-3. By competing 109 TU of clone 1-3 with serial dilutions of L. monocytogenes serovar 1/2a cells, the optimum inhibition was determined to be between 107 to 109 CFU. Later repeated experiments that utilized this same format and conditions never achieved


53 R2 = 0.9979R2 = 0.45260 10 20 30 40 50 60 70 80 90 100 789 log CFU/mL competing cells% inhibition anti-Listeria WC phage + listeria cells anti-Listeria WC phage + E. faecalis anti-BSA phage + listeria cells anti-BSA clone + E. faecalis Linear (anti-Listeria WC phage + listeria cells) Linear (anti-Listeria WC phage + E. faecalis) Figure 3-3. Competition effect of mixing pha ges and cells in an inhibition ELISA. Competing cells were serially diluted from 109 to 107 CFU/mL and mixed with 109 TU/mL of whole cell phage 13 or BSA phage. Following competition, bound phages were detected with -M13-HRP. % inhibition = 100-[(inhibited wells/uninhi bited wells)*100]. Linear regression trendlines indicated for whole cell cl one 1-3 with respective R2-values. (n=2 wells). linear regression (WC phage + listeria cells) linear regression (WC phage + E.faecalis cells )


54 inhibition of the antiListeria clones above 80%. The competition ELISA was then repeated using this range of diluted cells . Inhibition occurred for both the anti-BSA phage and antiListeria phage when pre-mixed with L. monocytogenes but at different levels. At 107 listeria cells, the anti-BSA phage reactivity dropped to 74% (inhibited 26%). The inhibition effect increased an additional 6% when the quantity of cells increased to 108 CFU. In comparison, the inhibition effect of clone 13 increased from 30% to 46% when the cell quantity increased from 107 to 108 CFU. The whole cell clone reactivity dropped to 40% (60% inhibition) wh en the cell concentration increased to 109 CFU/mL. The difference between the antiListeria clone and anti-BSA clone reactivity to 108 listeria cells was significant ( p =0.008) as measured using the two-sample t-test. This suggested that the antiListeria phage had higher specificity for L. monocytogenes cells than the anti-BSA phage. To further understand the interaction of the anti-whol e cell phages with L. monocytogenes cells, we measured the inhibition eff ect in the presence of a unrelated bacterium. Less inhibition was observed in the presence of E. faecalis . Using 107 E. faecalis cells, the anti-BSA phage retained 92% (8% inhibition) of its binding reactivity to its target antigen, BSA, and retained 80% of its binding reactivity when the competing cell concentration was increased to 108 cells. This result contra sted to the results using whole cell clone 1-3 when competed with E. faecalis. Using 107 cells, clone 1-3 retained only 62% of its binding reactivity, and d ecreased to approximately 50% using 109 cells. When we compared the inhibition of the whole cell clone using 108 listeria cells to 108 enterococcus cells, the p -value was 0.064.


55 These results suggested that the inhib ition of the whole cel l clone observed to L. monocytogenes might not be significant when compar ed with other bacterial species. However, a significant difference ( p =0.0041) was seen when we compared the antiListeria clone and the anti-BSA clone in the presence of 108 E. faecalis cells. Therefore, although the antiListeria phage could be absorbed out with listeria cells (no more than 80%) as expected, almost half could also be absorbed out by cells that were not used in the original panning. The ability of the antiListeria phages to bind to protei n or carbohydrate targets on L. monocytogenes serovar 1/2a cells was tested by tr eating the listeria cells with either proteinase K or sodium meta-periodate, respec tively. Proteinase K preferentially cleaves at aliphatic aromatic amino acid residues, al though this cleavage is sometimes ambiguous if the protein being cleaved is not known. Periodate oxidati on has been widely used to oxidize hydroxyl groups to aldehydes. Since th e target antigens were most likely proteinaseous or carbohydrate, following tr eatment by one or the other reagent, the mono-specific antiListeria phages should not have bound to the altered antigen target. This would assist in determining if the ep itope recognized was prot ein or carbohydrate in nature. We devised a simple direct ELISA a nd coated a microtiter plate (96 well) with L. monocytogenes serovar 1/2a cells or BSA, as we ll as no antigen (blank wells). The ELISA was performed with standard pro cedures using the PAb, anti-BSA phage, or whole cell phages as primar y detecting antibodies. The results of the direct treatment of cells in the ELISA are in Table 3-1.


56 Table 3-1. Detection of L. monocytogenes serovar 1/2a with whole cell phages by ELISA following periodate and protei nase treatment of cells. antibody acell treatment bPositive well cBlank wellSignal to Noise (A630)(A630) PAbuntreated 0.90 0.0713 sodium metaperiodate treated 0.70 0.135.3 proteinase K treated 0.84 0.0712BSA phageuntreated 1.4 0.0624 sodium metaperiodate treated 1.4 0.0722 proteinase K treated 1.3 0.0523clone 1-3untreated 0.45 0.114.1 sodium metaperiodate treated 1.2 1.061.1 proteinase K treated 0.15 0.081.9clone 17untreated 0.23 0.064.0 sodium metaperiodate treated 1.2 0.701.7 proteinase K treated 0.07 0.061.1clone 23untreated 0.31 0.065.0 sodium metaperiodate treated 0.85 0.621.4 proteinase K treated 0.09 0.071.4clone 53untreated 0.45 0.085.5 sodium metaperiodate treated 1.2 0.851.4 proteinase K treated 0.22 0.092.4clone 8untreated 0.28 0.073.9 sodium metaperiodate treated 0.90 0.581.6 proteinase K treated 0.09 0.071.3clone 23untreated 0.15 0.062.3 sodium metaperiodate treated 0.69 0.421.6 proteinase K treated 0.08 0.061.3clone 31untreated 0.38 0.073.4 sodium metaperiodate treated 0.82 0.631.3 proteinase K treated 0.11 0.071.5clone 43untreated 0.32 0.074.8 sodium metaperiodate treated 0.32 0.251.3 proteinase K treated 0.07 0.071.0clone 82untreated 0.19 0.063.4 sodium metaperiodate treated 0.12 0.111.1 proteinase K treated 0.08 0.061.3aphages or rabbit antiL. monocytogenes polyclonal antibody as primary antibodybantigens (cells or BSA) treated with PBS, sodium m-periodate, or proteinase K prior to ELISAcmean and standard deviation for n=2 wells.


57 As expected, absorbance signals of the PAb to treated or untreated test wells were positive and unchanged, with only a slight in crease in the background absorbance for periodate-treated blank test wells. There wa s also no major observable absorbance signal variation obtained by pre-treating BSA coated test wells and probing with the anti-BSA phage. Untreated, BSA-coated wells had an absorbance signal of 1.36 and slightly increased following periodate treatment to 1.44. Unexpectedly, there was only a slight decrease following proteinase treatment to 1.25. Proteinase treatment should have removed the BSA. Background signals (blank we lls) were equivalent to untreated blank wells in all cases. These results contrast ed to the results obt ained using the antiListeria whole cell clones (except clones 43 and 82) where their respective absorbance signals significantly increased following periodate treatment, three to five times that of their untreated wells. The respective S:N ratios for these test wells were low, however, reflecting a parallel increase in the background signals. Thes e results were only limited to wells that were incubated with the whole cell clones and not the anti-BSA clone. This was an unexpected discrepancy since all of the phages are essentially the same except for small peptide sequences contained within thei r respective scFvs. Although the periodatetreated test wells for clones 43 and 82 either remained equal to or only slightly lower than the untreated wells, background signals also increased. Following proteinase K treatment, signal absorbances were either reduced to background (S:N=1) or reduced to less than half of the absorbances for untreated wells. Elevation of background signals following pe riodate treatment, therefore, was only limited to the phages recovered from the w hole cell panning and not to the anti-BSA phage. We considered the possibility that prot einase treatment of the listeria cells in the


58 wells directly may have led to cells b ecoming detached from wells, and repeat experiments using the direct treatment of listeria-coated microwells were also inconclusive, indicating that the assay n eeded further optimization. From these experiments, we could not conclude what the nature of the antigen target for the whole cell clones was. The S:N ratios decreased for all clones whether test wells were treated with periodate or proteinase K which suggested that the antigen coul d be a glycoprotein. Several attempts were made to collect soluble scFv from the whole cell phage clones as well as the anti -BSA clone by infecting th e nonsuppressor strain of E. coli , HB2151, following the procedures outlined in the Materials and Methods. The production of soluble scFvs was tested by ELISA by coating a microtit er plate (96 well) with whole L. monocytogenes serovar 1/2a cells. The ELI SA was carried out using the PAb, whole cell phages, their respective sc Fvs, and anti-BSA phage as primary antibodies. Standard ELISA pr otocols were followed to detect scFvs bound to cells. As can be seen in Fig. 3-4, controls were all po sitive, but we did not achieve detection of L. monocytogenes with the whole cell antiListeria scFvs. There was successful detection of L. monocytogenes cells using phages (excep t for clones 23 and 43). We plasmid extracted the phagemid vectors from the E. coli HB2151 cells to first confirm that the pIT2 vector was presen t for each clone and ran the DNA extracts on agarose gels. After confirming the presence of the plasmids, we repeated the growth and induction of the cultures with 1mM IP TG. Following overnight induction, we concentrated the scFv supernatants 10X in Amicon spin concentrators and suspended the concentrates to one-tenth their original volumes in PBS. We tested both the concentrated and unconcentrated supernatants as primary an tibodies in the detection of whole listeria


59 0 5 10 15 20 25signal to noise Figure 3-4. Comparison of phage and scFv reactivities to L. monocytogenes cells by ELISA. Rabbit anti-Listeria polyclonal antibody (PAb), phages (109 TU/mL, and scFvs (culture supernat ant diluted 1:2) were diluted in casein blocking buffer then incubated with li steria-coated microwells. Bound PAb, scFv, and phages were detected with e ither goat anti-rabbit Ig-HRP, protein LHRP or -M13-HRP secondary reagents, re spectively. Anti-BSA scFv and phage signal to noise ratios are based on reactivity to BSA coated microwells. (mean + S.D., n=2 wells) 1-3 17 24 53 8 23 31 43 82 BSA PAb


60 cells by ELISA. There was no signal generated in the ELISA for either the concentrated or the unconcentrated samples for the antiListeria clones; signals for all clones were equal to blank wells. Finally, we asked if the scFv was either pr esent in the supernatant or not secreted and still contained within the E. coli host cells. Although the scFv gene is preceded by a pelB leader sequence to assist in secreting the scFv protein to the periplasm, we anticipated that a fair percentage of the scFv would still be found in the cell cytoplasm. Induced cell supernatants and cells were so lubilized in 2% SDS (w/v) and analyzed by SDS-PAGE. Western blotting was used to an alyze for the presence of scFv in both fractions. The results of the We stern blot are in Figure 3-5. Western blot analysis showed labeling of the 31-kDa scFv in both the cell pellet and supernatant for the anti-BSA scFv using ei ther the protein-L or the anti-c-myc HRPconjugated reagents. There was no apparent sc Fv labeling for the whole cell clone 1-3 in either fraction. These data verified the re sults obtained in the ELISA. Because we confirmed the presence of the pha gemid vector within the cells previously, this suggested that the scFv gene was either not being transcribed or translated for the whole cell clone. Although we could isol ate phages to whole L. monocytogenes by panning in suspension, we found it difficult to determine th eir antigen targets due to the complexity of the bacterial surface. Because of the diffi culty in obtaining a soluble scFv from the whole cell clones, compiled with their appare nt lack of outstanding specificity to any Listeria spp., we decided to pursue the devel opment of clones to known listeria surface antigens.


61 Figure 3-5. Expression and detectio n of clone 1-3 and BSA scFv in E. coli HB2151 supernatant and cell fractions. scFv s were expressed using procedures described in the Materials and Methods. Cell pellet and culture supernatant antigens were separated by SDS-PAGE a nd transferred to a nitrocellulose membrane for Western blotting. (1) BS A scFv culture supernatant, (2) BSA scFv, cell pellet, (3) clone 1-3 culture supernatant, (4) clone 1-3 cell pellet. Antigens were probed with either (A ) protein L-HRP or (B) anti-c-myc monoclonal antibody conjugated to HRP. Molecular weight markers (MWM) are indicated for each blot. A B 25 kD 37 kD 12341234 MWM MWM


62 scFv-Phage Antibodies to L. monocytogenes Flagella Recognize Both Glycosylated and Unglycosylated Epitopes The immobilization of specific surface com ponents on solid supports is an effective strategy for isolating scFv from phage displa yed libraries. As an advantage of over whole cell panning, the antigen in this case is known a nd can therefore be used effectively in several assay forms. Muri ne monoclonal antibodies to the flagellar Hantigens of bacteria have been used extens ively in bacterial identification (112,113). The isolation of bacterial flagella can be performed easily, and the crudely purified flagella obtained can be passively absorbed onto polystyrene immunotubes for the panning of phage libraries. Following the procedures described in the Material and Methods, a stock preparation of flagella from L. monocytogenes serovar 1/2a was produced (Fig. 6). The flagellar preparation was analyzed by SDS-PAGE, using Coomassie blue stain and Tsai and Frasch silver stai n (Fig. 3-7A and 3-7B). The molecular weight of flagellin is seen at 29 kDa, in agreement with result s of others (60). Tsai and Frasch silver staining demonstrated non-staini ng of the flagella band a nd visualization of several proteins of various molecular weights in low concentrations. For panning, immunotubes were coated with the flagella r preparation and blocked with casein blocking buffer. Three rounds of panning without am plification were performed. The ability of selected clones to detect flagella was test ed by ELISA (Fig. 3-8A) using the phages as primary antibodies. Of the 31 phages tested, 19 (61%) were positive in their detection of flagella, having S:N ratio s ranging between 10 to 20. Because of the high percentage of positive phages, we did not test more than the original 31 clones and focused on the 19 clones that demonstrated positive reactivity (S:N > 2).


63 Figure 3-6. Electron micrograph of uranyl acetate negative stain of L. monocytogenes serotype 1/2a flagella preparation. Fl agella were sheared from cells in a blender and harvested by ultracentrifugati on. This preparation was used in all applications for both the isolation and screening of listeri a-specific phages. Black bar indicates 0.5 M. EM pi cture by Henry C. Aldrich, Ph.D., Professor Emeritus, Department of Micr obiology and Cell Science, University of Florida.


64 25 kD 37 kD 25 kD 37 kD Figure 3-7. Analysis of L. monocytogenes serovar 1/2a flagella by SDS-PAGE. Following ultracentrifugation, both the pelleted flagella (lanes 2-5) and supernatant (lanes 8-10) were serially diluted and resolved by SDS-PAGE (12% (w/v) polyacrylamide) . Included on gel are mo lecular weight markers (lanes 1 & 6) and BSA protein (12 g, la ne 7). (A) Coomassie blue-stain, (B) Tsai-Frasch silver stain. B A 1 2 3 45 6 78 9 10 1 2 3 45 6 78 910


65 29 12 11 31 5 4 21 20 19 0 5 10 15 20 25 30 clonesignal to noise A 12 29 11 31 5 4 21 20 19 0 5 10 15 20 25 30 clonesignal to noise B Figure 3-8. ELISA of anti-Li steria flagella phages following panning on immunotubes. Signal to noise ratios were calculated fo r reactivity to microwells coated with listeria flagella (A) or whole flagel lated cells (B). Bound phages were detected with -M13-HRP secondary antibody. We lls were developed with TMB substrate, and their absorbances r ead at 630 nm. S:N were calculated for 31 phage clones (1-31). (n=2 wells)


66 We tested these clones further by ELISA to whole L. monocytogenes serovar 1/2a cells grown in flagella-inducing conditions (2 4C). As can be seen in Figure 3-8B, a majority of the phages that were positive to purified flagella were also positive to flagellated listeria cells. Fourteen of th e nineteen clones were positive to whole cells with S:N ratios ranging from 2 to 13. Thr ee clones (18, 19, and 20) that were negative to flagella in the ELISA were positive to whole cells (S:N of 10 to 12), higher than most clones that exhibited positive reactivity to bot h flagella and whole cells. The ability of the anti-flagella phages to discriminate between the L. monocytogenes and Listeria spp. was tested by ELISA, and as with the whol e cell clones, the results showed high crossreactivity with all serovars and species (dat a not shown). These data may reflect the antigenic homology shared be tween flagellin proteins on all listeria species. To further understand the reactiv ity of the anti-flagella phage s, we additionally tested them to a L. monocytogenes mutant lacking flagella by ELISA. Of the 14 phages positive to whole flagellated cells , except in two cases, all react ed to the flagella mutant cells to varying degrees. The results for si x of the fourteen clones tested are shown in Figure 3-9. Clone 5 had a S:N ratio for the mutant strain equal to the wild-type, while most other clones had S:N ratios significantl y higher to wild-type listeria cells. DNA sequence analysis of all seve n clones showed that they were separate clones. Western blot analysis of the flagel lar preparation and flagellated wild-t ype L. monocytogenes cells was carried out using all 31 phages (Figs. 3-10A and 10B).


67 4 5 11 12 29 31 PAb 0 2 4 6 8 10 12 14 16 18 20 22signal to noise wt, 10403S flaA Figure 3-9. Comparison of ELISA signal to noise ra tios of anti-listeria flagella phages to flagellated and non-flagellated Listeria . Microwells were coated with L. monocytogenes 10403S (Fla+) or the isogenic mutant HEL-403 (Fla-) cells and probed with either PAb or anti-listeri a flagella phage clones (clones 4, 5, 11, 12, 29, and 31). Bound phages and PAb were detected with -M13-HRP or goat anti-rabbit-HRP secondary antibodi es, respectively. Signal to noise ratios are shown with standard devi ations. *Significant differences ( p <0.05) were calculated from signal-to-noise ra tios using the Student t-test. (n=2 wells) * * * *


68 Figure 3-10. Reactivity of an ti-listeria flagella scFv phages to flagella or whole L. monocytogenes serovar 1/2a cells by Western bl ot. (A) flagella or (B) cells were probed with anti-flagella phage clones (1-31) or PAb. Bound antibody or phages were detected with goat an ti-rabbit or mouse anti-M13 secondary monoclonal antibodies conjugated to HR P, respectively. Both positive and negative detection of flagella are pi ctured (29 kDa indicated by arrow). B PAb 4 5 18 19 20 4 A PAb 18 19 20 PAb 5 4 PAb 5 4 18 19 20 A B


69 Nineteen of the 31 phages ha d identical Western blot profiles, showing correct labeling of the 29-kDa flagella band. The We stern blot confirmed the previous ELISA test of these 31 clones. Consequently, this also repeated for clones 18, 19, and 20, which did not react to flagella in the ELISA but inst ead reacted to whole ce lls. By Western blot analysis, clones 18, 19, and 20 did not label the flagella band nor any other protein band. A repeat analysis using We stern blotting to flagellate d and non-flagellated cells demonstrated that, except for a high molecular weight banding pattern observed using the PAb, no antigenic bands were observed when non-flagellated L. monocytogenes cells were reacted with anti-flagella phages (Figure 3-11). Because of the apparent contradiction to the ELISA results that demonstrated recognition of the non-flagellated strain and th is Western blot, we further tested the antiflagella scFvs. Because L. monocytogenes flagellin are glycosylated (36) and because we used intact flagella in our pannings, there ex isted the possibility that some of the scFvphages were recognizing the glycosylation modifications on the flagella and these determinants could be present on other nonf lagellar components. The results of a Western blot using anti-flagella clones 31 (lea st cross-reactive clone to non-flagellated cells) and 5 (most cross-reactiv e) to periodate-treated flage lla are shown in Figure 3-12. Periodate treatment clearly removed recogni tion of the flagella antigen by clone 5, whereas recognition was preserved with clone 3 1. This result suggested that clone 5 was indeed recognizing a glycosylated site. We selected 20 clones to express their scFv and tested their react ivity to flagella by ELISA using the PAb and scFvs as primary anti bodies. Eleven of the twenty scFvs tested


70 250 150 100 75 50 37 25 20 151 2 3 1 2 3 1 2 3MWM1 2 3 3 2 1 Figure 3-11. SDS-PAGE and West ern blot analysis of listeri a flagella, whole flagellated cells, and non-flagelleted cells with antiflagella phages. (1) listeria flagella, (2) L. monocytogenes 104030S wild-type cells, (3) L. monocytogenes HEL-304 ( flaA ) cells. (A) SDS-PAGE (4 -20% w/v polyacrylamide, Coomassie stain), (B) Western blotting of antigens using PAb or phage clones 5, 31, and 4. Bound PAb and phages were detected with goat anti-rabbit IgHRP or -M13-HRP, respectively. A B PAb 5 31 4


71 25 kD 37 kD 25 kD 37 kD Figure 3-12. SDS-PAGE and West ern Blot analysis of period ate-treated listeria flagella using anti-listeria phages. (A) Coomassie blue stain of 10 g of listeria flagella (lane 1) or listeria flagella treated with 4 mM sodium meta-periodate (lane 2). (B) Western blot. Flage lla antigens were transferred to a nitrocellulose membrane and probed with anti-listeria flag ella phage clone 31 (lanes 1 &2) or clone 5 (lanes 3 & 4). Phages were detected with -M13HRP. A B 1 2 MWM MWM 1 2 34


72 had positive S:N ratios above 2 with the majority demonstrat ing S:N ratios from 4 to 8 (Fig. 3-13), relatively low compared to the PAb (S:N=26) or anti -BSA scFv (S:N=24). Except for the PAb, there was no positive detecti on of whole listeria cells with any of the scFvs by whole cell ELISA (data not shown). We selected a single scFv (anti-flagella clone 4) based on it s high S:N value to purified flagella to test furt her as primary antibody to whole L. monocytogenes cells. We concentrated the scFv ten-fold using Amicon ultra spin concentrators, and additionally nickel-affinity purified 500 mL of clone 4 scFv supernatant. From the 500 mL of scFv cell supernatant, approximately 30 mg of scFv was recovered. We repeated the ELISA using the unconcentrated scFv, 10X c oncentrate, and purified scFv (at 10 g/mL) as primary detecting antibodies, as well as anti-f lagella phage and PAb to flagella and whole cells. The results of our analysis ar e summarized in Figures 3-14 and 3-15. By concentrating the scFv ten-fold, the S: N ratio to purified flagella more than doubled from 7 to 15 (n=2, p =0.025). The S:N of 7 was not significantly different (n=2, p =0.82) when we used the purified scFv at 10 g/mL compared to the test wells that received unpurified, unconcentr ated scFv. The S:N fell significantly (S:N=1, n=2, p =0.32) to background levels when the concentration was diluted to 1 g/ml or less. Although concentrating the scFv raised the S:N values to flagella, these results did not repeat in the detection of flagella on whole L. monocytogenes cells. The results showed that at any concentration level of clone 4 scFv, including high amounts (10 g/mL), there was no detection of whole flagellated cells (Fig. 3-15). Conversely, controls were positive in their detection of whole cells. When using scFv, we were not able to achieve S:N ratios close to that obtained with phage.


73 Figure 3-13. Anti-flagella scFv reactivitie s to listeria flagel la by ELISA. Bound PAb and scFvs (crude supernatant diluted 1: 2 casein blocking buffer) were detected with goat anti-rabbit Ig -HRP or protein L-HRP secondary reagents, respectively. Signal to noise ratio s were calculated based on mean absorbances. Anti-BSA scFv signal to noise ratio is based on reactivity to BSA coated microwells. (mean S.D., n=2 wells) 12 13 16 BSA 2 4 10 5 PAb 1128 17 31 0 4 8 12 16 20 24 28 32signal:noise


74 0510152025303540 unconcentrated scFv 10X concentrated scFv 10 g/ml purified scFv 1 g/ml purified scFv 0.1 g/ml purified scFv phage anti-L. monocytogenes PAb signal to noise Figure 3-14. Effect of concentrating anti-f lagella clone 4 scFv for the detection of flagella by ELISA. scFvs were expressed from E. coli HB2151 cells using procedures described in the Materials and Methods. Culture supernatant was either concentrated 10X in Amicon Ultra spin concentrators or nickel-columnpurified. Purified scFv was serially d iluted in ten-fold dilutions prior to addition to microwells. Bound PAb, clone 4 phages, and scFvs were detected with either goat anti-rabbit Ig-HRP, -M13-HRP, or protein L-HRP secondary reagents, respec tively. Microwells were coated with 1 g of crudely purified flagella. (mean S.D., n=2 wells)


75 Figure 3-15. Effect of concentrating anti-f lagella clone 4 scFv for the detection of L. monocytogenes serovar 1/2a cells by ELISA. scFvs were expressed from E. coli HB2151 cells using procedures describe d in the Materials and Methods. Culture supernatant was either concentrated 10X in Amicon Ultra spin concentrators or nickel-columnpurified. Purified scFv was serially diluted in ten-fold dilutions prior to addition to microwells. Bound PAb, clone 4 phages, and scFvs were detected with ei ther goat anti-rabbit Ig-HRP, -M13-HRP, or protein L-HRP secondary reagents, resp ectively. Microwells were coated with 1 g of crudely purified flagella. (mean S.D., n=2 wells) 02468101214 unconcentrated scFv 10X concentrated scFv 10 g/ml purified scFv 1 g/ml purified scFv 0.1 g/ml purified scFv phage anti-L. monocytogenes PAb signal to noise


76 The panning experiments to flagella in immunotubes demonstrated a successful strategy to isolate both phag es and their soluble scFvs qui ckly without the use of amplification. We employed a variation of this panning proc edure to obtain our goal of isolating recombinant an tibodies that recognize L. monocytogenes by combining both panning on immunotubes and panning to cells in suspension. Recombinant lmo1076 (Auto) from L. monocytogenes is Suitable as Antigen for Phage Display Panning Auto is a listeria autolysi n encoded by a 1,716-bp gene ( lmo1076 ) that encodes a protein of a predicted molecular weight of 62 kD a. This protein consists of three distinct regions, (i) an N-terminal 26-amino acid sign al sequence, (ii) an autolysin hydrolase domain from amino acids 75 to 243, (iii) a nd four G-W modules found on the C-terminal end that assist in its association to the b acterial surface. We constructed primers to amplify only the hydrolase domain of Auto us ing PCR. Our panning strategy included panning to whole L. monocytogenes cells, which selected for phages that could bind the amino acids exposed on both recombinant protei n as wells as the native protein on whole cells. This strategy would not require our recombinant protein to necessarily refold properly into a functional enzyme. Additiona lly, because we were subcloning the AutoPCR product into an expression vector in E. coli without the signal sequence, the recombinant protein (rAuto) would be containe d within the cytoplasm where it could be recovered from a cell pellet with relative ease. The PCR-amplified recombinant Auto gene (668 bp) was cloned as an Xho I/ BamH I fragment into a pET19b expressi on vector (forming pGTR1499) and electroporated into E. coli Tuner cells using procedur es stated in the Material s and Methods. Restriction enzyme digestions of pGTR1499 confir med the correct Auto DNA product.


77 Recombinant protein expression in the E. coli Tuner cells was induced with 1 mM IPTG for four hours. To confirm the presence of rAuto, aliquots of cells were removed every hour just prior to and during induction. Cell culture supernatants and cells were examined by SDS-PAGE. The rAuto protein migrated with an apparent molecular weight of 34 kDa and corresponded to the a pproximate calculated molecular weight. rAuto could be visualized w ithin the cell fractions follo wing 4 hour IPTG-induction. The same antigens were transferred to a nitrocellulose membrane and probed initially with an anti-hexa-His monoclonal antibody (specific for His-tagged proteins) conjugated to horseradish peroxidase (Novage n). Additionally, hexa-His-t agged scFv and hexa-Histagged recombinant vaccinia viral protei n, A27L, were included as controls. The anti-hexa-His monoclonal antibody labeled all controls as well as the His-tagged rAuto (Fig. 3-16). Because the protein was not secreted (no signal sequence was cloned in frame with the recombinant protein) th e recombinant protein would have to be harvested from the cell cytoplasm. We induced a larger 1 L volume of cells using 1 mM IPTG for four hours. We then used a comm ercial lysis detergent mixture for extraction of soluble and insoluble protein fractions from the frozen cell pellets as described in the Materials and Methods. Antigens were exam ined by SDS-PAGE (Fig. 3-17). The rAuto protein was contained within the first solubl e fraction collected following treatment with the detergent mixture. We purified this frac tion over a nickel affinity column to purify the His-tagged recombinant protein using step-w ise gradient elutions with imidazole. The majority of recombinant protein eluted at a concentration of 300 mM imidazole with lesser quantities eluted at 500 mM im idazole (Fig. 3-18). Following nickel-column


78 purification, three major ba nds could be visualized corresponding to approximate molecular weights of 34 kDa, 20 kDa, and 17 kDa, as well as several higher and lower 127 80 53 41 27 21 15 11 21 127 80 53 27 15 41 11MWM scFvantibody rA27L vacciniaprotein rAutofirst elution rAutosecond elution MWM scFvantibody rA27L vacciniaprotein rAutofirst elution rAutosecond elution Figure 3-16. SDS-PAGE and Western Blot an alysis of recombinant Auto protein, postaffinity column purification. (A) Coomassi e blue stain of Histagged proteins: scFv antibody fragment (scFv), recombinant vaccinia surface protein (rA27L), and recombinant L. monocytogenes EGDe Auto surface protein (rAuto) and resolved by SDSPAGE (4-20% gradient polyacrylamide). (B ) Antigens were anal yzed by Western Blot using a mouse anti-hexa-histidine m onoclonal antibody conjugated to HRP.


79 100 20 250 150 75 50 37 25 15cells, pre-IPTG cells, post-IPTG supernatant, pre-IPTG supernatant, post-IPTG MWM MWM Figure 3-17. SDS-PAGE analysis of CellLytic extracted recombinant Auto protein from E. coli DE3 Tuner cells. Samples of cells containing the pGTR1499 expression vector and culture supernat ant were collected prior to IPTG induction and at 4 hours post-IPTG induc tion. Cells were frozen prior to extraction with CellLytic lysis reagent. Total cellular material (soluble and insoluble) following lysis (lane 1) was spun down to separate soluble protiens (lane 2). The pellet was further treated with CellLytic detergent supplemented with lysozyme and spun down to collect proteins in the second soluble fraction (lane 3). Repeat tr eatment and centrifugation was used to collect the third soluble fraction (lane 4) . The remaining pellet of insoluble material was suspended in water (lane 5) . The first soluble fraction obtained, containing the Auto protein (34 kDa), was used for nickel-affinity column purificaiton. 1 2 3 4 5


80 127 80 53 41 27 21 15 11 190 1 2 3 4 5 6 7 cells, post-IPTG MWM Figure 3-18. SDS-PAGE analysis of nickel -affinity-column purified recombinant Auto protein following Cell Lytic extraction. Large s cale purification and collection of recombinant Auto protein was obtained by IPTG induction of one liter of E. coli Tuner cells containing pG TR1499 protein expression vector. Following IPTG induction, ce lls were frozen and lysed with CellLytic reagent and centrifuged to obtain the first soluble fraction (column load, lane 1). The recomb inant protein is not predominantly contained within the column flow-thru (l ane 3) or wash fraction (lane 4). A step gradient of 100 mM imidazole (l ane 5), 300 mM imidazole (lane 6), and 500 mM imidazole (lane 7) was used to elute bound proteins from the column matrix. The 300 mM and 500 mM imidazole fractions, containing recombinant Auto protein at 34 kDa, were pooled. The final quantity of recombinant protein obtained was appr oximately 7 mg, determined by DC protein assay. Lanes were loaded with 40 L of samples diluted 1:2 in LSB.


81 molecular weight proteins in relatively low concentrations. The 300 mM imidazole elution was pooled with the 500 mM eluti on and buffer exchanged with PBS using Amicon Ultra spin concentrators (Millipore) . The final quantity of purified rAuto recovered from a liter of induced Tuner cells was approximately 7.6 mg, as determined using the DC Protein assay method. scFv-Phage Antibodies to rAuto Discriminate L. monocytogenes from L. innocua Panning was performed to the rAuto pr otein by coating two immunotubes with 10 g/mL of rAuto suspended in PBS. Panning wa s carried out initially with amplification of eluted phages following the first round. Am plified phages were th en used in a second round of panning to rAuto in an immunotube. We used a third panning round to whole L. monocytogenes cells in suspension to isolate clones that recognized native Auto on whole cells. We chose the genomic DNA sequenced strain, EGDe, from which we originally amplified the rAuto protein for this purpos e. Following a single round of panning to whole cells in suspension, we then perf ormed a negative selec tion by incubating the eluted phages with L. innocua in a second parallel experiment to remove phages that could have nonspecifically bound. From these two attempts, we collected 117 phage s. We tested the ability of 42 clones, 20 clones from the first attempt and 22 cl ones from the second, to recognize rAuto, EGDe cells, and L. innocua cells by ELISA using the PA b as well as the phages as primary antibodies. Using an arbitrary S:N ratio cut-off value of two, 31 of 42 clones were positive to the rAuto pr otein (Fig. 3-19A and Fig. 3-19B). Twenty clones were positive to whole L. monocytogenes , and six clones were positive to L. innocua . Eleven of 20 clones or 70% of phages from the first pool recognized whole L. monocytogenes.


82 12N-1 2N-1 1N-1 Pab 0 2 4 6 8 10 12 14 16 18 20signal to noise rAuto EGD-e cells L.innocua cells clone 20N-3 18N-3 16N-3 2N-3 Pab 0 2 4 6 8 10 12 14 16 18 20 clonesignal to noise rAuto EGD-e cells L.innocua cells Figure 3-19. ELISA of anti-Auto phages following panning to recombinant Auto and whole cells. Phages were tested from two separate pools following panning to recombinant Auto on immunotubes: (A) panning to whole L. monocytogenes EGDe cells and (B) negative panning to L. innocua cells. Bound PAb and phages were detected with either goat anti-rabbit Ig-HRP or -M13-HRP secondary reagents, respectiv ely. Signal-to-noise ratios were calculated based on mean absorbances. Microwells were coated with purif ied rAuto or whole cells at 109 CFU/mL. (n=2 wells) A B


83 cells and not L. innocua . Six of the 20 phages (30%) from the first panning attempt recognized L. innocua . No phages were positive to L. innocua following negative selection. The quantity of phages that recognized L. monocytogenes also decreased following negative selection to 4 out of 22 phages (18%) when we performed the negative selection. Since we ultimately wanted to use scFv and not phage in the detection device, we selected those 20 clones that were positive to L. monocytogenes and not to L. innocua and infected the E. coli nonsuppressor strain HB2151 to produce soluble scFv. The cells were induced with IPTG, and supernatants were collected accordi ng to the procedures described in the Materials and Methods and used in an ELISA as primary detecting antibody. The phage from clone 1N-1 was included as a positive control. As can be seen in Fig. 3-20, seven of the twenty clones made functional scFvs that were positive for rAuto but where negative to whole cells. Western blot analysis was carried out by reacting the anti-Auto scFvs against L. monocytogenes and rAuto (Fig. 3-21). One of the se ven clones (1N-1) labeled the rAuto protein at 34 kDa and the smaller 17-kDa band. None of the seven scFvs detected native Auto on whole cells. The PAb reacts to rA uto, and a band correspond ing near the native 62-kDa Auto protein could be visualized in the whole cell preparation. Without a monoclonal antibody to verify th e native 62-kDa Auto protein in a Western blot, this is only speculative. Conclusion of Specific Aim 1 There was a distinct drawback at panning to unknown targets on whole cells, inherent in whole cell panning. Most notably, because whole cell panning used cells suspended in a buffer, proteins were preser ved in their native states on th e cell’s surface, and therefore


84 0 2 4 6 8 10 12 14 16 18 20 1N-12N-112N-12N-316N-318N-320N-31N-1 rAuto L.monocytogenes Figure 3-20. ELISA of anti-Auto scFv to recombinant Auto and L. monocytogenes EGDe cells. scFvs were expressed from E. coli HB2151 cells using procedures described in Materials and Methods. Bound scFv and phage ( ) were detected with protein L-HRP or -M13-HRP secondary reagents, respectively. Signal to noise ratios we re calculated based on measured mean absorbances. (n=2 wells).


85 25 37 50 75 20 15 25 37 50 75 20 15 Figure 3-21. Western blot analysis of antiAuto scFv clones to both recombinant Auto and L. monocytogenes EGDe cells. (A) SDS-PAGE analysis of recombinant Auto (12% w/v polyacrylamide, Coomassi e stain). (B) Western blot analysis of 2.5 g of recombinant Auto or (C) 107 CFU/mL of whole L. monocytogenes EGDe cells (2% w/v SDS extraction) . Membranes were probed with anti-Auto scFv clones 1N -1, 2N-1, 12N-1, 2N-3, 16N-3, 18N-3, or 20N-3 (lanes 1-7 respectively) or PA b. Bound scFv or PAb were detected with protein L-HRP or goat anti -rabbit Ig-HRP, respectively. A B C PAb 1 2 3 4 5 6 7 1 2 3 4 5 6 7 PAb


86 there existed a higher probability of rec overing scFv that recognized conformational binding domains. We could circumvent this potential problem by selecting specific antigen targets (flagella or Au to) to pan against. Using th is strategy, we were able to demonstrate successful isolation of phages that not only recognized th eir target antigens but also whole L. monocytogenes cells. Ultimately, the use of soluble scFvs in the detection of whole L. monocytogenes cells would require optim ization. In all cases, scFvs obtained from our panning attempts demonstrated lower affinities compared to scFv-phages or whole immunoglobulin. Specific Aim 2: Optimization of Immunological Reagents The sensitivity of the fiber optic biosen sor ultimately relies on specificity and robustness of the immunological reagents it utilizes. Although monoclonal antibodies can be highly specific in th eir interactions with thei r respective antigens, antibody fragments, such as scFvs, often demonstrat e a lowered avidity to antigen compared to whole Ig or Fab (114-116). The collective on-off binding rates of scFvs have an additive effect (avidity) when displayed on the phage end but can be substantially reduced when presented as singular, soluble fragments. The lowered affinity exhibited by scFvs compared to whole Ig can compromise both th e detection limits and sensitivity of an immunoassay regardless of specificity. The low affinity of scFvs can be increased through genetic manipulation and affinity matu ration. Alternatively, simpler methods to enhance the signal generated in immunoassays use antibodies or antibody fragments that have been chemically or genetically tagged with other proteins, enzymes, or reporter molecules. Biotin, horseradish peroxidase , fluorophores, and nanoparticles have all been utilized for this purpose (71).


87 scFv Antibodies Can be Genetically Fused to a Biotin-receiving Peptide (Avitag) Biotinylation of antibody mol ecules with detection with streptavidin conjugates is a common strategy to amplify the signal obtai ned in immunological analysis. Proteins such as scFv or monoclonal antibodies can be ch emically labeled with biotin, albeit in an uncontrolled reaction, to any available amines , even within the an tigen-binding domain. Therefore, we chose to take advantage of the ability of genetically modifying scFv antibodies to enable enzymatic ally me diated and site-directed biotinylation. Avitag is a 15-amino acid sequence that is specifically recognized and biotinylated by the BirA enzyme within E. coli AVB100 cells (Avi dity) where the birA gene is stably integrated into the chromosome. We subcl oned the scFv gene for whole cell clone 1-3, anti-flagella clone 5, a nd anti-BSA scFv as Hin dIII/ Not I fragments into the 3’ multiple cloning site of the pAC5 vector in E. coli AVB100 cells. Production of scFv was induced with IPTG and L-arabinose after adding d-biotin to the culture medium. Antigens were analyzed by SDS-PAGE, transf erred to a nitrocellulose membrane, and probed with SA-HRP. The results are shown in Fig. 3-22. The media containing the expressed scFv contained many proteins of various molecular weights as visualized by SDS-P AGE. A band corresponding to approximately 31 kDa was faintly visible, possibly representi ng the scFv protein. Western blot analysis of the same antigens using streptavidin-hor seradish peroxidase as a primary reagent clearly showed labeling of the 31-kDa scFv for clone 1-3 and BSA clone, confirming the presence of expressed recombinant antibody. There was no apparent band detection of biotinylated scFv clone 5.


88 40 82 125 190 31 17 40 82 125 190 31 17 Figure 3-22. SDS-PAGE and West ern blot analysis of biotin ylated Avitag-scFvs. Whole cell clone 1-3, anti-flagell a clone 5, and anti-BSA sc Fv genes were cloned and expressed fused to an Avitag peptide using the pAC5 expression vector in E. coli AVB100 cells using the procedures described in the Materials and Methods. (1) culture supernatant of anti-Listeria whole cell clone 1-3, (2) anti-flagella clone 5, and (3) anti-B SA clone. (A) SDS-PAGE (10% w/v polyacrylamide, Coomassie stain), (B ) Western blot analysis of Avitag supernatants detected using streptavid in-HRP. Arrow indicates expected size of scFv at approximately 31 kDa. 1 2 3 1 2 3 MWM MWM A B

PAGE 100

89 The abilities of the Avitag scFvs to detect flagella and L. monocytogenes cells were tested by ELISA. Microtiter plates (96 well) were coated with flagella, wild-type L. monocytogenes , and the isogenic flagellar mutant. An ELISA was carried out using PAb well as the Avitag anti-BSA scFv. The result s of the Avitag scFvs used in an ELISA are summarized in Table 3-2. The results show that both clone 1-3 and cl one 5 were positive to whole cells, both the wild-type and mutant, with signal to noi se ratios from 5 to 8. Both clones were negative to flagella. This would seem to veri fy that functional scFv was expressed for the anti-whole cell clone 1-3. The reactivity observed for clone 5 to whole cells was unexpected given both the lack of detection of a scFv on the We stern blot and the lack of a signal to flagella in the ELISA test wells. It is unlikely that a low yield of scFv (which could explain lack of reactivity in both prev ious cases) could also explain the S:N of 6 and 5 to flagellated and non-fl agellated cells, respectively. Th erefore, this would seem to indicate a false positive result in the ELI SA to whole cells. Possible causes were considered to be either an unknown artifact of the induction, such as excessive, unused biotin in the media that could bind to the list eria cells, or a simple reagent problem could be producing the false positive signal. We repeated the ELISA to understand the a pparent false positive result. Controls were carried out which again included the ch emically biotinylated PAb, the Avitag antiBSA scFv, as well as supernatant from cells containing the pAC5 vector without the scFv insert. As can be seen from the Table 3-3, we were able to reproduce results similar to the previous ELISA. All controls were pos itive including the anti -BSA scFv to cells including the vector only supern atant. A repeat ELISA (Table 3-4) also demonstrated

PAGE 101

90 Table 3-2. Detection of liste ria flagella and whole cells by ELISA using Avitag scFvs. Pab-B amean+s.d. dsignal to noise bL. monoc y to g enes fla g ella 0.33 6.0bL. monoc y to g enes serovar 1/2a 0.92 13bL. monoc y to g enes ( flaA ) 0.88 12Avitag BSA cBSA, 1 g/mL 0.75 12Avitag 1-3 cL. monocytogenes flagella 0.07 1.2 L. monocytogenes serovar 1/2a 0.51 7.8 L. monocytogenes ( flaA) 0.53 7.4Avitag 5 cL. monocytogenes flagella 0.07 1.2 L. monocytogenes serovar 1/2a 0.48 5.6 L. monocytogenes ( flaA) 0.49 4.5achemicall y biotin y lated PAbb96-well plate coated with listeria fla g ella or 108CFU/mL of whole L. monoc y to g enes cellscAvita g culture supernatants diluted 1:2 in casein blockin g buffer prior to assa y and detected with SA-HRPdmean and standard deviation for n=2 wells

PAGE 102

91 Table 3-3. Detection of listeria flag ella and cells by ELISA with Avitag fused scFv. Pab-B amean+s.d. dsignal to noise p -value f L. monocytogenes flagella 1.1 17 L. monocytogenes serovar 1/2a 1.1 19 L. monocytogenes ( flaA) 1.3 18Avitag BSA cBSA, 1 g/mL 0.62 9Avitag 1-3 cL. monocytogenes flagella 0.12 20.022754 L. monocytogenes serovar 1/2a 0.39 60.020313 L. monocytogenes ( flaA) 0.44 70.019839Avitag 5 cL. monocytogenes flagella 0.12 20.000107 L. monocytogenes serovar 1/2a 0.36 40.966041 L. monocytogenes ( flaA) 0.39 50.314318vector only eL. monocytogenes flagella 0.11 1 L. monocytogenes serovar 1/2a 0.34 4 L. monocytogenes ( flaA) 0.38 5achemically biotinylated PAb detected with SA-HRPb96-well plate coated with listeria fla g ella or 108CFU/mL of whole L. monoc y to g enes cellscAvita g culture supernatants diluted 1:2 in casein blockin g buffer prior to assa y and detected with SA-HRPdmean and standard deviation for n=2 wellseculture supernatant of E. coli Tuner cells containing pAC5 vector without scFv insertfp-value based on comparison to respective vector onl y data

PAGE 103

92 Table 3-4. Detection of listeria flag ella and cells by ELISA with Avitag fused scFv. Pab-B amean+s.d. dsignal to noise p -value f L. monocytogenes flagella, 1 g 1.1 18 L. monocytogenes serovar 1/2a 1.3 19 L. monocytogenes ( flaA) 1.3 20Avitag BSA cL. monocytogenes flagella, 1 g 0.10 10.007678 L. monocytogenes serovar 1/2a 0.61 110.008054 L. monocytogenes ( flaA) 0.58 100.000264Avitag 5 cL. monocytogenes flagella, 1 g 0.13 20.387372 L. monocytogenes serovar 1/2a 0.61 110.014855 L. monocytogenes ( flaA) 0.59 110.079490blocking buffer alone eL. monocytogenes flagella, 1 g 0.10 1 L. monocytogenes serovar 1/2a 0.63 9 L. monocytogenes ( flaA) 0.60 9achemicall y biotin y lated PAb detected with SA-HRPb96-well plate coated with listeria fla g ella or 108CFU/mL of whole L. monoc y to g enes cellscAvita g culture supernatants diluted 1:2 in casein blockin g buffer prior to assa y and detected with SA-HRPdmean and standard deviation for n=2 wellseno primary antibody addedfp-value based on comparison to respective buffer onl y data

PAGE 104

93 that there was clear evidence that the abse nce of primary antibody (blocking buffer alone) also created a positive result to whole cells but not flagella alone with a S:N of 9,close to that of clone 5 scFv (S:N=11). Although we could induce expression of the anti-BSA and clone 1-3 scFv in the Avitag system, we cannot confirm that cl one 1-3 was functional until the ELISA is optimized. These data suggested that although cloni ng was successful as evident from the functional BSA clone, there was little to no pr esence of functional, soluble biotinylated clone 5 scFv. It was therefore inconclusive if the whole cell cl one 1-3 was functional using this assay format, given its positive reac tivity to whole cells, a scenario we could repeat by using Avitag BSA scFv supernatant to whole cells. We demonstrated that reactivity to whole cells coul d be reproduced without the presence of antibody by loading diluted SA:HRP alone into test wells. The a ssay for using Avitag biotinylated reagents to whole listeria cells would ha ve to be optimized, and there needs to be better understanding of the surface antigens that could possibly be interactive with streptavidin or HRP to gain further insight into this phenomenon. Improved Detection of L. monocytogenes Cells Through Chemical Biotinylation of scFv We attempted the chemical addition of bio tin onto the scFv in parallel work to genetically fusing the scFv gene to the biotinylation cassete utilizing the Avitag in vivo biotinylation system. Chemical biotinylati on, though not as precise as the Avitag system, would allow us to easily modify and test qua ntified amounts of anti body. We used in our biotinylation reactions the esterified reactive form of biotin, (+)-biotin N hydroxysuccinimide ester (NHS-biotin). In th is form and in the presence of a high pH buffer, primary amines on target proteins can be covalently reacted with biotin. Excess

PAGE 105

94 biotin is easily removed, either through dialysis or through re peated suspensions in filter devices. Stoichiometry of biotinylation can be then measured per target using commercial kits that use HABA, a natural av idin ligand, as a competitor for biotin. Those scFv clones that demonstrated utility in previous assays were grown in large volumes and nickel column purified. Followi ng purification they were buffer exchanged with a carbonate-bicarbonate buffer to rem ove excess imidazole and to increase the pH that allows for biotinylation. Both the scFv and the PAb were biotinylated following the procedure outlined in the Methods and Materials. We first attempted to enhance absorbance signals in the whole cell and flagellar ELISAs by using chemically biotinylated anti-f lagella clone 4 scFv antibody. We diluted both the unmodified and biotinylated scFv antibodies to 50 g/mL and 10 g/mL. Protein L-HRP was used as secondary reagen t to detect unmodified scFv antibodies, while streptavidin-HRP was us ed as secondary antibody to detect biotinylated scFv antibodies (scFv-B). The results of the modi fied ELISA using concentrated scFv-B are summarized in Table 3-5. In test wells coat ed with purified flagella, at 50 g/mL the unmodified scFv yielded absorbances (0.705) hi gher than scFv-B (0.459). This was also true using a lower concentrati on of unmodified scFv at 10 g/mL, implying that although biotinylation of the scFv was accomplished su ccessfully, the scFv-B did not significantly improve detection. There was a measurable di fference seen when the scFv-B was used in test wells coated with listeria whole cells. In these test wells, the scFv-B had higher absorbances at 50 g/mL and 10 g/mL con centrations than unmodifed scFv with absorbance signals of 0.626 and 0.381 respectiv ely, compared to unmodified scFv, at

PAGE 106

95 Table 3-5. Detection of liste ria flagella by ELISA using chemically-biotinylated antiflagella scFv. mean + s.d. dsignal to noise unmodified anti-Listeria flagella scFvb, 50 ug/mL0.10 0.001 2 0.08 0.003 1biotinylated anti-Listeria flagella scFvb, 50 ug/mL0.63 0.016 5biotinylated anti-Listeria flagella scFvb, 10 ug/mL0.38 0.006 5unmodified rabbit anti-L. monocytogenes polyclonalb1.8 0.029 23biotinylated rabbit anti-L. monocytogenes polyclonalb1.5 0.042 24anti-BSA scFv c unmodified0.33 0.024 5anti-BSA phage c 1.4 0.053 17a96-well plate coated with 1 g of listeria fla g ellabchemicall y biotin y lated antibodies detected with SA-HRPcanti-BSA scFv culture supernatant and pha g es are used unmeasured and untitered, respectivel y dmean and standard deviation for n=2 wells unmodified anti-Listeria flagella scFvb, 10 ug/mL

PAGE 107

96 0.103 and 0.083 respectively. Alth ough the absorbance signals we re high to whole cells using scFv-B (S:N = 5) at a concentration of 10 g/mL, the calculated S:N ratios demonstrated a plateau effect when the scFv -B concentration was increased 5-fold to 50 g/mL. S:N ratios did not increase linearly with increasing concentrations of scFvB,starting at a S:N of 5 (at 10 g/mL) and ending with a S:N of 5 (at 50 g/mL). Therefore a saturation level of bound scFv-B antibody was achieved. Simply increasing the quantity of scFv-B did not raise S:N levels above what could be achieved with less antibody, shown in this assay to be ar ound 10 g/mL. Although we achieved higher absorbances with chemically biotinylated scFv, these S:N ratios and absorbances were not higher than using the identical scFv still fused to phage (A630=1.14, S:N=17). Anti-Listeria Auto Streptabody Has Improved Detection of rAuto Multivalent immunoreagents can be construc ted as a means to increase the binding efficiencies of scFv antibodies to their re spective antigens. Because scFv antibodies demonstrate lowered binding a ffinity, many methods have been utilized to increase it through affinity maturation, chain shuffling, or ligand fusions. Most of these methods, though useful, require a high amount of manipula tion and therefore prone to high rates of error. Because the binding inter action of biotin to tetrameric streptavidin is one of the highest noncovalent bonds found naturally (Kd=10-15 M), biotinylated proteins can be tetramerized to streptavidin easily in several ways. We c onstructed a streptabody (StAb) using chemically biotinylated scFv that wa s derived from our phage displayed panning experiments to L. monocytogenes Auto protein. The StAb clone generated, st18N-3B, was applied in an ELISA to analyze utility.

PAGE 108

97 We chemically biotinylated antiListeria Auto clone 18N-3 scFv using NHS-biotin ester as described in the Materials and Me thods, and reactivity wa s confirmed to rAuto via ELISA (data not shown). We used a four-m olar excess of biotinylated scFv (scFv-B) to commercial streptavidin-c onjugated horseradish peroxi dase (SA-HRP) based on the molecular weights of both scFv protein (~31 kDa) and SA-HRP (> 173 kDa). Cloutier et al, demonstrated that 3:1 to 6:1 molar ratios of scFv-B to SA:HRP will still form useable StAbs (117). We tested the ability of clone 18N-3 as phage, unmodified scFv, biotinylated scFv, or StAb, as well as no antibody (SA-HRP alone) to recognize rAuto, whole L. monocytogenes EGDe cells, and L. innocua by ELISA. Since the StAb essentially constitutes a one-step reagent, it was added in the final 30 minutes of incubation during the ELISA when all sec ondary reagents were added. Absorbance readings were taken every 10 minutes for 30 mi nutes and also included a final reading at 60 minutes. The results of the StAb reactivit y to rAuto by ELISA are summarized in Fig. 3-23. The StAb showed excellent binding to the recombinant Auto protein by ELISA. To recombinant Auto, st18N-3B had absorbance signals higher than scFv, scFv-B, or phage through all time points up to 30 minutes at which time it plat eaus and begins to decline. At 30 minutes the StAb had reach ed a maximum absorbance of 1.58 compared to the phage at 1.35 ( p =0.0035). This absorbance valu e was significantly higher than scFv (A630=1.21, p =4.1x10-5) or scFv-B (A630=1.26, p =5.9x10-5). At 60 minutes, scFv, scFv-B, and phage all have absorbance values th at continued to rise without increases in background noise in the blank test wells. Although these values are significantly different, they only represent 20% differences.

PAGE 109

98 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0102030405060 (minutes)Absorbance, 630nm phage scFv scFv-B stAb SA:HRP Figure 3-23. Detection of recombinant Auto protein using using anti-Auto phages, scFv, biotinylated scFv (scFv-B), and stre ptabody (stAb) via an ELISA. The ELISA was performed as described in the Materials and Methods. One microgram of recombinant Auto was coat ed microwells and probed with antiAuto immunoreagents standardized to 5 g/mL. Untitered phage preparation was diluted 1:2 prior to in cubation. scFvs and scFv-B detected with SA-HRP secondary reagent. Ph ages detected with -M13-HRP. Absorbance readings were taken every ten minutes up to one hour. SA-HRP in the legend indicates that no primary antibody was us ed. *Significant differences ( p <0.05) were calculated from signal-to-noi se ratios using the Stude nt t-test. (mean S.D., n=4 wells). * *

PAGE 110

99 Background noise was taken into account for th e baseline reactivity of SA-HRP to rAuto and whole cells in this ELISA, and a slight increase was observed over the required time period. There was an increase observed fr om 0.055 to 0.160 within 30 minutes of development time, approximately equal to the background noise for the StAb at 30minutes. Although the actual co ncentration of unreacted SA-H RP in the StAb prep is indeterminable in the format of this assay, it is assumed to be less than the 5 g/mL because unreacted SA-HRP only constitutes a fraction of the StAb, used here at a working dilution of 5 g/mL. The absorbance values obtained for the StAb therefore may actually be closer in value to the abso rbance values of the phage to rAuto. In a parallel ELISA, we used all r eagents in the detection of whole L. monocytogenes and L. innocua (Fig. 3-24A and 3-24B). The StAb did not perform as well in the detection of whole cells comp ared to the other reagents used. To L. monocytogenes EGDe cells, we demonstrated that th e phage had significant absorbance values higher than scFv, scFv-B, or StAb through all time points. At 60 minutes, the mean phage absorbance had increased to 1.18 compared to the StAb (A630=0.33), scFv-B (A630=0.21), or scFv (A630=0.18). It can be seen that alt hough it appears that the StAb is functional (significant reactivity to rAuto), we could not de monstrate outstanding utility in the detection of whole cells. For L. innocua , all absorbance signals were not significantly different to SA-H RP alone, and signal-to-noise ratios never rose to more than 1.5 that of background noise fo r all reagents, as expected ( L. innocua lacks native Auto). Although the StAb did not e nhance absorbance signals to whole L. monocytogenes significantly compared to unmodif ied scFv, its reactivity to rAuto following only 30 minutes of incubation in the ELISA was substantial.

PAGE 111

100 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0102030405060 (minutes)Absorbance, 630nm phage scFv scFv-B stAb SA:HRP 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0102030405060 (minutes)Absorbance, 630nm phage scFv scFv-B stAb SA:HRP Figure 3-24. Detection of L. monocytogenes EGDe and L. innocua cells via ELISA using phages, scFvs, scFv-B, and stAbs. The whole cell ELISA was performed as described in the Materials and Methods to 108 CFU/mL of both (A) L. monocytogenes EGDe and (B) L. innocua cells. Whole cell-coated microwells were probed with antibody fr agments standardized to 5 ug/mL in casein blocking buffer. The untitered pha ge preparation was diluted 1:2 prior to incubation. scFvs and scFv-B detect ed with SA-HRP secondary reagent. Phages detected with -M13-HRP. Absorbance readings were taken every ten minutes up to one hour. SA-HRP in the legend indicates that no primary antibody was used. *Significant differences ( p <0.05) were calculated from signal-to-noise ratios using the Stude nt t-test. (mean S.D., n=4 wells). B A * * * *

PAGE 112

101 Phage displayed technology is a fast and e fficient method used for the isolation of recombinant monoclonal antibody fragments by panning the phage library on a desired antigen. In this procedure antibodies can be isolated very rapidly, when compared to using classic hybridoma technology. The To mlinson I scFv phage display library wasused for panning to whole cells as well as purified listeria surf ace antigens to isolate antibodies for recognizing L. monocytogenes . We presented data that although we could isolate phages to whole listeria , these phages did not demonstr ate outstanding specificity. Because of the complex nature of the bacter ial surface, it was difficult to determine their target antigen. Panning for phages to purified antigen such as flagella or the murein hydrolase, Auto, proved to be more successful , and soluble scFvs obtained demonstrated an improved specificity to their target antig ens than our whole ce ll phages. Although the scFvs in the scope of our research possesed ge nerally low affinities in their detection of whole bacterial cells compared to their respective phages or whole Ig, genetic or chemical biotinylation showed poten tial in improving assay sensitivity.

PAGE 113

102 CHAPTER 4 DISCUSSION The Foodborne Diseases Active Surveill ance Network (FoodNet), operated under the Centers for Disease Control and Preventi on (CDC, Atlanta, GA), estimates that there are over 76 million cases of illnesses, 352,000 hos pitalizations, and 5,000 deaths attributo food borne diseases in the United States a nnually (118). Through an active surveillance in 1997 by Food Net and the CDC, 2,500 cases of listeriosis and 500 deaths were reported annually in this country (11). A high mortality rate is in part due to immunocompromised patients, who are in fact some of the most at risk to Listeria , succumbing to infection. It is the ubiquitous natu re and the relative ease of isolating some of these organisms, including L. monocytogenes , from the environment that mark them as important targets for food safety. Although human cases of listeriosis are sporadic and trea , detection of Listeria species remains important in the improvement of public health, especially among those most susceptible to this disease. To address the growing need for quick pathogen detection, the Naval Research Laboratory's Center for Bio/Molecular Scie nce and Engineering developed the first commercially available antibody-based fiber optic biosensor for the detection of pathogens and their toxins, the RAPTOR (75) (Research Internati onal, Woodinville, WA) (111). This biosensor is a fully automated, field-ready, path ogen detection system based on the principles of a sandwich fluoroimm unoassay. The RAPTOR utilizes fiber optic probes coated with antibodies that bind target molecules and excite bound fluorophorelabeled secondary antibodies with an evanes cent wave emitted from a laser. On-board

PAGE 114

103 software discriminates signal generated from the bound fluorophores and provides the operator with easily readable, un-biased resu lts displayed on an LCD screen within 20 minutes. Systems such as the biosensor provi de an important tool that is rapid and reliable in the detecti on of food pathogens. The fiber optic biosensor ultimately relies on the use of highly specific antibodies. Phage-displayed technology is a fast and e fficient method used for the isolation of recombinant monoclonal antibody fragment s bound to antigen by panning the phage library to a desired antigen. In this proce dure, antibodies can be isolated very rapidly when compared to using classic hybridoma technology. The Tomlinson I scFv phage display library was used for panning whole Listeria species and purified listeria surface antigens to isolate antibodies th at recognize listeria epitopes. This thesis describes efforts to isolate r ecombinant antibodies using bacterial cells, flagella, and a recombinant listeria surface protein for potential use in the biosensor system using phage display. Specific Aim 1: Isolation of Phage Display Antibody Fragments to Listeria Phage display has proven to be a useful t ool for the isolation of antibody fragments with high specificity (119-121). The re sults discussed here represent panning experiments that combined factors that woul d lead to quick and successful isolation of highly specific antibody fragments using phage display. Whole cell panning in suspension took advantage of the easy manipul ations that could be performed in a panning experiment using both amplification and non-amplification of phages. Phages displaying specific scFv could of ten be recovered in a single day without amplification or two weeks if performing amplif ication. Cells with bound phages could be easily pelleted by centrifugation, and the unbound phages remove d; thus allowing for titers to be

PAGE 115

104 calculated. These phage titers we re useful in the determining the efficiency of a panning experiment, as well as estimating whethe r enrichment had been achieved following amplification rounds. The ti ter data collected from ite rative panning rounds provided indirect evidence that enrichment was ach ieved in our panning e xperiments to whole cells. After the first rou nd of panning, approximately 104 TU of phage were recovered. But following the third round for panning, this number increased to 107 TU. Thus, a clone that could bind to the listeria cell surface was elute d, amplified, and its quantity expanded; therefore the titer increa sed. Theoretically, the original 104 phages now existed as 1,000-fold copies, increa sing the probability of selecting a L. monocytogenes specific phage. There were other additional advantages of phage display panning to whole cells in suspension. Cells did not have to be bound to solid supports, which could potentially limit available exposed epitopes. Therefor e, panning in suspension to whole cells targeted only surface-exposed antigens. There have been dozens of papers attempting to accomplish the generation of monoclonal antibodies to L. monocytogenes over the past decades by using classic hybridoma techniques. Because success has been elusive, this may suggest that L. monocytogenes lacks surface epitopes that are unique and antigenic, or that such epitopes are not effec tively processed and presented during in vivo antibody generation and maturation. Consider ing the intracellular nature of L. monocytogenes , phage display panning to whole cells demonstrated that antib odies could be generated to antigenic-surface targets potentially not addressed in a humoral immune response (122,123).

PAGE 116

105 We attempted three panning experiments to L. monocytogenes in suspension using both amplification and non-amp lification of recovered phages. The results reported here provide evidence that although we could is olate phage clones th at could identify L. monocytogenes , their reactivity was not specific, thus limiting their usefulness in the biosensor. In our panning efforts to L. monocytogenes we did not measure the increases in specificity to a known serotype followi ng each panning round, and therefore it would be difficult to ascertain whether amplificati on enriched for specific phages in our panning effort, although, as stated above, we could measure an increase in the quantity of phages following each amplified round, implying that enrichment was achieved (to an unknown target). By including amplification steps in our experiments, we had hoped to select phage clones that were enriched to pathogenic L. monocytogenes following multiple rounds of panning. Remarkably, increasing the number of panning round s past three in a whole cell suspension, such as we used, does not necessarily increas e the probability of isolating phages more specific to pathogenic L. monocytogenes (119). Thus, panning is an attempt to select for phages (using amplif ication or not) within a given population that have higher biding affinities to a target than among other phage s in the library or pool. To provide more direct evidence of en richment, we screened 160 clones following three rounds of panning from both amplified and non-amplified pools, and all clones showed mediocre (S:N of 2) to no reactivity to L. monocytogenes (Fig. 3-1). Our data show that amplification steps, though increa sing the total number of phages isolated, did not improve our probability of finding a clone specific to even the serotype used in the panning experiments, perhaps indicative of the shared surface homology between serotypes. It was not unusual to observe ELISA absorbance signa ls to non-pathogenic L.

PAGE 117

106 seemlier , L. innocua , and L. welsimeri as high for L. monocytogenes serovar 1/2a (the serovar used in our pannings) from our five best clones (Fig. 3-2) . Additionally, ELISA absorbance readings to L. monocytogenes were generally low in value. In our hands and under the conditions of the ELISA we used, we did not observe S:N greater than 5 (based on absorbances) for phages with binding to w hole cells. Day-to-day assay variability could either raise or lower these values but not significantly. Lo w reactivity to whole cells could be attributed to a low quantity of available binding antigens on the surface of L. monocytogenes ; however several other factors need to be accounted for, including the listeria preparation used for the assay, the av ailability of epitopes in the ELISA format used, and the health of the cells in the microwells pr ior to using the ELISA. Furthermore, there should be further ch aracterization and better knowledge of the antigenic structure of L. monocytogenes in relation to serovars and species, though some attempts have been made (124). Using 2D-SDS PAGE and MALDI-TOF MS, Trost et al. proposed that, there are 138 secreted proteins of L. monocytogenes , a potential pool not addressed by whole cell panning (7). Wh en these factors, among others, are taken into consideration, phage display could perh aps isolate scFv recognizing surface antigens expressed under controlled conditions in vitro . Ultimately, panning to whole cells in suspension proved to be practical. The useful ness of this sort of panning strategy is illustrated well by Brewster and colleagues (119) , who included a final subtractive round of panning to L. innocua and L. ivanovii in their successful isolation of an L. monocytogenes -specific scFv to address cross-reac tive antigens from a phage displayed library.

PAGE 118

107 There was a distinct drawback, inherent in whole cell panning, to panning to unknown targets on whole cells. Most notabl y, because whole cell panning used cells suspended in a buffer, proteins were preserved in their native states on the cell surface, and therefore there existed a higher probabi lity of recovering scFv that recognized conformational binding doma ins. Though detection of L. monocytogenes was achieved via ELISA using the phages obtained, further investigative techniques used to identify the target antigen were negative. Except for our polyclonal antibody c ontrol, Western blots to listeria surface extracts using phage as pr imary antibody were visually negative (data not shown). Such a result, at best, could confirm the af orementioned point; indeed the phages obtained in our panning experiments we re to native surface exposed antigens. Thus, we could not identify the target antig en of the phages isolated from our panning attempts, as the Western blot procedures we used intrinsically used denatured protein (protein in the presence of SDS and reducing agents). A dditionally, the reactivity of monoclonal antibodies in a Western blot to lis teria surface proteins can be variable based on the simple addition of detergent or salt to cell suspension steps prior to loading a polyacrylamide gel in an SDSPAGE. Extraction methods can be optimized if the target protein is known. Thus, we investigated use of other methods, other than a Western blot. To take advantage of the native nature of the antigen exposed on the surface of the cells, an inhibition ELISA was used in whic h the anti-listeria phages were mixed with whole listeria cells in serial fold dilutions pr ior to their addition to listeria-coated ELISA microwells (Fig. 3-3). As expected, we could effectively inhi bit the binding of antilisteria phage clones when mixed with Listeria cells, and that this effect (approximately 50% inhibition of phages when mixed with 108 cells) was significantly higher (as

PAGE 119

108 determined by Student t-test evaluations of S:N values) than the inhibition seen when we used phages not specific for Listeria but BSA (approximately 20% inhibition of phages when mixed with 108 cells). Our data would seem to i ndicate then that a specific binding did occur for the antiListeria phages and that the bindi ng was not an aberrant nonspecific side effect due to the random select ion of low affinity scFvs within the phage library. However, this inhibition effect could be repeated by using Enterococcus faecalis , a non-related, gram-positive bacterium, demonstra ting that there is effective inhibition of the anti-listeria phage clones (o r their respective soluble scFv antibodies). These results implies that the specificity of the phage clones is to a common or homologous antigen present on both E. faecalis and L. monocytogenes . Such a result could be problematic in a real world sample (such as food) likely c ontaminated or enriched with other bacteria and therefore of limited use for the biosensor system. Because the surface of Listeria cells is highly decorate d with lipoteichoic and teichoic acids and glycoproteins, there was th e possibility that the phages isolated from the whole cell panning recognized carbohydr ates or carbohydrate modifications on proteins. This argument could explain the abse nce of a Western blot result as well, as carbohydrates may not have been extracted effi ciently and therefore were unavailable for the phage to bind to. Therefore, to further investigate the binding conditions of the antilisteria phages, we used a st rategy that would se lectively destroy the target antigen by treating the listeria cells w ith either proteinase K or sodium meta-periodate. By destroying the antigen with either of these tw o reagents, we could then determine whether the target recognized by the phages wa s either a protein or carbohydrate.

PAGE 120

109 Proteinase K preferentially cleaves ali phatic residues of aromatic amino acids, although the specific cleavage locations were ambiguous in our experiments because we did not know what the target pr otein was. ELISA absorbance signals were dramatically decreased following treatment of listeria-coated wells with proteinase K, suggesting that the target epitope recogni zed by the phages was a protei n. However, because the absorbance signals were reduced to the level of blank, untreated wells , we could not rule out the possibility that the proteins anchoring the cells to the plate were cleaved, therefore releasing the most if not all of the cells; thus the cells remaining in the wells were reduced to an undeteclevel with the phages using ELISA. We combined the proteinase K strategy with periodate oxidation of carbohydrates, which is widely used to oxidize hydroxyl groups to aldehyde s (125,126). Under mild conditions with a reduced concentration of periodate and shorter reaction time, oxidation can be restricted to select hydr oxyl groups of carbohydrates or glycoproteins. Under normal periodate oxidation conditions followed by a mild acid treatment (also known as Smith Degradation) all residues with adjacent hydroxyl groups are oxidized to aldehydes. However, the use of periodate reagent should be used in a controlled manner, especially if the carbohydrate make-up of th e antigen in question is unde termined. If an excess of sodium metaperiodate is adde d or incubated too long, the ox idation can begin to affect the amines on the protein and may lead to pr otein degradation or even favor an amine linkage to the polystyrene. The principles that can affect adsorp tion to polystyrene include non-covalent interactions of macromolecule s to its surface, such as weak van der Waals bonds and strong hydrogen bonds. These intermol ecular bindings are dependent on the

PAGE 121

110 intramolecular electric polarities that reside on the adsorbi ng molecule. Thus, in our immunoassays, treatment of coated-microwells with periodate could have affected the critical adsorption of macromolecules, especi ally the scFv. Was this evident in the significantly increased background levels in periodate-treated wells that were not coated with antigen but incubated with only phages, appearing to cause sticky blank wells (Table 3-1). Reports of elevated backgrounds follo wing periodate treatment are rare in current literature and may very well be unreported. Such may be the case from a very similar method of periodate treatment of cells direc tly in microwells used by Brigmon et al. (127) to eliminate the antigenic structur e of the epitope reco gnized by a monoclonal antibody to Fibrobacter succinogenes . For our purposes, optimization of the ELISA using direct periodate treatment would be desirable. Without further research, the usefulness of the results reported in our ELI SA using periodate is questionable in its current state. A reasonable and simple modification could employ treatment of the listeria cells in a microcentrifuge tube prior to their coating in the microwells. In this scenario, extended incubations of periodate in the blank microwells is eliminated wholly, and excess proteinase K or peri odate can be washed from the cells prior to their addition to microwells. However, it is unknown how ef ficiently proteinase K treated cells would coat a polystyrene microwell. Because we ultimately wanted to use the so luble form of the scFv and not its form fused to the phage, we made several attemp ts at inducing scFv expression in HB2151 TAG (Amber) stop non-s uppressor strain of E. coli infected with our whole cell phage clones. Although we could confirm the presence of the pIT2 phagemid vector in the cells by extracting the plasmids and verifying their presence in agarose ge ls (data not shown),

PAGE 122

111 we could not verify the presence of scFv in either the induced cu lture supernatant or within the cells. In all cases we used reag ent probes specific for recombinant scFv, either an anti-c-myc HRP conjugated antibody or protein L-HRP. Additionally, our control scFv (to BSA) was found in both the supernatan t and within the cells, as expected (Fig. 35). Our observations led us to examine what possibilities could explain the lack of soluble scFv in an induced culture, a common problem in phage display. Suzuki et al. (120) observed the phenomenon of scFv that bind to antigen targets as phage but not as soluble scFv, and addressed the issue of apparent variability in scFv expression in non-suppressor cells. Upon insp ection of the sequenced phagemid vectors from their panning experiments, they noted nonsense frameshift mutations within the complementary binding domains or within the framework of the scFv genes. Furthermore, these mutations were supressed in the E. coli suppressor strain TG1, and thus scFv-gpIII genes could be properly tr anscribed and translat ed into a functional antibody (or even scFv-pIII) (120). We reasone d that a transcripti onal or translational error could have occurred with the anti-l isteria whole cell clones but we did not investigate this further in light of the poor reactivity of the phages to L. monocytogenes . This highlights an important point though; expression of selected clones is normally ensured in the TG1 suppressor strain of E. coli in those scFvs that contain TAG stop codons, regardless of the location of the stop c odon (TG1 is able to suppress termination and introduce a glutamate residue at thes e positions). When these phage clones containing multiple stop codons within thei r scFv are used to infect HB2151 (a nonsuppressor strain) to give soluble expressi on of antibody fragments, scFvs that contain TAG stop codons will not produce any soluble sc Fv. Thus, inspection of the sequenced

PAGE 123

112 whole cell scFvs for TAG stop codons outside of the scFv-pIII fram e, could potentially answer questions as to why there was apparent lack of translation of the scFv gene for the whole cell clones in our experiments; perh aps they possessed multiple stop codons, and thus no scFv was expressed. We showed that, although we could quick ly recover phages by panning to whole cells, further research needs to optimize th e protocols involving formation of soluble scFv. By using panning selections to w hole cells, we wanted to isolate antibody fragments with high specificity to L. monocytogenes, but in fact, the complex nature of the bacterial cell surface impeded further char acterization of antibody phage clones in the framework of our research. Given these resu lts, successful selection of specific scFv would require selection of a specific surface antigen. As stated above, there are many factors th at contribute surface proteins and other bacterial surface components, such as flagella . Because bacterial fl agellar proteins are often antigenically heterogeneous among bacter ial species, and because they have been extensively characterized using monoclonal antibodies (128-131), we selected listeria flagella protein as a target for panning to achieve listeria-specific scFvs. We isolated listeria flagella by a simple shearing procedure using a blender while confirming the purity of the preparation by pol yacrylamide gels stained with Coomassie blue. These results agreed with the results of Peel et al. (60,130) , showing a protein band at 29 kDa. The Tsai-Frasch silver stain revealed that many smaller quantities of carbohydrates or glycoprotein contaminants of various molecular weights were within the flagella preparation (Fig. 3-7B). We were not concerned, however, about these contaminants; theoretically, if a phage were to be isolated to one of these minor

PAGE 124

113 components, the antibody would still have potential utility in the detection of L. monocytogenes in the biosensor, because all the co mponents in the flagella preparation originated from L. monocytogenes . After adsorbing the flagellum prepar ation to immunotubes and panning using amplification steps, we isolated and tested 31 phages. A high number of these phages (61%) bound to flagella in an ELISA with S: N much higher than what we observed with phages detecting whole bacterial cells. Howe ver, we also observed three phages that reacted to L. monocytogenes (but not to flagella), pres umably to any of the minor contaminants visualized in the Tsai-Frasc h polyacrylamide gel stain. Low detection levels of antigenic determinants on cells preselected during panni ng may highlight an important aspect of immunodetection assays wh en applied to detection of whole bacterial cells or non-charged targets: it is not enough to posses a specific antibody, it must also posses high affinity or avidity, a shortcoming of antibody fragments such as scFvs (132). Our ELISA results demonstrate that detecti on was absent to whol e cells using soluble scFv. We considered the possi bility that the microwell plat es were not coated optimally, and although wells received 109 CFU/mL of cells, the actual number of cells binding to the polystyrene may in fact have been lower. Brewster et al. (119) addressed the issue of coating variability in microwells when screening scFvs to Listeria spp. by first standardizing cell quantities (both L. monocytogenes and others) to a known concentration in buffer using a spectrophotomet er then performing a suspended-bacteria binding specificity assay where all i mmunoassay steps were performed in a microcentrifuge tube, including substrate deve lopment. Following simple centrifugation to pellet the cells, the substrate was transf erred to ELISA microw ells and absorbance

PAGE 125

114 readings were taken. This technique pr oved to be successful in determining the specificity of anti-listeria scFv phages to a panel of listeria species and could be applied to our experiments. We demonstrated that both anti-flagella phages and their solubl e scFvs could detect the flagellum band at 29 kDa in a Western bl ot (Figures 3-10 and 3-11), although Peel et al. (130) noted that low voltage SDS-PAGE se paration of listeria fl agella separates the protein into multiple, antigenically heteroge nic bands. Additionally, Dons et al. (133) noted that there is a discrepa ncy in the predicated and actua l molecular weights of listeria flagellin, possibly due to glycos ylation (36). ELISA cross-re active profiles of the phages demonstrated variable reactivity to all lis teria species (data no t shown), a phenomenon not uncommon when generating MoAbs to liste ria flagella (128,129). It was not fully unexpected, then, that the antibodies we ge nerated would be cross-reactive among the listeria species. Surprisingly, however, some of the anti-flagella antibodies cross-reacted to the L. monocytogenes wild-type isogenic flagella muta nt (containing a deletion of the flaA gene) in an ELISA. Dons et al. (133) showed that th ere is a single flagellin gene in L. monocytogenes and more importantly, recently Schirm (36) showed that there is post-translational modification of the CLIP23485 flagellin proteins at the GlcNAc backbone residue with a O -glycosidic linkage as they are exported from the bacteria. This led us to consider that there might be other O -linked glycoproteins or carbohydrates on the surface of L. monocytogenes , including the flagellum mutant. It has been suggested that the discrepancies in the predicted and actual flag ellin, molecular weights that Dons observed can be attributed to post-tran slational modifications (36). The cell walls of many gram-

PAGE 126

115 positive bacteria are modified by O -acetylation of MurNac (3 3), as well as being decorated with many carbohydrates such as teicho ic acids. Therefore, in an attempt to understand the cross-reactivity of our anti-flagella scFvs to non-flagellated cells, we chemically modified the O -glycosidic linkages on the flagel la with periodate reagent. We tested six anti-flagella scFv clones using a Western blot to listeria flagella that were pre-treated with periodate and found that only a single clone of the six tested would not bind to the flagellin after treatment (Fig. 3-12) . This suggested that the scFv for this clone was specific to glycosylation on the fl agella, and would reasona bly explain, in part, why we observed reactivity of this scFv to nonflagellated cells. Th erefore, these initial results indicated that the scFv was binding to carbohydrate molecules on the surface of the non-flagella ted cells. Because our panning was to intact whole flagella sheared from cells, we could assume that the reported glyc osylated modifications existe d on the flagella that we adsorbed to the immunotubes. Because the glycosylation modification sites are in a central, surface-exposed region on L. monocytogenes flagella (36), there existed in our experiment the real possibility that phages could then bind to these exposed areas in a panning procedure. Although H-antigens can be serospecific (128-131), the modification of listeria flagella is identical between L. monocytogenes serotypes 1/2a, 1/2c, and 4b (36). These modifications may represent shared common epitopes among all listeria Hantigens and may also explain the non-serosp ecific reactivity of our antibodies isolated from our panning experiments beyond day-to-day assay variability. We were able to successfully express solubl e anti-flagella scFv for at least 10 of the 31 phage clones by infecting a non-suppressor E. coli strain, HB2151, and verified their

PAGE 127

116 functionality by using them as primary anti bodies to flagella in ELISAs and Western blots. Although detection of flagella was robust (sensitive detection of a microgram of flagellin protein), we observed dramatic decr eases in their binding to whole cells. Signal to noise ratios dropped from approximately 5 (to flagella) to 1. Attempts at concentrating the supernatants containing th e scFv ten-fold or purifying concentrated scFv did not improve signal detection to whole cells (Fig. 3-15), reflecting the low affinity of the molecule and the potentially lo w quantity of flagella on the cells (proposed to be between one to five of pe ritrichous flagella). This led us to ask what issues then are present regarding successful detection using scFv besi des possessing specificity and affinity. Because we had ten clones that produced high quantities of soluble scFv, this meant that we also had 21 clones (almost two-thirds of our pool) that did not. Das et al. (121) demonstrated that within a pool of expressed scFv, there are both functional and nonfunctional fo rms for unknown reasons. Reports of optimization of renaturation conditions (121,134) allowed for a 50% increase of solubl e, functional scFv, but these scFv still exhibited less binding activity to whole vi rus than to purified protein compared to Fab. Furthermore, it is well known that refolded an tibody may not retain binding activity upon expression and can a ggregate easily. Although we did not look further into these issues with our anti-flagell a scFvs, it would be useful to show a direct comparison of active scFv to non-active within an expressed pool. Simply concentrating or purifying scFv would potentially be usef ul if the scFv existed in a mixed pool containing both denatured and renatured fo rms. Although independent of volume, the concentrated pool would still contain only a fr action of functional scFv; the concentration of functional scFv would s till increase. This potentia l problem can be overcome by

PAGE 128

117 screening a high number of phages, until a phage clone is obtained that is specific and able to produce soluble scFv. Although we screened 160 clones from panning to whole cells, only a very small fraction demonstrated some affinity to Listeria spp . Later, this smaller pool of phages limited the chance of se lecting a clone able to produce scFv; none of the clones did. The ability of finding a th ird of clones able to form soluble scFv derived from a relatively small pool in our flag ella panning attempts may be attributed to using a specific antigen (flagella) for our pannings, although it is unknown how this specifically contributed to obt aining clones exhibiting successf ul scFv expression except for having a known target to sc reen to in various assays. Surface proteins serve as excel lent targets for the devel opment of antibodies in the detection of pathogenic organisms. Bacter ial proteins can serve as major virulence factors needed for bacterial entr y and persistence within cells. L. monocytogenes is an intracellular organism dependent on several prot eins to be coordinately expressed, thus allowing for invasion and subsequent escape from host cells, and therefore there have been many attempts at developing monoc lonal antibodies to surface targets of Listeria (119,135,136) including its virulence factors, with downstream applications in the detection of whole cells. Ma ny of these antibodies, however, have cross-reactivity to not only non-pathogenic L. innocua, but also to other bacteria species. Furthermore, the expression of several of these proteins is re stricted by their interactions with host cells during an infection or the an tigen is expressed only dur ing growth at 37C (and not 24C). Antibodies that target virulence f actors, such as ActA, InlB, and LLO, though specific for L. monocytogenes, would have limited usefulness in the sampling of foodstuffs in the biosensor system.

PAGE 129

118 Recently, Cossart and colleagues identifie d and described an autolytic murein hydrolase of L. monocytogenes named lmo1076, or Auto, important in invasion and entry into host cells. The Auto protein, is not contro lled by the listeria virulence transcriptional gene activator, PrfA, and is, to da te, the only murein hydrolase found in L. monocytogenes that is absent from non-pathogenic L. innocua (37). That Auto is not controlled by PrfA is, in fact, an advantage for our purposes. Cells would not have to be grown under complex growth conditions to i nduce its expression, such as cell culture infections. Auto is a murein hydrolase, impor tant in the hydrolysis of the peptidoglycan at certain time points during cell growth. Cells could be easily grow n in BHI medium to exponential phase and harvested for immunoa ssay applications. Hydrolases, unlike surface proteins associated with virulence, are necessary for cell growth and survival. The importance of having hydrolases presen t during cell growth was demonstrated in studies by Pilgrim et al. (137). This work de monstrated that the de letion of the critical hydrolase P60 iap gene in Listeria led to abnormal septum fo rmation and short cellular forms, as well as attenuating in virulence. Targeting of the Auto hydrolase for panning provided us with an opportunity to generate a highly specific anti body fragment useful in the detection of L. monocytogenes . We successfully constructed PCR primers to amplify the hydrolase domain of Auto from the genomic strain of L. monocytogenes EGDe, constituting a recombinant form of the gene of 668 base pairs (bp). This r ecombinant form did not contain the sequence encoding the N-terminal signal sequence necessar y for export of the protein from within a bacterial cell or the C-terminal G-W module sequence necessary for attachment of Auto to the outer surface. The recombinant Auto genes were translated into a polypeptide

PAGE 130

119 within the expression host ce ll cytoplasm, where the protei n could then be collected following lysis of the cell. For our purposes and without in depth proteomic structural analysis of the recombinant form of th e Auto protein (such as gel filtration chromatography, x-ray crystallography, or circul ar diachromism) it would be difficult to ascertain what impact dele tion of the G-W module had on folding or conformation of rAuto. The majority of monoc lonal antibodies bind to linear stretches of amino acids on their targets and rarely to conformational forms (138,139), and therefore generation of scFvs to the recombinant form would most lik ely be to stretches of amino acids found on the native Auto protein. Potential bindi ng domains on the recombinant protein may not necessarily be exposed on the native protein, however. To circumvent this potential problem, we included a panning selec tion round to native Auto on whole L. monocytogenes cells, following selection of phage s to the recombinant protein in immunotubes. We were able to subclone the Auto-P CR product into a high-expression vector, pET19b, and induce protein expression in the host E.coli BL21 (DE3) Tuner cells. Following expression, we recovered recombinant protein from the cytoplasm of the cells by lysing and solubilizing the membrane protein using a commercial lysis reagent containing detergent and lysozyme. The pET19b vector contains codons encoding six histidine amino acids (Fig. 2-2), and cloning of the Auto protein in frame with this sequence forms a hexa-histidine fusion tag (6Hi s-tag). The 6His-tag has high affinity for nickel ions and passage of His-tagged proteins over agar ose matrices charged with nickel-ions (metal chelate affinity chro matography (MCAC)) allows for binding and elution of tagged recombinant proteins with relative ease using imidazole. Untagged

PAGE 131

120 proteins pass through the column. Using MCAC , we were able to purify approximately 7 mg of rAuto from a liter of induced host expression E. coli cells. We verified the recombinant protein at 34 kDa following its electrophoresis using SDS-PAGE, in agreement with the predicted molecular wei ght. Two lesser weight proteins at 20 kDa and 17 kDa, presumably either degradation products or host cell pr oteins bound to rAuto during elution, co-purified with the rAuto protein. We furt her verified the protein as rAuto and not an expression sy stem artifact through detec tion with an anti-6His-tag monoclonal antibody in a Western blot, which cl early identified the band at 34 kDa (Fig. 3-13). To a lesser degree, a continuous banding pattern was observed from 34 kDa and below, indicating that some minor protein degradation had occurred. For our panning experiments, we perfor med two parallel panning procedures both initially on rAuto-coated immunotubes and amp lification of phages. After two rounds of panning, the pool of amplified phages was then used for a third panning round to whole cells in suspension of L. monocytogenes EGDe, the genomic strain used to PCR-amplify the Auto gene. In the second parallel expe riment, we included a negative selection panning round to L. innocua . Although we observed phages that could recognize L. moncoytogenes cells from both parallel panning pool s, there were nodifferences in their detection of both rAuto and whole cells (Fig. 319). Surprisingly, we recovered a higher quantity of phages that met our initial goal of detecting L. monoctyogenes without including negative selection (85% positive de tection of rAuto, 85% positive detection of L. monocytogenes EGDe cells, 30% positive detection of L. innocua cells). By incorporating a negative selection, we could wholly eliminate the phages’ detection of L. innocua cells, although their collec tive reactivity also fell (5 0% positive detection of

PAGE 132

121 rAuto, 18% positive detection of L. monocytogenes EGDe cells, 0% positive detection of L. innocua cells). The data we present from these panning at tempts demonstrate at least two important points. Under the conditions we used in our experiments, negative selection used in a phage display system worked at elimina ting unwanted cross-reactivity. Secondly, by incorporating panning procedures to both a highly selective ta rget and to whole cells, we could quickly isolate highly specific phage s (and their soluble scFvs). The initial screening of these phages demonstrated a profound difference in their reactivity compared to our initial panning experiments to whole cells alone by using more stringent panning parameters as evidenced in other work (119). For the first time in our work, we were able to demonstrate phages th at were not only specific to L. monocytogenes (and not non-pathogenic L. innocua ) but that were also non-reactiv e to non-related bacteria such as E. faecalis , E. coli 0157:H7, or S . enterica serovar Typhimurium, unlike any phages from our previous panning attempts. Our initial results seem to also verify the results of Cossart et al. that both L. innocua and L. monocytogenes serovar 4b lack native Auto protein (37), although more in de pth research would need to be performed such as testing the hybridization of our aut PCR product with genomic ex tracts of the individual L. monocytogenes serovars in a Southern blot. L ack of hybridization of the PCR product with either L. innocua or L. monocytogenes serovar 4b genome would help verify our panning results by demonstrating the absence of Auto from these cells. It is not known, though likely, that Auto ho mologues exist in these Listeria spp . In the scope of our researc h, our results also demonstrat e drawbacks and contradictory data following panning to rAuto. Seven phages that exhibited the highest positive

PAGE 133

122 reactivity to both rAuto and L. monocytogenes cells in the ELISA used as primary antibody in a Western blot. We could not de monstrate detection of the native Auto protein, a 62-kDa protein, from SDS-extracte d cells (37) with our scFv or phages in a Western blot, although this was most likely due to denaturation of the protein during extraction in the presence of SDS and a reducing agent. Furthermore, only a single clone, 1N-1, detected the rAuto protein in the West ern blot, an apparent contradiction of the ELISA where all seven clones we re clearly positive to rAuto (Figures 3-20 and 3-21). How are these antibodies able to detect rAuto in an ELISA but not in a Western blot? It is possible that scFvs are unexpectedly r ecognizing conformationa l epitopes, although this is difficult to ascertain from this initial data. The scenario that our data present would require the recombinant form of the protein to have retained some structural folding present in our panning experiments when us ed as antigen. Not an unusual issue, monoclonal antibodies that bind to both recomb inant and native listeria hydrolases have been studied, such as the listeria P60 hydrolase (140). If the phages tested in our research recognize the native Auto protein, this led us to question the possibility that the recombinant form of Auto has antigenical ly homologous structure to its native form present on the cell surface, which can only be answered directly through structural research or indirectly by testing functionality. Carroll et al. demonstrated that a recombinant murein hydr olase, MurA, could be PCR amplified from L. monocytogenes EGDe and retain its lytic activity (140) indicating that the recombinant form was properly folded. Although we cloned the hydrolytic domain of Auto, we were unable to verify activity of the enzyme, which would have enabled verification of proper folding of structural domains. But as stated earlier, without the G-

PAGE 134

123 W module present on rAuto, it is unlikely (though unproven in its current state) that the recombinant protein is properly folded and active, which is also highly dependent on conditions the enzyme resides in (such as prope r pH and the presence of select metal ions like Mg2+ or Ca2+). In our study, we were able to isol ate scFv that could recognize only L. monocytogenes cells , when fused to the phage, repeating our previous ELISA results from whole cell panning and panning to flagella. Soluble anti -Auto scFvs exhibited high specificity and sensitivity, positive in their detection of 1 g of rAuto in an ELISA (S:N from 10 to 16), but failed to generate S:Ns a bove 1 in their detection of L. monocytogenes cells. Further characterization of the antirAuto monoclonal antibodies could further advance our knowledge of the homology shared among the many hydrolases of Listeria spp . In their current state, these antibodies have potential use in the biosensor, possibly as a specific secondary antibody in a sandwich type assay. They could also be additionally combined with the anti-flagella antibodies as a “polyclonal” detection re agent or in capture/detector combos. Specific Aim 2: Optimization of Immunological Reagents There have been many studies on the usef ulness of scFvs and the improvement of their low affinities to increase binding to antigens (114,117,121,141). Biotinylation of antibody molecules and detection with strept avidin conjugates is a common strategy to amplify the signal obtained in immunological an alysis. In these stud ies, we described a way to increase assay signals following site-s pecific or chemical biotinylation of scFvs recognizing L. monocytogenes .

PAGE 135

124 Expression vectors were constructed by fu sing a 15-peptide biotinylation acceptor domain to the C-terminal end of selected scFvs. The genes coding for whole cell clone (1-3) scFv, anti-flagella scFv, and anti-BSA scFv were cloned into the pAC5 Avitag vector. Anti-BSA antibodies were biotinylat ed and tested as a model for the scFv antibodies. Recombinant scFvs were biotinylated and expressed in E. coli AVB100 containing the BirA enzyme ( birA is stably integrated into the chromosome under the control of AraC). Restriction digestion was done with Hin dIII and Not I to obtain the fragment containing the ribosome binding site, the pelB leader sequence, and the entire scFv gene (Figure 1-1). In a Western blot , we could demonstrate successful expression of the whole cell and BSA scFv clones fused to the biotinylated peptide following their transfer to a nitrocellulose membrane and by simple reaction with streptavidin conjugated to HRP. This was the first time we coul d demonstrate successful expression of an antiwhole cell scFv (clone 1-3) that could not be previously expressed. Coincidentally, we also demonstrated that an anti-flagella clon e previously expressed as soluble scFv in E. coli HB2151 cells could not be expressed in th e Avitag system. Our current data are unable to provide a reasonable explanation for these observations. As a drawback, in the scope of our experiments we were not able to prove functionality of the expressed Avitag biotinylated scFvs via ELISA. We observed reactivity of SA-HRP to listeria-coated microwells (microwells without the additi on of primary antibody), and therefore could not verify that these antibody fragments were functional and therefor e responsible for the absorbances generated. In fact, wells not receiving primary anti body would generate S:N from 5 to 9. We did not make this observa tion in BSA-coated microwells or with the anti-BSA Avitag scFv, seeming to indicate f unctionality for this clone. The ELISA

PAGE 136

125 results suggest that the assa y needs troubleshooting. Biotin, biotin-like, and peroxidaselike surface molecules ar e not known to exist on L. moncoytogenes , and it is therefore difficult to explain this unusual and unlikely interaction of the SA-HRP with the listeria cells. Our studies would seem to verify also that expressed antibody is not always refolded properly or remain fully active. Suzuki (142) verified the failure of phagedisplayed scFvs to demonstrate significant bi nding to target antig ens in a conventional E. coli expression system due to translational errors or mutations within certain scFv genes. There have been reports demonstrating that the successful expression of scFv fused to a biotin mimic tag is often the result from the contribution of variable expression conditions (121). The variable expression conditions could be improved by optimizing the renaturation and solubilization conditions during expression of soluble scFv in the presence of L-arginine. L-arginine can increase the yiel d of refolded proteins by decreasing the quantity of aggregated forms, an apparent common problem with scFv expression (121,143). These reported results ar e particularly importa nt in studies using scFvs because there is no universal renatura tion protocol that can be used for proper refolding of all proteins (144). Optimizati on allowed for a 50% increase in the quantity of soluble scFv, although in th eir studies the scFv still exhi bited less bindin g activity to whole virus compared to Fab. This phenome non may have, in fact, been evidenced with both scFvs that we cloned into the pAC5 e xpression vector system. Although different strategies have been adopted to form prope rly refolded scFvs, such as the critical disulfide bond formation demonstrated by Das (121), there needs to be more consideration of the factors that allow for optimum refolding, expression vectors used,

PAGE 137

126 and the influence of the amino acids that co mprise the scFv that allow for localization and successful export. Despite the advantages of in vivo expression of recombinant scFv, we observed that the binding activity was compromised compared to whole Ig or even scFv fused to the phage (F igures 3-14, 3-15, and 3-20). We attempted chemical addition of biotinyla tion of scFvs in parallel to genetically fusing them to biotin accepto r sequences using the Avitag in vivo biotinylation system. Chemical biotinylation, though not as precise as the Avitag system, allowed us to treat purified and quantifiable amounts of antibody with relative ease . Using set quantities of scFv in our assays is an important step towards optimizing a protocol in a detection system, especially when used in a sandwich or capture format such as used in the fiber optic biosensor. We successfully chemically treated select scFvs and a rabbit antiL. monocytogenes polyclonal antibody with an esterified form of biotin to predominantly label peptide amines. We used these reagents in the detection of listeria flagella in both ELISAs and Western blots using SA-HRP for the biotinylated scFv or protein L-HRP for unmodified scFv. Our initial results indicated that the scFv -B remained functional following chemical biotinylation but did not signifi cantly improve absorbance signals to purified flagella in an ELISA (Table 3-5). In fact, unmodifi ed scFv had higher S:N ratios from 7 to 12 compared to scFv-B (S:N from 2 to 4). Un expectedly, signals were greatly enhanced in the detection of whole cells co mpared to unmodified scFv; biotinylated scFv had a S:N of approximately 5 and unmodified scFv had a S:N of approximately 2. We could not ignore, however, the ELISA results that demons trated false positive reactivity of our SAHRP to whole listeria cells (Table 3-4). Th ere was a clear discrepancy with the results

PAGE 138

127 that we obtained by using chemically biotinyl ated scFv. Using chemically biotinylated scFv, we observed a linear increase in absorb ance signals by increasing the concentration of scFv-B to whole cells from 10 g/mL to 50 g/mL. This was not observed with the Avitag cloned scFvs, where the presence (Avitag scFv) or absence of the primary antibody (buffer alone) created equal S:N impl ying an interaction of the SA-HRP with the cells. It is not known in the scope of our work how the difference in bitoinylation of the primary antibodies (scFvs) while using the same SA-HRP secondary reagent effected increased backgrounds. Although we coul d demonstrate successful chemical biotinylation, there needs to be further char acterization of the inte raction of biotin-scFv and streptavidin with the surf ace of listeria. A step in op timization of the assay could incorporate the use of a different conjugated r eagent, such as alkaline phosphatase instead of SA-HRP. Long-term research could incor porate computational analysis of amino acid sequences in streptavidin-lig and interactions and surface proteins found on L. monocytogenes (145,146). We adopted chemically derivatized biotinyl ated scFv tetramerized to streptavidinHRP to form a single-use iummnodetection reag ent called a streptabody (stAb) to address decreased multivalency exhibited by scFvs to whole L. monocytogenes (Figures 3-23 and 3-24). The failure of soluble scFvs to recogni ze an antigen with high affinity is generally a common problem encountered with pha ge display antibody systems (116,117,142). Many evolutionary engineering or antibody re shaping methods have been successfully applied to increase st ability and affinity of scFv by using DNA chain-shuffling (147), error prone PCR (142), or affinity ma turation (141,148,149). Thes e methods require numerous manipulations that could lead to high error rates. Many simpler methods to

PAGE 139

128 improve detection using scFv have also been attempted. Wang and Ferrone (116) applied a simple method of improving a binding assa y by mixing low affinity scFv fragments fused with c-myc tags with an anti-c-myc monoclonal antibody to form sdimeric antibody fragment-anti tag monoclonal antibody complexe s that demonstrated a two-fold increase in their detection assays. Relatively low affinity scFvs can be fu rther manipulated by tertamerizing their biotinylated forms to streptavidin (117) forming high avidity stAbs. StAbs were successfully created to detect small qua ntities of carcinoembryonic antigen at concentrations 100 times lower than using unmodified scFv (117). Studies of the tetramerized molecule multivalent and avidity effects measured in surface plasmon resonance (SPR) demonstrated hi gher affinity constants, KA, of the stAbs compared to scFvs (117). We performed a simple proced ure of mixing biotinyl ated anti-rAuto scFv, 18N-3B, with SA-HRP and measured its ac tivity in an ELISA to a purified listeria antigen (recombinant Auto protein) and to whole listeria cells. Using standardized concentrations of immunoreagents, we demonstrated that the 18N-3B stAb had significantly higher detection of rAuto than its respective pha ge, scFv, or scFv-B. The reactivity of the stAb, however, did not signi ficantly demonstrate signal amplification in the detection of listeria cel ls unexpectedly. Based on th ese results, more research regarding whether proper formation of the 18N-3 stAb was achieved is needed to confirm the correct molecular weight (approximately 188 kDa) by using either size-exclusion gel chromatography or non-denaturing gel electrophoresis. There are potential problems in the in vitro manufacturing of tetramerized imunnodetection reagents, especially regarding the orientation of the scFv-B in the SA-

PAGE 140

129 HRP tetramer. Because the addi tion of biotin chemically to a protein is not as controlled as site-directed in vivo biotinylation of the Avitag peptide to the C-terminal end of a scFv, chemically biotinylated scFv may not be orie nted with the proper antigen binding sites of the antibody exposed outwardly from the stAb complex (117). Furthermore, it would be useful to investigate the effect of using variable length biotin arms when labeled to scFv. The NHS-biotin used in our biotinylation pro cedure contains a relativ ely short and simple linker arm when bound to amines on the protei n (13.5 ). Other forms of NHS-biotin are available that have longer linke r arms (39 ) that would allow for a more flexible stAb to promote multivalent binding to repetitive or complex epitopes. StAbs used in the conditions of our research may be at best useful in common immunodetection methods such as ELISAs and Western blots. Conventional methods have not been able to obtain monoclonal antibodies with both high specificity and avidity required by rapid detection systems for L. monocytogenes and other pathogens. Phage displa y allows for achievement of this goal. The goal of our research was to use a nave phage display scFv library in obtaining specific monoclonal antibody fragments for use in a fiber optic biosensor in the detection of the food borne pathogen, L. monocytogenes , as well as improving on current methodologies used in phage display that can be applied to other bacterial species. At the current state of our goal, we present results that demonstrate that scFvs can be developed quickly to not only whole listeri a cells, but also to its spec ific cellular components, such as flagella and surface hydrolases. Furthe r research into the improvement of these immunological reagents is needed, although we present initial work demonstrating practical methods that can be further optimized into a capture assay format. In summary,

PAGE 141

130 we feel that research should continue on the development of scFv antibody technology for use in new and existing detection platfo rms, because recombinant antibodies offer numerous advantages over conventional animal-based methods.

PAGE 142

131 LIST OF REFERENCES 1. Murray, E.G.D. 1926. A disease of rabb its characterized by a large mononuclear leucocytosis caused by a hitherto undescribed bacillus Bacterium monocytogenes (n.sp.). J. Pathol. Bacteriol. 29:407-439. 2. Hof, H. 2003. History and epidemiology of listeriosis. FEMS Immunol. Med. Microbiol. 35:199-202. 3. Pirie, J.H. 1940. Listeria: change of name for a genus of bacteria. Nature 145:264. 4. Reiss, H.J., J.P.a.H.K. 1951. Gr anulomatosis infantiseptica. Eine Allgemeininfektion bei Neugeborenen und Suglingen mit miliaren Granulomen. Z. Ges. Inn. Med. 6:451-457. 5. Seeliger,H.P.R. 1961. Listerio sis. Karger Verlag, Basel. 6. Rocourt, J., A.S.a.H.P.R.S. 1983. Diffrentiation biochimique des groupes gnomiques de Listeria monocytogenes (sensu lato). Ann. Microbiol. 134 A:6571. 7. Trost, M., Wehmhoner, D., Karst, U., Dieterich, G., Wehland, J., and Jansch, L. 2005. Comparative proteome analysis of s ecretory proteins from pathogenic and nonpathogenic Listeria species. Proteomics. 5:1544-1557. 8. Schlech, W.F., III, Lavigne, P.M., Bortolussi, R.A., Allen, A.C., Haldane, E.V., Wort, A.J., Hightower, A.W., Johnson, S. E., King, S.H., Nicholls, E.S., and Broome, C.V. 1983. Epidemic listerios is--evidence for transmission by food. N. Engl. J. Med. 308:203-206. 9. CDC, 2. Centers for Disease Contro l and Prevention (CDC), Outbreak of Listeriosis associated with homemade Mexican-style cheese-North Carolina, October 2000-January 2001. 50 26, 560-562. 2001. MMWR. 10. Centers for Disease Control and Prev ention (CDC). Outbreak of Listeriosis associated with homemade Mexican-style cheese--North Carolina, October 2000January 2001. JAMA 286:664-665. 11. CDC, 1. Centers for Disease Control and Prevention (CDC), Multistate outbreak of Listeriosis. 47 50, 1085-1086. 1998. MMWR.

PAGE 143

132 12. Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee, J.S., Shapiro, C., Griffin, P.M., and Tauxe, R.V. 1999. Food -related illness and death in the United States. Emerg. Infect. Dis. 5:607-625. 13. Salvat, G. and Fravalo, P. 2004. Risk a ssessment strategies for Europe: integrated safety strategy or final product control: Example of Listeria monocytogenes in processed products from pork meat industry. Dtsch. Tierarztl. Wochenschr. 111:331-334. 14. Mengoni, G.B. and Apraiz, P.M. 2003. [Monitoring of a HACCP (Hazard Analysis Critical Control Point) plan for Listeria monocytogenes control]. Rev. Argent Microbiol. 35:224-227. 15. Smith, J.L. 1999. Foodborne infections during pregnancy. J. Food Prot. 62:818829. 16. Abram, M., Schluter, D., Vuckovic, D ., Waber, B., Doric, M., and Deckert, M. 2002. Effects of pregnancy-associated Listeria monocytogenes infection: necrotizing hepatitis due to impaired ma ternal immune response and significantly increased abortion rate. Virchows Arch. 441:368-379. 17. Antal, E.A., Dietrichs, E., Lober g, E.M., Melby, K.K., and Maehlen, J. 2005. Brain stem encephalitis in listeriosis. Scand. J. Infect. Dis. 37:190-194. 18. Otter, A. and Blakemore, W.F. 1989. Observation on the presence of Listeria monocytogenes in axons. Acta Microbiol. Hung. 36:125-131. 19. Campero, C.M., Odeon, A.C., Cipolla, A.L., Moore, D.P., Poso, M.A., and Odriozola, E. 2002. Demonstration of Listeria monocytogenes by immunohistochemistry in formalin-fixed br ain tissues from natural cases of ovine and bovine encephalitis. J. Vet. Med. B Infect. Dis. Vet. Public Health 49:379383. 20. Jin, Y., Dons, L., Kristensson, K., and Rottenberg, M.E. 2001. Neural route of cerebral Listeria monocytogenes murine infection: role of immune response mechanisms in controlling bacterial neuroinvasion. Infect. Immun. 69:1093-1100. 21. Neumann, K., Eppler, E., Filgueira, L ., Groscurth, P., Gasal, E., Schaffner, A., Schoedon, G., and Schneemann, M. 2003. Listeria species escape from the phagosomes of interleukin-4-deactivat ed human macrophages independent of listeriolysin. Immunol. Cell Biol. 81:431-439.

PAGE 144

133 22. Vazquez-Boland, J.A., Kuhn, M., Be rche, P., Chakraborty, T., DominguezBernal, G., Goebel, W., Gonzlez-Zor n, B., Wehland, J., and Kreft, J. 2001. Listeria pathogenesis and mol ecular virulence determinants. Clin. Microbiol. Rev. 14:584-640. 23. Busch, D.H., Kerksiek, K., a nd Pamer, E.G. 1999. Processing of Listeria monocytogenes antigens and the in vivo T-cell response to bacterial infection. Immunol. Rev. 172:163-169. 24. Haschtmann, D., Gerber, H.J., and Mielke, M.E. 2005. Cytotoxic activity of murine resident peritoneal cells against Listeria monocytogenes -infected hepatocytes in vitro. Microbes Infect. 25. Pamer, E.G. 2004. Immune responses to Listeria monocytogenes . Nat. Rev. Immunol. 4:812-823. 26. Berg, R.E., Crossley, E., Murray, S., and Forman, J. 2005. Relative Contributions of NK and CD8 T Cells to IFN-{gamma} Mediated Innate Immune Protection against Listeria monocytogenes . J. Immunol. 175:1751-1757. 27. Conlan, J.W. 1997. Critical roles of neutrophils in host defense against experimental systemic infections of mice by Listeria monocytogenes , Salmonella typhimurium , and Yersinia enterocolitica . Infect. Immun. 65:630-635. 28. Schlech, W.F. 1996. Pathogenesis and immunology of Listeria monocytogenes . Pathol. Biol. (Paris) 44:775-782. 29. Campbell, D.J. and Shastri, N. 1998. Bacterial surface proteins recognized by CD4+ T cells during murine infection with Listeria monocytogenes . J. Immunol. 161:2339-2347. 30. Schwarzkopf, A. 1996. Listeria monocytogenes --aspects of pathogenicity. Pathol. Biol. (Paris) 44:769-774. 31. Braun, L. and Cossart, P. 2000. Interactions between Listeria monocytogenes and host mammalian cells. Microbes Infect. 2:803-811. 32. Cabanes, D., Dehoux, P., Dussurget, O., Frangeul, L., and Cossart, P. 2002. Surface proteins and the pathogenic potential of Listeria monocytogenes . Trends Microbiol. 10:238-245. 33. Navarre, W.W. and Schneewind, O. 1999. Surface proteins of gram-positive bacteria and mechanisms of their targeting to the cell wall envelope. Microbiol. Mol. Biol. Rev. 63:174-229. 34. Wemekamp-Kamphuis, H.H., Wouters, J.A., de Leeuw, P.P., Hain, T., Chakraborty, T., and Abee, T. 2004. Identification of sigma factor sigma Bcontrolled genes and their impact on aci d stress, high hydrostatic pressure, and

PAGE 145

134 freeze survival in Listeria monocytogenes EGD-e. Appl. Environ. Microbiol. 70:3457-3466. 35. Ferreira, A., Gray, M., Wiedma nn, M., and Boor, K.J. 2004. Comparative genomic analysis of the sigB operon in Listeria monocytogenes and in other Gram-positive bacteria. Curr. Microbiol. 48:39-46. 36. Schirm, M., Kalmokoff, M., Aubry, A., Thibault, P., Sandoz, M., and Logan, S.M. 2004. Flagellin from Listeria monocytogenes is glycosylated with betaO linked N -acetylglucosamine. J. Bacteriol. 186:6721-6727. 37. Cabanes, D., Dussurget, O., Dehoux, P., and Cossart, P. 2004. Auto, a surface associated autolysin of Listeria monocytogenes required for entry into eukaryotic cells and virulence. Mol. Microbiol. 51:1601-1614. 38. Abachin, E., Poyart, C., Pellegrini, E., Milohanic, E., Fiedler, F., Berche, P., and Trieu-Cuot, P. 2002. Formation of D-alan yl-lipoteichoic acid is required for adhesion and virulence of Listeria monocytogenes . Mol. Microbiol. 43:1-14. 39. Braun, L., Dramsi, S., Dehoux, P., Bier ne, H., Lindahl, G., and Cossart, P. 1997. InlB: an invasion protein of Listeria monocytogenes with a novel type of surface association. Mol. Microbiol. 25:285-294. 40. Lalic-Multhaler, M., Bo hne, J., and Goebel, W. 2001. In vitro transcription of PrfA-dependent and -independent genes of Listeria monocytogenes . Mol. Microbiol. 42:111-120. 41. Lecuit, M., Ohayon, H., Braun, L., Mengaud, J., and Cossart, P. 1997. Internalin of Listeria monocytogenes with an intact leucine-rich repeat region is sufficient to promote internalization. Infect. Immun. 65:5309-5319. 42. Lecuit, M. 2005. Understanding how Listeria monocytogenes targets and crosses host barriers. Clin. Microbiol. Infect. 11:430-436. 43. Lecuit, M., Dramsi, S., Gottardi, C., Fedor-Chaiken, M., Gumbiner, B., and Cossart, P. 1999. A single amino acid in E-cadherin responsible for host specificity towards the human pathogen Listeria monocytogenes . EMBO J. 18:3956-3963. 44. Freiberg, A., Machner, M.P., Pfe il, W., Schubert, W.D., Heinz, D.W., and Seckler, R. 2004. Folding and stability of the leucine-rich repeat domain of internalin B from Listeri monocytogenes. J. Mol. Biol. 337:453-461. 45. Marquis, H., Goldfine, H., and Port noy, D.A. 1997. Proteolytic pathways of activation and degradation of a bacterial phospholipas e C during intracellular infection by Listeria monocytogenes . J. Cell Biol. 137:1381-1392.

PAGE 146

135 46. Marquis, H. and Hager, E.J. 2000. pH-regulated activati on and release of a bacteria-associated phospholipase C during intracellu lar infection by Listeria monocytogenes . Mol. Microbiol. 35:289-298. 47. Smith, G.A., Marquis, H., Jones, S., Johnston, N.C., Portnoy, D.A., and Goldfine, H. 1995. The two distinct phospholipases C of Listeria monocytogenes have overlapping roles in escape from a vacuole and cell-to-cell spread. Infect. Immun. 63:4231-4237. 48. Vazquez-Boland, J.A., Kuhn, M., Be rche, P., Chakraborty, T., DominguezBernal, G., Goebel, W., Gonzalez-Zor n, B., Wehland, J., and Kreft, J. 2001. Listeria pathogenesis and mol ecular virulence determinants. Clin. Microbiol. Rev. 14:584-640. 49. Alvarez-Dominguez, C., Roberts, R., and Stahl, P.D. 1997. Internalized Listeria monocytogenes modulates intracellular traffickin g and delays maturation of the phagosome. J. Cell Sci. 110 ( Pt 6):731-743. 50. Bielecki, J., Youngman, P., Connelly, P., and Portnoy, D.A. 1990. Bacillus subtilis expressing a haemolysin gene from Listeria monocytogenes can grow in ammalian cells. Nature 345:175-176. 51. Dussurget, O., Cabanes, D., Dehoux, P ., Lecuit, M., Buchrieser, C., Glaser, P., and Cossart, P. 2002. Listeria monocytogenes bile salt hydrolase is a PrfAregulated virulence factor involved in the intestinal and hepatic phases of listeriosis. Mol. Microbiol. 45:1095-1106. 52. Suarez, M., Gonzalez-Zorn, B., Vega, Y., Chico-Calero, I., and Vazquez-Boland, J.A. 2001. A role for ActA in epithelial cell invasion by Listeria monocytogenes . Cell Microbiol. 3:853-864. 53. Monack, D.M. and Theriot, J.A. 2001. Actin-based motility is sufficient for bacterial membrane protrusion fo rmation and host cell uptake. Cell Microbiol. 3:633-647. 54. Cossart, P., Pizarro-Cerda, J., and L ecuit, M. 2003. Invasion of mammalian cells by Listeria monocytogenes : functional mimicry to s ubvert cellular functions. Trends Cell Biol. 13:23-31. 55. Boujemaa-Paterski, R., Gouin, E., Ha nsen, G., Samarin, S., Le Clainche, C., Didry, D., Dehoux, P., Cossart, P., Kocks, C., Carlier, M.F., and Pantaloni, D. 2001. Listeria protein ActA mi mics WASp family proteins: it activates filament barbed end branching by Arp2/3 complex. Biochemistry 40:11390-11404. 56. Auerbuch, V., Loureiro, J.J., Gertler, F.B., Theriot, J.A., and Portnoy, D.A. 2003. Ena/VASP proteins contribute to Listeria monocytogenes pathogenesis by controlling temporal and spatial persistence of bacterial actin-based motility. Mol. Microbiol. 49:1361-1375.

PAGE 147

136 57. Samarin, S., Romero, S., Kocks, C., Di dry, D., Pantaloni, D., and Carlier, M.F. 2003. How VASP enhances actin-based motility. J. Cell Biol. 163:131-142. 58. Cossart, P. 2000. Actin-based motility of pathogens: the Arp2/3 complex is a central player. Cell Microbiol. 2:195-205. 59. Way, S.S., Thompson, L.J., Lopes, J.E., Hajjar, A.M., Kollmann, T.R., Freitag, N.E., and Wilson, C.B. 2004. Characterization of flagellin expre ssion and its role in Listeria monocytogenes infection and immunity. Cell Microbiol. 6:235-242. 60. Peel, M., Donachie, W., and Shaw, A. 1988. Temperature-dependent expression of flagella of Listeria monocytogenes studied by electron microscopy, SDS-PAGE and western blotting. J. Gen. Microbiol. 134:2171-2178. 61. Way, S.S., Thompson, L.J., Lopes, J.E., Hajjar, A.M., Kollmann, T.R., Freitag, N.E., and Wilson, C.B. 2004. Characterization of flagellin expre ssion and its role in Listeria monocytogenes infection and immunity. Cellular Microbiology 6:235242. 62. Grundling, A., Burrack, L.S., B ouwer, H.G., and Higgins, D.E. 2004. Listeria monocytogenes regulates flagellar motility gene expression through MogR, a transcriptional repressor required for virulence. Proc. Natl. Acad. Sci. U. S. A 101:12318-12323. 63. Hayashi, F., Smith, K.D., Ozinsky, A., Hawn, T.R., Yi, E.C., Goodlett, D.R., Eng, J.K., Akira, S., Underhill, D.M., a nd Aderem, A. 2001. The innate immune response to bacterial flagellin is mediated by Toll-like receptor 5. Nature 410:1099-1103. 64. Coupland, R. and Leins, K.R. 2005. Science and prohibited weapons. Science 308:1841. 65. Woodall, J.P. 2005. WHO and bi ological weapons investigations. Lancet 365:651. 66. Szinicz, L. 2005. History of chem ical and biological warfare agents. Toxicology . 67. Kron, S. and Mendlovic, S. 2002. Mental health consequences of bioterrorism. Israel Medical Association Journal 4:524-527. 68. Shoham, D. and Wolfson, Z. 2004. Th e Russian biological weapons program: vanished or disappeared? Crit Rev. Microbiol. 30:241-261. 69. Rasooly, A. 2001. Surface plasmon resonance analysis of staphylococcal enterotoxin B in food. J. Food Prot. 64:37-43.

PAGE 148

137 70. Taitt, C.R., Golden, J.P., Shubin, Y.S., Shriver-Lake, L.C., Sapsford, K.E., Rasooly, A., and Ligler, F.S. 2004. A porar ray biosensor for detecting multiple analytes in complex samples. Microb. Ecol. 47:175-185. 71. Mao, X., Yang, L., Su, X.L., and Li, Y. 2005. A nanoparticle amplification based quartz crystal microbalance DNA sensor for detection of Escherichia coli O157:H7. Biosens. Bioelectron. 72. Prusak-Sochaczewski, E., Luong, J.H., and Guilbault, G.G. 1990. Development of a piezoelectric immunosensor for the detection of Salmonella typhimurium . Enzyme Microb. Technol. 12:173-177. 73. Homola, J., Dostalek, J., Chen, S., Rasooly, A., Jiang, S., and Yee, S.S. 2002. Spectral surface plasmon resonance biosen sor for detection of staphylococcal enterotoxin B in milk. Int. J. Food Microbiol. 75:61-69. 74. Little, M., Kipriyanov, S.M., Le Gall, F., and Moldenhauer, G. 2000. Of mice and men: hybridoma and recombinant antibodies. Immunol. Today 21:364-370. 75. George P.Anderson, C.A.R.-T.a.F.S.L. RAPTOR: A Por , Automated Biosensor. 2005. Center for Bio/Molecular Scien ce and Engineering, Naval Research Laboratory, Washington DC 20375, Proceedings of the First Conference on Point Detection for Chemical a nd Biological Defense. 76. Kohler, G. and Milstein, C. 1975. Con tinuous cultures of fused cells secreting antibody of predefined specificity. Nature 256:495-497. 77. Smith, K.A., Nelson, P.N., Warren, P ., Astley, S.J., Murray, P.G., and Greenman, J. 2004. Demystified... recombinant antibodies. J. Clin. Pathol. 57:912-917. 78. Zanella, I., Verardi, R., Negrini, R., Poiesi, C., Ghielmi, S., and Albertini, A. 1992. New heteromyeloma cell lines for the production of human monoclonal antibodies. J. Immunol. Methods 156:205-215. 79. Edelman, L., Margaritte, C., Chaabihi, H., Monchatre, E., Blanchard, D., Cardona, A., Morin, F., Dumas, G., Petr es, S., and Kaczorek, M. 1997. Obtaining a functional recombinant an ti-rhesus (D) antibody usi ng the baculovirus-insect cell expression system. Immunology 91:13-19. 80. Radomska, H.S. and Eckhardt, L.A. 1995. Mammalian cell fusion in an electroporation device. J. Immunol. Methods 188:209-217. 81. Falkenberg, F.W. 1998. Production of monoclonal antibodies in the miniPERM bioreactor: comparison with ot her hybridoma culture methods. Res. Immunol. 149:560-570.

PAGE 149

138 82. Bruce, M.P., Boyd, V., Duch, C., and White, J.R. 2002. Dialysis-based bioreactor systems for the production of monoclona l antibodies--alternatives to ascites production in mice. J. Immunol. Methods 264:59-68. 83. Ma, J.K.C., Hiatt, A., Hein, M., Vine, N.D., Wang, F., Stabila, P., Van Dolleweerd, C., Mostov, K., and Lehner, T. 1995. Generation and assembly of secretory antibodies in plants. Science 268:716-719. 84. Geisse, S. and Kocher, H.P. 1999. Pr otein expression in mammalian and insect cell systems. Methods In Enzymology 306:19-42. 85. Green, L.L. 1999. Antibody engineering vi a genetic engineering of the mouse: XenoMouse strains are a vehicle for the f acile generation of therapeutic human monoclonal antibodies. Journal of Immunological Methods 231:11-23. 86. Zanella, I., Verardi, R., Negrini, R., Osio, L., Albertini, A., and Ghielmi, S. 1993. Selection of new heteromyeloma cell lines for the obtainment of human monoclonal antibodies. Year Immunol. 7:213-219. 87. Mendez, M.J., Green, L.L., Corvalan, J.R., Jia, X.C., Maynard-Currie, C.E., Yang, X.D., Gallo, M.L., Louie, D.M., Lee, D.V., Erickson, K.L., Luna, J., Roy, C.M., Abderrahim, H., Kirschenbaum, F., Noguchi, M., Smith, D.H., Fukushima, A., Hales, J.F., Klapholz, S., Finer, M.H., Davis, C.G., Zsebo, K.M., and Jakobovits, A. 1997. Functional transplant of megabase human immunoglobulin loci recapitulates human antibody response in mice. Nat. Genet. 15:146-156. 88. Hiatt, A., Cafferkey, R., and Bo wdish, K. 1989. Production of antibodies in transgenic plants. Nature 342:76-78. 89. Hoogenboom, H.R., Marks, J.D., Griffiths, A.D., and Winter, G. 1992. Building antibodies from their genes. Immunol. Rev. 130:41-68. 90. Marks, J.D., Hoogenboom, H.R., Bonnert , T.P., McCafferty, J., Griffiths, A.D., and Winter, G. 1991. By-pa ssing immunization. Human an tibodies from V-gene libraries displayed on phage. J. Mol. Biol. 222:581-597. 91. Smith, G.P. and Petrenko, V.A. 1997. Phage Display. Chem. Rev. 97:391-410. 92. Winter, G. 1990. Gene tec hnologies for antibody engineering. Behring Inst. Mitt. 10-20. 93. Robert Webster. 2001. Filamentous Phage Biology. In Phage Display: A Laboratory Manual. D.R.B.J.K.S.G.J.S.Carl os F.Barbas III, editor. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 1.1-1.37. 94. Zeri, A.C., Mesleh, M.F., Nevzorov, A. A., and Opella, S.J. 2003. Structure of the coat protein in fd filamentous bacteri ophage particles determ ined by solid-state NMR spectroscopy. Proc. Natl. Acad. Sci. U. S. A 100:6458-6463.

PAGE 150

139 95. Zeri, A.C., Mesleh, M.F., Nevzorov, A. A., and Opella, S.J. 2003. Structure of the coat protein in fd filamentous bacteri ophage particles determ ined by solid-state NMR spectroscopy. Proc. Natl. Acad. Sci. U. S. A 100:6458-6463. 96. Martin, A. and Schmid, F.X. 2003. A pr oline switch controls folding and domain interactions in the gene-3-prote in of the filamentous phage fd. J. Mol. Biol. 331:1131-1140. 97. Ng, T.W., Akman, L., Osisami, M ., and Thanassi, D.G. 2004. The usher N terminus is the initial targeting s ite for chaperone-subunit complexes and participates in subsequent pilus biogenesis events. J. Bacteriol. 186:5321-5331. 98. Lubkowski, J., Hennecke, F., Pluckthun, A., and Wlodawer, A. 1999. Filamentous phage infection: crystal structure of g 3p in complex with its coreceptor, the Cterminal domain of TolA. Structure 7:711-722. 99. Armstrong, J., Hewitt, J.A., and Perham , R.N. 1983. Chemical modification of the coat protein in bacteriophage fd and orie ntation of the virion during assembly and disassembly. The EMBO Journal 2:1641-1646. 100. Click, E.M. and Webster, R.E. 1998. The TolQRA proteins are required for membrane insertion of the major capsid protein of the filamentous phage f1 during infection. J. Bacteriol. 180:1723-1728. 101. Gray, C.W. 1989. Three-dimensional st ructure of complexes of single-stranded DNA-binding proteins with DNA. IKe and fd gene 5 proteins form left-handed helices with single-stranded DNA. Journal of Molecular Biology 208:57-64. 102. Wen, J.D. and Gray, D.M. 2004. Ff Gene 5 SingleStranded DNA-Binding Protein Assembles on Nucleotide s Constrained by a DNA Hairpin. Biochemistry 43:2622-2634. 103. Endemann, H. and Model, P. 1995. Lcoa tion of Filamentous Phage Minor Coat Proteins in Phage and in Infected Cells. Journal of Molecular Biology 250:496506. 104. McDonnell, P.A., Shon, K., Kim, Y., and Opella, S.J. 1993. fd Coat Protein Structure in Membrane Environments. Journal of Molecular Biology 233:447463. 105. Rakonjac, J., Feng, J.n., and Model, P. 1999. Filamentous phage are released from the bacterial membrane by a two-step mechanism involving a short C-terminal fragment of pIII. Journal of Molecular Biology 289:1253-1265. 106. Russel, M. 1995. Moving through th e membrane with filamentous phages. Trends in Microbiology 3:223-228.

PAGE 151

140 107. Stopar, D., Spruijt, R.B., Wolfs, C.J.A.M., and Hemminga, M.A. 2003. Proteinlipid interactions of bacteri ophage M13 major coat protein. Biochimica et Biophysica Acta (BBA) Biomembranes 1611:5-15. 108. Human Single Fold scFv Libraries I + J (Tomlinson I + J). 2002. Cambridge, UK, MRC Laboratory of Molecular Biology, MRC Centre for Protein Engineering. 109. Broders, O., Breitling, F., and D ubel, S. 2003. Hyperpha ge. Improving antibody presentation in phage display. Methods Mol. Biol. 205:295-302. 110. Rondot, S., Koch, J., Breitling, F., and Dubel, S. 2001. A helper phage to improve single-chain antibody presen tation in phage display. Nat. Biotechnol. 19:75-78. 111. DeMarco, D.R., Saaski, E.W., Mc Crae, D.A., and Lim, D.V. 1999. Rapid detection of Escherichia coli O157:H7 in ground beef using a fiber-optic biosensor. J. Food Prot. 62:711-716. 112. Wolf, V. and Lange, R. 1994. Charact erization of antiser a raised against Treponema denticola (ATCC 33521) whol e cell, outer sheath, protoplasmic cylinder, and axial flagella. Zentralbl. Bakteriol. 280:325-331. 113. Krysinski, E.P. and Heimsch, R.C. 1977. Use of enzyme-labeled antibodies to detect Salmonella in foods. Appl. Environ. Microbiol. 33:947-954. 114. Razai, A., Garcia-Rodriguez, C., Lou, J., Geren, I.N., Forsyth, C.M., Robles, Y., Tsai, R., Smith, T.J., Smith, L.A., Siegel , R.W., Feldhaus, M., and Marks, J.D. 2005. Molecular evolution of antibody a ffinity for sensitive detection of botulinum neurotoxin type a. J. Mol. Biol. 351:158-169. 115. Hao, H.J., Jiang, Y.Q., Zheng, Y.L ., Ma, R., and Yu, D.W. 2005. Improved stability and yield of Fv targeted su perantigen by intr oducing both linker and disulfide bond into the targeting moiety. Biochimie . 116. Wang, X., Campoli, M., Ko, E., Luo, W., and Ferrone, S. 2004. Enhancement of scFv fragment reactivity with target antigens in binding assays following mixing with anti-tag monoclonal antibodies. J. Immunol. Methods 294:23-35. 117. Cloutier, S.M., Couty, S., Terskikh, A., Marguerat, L., Crivelli, V., Pugnieres, M., Mani, J.C., Leisinger, H.J., Mach, J.P ., and Deperthes, D. 2000. Streptabody, a high avidity molecule made by tetrameri zation of in vivo biotinylated, phage display-selected scFv fragments on streptavidin. Mol. Immunol. 37:1067-1077. 118. Centers for Disease Control and Pr evention (CDC). Food Related Diseases. Atlanta, GA. 12-5-2003. 119. Paoli, G.C., Chen, C.Y., and Brewst er, J.D. 2004. Single-chain Fv antibody with specificity for Listeria monocytogenes . J. Immunol. Methods 289:147-155.

PAGE 152

141 120. Suzuki, Y., Ito, S., Otsuka, K., Iwasawa, E., Nakajima, M., and Yamaguchi, I. 2005. Preparation of functi onal single-chain antibodies against bioactive gibberellins by utilizing randomly mu tagenized phage-display libraries. Biosci. Biotechnol. Biochem. 69:610-619. 121. Das, D., Kriangkum, J., Nagata, L.P., Fulton, R.E., and Suresh, M.R. 2004. Development of a biotin mimic tagged ScFv antibody against western equine encephalitis virus: bacteria l expression and refolding. Journal of Virological Methods 117:169-177. 122. Tvinnereim, A.R., Hamilton, S.E., a nd Harty, J.T. 2004. Neutrophil Involvement in Cross-Priming CD8+ T Cell Re sponses to Bacterial Antigens. J. Immunol. 173:1994-2002. 123. Pamer, E.G. 2004. Immune Responses to Listeria monocytogenes . Nat. Rev. Immunol. 4:812-823. 124. Tabouret M, d.R.J.D.G. 1992. Analysis of surface proteins of Listeria in relation to species, serovar and pathogenicity. J. Gen Microbiol 138:743-753. 125. Goso, Y., Ikezawa, T., Kurihara, M., Endo, M., Hotta, K., and Ishihara, K. 2003. Characterization of Rat Gastric Mucins Using a Monoclonal Antibody, RGM23, Recognizing Surface Mucous Cell-Type Mucins. J. Biochem (Tokyo) 133:453460. 126. Pazur, J.H., Jensen, P.J., and Murray, A.K. 2000. Oligosaccharides as Immunodeterminants of Erythropoieti n for Two Sets of Anti-Carbohydrate Antibodies. Journal of Protein Chemistry 19:631-635. 127. Brigmon RL, L.D.A.H.Y.J.Z.SG. 1997. Production of a cell wall-specific monoclonal antibody to Fibrobacter succinogenes . Lett. Appl. Microbiol. 25:261264. 128. Kim, S.H., Park, M.K., Kim, J.Y ., Chuong, P.D., Lee, Y.S., Yoon, B.S., Hwang, K.K., and Lim, Y.K. 2005. Development of a sandwich ELISA for the detection of Listeria spp. using specific flagella antibodies. J. Vet. Sci. 6:41-46. 129. Skjerve, E., Bos, W., and van der, G.B. 1991. Evaluation of monoclonal antibodies to Listeria monocytogenes flagella by checkerboard ELISA and cluster analysis. J. Immunol. Methods 144:11-17. 130. Peel, M., Donachie, W., and Shaw, A. 1988. Physical and antigenic heterogeneity in the flagellins of Listeria monocytogenes and L. ivanovii. J. Gen. Microbiol. 134:2593-2598. 131. Allerberger, F. 2003. Listeria: grow th, phenotypic differentiation and molecular microbiology. FEMS Immunology and Medical Microbiology 35:183-189.

PAGE 153

142 132. Ravn, P., Danielczyk, A., Bak Jensen, K ., Kristensen, P., Astrup Christensen, P., Larsen, M., Karsten, U., and Goletz, S. 2004. Multivalent scFv Display of Phagemid Repertoires for the Selection of Carbohydrate-specific Antibodies and its Application to the Thom sen-Friedenreich Antigen. Journal of Molecular Biology 343:985-996. 133. Dons, L. 1992. Cloning and characteriza tion of a gene encoding flagellin of Listeria monocytogenes . Mol Microbiol. 6:2919-2929. 134. Brockmann, E.C., Cooper, M., Stroms ten, N., Vehniainen, M., and Saviranta, P. 2005. Selecting for antibody scFv fragments with improved stability using phage display with denaturatio n under reducing conditions. Journal of Immunological Methods 296:159-170. 135. Lathrop, A.A., Jaradat, Z.W., Haley, T., and Bhunia, A.K. 2003. Characterization and application of a Listeria monocytogenes reactive monoclonal antibody C11E9 in a resonant mirror biosensor. J. Immunol. Methods 281:119-128. 136. Yu, K.Y., Noh, Y., Chung, M., Park, H.J., Lee, N., Youn, M., Jung, B.Y., and Youn, B.S. 2004. Use of monoclonal antibodies that recognize p60 for identification of Listeria monocytogenes . Clin. Diagn. Lab Immunol. 11:446-451. 137. Pilgrim, S., Kolb-Maurer, A., Ge ntschev, I., Goebel, W., and Kuhn, M. 2003. Deletion of the Gene Encoding p60 in Listeria monocytogenes Leads to Abnormal Cell Division and Loss of Actin-Based Motility. Infection and Immunity 71:34733484. 138. Muyldermans, S., Cambillau, C., a nd Wyns, L. 2001. Recognition of antigens by single-domain antibody fragments: the s uperfluous luxury of paired domains. Trends in Biochemical Sciences 26:230-235. 139. MacCallum, R.M., Martin, A.C.R., and Thornton, J.M. 1996. Antibody-antigen Interactions: Contact Analys is and Binding Site Topography. Journal of Molecular Biology 262:732-745. 140. Carroll, S.A., Hain, T., Technow, U., Da rji, A., Pashalidis, P., Joseph, S.W., and Chakraborty, T. 2003. Identification and characterization of a peptidoglycan hydrolase, MurA, of Listeria monocytogenes , a muramidase needed for cell separation. J. Bacteriol. 185:6801-6808. 141. Yau, K.Y., Dubuc, G., Li, S., Hirama, T., MacKenzie, C.R., Jermutus, L., Hall, J.C., and Tanha, J. 2005. Affinity matura tion of a V(H)H by mutational hotspot randomization. J. Immunol. Methods 297:213-224. 142. Suzuki, Y., Ito, S., Otsuka, K., Iwasawa, E., Nakajima, M., and Yamaguchi, I. 2005. Preparation of functi onal single-chain antibodies against bioactive gibberellins by utilizing randomly mu tagenized phage-display libraries. Biosci. Biotechnol. Biochem. 69:610-619.

PAGE 154

143 143. Swaminathan, S. and Khanna, N. 1999. Affinity Purification of Recombinant Interferon-[alpha] on a Mimetic Ligand Adsorbent. Protein Expression and Purification 15:236-242. 144. Verma, R., Boleti, E., and Ge orge, A.J.T. 1998. Antibody engineering: Comparison of bacterial, yeast, inse ct and mammalian expression systems. Journal of Immunological Methods 216:165-181. 145. Popowska, M. and Markiewicz, Z. 2004. Classes and functions of Listeria monocytogenes surface proteins. Pol. J. Microbiol. 53:75-88. 146. Chakrabarti, R., Klibanov, A.M., and Friesner, R.A. 2005. Computational prediction of native protein ligand-bindi ng and enzyme active site sequences. PNAS 102:10153-10158. 147. Wang, X.B., Zhou, B., Yin, C.C., Lin, Q., and Huang, H.L. 2004. A New Approach for Rapidly Reshaping Single-Chain Antibody In Vitro by Combining DNA Shuffling with Ribosome Display. J. Biochem (Tokyo) 136:19-28. 148. Bose, B., Chugh, D.A., Kala, M., Acha rya, S.K., Khanna, N., and Sinha, S. 2003. Characterization and molecular modeling of a highly santi-Hepatitis B surface antigen scFv. Molecular Immunology 40:617-631. 149. Irving, R.A., Kortt, A.A., and Hudson, P.J. 1996. Affinity maturation of recombinant antibodies using E. coli mutator cells. Immunotechnology 2:127-143.

PAGE 155

144 BIOGRAPHICAL SKETCH Harald Messer Jr. was born on Travis A.F.B. outside Sacramento, CA, in 1974. Following brief periods of residence with hi s family at Clark A.F.B., Philippines, and Warren A.F.B. in Cheyenne, WY, he moved with his family in 1981 to Eglin A.F.B., in the panhandle of Northwest Florida. He gr aduated from Niceville Senior High School in 1993. In August of 1993, he began his unde rgraduate studies at Okaloosa-Walton College and in 1997 he received his Bachelor of Science degree from the University of Florida in microbiology and cell science. In August of 2003 he began his graduate studies at the University of Florida’s Co llege of Medicine in the Department of Molecular Genetics and Microbiology.