MOLECULAR CHARACTERIZATION OF EDWARDSIELLA SPP. AND FLAVOBACTERIUM COLUMNARE Except where reference is made to the work of others, the work described in this dissertation is my own or was done in collaboration with my advisory committee. This dissertation does not include proprietary or classified information. Yinfeng Zhang Certificate of Approval: John M. Grizzle Covadonga R. Arias, Chair Professor Associate Professor Fisheries and Allied Fisheries and Allied Aquacultures Aquacultures Craig A. Shoemaker Stuart B. Price Affiliate Assistant Professor Associate Professor Fisheries and Allied Pathobiology Aquacultures Kenneth M. Halanych Joe F. Pittman Associate Professor Interim Dean Biological Sciences Graduate School MOLECULAR CHARACTERIZATION OF EDWARDSIELLA SPP. AND FLAVOBACTERIUM COLUMNARE Yinfeng Zhang A Dissertation Submitted to the Graduate Faculty of Auburn University in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy Auburn, Alabama August 4, 2007 iii MOLECULAR CHARACTERIZATION OF EDWARDSIELLA SPP. AND FLAVOBACTERIUM COLUMNARE Yinfeng Zhang Permission is granted to Auburn University to make copies of this dissertation at its discretion, upon request of individuals or institutions at their expense. The author reserves all publication rights. Signature of Author Date of Graduation iv VITA Yinfeng Zhang, daughter of Liren Zhang and Fengling Wang, was born on April 29, 1979, in Shuangcheng, Hei Long Jiang province, China. She graduated from Zhaolin High school of Shuangcheng city, Hei Longjiang Province, China in 1998. In the same year, she attended Dalian Fisheries University, Dalian, Liaoning province, China. In 2002, she received a Bachelor degree in Agricultural Sciences from Dalian Fisheries University. Upon graduation, she entered Auburn University and became a graduate research assistant in the Department of Fisheries and Allied Aquacultures. After studying for two years under the supervision of Dr. Grizzle, she graduated with a Master of Science degree majoring in fish diseases in 2004. She continued her study in the same Department pursuing a Doctor of Philosophy degree. During her PhD, she got married with Mingkang Jiang, a graduate student in the Department of Fisheries and Allied Aquacultures. They have a baby boy Austin N. Jiang. v DISSERTATION ABSTRACT MOLECULAR CHARACTERIZATION OF EDWARDSIELLA SPP. AND FLAVOBACTERIUM COLUMNARE Yinfeng Zhang Doctor of Philosophy, August 4, 2007 (M.S. Auburn University, 2004) (B.A., Dalian Fisheries University, 2002) 144 Typed Pages Directed by Covadonga R. Arias Bacterial diseases are responsible for large economic losses in aquaculture around the world. Flavobacterium columnare and Edwardsiella spp. negatively impact the channel catfish industry in southern USA. However, limited biological information is available for F. columnare and Edwardsiella ictaluri, which has hampered the development of accurate detection methods. In addition, the mechanisms of virulence of these pathogens are poorly understood which has prevented us to propose effective control/treatment methods for them. In the present study, new signature sequences for these pathogens were identified and evaluated as target candidates in PCR-based approaches. Unfortunately, multiplex PCR and real-time PCR tests developed in this vi study failed to provide the required specificity and sensitivity and could not be implemented as diagnostic tools. However, an intervening sequence (IVS) was discovered in E. ictaluri when sequencing the 23S rRNA gene. IVS are seldom found in bacteria and this is the first time an IVS has been described in the genus Edwardsiella. This IVS exhibited 97% similarity to the IVS in Salmonella typhimurium. A 23S rRNA gene-based phylogenetic tree was constructed placing E. ictaluri and E. tarda in context with other enterobacteria. This tree showed that Edwardsiella spp. are phylogenetically closer to the genus Erwinia than to the core members of the Enterobacteriaceae family. A phenotypic and genotypic comparison among F. columnare strains with different degrees of virulence was carried out in order to identify virulence markers. Lipopolysaccharide (LPS) and total protein profiles were characterized to illustrate the phenotypic differences between virulent and avirulent F. columnare strains. A F. columnare avirulent mutant lacked the high molecular weight bands of the LPS but showed two low molecular weight proteins that were absent in virulent strains. Four putative virulence genes were identified (gtf, hemH, norB, trx) by partially sequencing a shotgun genomic library from a virulent F. columnare strain. Nucleotide sequences of these genes divided the F. columnare strains analyzed into two populations that correlated well with previously described genomovars. Expression of these genes differed among strains under the same conditions. Differential gene expression was also observed when cells were grown under iron-restricted conditions and in the presence of catfish skin explants. These results provide new insights into the understanding of genetics and pathogenesis of F. columnare. vii ACKNOWLEDGMENTS I thank Dr. Arias for her supervision and encouragement. I also want to thank Dr. Klesius for allowing me to work at the Aquatic Animal Health Research Laboratory (ARS-USDA, Auburn, AL) and Dr. Shoemaker for his guidance and help during my research. Thanks also go to Dr. Olivares-Fuster for his technical advice and suggestions. I want to thank Paige Mumma and Dr. Shelby (Aquatic Animal Health Research Laboratory ARS-USDA, Auburn, AL) for technical laboratory assistance. Thanks to Dr. Xu (Aquatic Animal Health Research Laboratory ARS-USDA, Auburn, AL) for helping with the channel catfish skin explants preparation techniques. Thanks to Drs. John Grizzle and Jeff Terhune from the Department of Fisheries and Allied Aquacultures, Auburn University, Bill Hemstreet from the Alabama Fish Farming Center, and Dr. Al Camus from the Delta Research and Extension Center, Mississippi, for providing and/or isolating some of the bacteria used in this study. I would like to thank all my committee members for their suggestions and encouragement. Thanks are also due to family members Liren Zhang, Fengling Wang, and Mingkang Jiang for their support during the course of my research. viii Journal format used Journal of Aquatic Animal Health Computer software used Microsoft Word 2001 for Windows XP, Adobe Photoshop 7.0, Vector NTI ? Suite 8, Applied Biosystems Primer Express software, BioNumerics 4.0, ClustalW, mfold (prediction of RNA secondary structure (M. Zuker)) ix TABLE OF CONTENTS LISTS OF TABLES........................................................................................................... xi LISTS OF FIGURES ....................................................................................................... xiii CHAPTER I. INTRODUCTION.........................................................................................1 CHAPTER II. EVALUATION OF REAL-TIME PCR FOR DETECTION OF EDWARDSIELLA SPP. AND FLAVOBACTERIUM COLUMNARE ......21 Abstract.............................................................................................21 Introduction.......................................................................................23 Materials and Methods......................................................................25 Results...............................................................................................33 Discussion.........................................................................................35 Tables................................................................................................37 Figures ..............................................................................................40 CHAPTER III. IDENTIFICATION AND CHARACTERIZATION OF AN INTERVENING SEQUENCE WITHIN THE 23S RIBOSOMAL RNA GENE OF EDWARDSIELLA ICTALURI .......................................41 Abstract.............................................................................................41 Introduction.......................................................................................42 Materials and Methods......................................................................44 Results...............................................................................................48 Discussion.........................................................................................51 Tables................................................................................................55 Figures ..............................................................................................57 x CHAPTER IV. COMPARISON OF LIPOPOLYSACCHARIDE AND PROTEIN PROFILES BETWEEN FLAVOBACTERIUM COLUMNARE STRAINS FROM DIFFERENT GENOMOVARS .................................61 Abstract.............................................................................................61 Introduction.......................................................................................62 Materials and Methods......................................................................63 Results...............................................................................................66 Discussion.........................................................................................69 Figures ..............................................................................................73 CHAPTER V. IDENTIFICATION, CHARACTERIZATION AND EXPRESSION PATTERNS OF GTF, HEMH, NORB, AND TRX GENES IN FLAVOBACTERIUM COLUMNARE........................................................77 Abstract.............................................................................................77 Introduction.......................................................................................79 Materials and Methods......................................................................81 Results...............................................................................................85 Discussion.........................................................................................88 Tables................................................................................................92 Figures ..............................................................................................96 CHAPTER VI. CONCLUSIONS ......................................................................................99 CUMULATIVE BIBLIOGRAPHY ................................................................................104 xi LISTS OF TABLES Table II.1. Isolates used in the PCR detection study ........................................................37 Table II.2. Comparison of real-time PCR, microbial culture and IFA for detection of E. ictaluri. Tank 1-3 contained 9 challenged fish each. Only 5 fish in tank 3 were alive at the 14 d post-challenge. Dead fish were not tested. Non-challenged fish were alive and negative by all detection methods (data not shown) .............................................................................................39 Table III.1. Strains used in the IVS study..........................................................................55 Table III.2. Comparison of E. ictaluri helix-45 IVS with other Enterobacteriaceae helix-45 IVSs. Nucleotide sequence identities over 80% are shown in bold ...........................................................56 Table V.1. Strains of F. columnare used in the study showing genomovar (G) ascription and standard PCR amplification results of gtf (glycosyltransferase gene), hemH (ferrochelatase gene), norB (nitric oxide reductase gene), and trx (thioredoxin gene). ?+? represents positive PCR amplification of the gene, ?-? represents no amplification when DNA was used as template...............92 Table V.2. Primers for gene detection and gene expression of F. columnare.................93 xii Table V.3. RT-PCR results from nine strains of F. columnare. G represents genomovar ascription. ???represents no expression, ?w? represents weak expression, ?+? represents expression, ?++? represents strong expression. NA, no amplified product was obtained when genomic DNA was used as template.............................94 Table V.4. F. columnare ALG-00-530 gene expression under standard conditions, iron-limited conditions, and in the presence of catfish skin explants. C 0 is the baseline control for catfish skin explant experiment. ?-? represents no expression, relative intensity of the expressed genes is expressed by the number of ?+? symbols...............................................................................................95 xiii LISTS OF FIGURES Figure I.1. Modified Eubacteria domain phylogenetic tree based on 16S rDNA sequences (Madigan et al. 2003). Numbers reflect percentage of fish pathogenic bacterial species within each group to total bacterial fish pathogens (calculated based on the fish bacterial disease list by Austin and Austin (1999)) .............................................................................................13 Figure II.1. Specificity test of the multiplex PCR. Lane 1, E. ictaluri; lane 2, E. tarda; lane 3: F. columnare; lane 4, F. psychrophilum; lane 5, F. johnsoniae,; lane 6, Y. ruckeri; lane 7, S. typhimurium; lane 8, F. hydatis; lane 9, Escherichia coli. Primers used for top gel (A) were universal forward primer and EicEta-23S-R. Primers used for bottom gel (B) were universal forward primer with Fco-23S-R .............................40 Figure III.1. Parsimony tree based on 23S rRNA gene sequences. The tree shown is a consensus bootstrap tree based on 500 resampled parsimony trees. Significant (>50) bootstrap values for each branch are indicated. Species names and GenBank accession numbers are noted as well. Bar shows 5% sequence divergence.....................................57 xiv Figure III.2. Proposed secondary structure of IVS from E. ictaluri (A) and S. typhimurium (B). Secondary structure predictions based on free-energy minimization. Nucleotide differences in E. ictaluri IVS compared with S. typhimurium IVS are marked by asterisks. The proposed conserved site at which helix-45 was replaced by the IVS is indicated by a horizontal line ...........................................................58 Figure III.3. Agarose gel containing the PCR products from amplification of 23S rRNA gene containing the IVS. Lane M, 50 bp molecular marker. Lane C, no template DNA; lane 1, Escherichia coli; lane 2, E. tarda CECT 849; lane 3 to lane 9, E. ictaluri isolates: CECT 885, EILO, 195, 196, 151, 218 and 219, respectively. rrl operons containing IVS in helix-45 yielded a 152 bp amplified product. Operons lacking IVS generated a smaller amplicon of 54 bp. This 54 bp amplified should not be confused with the primer-dimer band observed in the no-template control.............................59 Figure III.4. rRNAs of E. ictaluri CECT 885 (1), EILO (2), E. tarda CECT 849 (3) and Escherichia coli (4). Total RNA was electrophoresed, blotted, and stained with methylene blue as described in Material and Methods..................................................................................60 Figure IV.1. Dendrogram based on AFLP patterns of four strains of F. columnare (ALG-00-530, FC-RR, ARS-1, and ALG-03-063). The tree was derived by UPGMA cluster analysis of the AFLP profiles.........................................................................................................73 xv Figure IV.2. Silver stained sodium dodecyl sulfate-polyacrylamide gel electrophoresis profiles of LPS preparation from F. columnare strains used in this study. Lane 1: molecular standard; lanes 2 through 5, LPS aqueous phase from: ALG-00-530 (2), FC-RR (3), ARS-1 aqueous (4), and ALG-03-063 (5); lanes 6 through 9, show LPS phenol phase from ALG-00-530 (6), FC-RR (7), ARS-1 (8) and ALG-03-063 (9)..............................................................................74 Figure IV.3. Immunoblot using anti-ALG-00-530 serum (A) and anti-FC-RR serum (B) with LPS preparation. Lane 1: molecular standard; lanes 2 through 5, LPS aqueous phase from: ALG-00-530 (2), FC-RR (3), ARS-1 aqueous (4), and ALG-03-063 (5); lanes 6 through 9, show LPS phenol phase from ALG-00-530 (6), FC-RR (7), ARS-1 (8) and ALG-03-063 (9)...............................................75 Figure IV.4. Western blot analysis of total protein with anti-ALG-00-530 serum (A) and anti-FC-RR serum (B). Lane 1, Molecular standard; lane 2, ALG-00-530; lane 3, FC-RR; lane 4, ARS-1; lane 5, ALG-03-063 ................................................................................................76 Figure V.1. Nucleotide sequence alignment of gtf, norB, and trx of F. columnare from genomovars I and II (GenBank assession numbers in brackets). 530 stands for ALG-00-530 (genomovar II) and 49512 is the abbreviation of ATCC 49512 (genomovar I). Nucleotides that are different between genomovars are indicated xvi by shaded/unshaded box. * indicates putative start condon. ?.? indicates missing nucleotides. Y=C+T, R=A+G, and W=A+T .................................96 Figure V.2. hemH, norB, and trx gene expression of F. columnare strain ALG-00-530 by reverse transcription PCR (RT-PCR). Expression patterns are shown under iron-limited conditions in the presence of 2, 2-dipyridyl and in Shieh broth at 0 h, 2 h, 4 h, 6 h, 8 h, and 24 h. 16S rDNA (16S) was used as an internal control........................................................................................98 1 I. INTRODUCTION Channel catfish (Ictalurus punctatus Rafinesque) is the most important fish species commercially cultured in the USA. Total foodsize catfish sales reached $460 million in 2005 (USDA 2005). Alabama, Arkansas, Louisiana, and Mississippi are the major catfish producing states in the USA, accounting for more than 90% of the total catfish sales in 2002 with more than 70% of all catfish operations (USDA 2003a). Channel catfish belongs to the family Ictaluridae within the order of Siluriformes (Tucker 1985). This species is characterized by having eight barbels around the mouth; deeply-forked tail; spots on body sides; and soft-rayed fins, although the dorsal and pectoral fins contain sharp, hard spines (Wellborn 1988). Channel catfish possess several desirable attributes for commercial production (Tucker 1985). This species is easy to spawn and usually do not reproduce in culture ponds. Channel catfish is hardy due to its tolerance to low oxygen, crowding, and its well adaptation to the various culture systems. It also presents an efficient feed conversion rate and accepts manufactured feed. All these characteristics make channel catfish a model species for aquaculture production. The main factor limiting expansion and profitability of the catfish industry is disease control. Currently, the two most prevalent diseases are: enteric septicemia of catfish (ESC) caused by Edwardsiella ictaluri and columnaris disease caused by Flavobacterium columnare (USDA 2003a). Other diseases affecting the catfish industry are caused by the protozoan Ichthyophthirius multifiliis, the triactinomyxid myxozoan 2 Aurantiactinomyxon ictaluri (proliferative gill disease (PGD)), trematodes, opportunistic bacterial pathogens, channel catfish virus, channel catfish anemia, and visceral toxicosis of catfish (USDA 2003a). Infectious disease is a main factor in fish losses. Sixty percent of foodsize fish operation experienced ESC and 50% experienced columnaris disease. In fingerling production farms, 50% of losses were caused by ESC and 45% were due to columnaris disease during 2001-2002 (USDA 2003a; USDA 2003b). Currently, E. ictaluri and F. columnare are considered the two most important bacterial pathogens for the channel catfish industry (Wagner et al. 2002). Although E. tarda does not impact channel catfish aquaculture production as much as E. ictaluri and F. columnare, it was included in this study for comparison purposes. Both E. tarda and E. ictaluri are thought to be genetically similar (Hawke et al. 1981); however, while E. ictaluri is a fairly specific pathogen for catfish, E. tarda presents a broader host range. Therefore, a genetic characterization of these two species might shed some light on their host-specificity. Enteric septicemia of catfish The genus Edwardsiella was first described in 1965, with E. tarda as the type species (Ewing et al. 1965). A second species was isolated from reptiles and birds, and was characterized by Grimont et al. (1980) as E. hoshinae. E. ictaluri, the causal agent of ESC was first isolated in 1976 (Hawke 1979); however, the bacterium was not characterized and classified until 1979 (Hawke et al. 1981). Characteristics ? The physiological and biochemical characteristics of E. ictaluri have been described in detail previously (Hawke 1979, 1981; Waltman et al. 1986). E. ictaluri is a short, Gram-negative rod with a dimension of about 0.8 X 3 ?m. It is cytochrome oxidase negative and ferments glucose. E. ictaluri does not produce H 2 S 3 from triple sugar iron agar and is negative for indole production as well. E. ictaluri is not able to grow with NaCl concentrations higher than 1.5%. Optimum growth temperature for E. ictaluri is between 25 ?C and 30 ?C. Regardless of origin, all E. ictaluri isolates contain plasmids (Speyerer and Boyle 1987). E. ictaluri is more related to E. tarda than to the other members in the family of Enterobacteriaceae based on DNA-DNA hybridization (Hawke et al. 1981). Epidemiology ? Enteric septicemia of catfish is a seasonal disease with high prevalence during May and June, and during September and October. E. ictaluri has a high infectivity rate in cultured channel catfish as all sizes of fish susceptible to ESC (Plumb 1999). Naturally occurring infections by E. ictaluri have been reported in walking catfish (Clarias batrachus Linnaeus) (Kasornchandra et al. 1987), white catfish (Ameiurus catus Linnaeus) (Hawke et al. 1981) , brown bullhead (Ameiurus nebulosus Lesueur) (Hawke et al. 1981), freshwater catfish (Pangasius hypophthalmus Sauvage) (Crumlish et al. 2002), green knifefish (Eigenmannia virescens Valenciennes) (Kent and Lyons 1982), danio (sind) (Devario devario Hamilton) (Waltman et al. 1985), rainbow trout (Oncorhynchus mykiss Walbaum) (Keskin et al. 2004), and tadpole madtom (Noturus gyrinus Mitchill) (Klesius et al. 2003). Experimental infections by injecting E. ictaluri cells intraperitoneally showed that tilapia (Sarotherodon aureus Steindachner) was slightly susceptible to E. ictaluri, while golden shiner (Notemigonus crysoleucas Mitchill), bighead carp (Aristichthys nolilis Richardson), largemouth bass (Micropterus salmoides Lacep?de), and blue catfish (Ictalurus furcatus Valenciennes) were resistant to this pathogen (Plumb and Sanchez 1983; Wolters et al. 1996). 4 Clinical signs and histopathology ? There are two forms of ESC: acute and chronic. In the acute form, diseased fish suffering from ESC hang with a head-up-tail-down posture and exhibit spinning swimming behavior. External lesions caused by E. ictaluri include petechial hemorrhages on skin, pale gills, exophthalmia, and small cutaneous lesions on the body surface (Hawke 1979). Internally, peritoneal cavity may contain bloody ascitic fluid, and the intestine may exhibit petechial hemorrhages (Hawke 1979). Multifocal necrosis of liver and swollen trunk kidney can be observed (Newton et al. 1989). Microscopically, hepatocytes are swollen and vacuolated and the exocrine pancrease around hepatic vessels is necrotic (Newton et al. 1989). In the chronic form, olfactory epithelia appear degenerated and granulomatous inflammation is found in olfactory lamellae. The brain is swollen and ulcerated (Newton et al. 1989; Morrison and Plumb 1994). An open lesion can develop through the skull giving the disease its common name ?hole in the head? disease. Diagnosis ? Classical fish pathogen identification relies on microbial culture techniques followed by biochemical characterization of the isolates. E. ictaluri can be isolated from infected organs by using general culture media such as brain heart infusion agar (BHI), trypticase soy agar (TSA) (Meyer and Bullock 1973), and blood agar. A selective medium (E. ictaluri medium or EIM) for the recovery of E. ictaluri was developed by Shotts and Waltman (1990). Colonies of E. ictaluri are smooth, circular, and slightly convex (Hawke 1979). An E. ictaluri colony on EIM appears clear and greenish, which can be distinguished from other Gram-negative bacteria, while the growth of Gram-positive bacteria is inhibited. Motility at 37?C, no indole production in 5 tryptone broth, and no H 2 S production on triple sugar iron differentiate E. ictaluri from E. tarda. Besides culture methods, some alternative techniques for direct detection of E. ictaluri have been described. Serology methods for detecting E. ictaluri have been developed by several groups. Enzyme immunosorbent assay (ELISA) has been used to detect E. ictaluri in decomposing channel catfish (Hanson and Rogers 1989). Indirect fluorescent antibody (IFA) test has been developed to diagnose E. ictaluri from artificially infected channel catfish fingerlings (Ainsworth et al. 1986). Though IFA had 90.3% correlation with culture methods, some false-negative results have been reported (Ainsworth et al. 1986). However, Panangala et al. (2006) described an efficient IFA test for E. ictaluri with no false negative results. In addition, indirect ELISAs have been developed to detect catfish serum antibodies against E. ictaluri (Waterstrat et al. 1989; Klesius et al. 1991). Fluorescent in situ hybridization (FISH) has also been used for diagnosis of E. ictaluri. A probe targeting an E. ictaluri plasmid was radiolabeled for detection of this bacterium in channel catfish. This probe did not hybridize with Escherichia coli, Aeromonas hydrophila or E. tarda (Speyerer and Boyle 1987). Reid and Boyle (1989) found that the same plasmid probe hybridized to all the tested E. ictaluri isolates from channel catfish and non-channel catfish except to an isolate from Maryland recovered from white catfish (Ameiurus catus Linnaeus). Polymerase chain reaction (PCR) is another diagnostic method widely used in clinical diagnosis. Since its description by Mullis and Faloona (1987), PCR has been proved to be a very useful tool for pathogen detection. Detection by PCR can be very 6 specific, sensitive and does not require pathogen isolation from the sample. PCR is a method by which nucleic acid sequences can be exponentially amplified in vitro (Mullis and Faloona 1987). This technique enables researchers to generate millions of copies of a single DNA molecule in a short period of time by using of a template DNA, DNA polymerase, nucleotides (A, T, G, C), and primers. A more advanced PCR technique, named real-time PCR (or quantitative PCR), is a modification of standard PCR. This technique simultaneously amplifies and quantifies target DNAs. Therefore, it will not only determine the presence of a specific DNA sequence but measure the number of copies of DNA. By monitoring fluorescence intensities in a real-time PCR apparatus, quantification of DNA can be accomplished by adding double-stranded DNA dyes such as SYBR Green (Yin et al. 2001), or using fluorescent labelled probes such as TaqMan probes (Heid et al. 1996), hybridization probes (Dietmaier and Hofstadter 2001), and molecular beacons (Tyagi and Kramer 1996). The main advantage of real-time PCR over standard PCR is that real-time PCR allows quantification of the target nucleic acids. In addition, real-time PCR data are collected during DNA amplification so that contamination risk is reduced. The use of real-time PCR for fish bacterial pathogen detection could be very useful in monitoring changes of pathogen loads in the host or the environment. When bacterial loads are low, accurate quantitative determination of the pathogen can be problematic. The use of real-time PCR might offer a better diagnostic sensitivity than standard PCR methods. Real-time PCR might allow accurate detection of fish pathogens under a preclinical stage. 7 Pathogen detection prior to the onset of disease will help to implement preventive measures and avoid infectious outbreaks. To date, no standard PCR protocol has been described for E. ictaluri. However, a real-time PCR protocol targeting a transposon was developed to detect E. ictaluri in fish samples. This real-time assay was sensitive enough to detect as few as two to three E. ictaluri cells from mixtures of noninfected catfish blood and E. ictaluri cells (Bilodeau et al. 2003). Interestingly, transposons are variable genetic elements that a priori are not considered good targets for PCR detection since they are mobile elements that might not be present in all E. ictaluri isolates. Loop-mediated isothermal amplification (LAMP), another nucleic acid amplification method, has been desribed for detection of E. ictaluri (Yeh et al. 2005) as well. By targeting a putative antigenic gene, eip18, this LAMP assay could detect as few as 20 colony forming units (CFU). The specificity of this method was tested and found to be satisfactory (Yeh et al. 2005). Pathogenesis ? Little is known about the virulence mechanisms of E. ictaluri. Possible virulence factors may include bacterial cell surface material, hemolysin, lipopolysaccharide (LPS), and hydrolytic enzymes. Virulent isolates of E. ictaluri contained greater amounts of capsular material and surface proteins and showed higher chondroitinase activity than avirulent isolates (Stanley et al. 1994). Futher work is needed to illustrate the role of surface protein and chondroitinase in E. ictaluri virulence. Although Williams et al. (2003) revealed that an attenuated E. ictaluri strain exhibited decreased hemolysin activity, hemolysin-deficient strains did not exhibit a reduction in virulence (Williams and Lawrence 2005). Arias et al. (2003) reported that a 8 rifampicin-mutant of E. ictaluri (RE-33) used as live vaccine, lacked the high molecular weight bands of the LPS compared with the parent virulent strain; however, RE-33 can still protect channel catfish against ESC. Compared to RE-33, another E. ictaluri strain without O-antigen of LPS can not efficiently protect channel catfish against E. ictaluri (Lawrence and Banes 2005). A recent study showed that a urease gene may play an important role in the intracellular replication of E. ictaluri in the vacuoles of channel catfish macrophages (Booth 2005). Further investigation is needed to clarify the role of urease in E. ictaluri pathogenecity. Treatment and prevention ? The first attempt to control ESC was to feed diseased fish with antibiotic medicated food (USDA 2003a). Ormetoprim- sulfadimethoxine, and Aquaflor (Schering-Plough, Kenilworth, NJ) are the approved drugs for treating ESC in catfish in the USA. Ormetoprim-sulfadimethoxine is known to inhibit bacterial growth. The effective ingredient in Aquaflor is florfenicol that inhibits protein synthesis in bacteria (Aquaflor product labeling http://www.aquaflor-usa.com). However, antibiotic therapy should not be considered the best management practice since inappropriate use of antibiotics may generate drug resistant bacterial isolates (Hawke et al. 1998). Vaccination is one of the best options to prevent infectious diseases. Classical vaccines can be made out of living, attenuated or inactivated microorganisms or purified macromolecular components derived from them (Abbas and Lichtman 2000). Vaccines have the benefit to induce protective immunity against microbial pathogens. Vaccination with inactivated E. ictaluri resulted in no strong acquired immunity to protect channel catfish against E. ictaluri infection (Thune et al. 1994). As an intracellular pathogen, a 9 modified-live E. ictaluri vaccine may be more effective in inducing acquired immunity in catfish. A modified-live vaccine developed by Klesius and Shoemaker (1999) is now commercially available to farmers (AQUAVAC-ESC, Intervet Inc. Millsboro, DE). This vaccine was developed using a rifampicin-resistant strategy that was previously used to generate the Brucella abortus live vaccine (RB51) against cattle brucellosis (Schurig et al. 1991). The rifampicin-resistant mutant was derived by repeated passage on nutrient agar supplemented with rifampicin at varying concentrations. This live E. ictaluri vaccine is unable to cause ESC but is capable of protecting catfish against ESC (Klesius and Shoemaker 1999). Channel catfish fry are vaccinated by being immersed in a bath containing the ESC vaccine (Shoemaker et al. 1999). In 2001-2002, 11.4% of fingerling operations vaccinated fry against ESC using AQUAVAC-ESC (USDA 2003b). Edwardsiella tarda septicemia The disease caused by E. tarda was first discovered in cultured eel (Hoshina 1962). This disease was first called emphysematous putrefactive disease in catfish (Meyer and Bullock 1973). However, it was referred as Edwardsiellosis by Plumb (1999) and appears as Edwardsiella tarda septicemia in the AFS-FHS blue book (Hawke 2003). Characteristics ?E. tarda is a small, straight, Gram-negative rod that is 1 ?m in diameter and 2-3 ?m in length. It is motile, catalase positive, cytochrome oxidase negative, lactose negative, and ferments glucose. E. tarda produces indole in tryptone broth and H 2 S on triple sugar iron slants. Epidemiology ? E. tarda has a broader host range than E. ictaluri, and it can infect both freshwater and marine fish species. Plumb (1993) reported a list of fish species infected by E. tarda such as goldfish (Carassius auratus auratus Linnaeus), 10 common carp (Cyprinus carpio carpio Linnaeus), grass carp (Ctenopharyngodon idella Valenciennes), and largemouth bass (M. salmoides). Miyazaki (1985) reported the infection in Japanese flounder (Paralichthys olivaceus Temminck & Schlegel), red seabream (Chrysophrys major Temminck & Schlegel), nile tilapia (Oreochromis niloticus niloticus Linnaeus), and Japanese eel (Anguilla japonica Temminck & Schlegel). E. tarda can also be isolated from other animals such as frogs (Sharma et al. 1974), reptiles (Sechter et al. 1983), swine (Owens et al. 1974), and humans (Sechter et al. 1983). The most prominent fish species infected by this organism are Japanese eels (A. japonica) and channel catfish (I. punctatus) (Meyer and Bullock 1973; Plumb 1999). Clinical signs ? E. tarda septicemia causes different clinical signs in different fish species. Diseased Japanese eels display petechial hemorrhages on the ventrum, and the anal region is swollen. Two forms of E. tarda septicemia occur in Japanese eels: nephric and hepatic (Miyazaki 1985). The nephric form is characterized by necrotic renal foci that can spread to spleen, liver, gills, stomach, and heart. The hepatic form is associated with microabscesses in the liver and can spread to other organs. E. tarda causes small cutaneous lesions, and abscesses can develop as malodorous gas and necrotized tissue-filled cavities within muscle in channel catfish (Meyer and Bullock 1973). Diagnosis ?Bacteriological or serological methods can be used to identify E. tarda. Isolation can be achieved on BHI agar or TSA agar. E. tarda forms small, green colonies with black center on EIM. Contrary to E. ictaluri, E. tarda produces indole, hydrogen sulphide and reduces methyl red. Standard PCR protocols for E. tarda have been developed, but they lacked the appropriate sensitivity and/or specificity (Aoki and 11 Hirono 1995; Chen and Lai 1998; Baird et al. 2003). A species-specific fragment cloned from a shotgun DNA genomic library was targeted to detect E. tarda; however, the analytical sensitivity of this PCR method was not well described (Aoki and Hirono 1995). Moreover, the authors failed to compare their PCR-based detection results with standard culture methods. Chen and Lai (1998) developed a PCR for E. tarda by using the hemolysin gene as the target, but the specificity of the primers was not tested. Finally, Baird et al. (2003) used the small ribosomal subunit as a target to identify Edwardsiella spp., but the assay sensitivity was not determined. Therefore, more effective and reliable PCR protocols are desirable for detection of E. tarda. Development of PCR protocols for the detection of E. tarda need to consider that E. ictaluri and E. tarda contain almost identical 16S rRNA gene sequences (> 99 %) and share a high similarity of 16S-23S intergenic spacer regions (Panangala et al. 2005). Specific molecular markers are necessary to differentiate E. tarda from E. ictaluri. Pathogenesis ?several virulence factors are believed to contribute to E. tarda pathogenesis. These include exotoxins (Ullah and Arai 1983), hemolysin production (Hirono et al. 1997), ability to resist serum-mediated killing, and invasion of epithelial cells (Janda et al. 1991). Several virulence genes have been identified in E. tarda by transposon mutagenesis using a fish infection model. These genes are fimA (fimbrial protein precursor), gadB (glutamate decarboxylase isozyme), katB (catalase precursor), pstS (phosphate binding protein), pstC (peripheral membrane protein C), ssrB (secretory system regulator) (Srinivasa-Rao et al. 2003). Treatment and prevention? E. tarda infection can be treated with oxytetracycline antibiotic. However, drug-resistant strains of E. tarda have already been reported (Aoki 12 et al. 1987). A live rifampicin resistant vaccine against E. tarda in fish has been developed, but it is not commercially available yet (USA Patent 7067122 http://www.freepatentsonline.com/7067122.html). Columnaris disease F. columnare belongs to the Cytophaga-Flavobacterium-Bacteroides (CFB) group. Flavobacteria are phylogenetically distant from other Gram-negative and Gram-positive pathogenic bacteria (Figure I.2). Flavobacteria account for 13 % of total bacterial fish pathogens, while most of the fish bacterial pathogens fall within the ?- proteobacteria group. Experimental methods (i. e. bacteria culturing, protein and DNA extraction, and genetic manipulation) used with other common pathogens are not suitable for most Flavobacterium spp. It is possible that virulence mechanisms of Flavobacterium spp. may also differ from other common fish bacterial pathogens. The disease caused by F. columnare is called columnaris disease and was first described by Davis (1922). However, the first successful isolation of the organism was described about 20 years later (Ordal and Rucker 1944). The causative organism has suffered several taxonomic reclassifications. It has been referred to as Bacillus columnaris (Davis 1922), Chondrococcus columnaris (Ordal and Rucker 1944), Cytophaga columnaris (Garnjobst 1945), Flexibacter columnaris (Leadbetter 1974), and again Cytophaga columnaris (Reichenbach 1989). Bernardet and Grimont (1989) maintained the name Flexibacter columnaris by presenting the DNA relatedness to the type strain of F. columnare. This was the first time this bacterial pathogen was officially described as a species. This organism was renamed as Flavobacterium columnare in 1996 (Bernardet et al. 1996). 13 Characteristics ? The characteristics of F. columnare have been described in detail by Plumb (1999) and Bernardet et al. (1989). F. columnare is a Gram-negative rod that is 2-10 ?m long and 0.5 ?m in diameter, motile by gliding, producing catalase, producing H 2 S, hybrolyzing gelatin, and no growth with more than 1% NaCl. Colonies on agar are spreading and more or less rhizoid. Colony colors range from pale yellow, greenish yellow, yellow to golden yellow (Griffin 1992). F. columnare presents a G + C content of about 32 %. Aquifex Thermodesulfobacterium Thermotoga Green nonsulfur bacteria Deinococci Spirochetes Green sulfur bacteria 13% Flavobacteria Defferibacter Cytophaga Verrucomicrobia Chlamydia Planctomyces/Pirella Cyanobacteria Gram-positive 32% bacteria Nitrospira Proteobacteria H ig h G C L o w G C ) ( % & 2% ' 53% F. columnare species exhibits variation in serotypes and genotypes. Four serotypes and one miscellaneous group have been described in the species (Anacker and Ordal 1959). Additionally, genotypic differences have been reported based on the analysis of 16S rDNA-RFLP, intergenic spacer region (ISR), amplified fragment length Figure I.1. Modified Eubacteria domain phylogenetic tree based on 16S rDNA sequences (Madigan et al. 2003). Numbers reflect percentage of fish pathogenic bacterial species within each group to total bacterial fish pathogens (calculated based on the fish bacterial disease list by Austin and Austin (1999)). 14 polymorphism (AFLP), and random amplified polymorphism DNA (RAPD) (Toyama et al. 1996; Triyanto and Wakabayashi 1999; Arias et al. 2004; Thomas-Jinu and Goodwin 2004). Epidemiology ? F. columnare is distributed worldwide in aquatic environments, being capable of infecting most freshwater fish species. In general, fish are considered as the main reservoir for F. columnare. However, this bacterium can be isolated from water (Rucker et al. 1953), especially during epizootics (McCarthy 1975). Transmission of this pathogen can occur through water without fish to fish contact (Welker et al. 2005). Several factors are involved in the spread of the disease. Injured or mechanical abraded fish are more likely to be infected by F. columnare (Davis 1922; Bader et al. 2003a). Channel catfish deprived of feed are more susceptible to F. columnare-induced mortality (Shoemaker et al. 2003a). Water temperature also plays a major role in columnaris disease epidemiology. Although columnaris disease occurs throughout the whole year, warm-weather favors F. columnare infection (Davis 1922). Other conditions favoring F. columnare infection include crowding (Suomalainen et al. 2005), low oxygen (Chen et al. 1982), high ammonia (Chen et al. 1982), and high nitrite (Hanson and Grizzle 1985). F. columnare often appears associated with other pathogens as mixed infections. Davis (1922) observed that large quantity of other bacteria present in columnaris lesions besides F. columnare. Hawke and Thune (1992) showed that out of 53 F. columnare infectionss, 46 involved E. ictaluri and/or Aeromonas spp. Currently, it is unclear whether F. columnare is a primary pathogen or an opportunistic one. Clinical signs ? Columnaris disease usually begins as an external infection of the fins, body surface, or gills. Small lesions can start as areas of pale discoloration at the 15 base of the dorsal fin or occasionally at the base of the pelvic fin, and lead to deterioration of the fins. Initial skin lesions appear as discrete bluish areas that evolve into depigmented necrotic lesions. Skin lesions can have yellowish mucoid material accompanied by mild inflammation. The lesions can rapidly develop to cover a greater portion of the body (Davis 1922). Lesions on gills are localized as white patches or yellow-orange-brown necrotic areas depending on the presence of debris in the lesion (Davis 1922; Plumb 1999). Although F. columnare starts as an external infection, the skin of the fish may be eroded completely to expose the underlying muscle (Pacha and Ordal 1970). F. columnare can also become systemic without obvious pathological changes in fish internal organs (Plumb 1999). Diagnosis ? Columnaris is often diagnosed by typical lesions on the body surfaces, and the presence of columns or ?hay stacks? of bacteria in a wet mount. F. columnare can be isolated on low nutrient agar plate such as Cytophaga agar (Anacker and Ordal 1959), Hsu-Shotts agar (Bullock et al. 1986), or Shieh agar (Decostere et al. 1997). A simple five-step method to distinguish F. columnare from the other yellow pigment producing bacteria was developed by Griffin (1992): (1) ability to grow in the presence of neomycin sulfate and polymyxin B; (2) colonies rhizoid and yellow pigmented; (3) production of gelatin degrading enzymes; (4) colonies absorbing Congo red; and (5) production of an enzyme that degrades chondroitin sulfate A. Chowdhury and Wakabayashi (1990) described an indirect fluorescent antibody technique to detect F. columnare in fish and in water. In their study, immunodetection of F. columnare was more sensitive than plate culture methods. Fatty acid methyl ester analysis (FAME) is another rapid and reliable method for identification of F. columnare 16 (Shoemaker et al. 2005a). Specific PCR methods have also been developed to identify putative F. columnare isolates (Toyama et al. 1996; Bader and Shotts 1998; Bader et al. 2003b; Darwish et al. 2004). However, all the PCR methods required either two-rounds of PCR amplification or restriction analysis after PCR amplification. A new PCR protocol developed by Welker et al. (2005) exhibited both high specificity and sensitivity. However, this protocol could not be adapted into real-time PCR due to its large amplicon size (about 500 bp). Recently, Yeh et al. (2006) developed a LAMP method to detect F. columnare. The assay can detect as low as 30 pg of genomic DNA and was able to detect F. columnare from experimentally infected channel catfish. Pathogenesis ? Although columnaris disease has been known for about 80 years (Davis 1922), little work has addressed the mechanisms of this disease in part due to the fastidious nature of this bacterium when cultured under laboratory conditions. Most of the bibliographical references for F. columnare are related to its taxonomic status rather than to its biological properties. To date, most F. columnare sequences deposited in GenBank (http://www.ncbi.nlm.nih.gov/Genbank/) correspond to ribosomal genes. Some non-ribosomal sequences include gliding motility genes (fjo17, fjo23, gldH, gldJ, and murF), a prolyl oligopeptidase (g4), a major outer membrane protein (momp), an alginate-O-acetylation protein (algI), and a partial sequence of a metalloprotease-like gene. Extracellular protease production in F. columnare has been described by Bertolini and Rohovec (1992) and Newton et al. (1997). The production of proteases may explain the necrotizing characteristics of F. columnare infection. F. columnare produces a chondroitin AC lyase, which can break down polysaccharides of connective tissue 17 (Griffin 1991). A recent study indicated that chondroitin AC lyase activity is related to F. columnare virulence (Suomalainen et al. 2006). Bacterial adhesion also appears to be related to virulence of F. columnare (Zaldivar 1985; Decostere et al. 1999a). A lectin- like carbohydrate-binding substance may be responsible for attachment of F. columnare to gill (Decostere et al. 1999b). Currently, there is limited information available on F. columnare genetics and pathogenesis. Treatment and prevention ? Because of the ubiquitous presence of F. columnare in aquatic environments, it is unrealistic that eradication from fish farms will occur. Improved water-management has been used to reduce physiological and environmental stress in fish. Increasing salinity to 1? in the culture systems may help to reduce fish losses (Altinok and Grizzle, 2001). Antibiotic treatment is not effective due to high costs and drug use restrictions. However, columnaris disease might be prevented through vaccination. A commercially available rifampicin-resistant live vaccine is now approved and has just started to be used by farmers (AQUAVAC-COL, Intervet). Our ability to suggest other health management strategies is limited by the lack of information related to F. columnare biology. The development of molecular detection/quantification methods for fish bacterial pathogens is a basic step needed to study the infection rate and transmission of the pathogens. Early detection of fish bacterial pathogens is critical to treat the disease at preclinical stage, and effective treatment methods have to be based on the knowledge of pathogen?s virulence mechanisms. In this dissertation, PCR-based detection methods were not successfully developed. However, an intervening sequence (IVS) was identified from E. ictaluri 23S rDNA. It was the first description in Edwardsiella genus and may 18 serve as a good maker for detection of E. ictaluri. Regarding F. columnare characterization, phenotypic and genotypic differences were identified within the species. Phenotypic differences were illustrated by characterizing the lipopolysaccharide (LPS) and protein profiles in both virulent and avirulent F. columnare strains. Four genes (gtf, hemH, norB, trx) were, for the first time, identified in F. columnare, and their sequences divided F. columnare strains into two populations. The dissertation has been arranged according to published manuscripts followed by a section of overall conclusions. 19 OBJECTIVES The main objectives of the current study are to develop a multiplex real-time PCR for simultaneous detection/quantification of E. ictaluri, E. tarda, and F. columnare and to identify virulence factors in F. columnare. Also, the study aims to evaluate and adapt a real-time PCR developed for E. ictaluri in our laboratory. Because molecular sequences are limited for E. ictaluri and F. columnare, ribosomal genes will be the first targets considered in order to develop a multiplex real-time PCR method. If the available molecular information does not provide enough differentiation, it will be necessary to identify new signature sequences in the three bacterial pathogens. Currently, there is lack of molecular pathogenesis data for F. columnare. The present study will attempt to identify and characterize virulence factors from both phenotypic and genotypic aspects. Lipopolysaccharide (LPS) should be first considered because of its important virulence role in a variety of Gram-negative bacterial species. Fortunately, a shotgun genomic library of F. columnare was generated recently, from which some useful information may be extracted. The hypotheses and objectives in the current study are listed as follows: Hypothesis 1: Ribosomal RNA genes contain specific signature sequences for E. ictaluri, E. tarda, and F. columnare that can be used in PCR based diagnosis Objective 1: Identify specific DNA sequences within 16S rDNA, 23S rDNA, and 16S-23S ISR and develop a real-time PCR and/or a multiplex real-time PCR (Chapter II and Chapter III) Hypothesis 2: Differences in virulence among F. columnare strains can be correlated with different phenotypic and genotypic markers 20 Objective 1: Compare protein and lipopolysaccharide profiles among F. columnare strains that differ in virulence (Chapter IV) Objective 2: Identify putative virulence genes from different F. columnare strains and to characterize the expression of these genes under different conditions (Chapter V) 21 II. EVALUATION OF REAL-TIME PCR FOR DETECTION OF EDWARDSIELLA SPP. AND FLAVOBACTERIUM COLUMNARE Abstract The objective of this study was to develop a multiplex real-time PCR for the simultaneous detection and quantification of Edwardsiella ictaluri, E. tarda and Flavobacterium columnare. Current PCR detection methods for F. columnare and E. ictaluri were tested in our laboratory for comparison purposes. We found that under our conditions, the real-time PCR protocol for E. ictaluri described by Bilodeau et al. (2003) did not provide us with the expected sensitivity. In fact, the sensitivity of this protocol was lower than classical culture methods. In addition, we failed to adapt the standard F. columnare PCR protocol developed in our laboratory into a real-time PCR method. Based on these preliminary data, additional signature sequences were needed to develop new or improved real-time PCR protocols for the detection of these pathogens. The 16S and the 23S rRNA gene sequences along with the 16S-23S intergenic spacer region (ISR) were used to design new primer sets. The ISR did not provide enough variability to differentiate E. ictaluri from E. tarda. A real-time PCR method was developed for F. columnare by using primers against the 16S rDNA; however, the specificity of these primers was not sufficient to discriminate F. columnare from F. aquatile. A multiplex real-time PCR protocol was developed to simultaneously detect E. ictaluri, E. tarda, and 22 F. columnare by targeting the 23S rDNA gene. Unfortunately, cross-reactivity with non- target sequences will not allow the use of these primers in an effective PCR protocol. Keywords: Flavobacterium columnare, Edwardsiella ictaluri, E. tarda, PCR, real-time PCR 23 Introduction Edwardsiella ictaluri, E. tarda, and F. columnare are main bacterial pathogens in aquaculture. In fact, F. columnare and E. ictaluri negatively impact the channel catfish industry in southeastern USA. Detection and identification of these pathogens are critical for disease management (Plumb 1999). Among all the detection and identification methods available to date, polymerase chain reaction (PCR)-based methods have the highest potential of being fast, specific and sensitive. Currently, there is only one PCR-based protocol for E. ictaluri detection. This real-time PCR protocol was developed for E. ictaluri quantification in catfish blood samples by Bilodeau et al. (2003). A transposon element was targeted in this protocol as signature sequence, although the transposon type was unknown. As a mobile genetic entity, such transposon may not be present in all E. ictaluri isolates, raising the question about the reliability of the protocol. Several PCR protocols have been developed for E. tarda. Aoki and Hirono (1995) targeted a species-specific fragment DNA of E. tarda with unknown function. However, the encoding information of this fragment was not provided. Labelled as a probe, this fragment hybridized only with E. tarda but not with other related fish pathogens. Primers based on such specific fragment provided desirable specificity in PCR; however the analytical sensitivity test was not well described. Additionally, concentrations of PCR reaction components were not given, making it hard to repeat and confirm their studies. Targeting a hemolysin gene of E. tarda, the PCR protocol developed by Chen and Lai (1998) amplified the expected amplicon from 40 E. tarda strains. Unfortunately, the specificity of the primers with other species was not tested. Baird et al. (2003) described a new PCR protocol using the small ribosomal 24 subunit gene as target to identify Edwardsiella spp. No sensitivity assay was performed in that study, though. Because of the above reasons, more research is needed to validate the already developed methods for Edwardsiella spp. detection. Specificity, and analytical and diagnostic sensitivity need to be established and validated before a PCR- based method can be applied for molecular diagnosis (Hiney and Smith 1998). Several PCR detection protocols targetting the 16S rRNA gene have been described for F. columnare detection (Toyama et al. 1996; Bader and Shotts 1998; Bader et al. 2003b; Darwish et al. 2004). The developed protocols either lacked sensitivity tests or needed two-round nested PCR. In addition, a misidentified F. johnsoniae isolate was thought to be F. columnare and was used in designing F. columnare specific primers, making these protocols questionable (Toyama et al. 1996; Bader and Shotts 1998; Bader et al. 2003b). Welker et al. (2005) developed a new protocol to detect F. columnare based on the intergenic spacer region (ISR) present between the 16S and the 23S rRNA genes. This new protocol exhibited desirable specificity and sensitivity, being able to detect as few as 100 F. columnare cells per sample. Unfortunately, the large size of the amplicon (~500 bp) in this protocol prevented the adaptation of this standard PCR to real- time PCR. Ideally, the size of the amplicon should be between 80-150 bp when using a real-time PCR protocol. PCR-based detection methods do not require pathogen isolation and recovery from the environment, shortening the time of pathogen detection. Compared to standard PCR, real-time PCR assays are faster and allow target quantification in a single assay. Typically, bacterial pathogens are present in natural environments in low numbers. Therefore, sensitive methods are needed to detect them in their aquatic reservoirs. 25 Moreover, early detection of fish pathogens in the preclinical stage can be critical in preventing outbreaks. Monitoring the changes of F. columnare load on fish body surface, internal organs, or the environment may provide important information about this pathogen?s infection rate and transmission. Welker et al. (2005) reported that F. columnare can be transmitted through water without direct contact with diseased fish. The authors also found a good survival of F. columnare in the biofilm from tank walls four days after inoculating the water with F. columnare cells. It will be interesting to determine whether F. columnare in biofilm may infect fish and what is the mimimum bacterial load to cause disease. In addition, F. columnare bacterial cells have the characteristic of adhering to each other, making it difficult to accurately enumerate by plate counting method. Real-time PCR, at this point, may provide better sensitivity to quantify changes of bacterial cell numbers. The objectives of this study were: (1) to develop a real-time PCR for F. columnare, and (2) to develop a multiplex real-time PCR protocol for simultaneous detection of E. ictaluri, E. tarda, and F. columnare. Materials and Methods Bacterial isolates, culture conditions and DNA extraction The type strains of the species E. ictaluri CECT 885, E. tarda CECT 849, and F. columnare ATCC 23463 as well as other reference strains and clinical and environmental isolates were used in this study (Table II.1). F. columnare was cultured in modified Shieh broth (Shoemaker et al. 2005b) at 28 ?C, while Edwardsiella strains were cultured in brain heart infusion broth (BHI) (Becton, Dickinson and Company, Sparks, MD) at 28 26 ?C. Genomic DNA was extracted using the DNeasy tissue kit (Qiagen, Valencia, CA) following the manufacturer?s instructions. Real-time PCR for E. ictaluri detection We adapted the real-time PCR protocol by Bilodeau et al. (2003) for E. ictaluri in our laboratory. Probe and primer sequences were identical to those originally described by Bilodeau et al. (2003). Primers and probe were ordered from Applied Biosystems. Probe was labelled with 6-carboxyfluorescein (6-FAM) at 5? end and 6-carboxy- tetramethyl-rhodamine (TAMRA) at the 3? end. Some of the reagents and equipment used in our protocol were different from those described by Bilodeau et al. (2003). Instead of using 1 X platinum quantitative PCR superMix-UDG (Invitrogen Life Technologies, Carlsbad, CA) for the real-time PCR reaction mixture, we used Taqman universal PCR master mix (Applied Biosystems, Foster, CA). The real-time PCR thermocyler we used was ABI PRISM 7000 Sequence Detection System (Applied Biosystems) not the iCycler (Bio-Rad Laboratory) used by Bilodeau et al. (2003). Bilodeau et al. (2003) described two methods for DNA isolation: one to extract DNA from erythrocytes and a second one to extract DNA from bacterial cells. We followed their procedure for erythrocyte DNA isolation. In addition, we also tried a commercial kit specific for blood (MO BIO, Carlsbad, CA) and the general DNeasy tissue kit (Qiagen) for purifying erythrocyte DNA. We did not try their method for bacterial DNA isolation, instead, DNeasy tissue kit was used for extraction DNA from bacterial cells, catfish kidney, and brain. 27 Primers and probe design for real-time PCR of F. columnare F. columnare 16S rRNA gene sequences were downloaded from Genbank. Sequences used were (with accession numbers in brackets): F. columnare (AJ491824 AB015480 AY842900 AY747592 AY561521 AY842901 AB016515 AY488506 AY550029 AY488507 AB015481 AY842899 D12659 AY095342 AB023660 AB010952 AB010951 AB078047 AY577821 AY635167 AJ831829 AJ831828 AJ831825 AJ831830 AJ831826 AJ831824 AJ831827), Flexibacter aurantiacus (AB078044 AB078045), Tenacibaculum maritimum (AB078057), F. psychrophilum (AF090991 AY034478 AB078060 AY577822 AY662493 AY662494), F. aquatile (M62797), F. hydatis (M58764), F. branchiophilum (D14017). Vector NTI ? Suite 8 was used to align all the sequences. Unique sequence areas were analyzed using the Applied Biosystems Primer Express software (Applied Biosystems) to design real-time PCR primers and probe. The probe labeling was the same as above (Fco-16probe: 6FAM-5?- TTTCTTCGGACGTTTTTCAAGGTGCTGC-TAMRA-3?; primer Fco-16F: 5?- GGAAACGACAGATTTGGAAACAG-3?; primer Fco-16R: 5?- GCACGAGCTGACAACAACCA-3?). The optimized PCR reaction included 1 X Taqman universal PCR master mix (Applied Biosystems), 0.2 ?M primers, and 0.25 ?M probe with final volume of 25 ?L. Quantified data were obtained by using ABI PRISM 7000 Sequence Detection System (Applied Biosystems). Amplification profile of PCR was as follows: one cycle of 2 min at 50 ?C, one cycle of 10 min at 95 ?C, and 40 cycles of 15 s at 95 ?C, 1 min at 60 ?C. 28 Specificity and sensitivity of the primers Specificity ? Real-time PCR specificity was tested with bacterial species listed in Table II.1. For real-time PCR of E. ictaluri, tested bacterial species were: F. columnare ATCC 49512, F. johnsoniae ATCC 43622, and all the listed isolates of E. ictaluri and E. tarda. For real-time PCR of F. columnare, tested bacterial species included E. ictaluri CECT 885, E. tarda CECT 849, Yersinia ruckeri ATCC 29473, and all the listed Flavobacterium spp. Bacterial DNA was used as template for PCR amplification. PCR amplification profile was as follows: one cycle of 2 min at 50 ?C, one cycle of 10 min at 95 ?C, and 40 cycles of 15 s at 95 ?C, 1 min at 60 ?C. Analytical sensitivity ? Several DNA preparation methods were used for E. ictaluri real-time PCR. Two strains of E. ictaluri were used: CECT 885 (type strain) and AL-93-75. Purified DNA: Purified genomic DNA and plasmid containing the cloned target transposon from E. ictaluri were 10-fold diluted, respectively. DNA amount ranged from 100 ng to 100 fg. Sensitivity test was performed in triplicate. Bacterial cells: Sensitivity was also tested by spiking boiled E. ictaluri cells into the PCR reaction. E. ictaluri CECT 885 and AL-93-75 were cultured in BHI broth overnight. Ten-fold dilution of the cultures was performed and plate counts were run in parallel. One hundred microliters of each dilution was boiled at 100 ?C for 5 min followed by 2 min centrifugation at 200 g. Five microliters supernatant was directly used as template for the real-time PCR. The sensitivity test was performed in triplicate. DNA from spiked blood: Further sensitivity tests were conducted by spiking known numbers of bacterial cells into catfish blood samples followed by DNA extraction. 29 Bacterial culture and dilution procedure was the same as above mentioned. Ten microliters of freshly obtained catfish blood was mixed with 25 ?L of each dilution followed by DNA extraction, and 50 ng of DNA was subjected to real-time PCR. Diagnostic sensitivity ? Channel catfish were challenged with E. ictaluri to establish the diagnostic sensitivity of our modified PCR approach. The challenge experiment was conducted by Craig A. Shoemaker (ARS-USDA, Auburn, AL) following published protocols (Klesius and Shoemaker 1997). Approximate ten-gram channel catfish were used in the study. The experiment setup included 9 unchallenged control fish, and 3 replicates of challenged fish in 3 separate tanks (9 fish per tank). Fish were immersed in 5 X 10 5 CFU/mL of E. ictaluri AL-93-75 strain in 57 L tank for 30 min. Water temperature was kept at 25 ?C ? 2 ?C for the duration of the experiment. Blood, brain, and kidney were sampled at 14d post-challenge for later analysis. DNA was extracted from kidney, brain, and 10 ?L of blood. Template DNA for real-time PCR included 300 ng of blood DNA, 50 ng of diluted DNA from blood, 50 ng of kidney DNA, and 50 ng of brain DNA. A standard curve was established with DNA from E. ictaluri CECT 885. The amount of DNA ranged from 100 ng to 100 fg by serial 10-fold dilution. As part of a collaborative effort, the above experiment was used to compare the diagnostic sensitivity between the real-time PCR protocol and a new method developed by Panangala et al. (2006) using an indirect fluorescent antibody assay (IFA). Samples for IFA were obtained from experimentally infected fish. Sample smears were made on glass slides and were air dried and heat fixed. Primary antibody was incubated on the slides for 45 min followed by rinsing with PBS and air dried. Fluorochrome-conjugated secondary antibodies were incubated on the slides for 45 min, rinsed, and air dried. Slides 30 were examined with a Nikon Eclipse 800 M epifluorescence microscope. In addition, samples from infected fish were plated onto E. ictaluri selective medium (EIM). Putative colonies were confirmed by fatty acid methyl ester analysis (FAME) (Shoemaker et al. 2005a). 23S rRNA gene and ISR amplification, cloning, and sequencing Amplification ?Internal spacer region within the ribosomal operon of E. tarda and E. ictaluri were amplified using two pairs of primers G1/L1 and 16S-14F/23S-1R (Arias et al. 1995) (G1:5?-GAAGTCGTAACAAGG-3?, L1:5?-CAAGGCATCCACCGT- 3?; 16S-14F: 5?-CTTGTACACACCGCCCGTC-3?, 23S-1R: 5?- GGGTTTCCCCATTCGGAAATC-3?). The ISR sequences of F. columnare had been already sequenced in our laboratory (GenBank assession numbers from AY754360 to AY754388). Welker et al. (2005) developed a standard PCR based on these ISR sequences; however, adaptation of the primers to real-time PCR failed due to the large size of the amplicon (~ 500 bp) (Thomas L. Welker, ARS-USDA, Auburn, Al, personal communication). The partial 23S rRNA gene of E. ictaluri, E. tarda, and F. columnare were amplified using universal primers 118V and 1037R (118V:5?- CCGAATGGGGAAACCCA-3?, 1037R: 5?-CGACAAGGAATTTCGCTAC-3?) (Arias et al. 1995). For the 3? end amplification of 23S rRNA gene, another three primers were used (23-3F: 5?-GGCGGCCGTAACTATAACG-3?, 23-2R: 5?- AGCCTCACGGTTCATTAGTACC-3?, 23-1R: 5?-GACCGAACTGTCTCACGACG-3?). 23-3F was used as the forward primer for E. ictaluri, E. tarda, and F. columnare. 23-2R was the reverse primer for E. ictaluri and E. tarda. 23-1R was the reverse primer for F. 31 columnare. Bacterial sequences for designing the three primers were obtained from the Comparative RNA Web Site (URL:http://www.rna.icmb.utexas.edu/). Sequences used for comparison were (GenBank accession number in bracket): Escherichia coli (AF053965 AJ278710), Salmonella enterica (U77919), S. bongori (U77927), Citrobacter freundii (U77928), Klebsiella pneumonia (X87284), Yersinia pestis (NC-004088), Y. enterocolitica (U77925), Erwinia carotovara (BX950851), Aeromonas hydrophila (X67946 X87281), Plesiomonas shigeloides (X65487), F. odoratum (M62807), Flexibacter flexilis (M62806), Chlorobium limicola (M62805), S. typhimurium LT2 (AE008895). The vector NTI ? software package (Invitrogen, Carlsbad, CA) was used to design primers. All reagents unless otherwise stated were purchased from Promega (Madison, WI). Amplification reactions were carried out in a final volume of 50 ?L containing 2.5 ?M MgCl 2 , 1 X Taq buffer, 0.2 ?M of both primers, 0.2 ?M of dNTPs, 1.7 unit of Taq polymerase, and 60 ng of template DNA. The amplification profile was as follows: hot- start for 10 min at 95?C, 30 cycles of 30 s at 94?C, 45 s at 55?C, and 1.5 min at 72?C. PCR products were electrophoresed on 1% agarose gel for 30 min at 100 V. The gel was visualized under ultraviolet light, and the proper bands were cut off and purified using the Geneclean kit (Q-BIO gene company, Irvine, CA). Purified PCR products were ligated into pGEM-T easy vector following manufacturer?s instruction. Cloning and sequencing ? Fifty microliters of competent Escherichia coli JM109 cells were used for transformation. Cell transformation was carried out as follows: cells were incubated for 20 min on ice, then hot shocked for 45 s at 42 ?C, and finally placed on ice for 5 min. The transformed Escherichia coli cells were plated on 32 Luria-Bertani (LB) agar containing ampicillin and IPTG/X-Gal (Invivogen, San Diego, CA). The plates were incubated at 37 ?C for 18 h. Transformants were cultured in LB broth at 37 ?C for 16-18 h. The aurum plasmid mini kit (Bio-Rad, Nercules, CA) was used to extract plasmids from clones. In order to check the size of the inserted amplicon, plasmids were restriction digested with EcoRI and resolved through a 1% agarose gel at 100 V for 30 min. Finally, the plasmid DNA was submitted to Auburn University Genetic Analysis Laboratory (AU-GAL) (http://www.auburn.edu/research/vpr/aurif/sequencing.htm) for sequencing. This laboratory uses ABI 3100 Genetic Analyser (Applied Biosystems) for DNA sequencing. Multiplex PCR Primers were designed to detect E. ictaluri, E. tarda, and F. columnare simultaneously. Primers flanking the intervening sequences (IVS) within the 23S rDNA of E. ictaluri were designed by aligning 23S rDNA sequences of related species. Comparative RNA Web Site (URL:http://www.rna.icmb.utexas.edu/) was used to obtain 23S rDNA data. These sequences were the same as those used for primer design against the 3? end of 23S rDNA above mentioned. The sequence of the universal forward primer for the three species was 5?-GACAGCYAGGATGYTGGCTT-3?. We based the specificity of the protocol on the reverse primer sequence. Therefore, two reverse primers were needed in the reaction. For E. ictaluri and E. tarda amplification, we used EicEta- 23S-R: 5?-CAGCAGCCCTCACAGGC-3? while the reverse primer for amplifying F. columnare was Fco-23S-R: 5?-CCAGAAATCCTCACGGAATC-3?. Due to the high nucleotide similarity between E. ictaluri and E. tarda, we targeted the intervening 33 sequence (IVS) since E. tarda does not present this sequence. Therefore, the species of Edwardsiella can be differentiated based on amplicon size. Combinations of the universal forward primer with each specific reverse primer were tested in the following bacterial species: E. ictaluri CECT 885, E. tarda CECT 849, F. columnare ATCC 49512, F. psychrophilum ATCC 49418, F. johnsoniae ATCC 43622, Yersinia ruckeri ATCC 29473, S. typhimurium (clinical isolate), F. hydatis ATCC 29551, and Escherichia coli K-12. Results The target areas for real-time PCR developed in the current study included the ISR, 16S rDNA, and 23S rDNA. The ISR of E. ictaluri shared high similarity with E. tarda; therefore, ISR was not suitable for real-time PCR development. E. ictaluri real-time PCR specificity The real-time PCR protocol for E. ictaluri in our laboratory modified from Bilodeau et al. (2003) amplified all 21 isolates of E. ictaluri but did not amplify any of the E. tarda, F. columnare, and F. johnsoniae strains tested. E. ictaluri real-time PCR analytical sensitivity By spiking known amount of DNA into real-time PCR reaction, the modified real-time PCR for E. ictaluri was able to detect 100 fg by using plasmid DNA containing the transposon, 100 pg of genomic bacterial DNA, and 4 X 10 6 CFU from boiled bacterial cells. The real-time PCR failed to detect any sample containing a mixture of E. ictaluri DNA and catfish erythrocyte DNA regardless of the extraction methods used. 34 E. ictaluri real-time PCR diagnostic sensitivity and comparison with cultural and immunological methods Table II.2 displayed the comparison of the real-time PCR results with the results from microbial culture and immunological methods. Negative control (i.e. non- challenged fish) samples were negative for E. ictaluri by all methods (data not shown). Only 23 challenged channel catfish were tested because 4 fish died before sample collection, and therefore could not be tested by any method. All positive samples by real- time PCR were also positive by culture methods (EIM). By contrast, real-time PCR did not amplify some positive samples identified by IFA and by selective medium (EIM). Real-time PCR failed to detect E. ictaluri in 23 blood samples, and only identified E. ictaluri from 5 out of 23 kidney samples and 3 out of 23 brain samples. Results obtained by IFA did not match the results obtained from culture method. Some samples that were positive by the culture method were negative by IFA method and vice versa. Specificity of real-time PCR for F. columnare detection All 32 F. columnare isolates tested yielded the expected positive amplification. No amplification was observed from F. psychrophilum, F. hydatis, F. johnsoniae, E. ictaluri, and E. tarda. However, the specific primers designed did amplify F. aquatile DNA under optimized real-time PCR conditions. To ensure F. aquatile was not contaminated with F. columnare DNA, F. aquatile cultures recovered from three different glycerol stock vials were identified by fatty acid methyl ester analysis (FAME) and confirmed as F. aquatile. No amplification was observed from F. aquatile DNA when the F. columnare specific primers by Welker et al. (2005) were used. Since 35 cross-reactivity could not be eliminated, no further tests, including sensitivity tests, were performed with these primers. Multiplex PCR The specificity of the multiplex PCR was first tested using a standard PCR protocol. The combination of forward primer with the Edwardsiella-specific reverse primer amplified E. ictaluri, E. tarda as well as F. columnare, Y. ruckeri, S. typhimurium and Escherichia coli DNAs. The combination of forward primer with F. columnare specific reverse primer amplified F. columnare as well as F. johnsoniae, Y. ruckeri, and S. typhimurium. Even after the annealing temperature was increased from 51?C to 58 ?C, similar results were obtained (Figure II.1). Therefore, this real-time multiplex PCR was not subjected to further trials. Discussion The real-time PCR protocol described by Bilodeau et al. (2003) failed to yield the expected results in our laboratory. Even though the primers and probe sequence were the same, the real-time PCR equipment we had available was different. Bilodeau et al. (2003) used iCycler by Bio-Rad, while we used the ABI PRISM 7000 Sequence Detection System. To my knowledge, both systems should be compatible. Bilodeau et al. (2003) reported a very high sensitivity level, since they were able to detect E. ictaluri from all the blood samples spiked with bacterial cells. However, we could not detect E. ictaluri from any of the challenged channel catfish blood samples tested, although we tried different methods for DNA extraction besides the one described by Bilodeau et al. (2003), including a commercial kit specifically designed for blood. Only less than 20% of kidney and brain samples were positive by the modified real-time PCR method. Overall, the 36 sensitivity of the E. ictaluri-specific protocol under our conditions was much lower than expected. On the other hand, when we used the same E. ictaluri specific primers in a standard PCR reaction, the sensitivity and specificity obtained were within the expected range (Welker et al. 2005). The combination of primers and probes designed to specifically amplify F. columnare also amplified F. aquatile DNA. This result was unexpected since the primers and probes used presented 11 mismatches with the F. aquatile 23S rDNA sequence. Several reasons may justify this result; e.g., the specific primers amplified was outside the target sequences. However, no conclusion could be inferred unless cloning and sequencing the non-specific band is attempted. Regarding the multiplex PCR protocol for F. columnare, E. ictaluri, and E. tarda, the designed primers could not differentiate target species from other non-target species. Concurrently with our research, Panangala et al. (2006) developed an indirect fluorescent antibody (IFA) test to simultaneously detect E. ictaluri and F. columnare. Moreover, the same authors were working on developing a multiplex PCR protocol for F. columnare, E. ictaluri and A. hydrophila (Panangala et al. 2007). At that point, based on our unsuccessful data and considering the ongoing research by Dr. Panangala?s group, we decided to end this project and redirect our research efforts. 37 Table II.1. Isolates used in the PCR detection study. Species Isolates Origin ALG-00-513, ALG-063, ALG-00-527 Greensboro, AL, USA MS 467, MS 463, MS 465, MS 475 Stoneville, MS, USA MO-02-23 Lake of the Ozarks, MO, USA BZ-05 Brazil ATCC 23463 Diseased chinook salmon, Oncorynchus tshawytscha, Snake River, WA, USA F. columnare ATCC 49512 Skin lesion of a brown trout fry, Salmo trutta, Basse-Normandie, France Fco152, Fco96, Fco107, Fco109, Fco145, Fco147, Fco103, Fco63 Channel catfish, Mobile River, AL, USA Fco150 Threadfin shad, Mobile River, AL, USA Fco115, Fco127, Fco113, Fco136, Fco106, Fco131, Fco142, Fco138 Blue catfish, Mobile River, AL, USA Fco97, Fco102, Fco143, Fco94 Freshwater drum, Mobile River, AL, USA F. psychrophilum ATCC 49418 Coho salmon kidney, Oncorhynchus kisutch, WA, USA F. johnsoniae ATCC17061 Soil, England ATCC 43622 Salmonid fish? F. aquatile ATCC 11947 Deep well in Kent, England F. hydatis ATCC 29551 Gills of diseased salmon, MI, USA Y. ruckeri ATCC 29473 Rainbow trout, Onchorhynchus mykiss, with redmouth disease, USA E. ictaluri 151,218, 219,195,196, RE33, AL-02-27, AL-93-75, AL- 93-58, AL-02-27, ALG-03- 275, ALG-03-278, ALG-03- 101, ALG-03-190 Catfish, AL, USA CECT 885 Catfish, GA, USA (cont.) 38 EILO Catfish, Thailand S94-1051, S94-1034, 1696,1963, 1760, Catfish, MS, USA E. tarda CECT 849 Human feces ATCC 15947 Human feces S. typhimurium Clinical isolate Escherichia coli K-12 Human feces CECT, Spanish Type Culture Collection ATCC, American Type Culture Collection 39 Table II.2. Comparison of real-time PCR, microbial culture and IFA for detection of E. ictaluri. Tank 1-3 contained 9 challenged fish each. Only 5 fish in tank 3 were alive at the 14 d post-challenge. Dead fish were not tested. Non-challenged fish were alive and negative by all detection methods (data not shown). Blood Kidney Brain Tank Fish Culture IFA Real- time Culture IFA Real- time Culture IFA Real- time 1 1 - - - - - - - - - 1 2 - - - + - - - - - 1 3 - - - - - - - - - 1 4 - - - + - - - + - 1 5 + + - + - + + - - 1 6 + - - + - + + - + 1 7 - - - - - - - - - 1 8 + - - + - - + - + 1 9 - - - - + - - - - 2 1 + - - + - + + + - 2 2 + - - - - - - - - 2 3 + - - + ? - + + - 2 4 - + - + ? - + ? + 2 5 - + - - + - - + - 2 6 + + - + - - + + - 2 7 - + - + + - - + - 2 8 - + - - + - - + - 2 9 + - - + + + - + - 3 1 - - - - - - + - - 3 2 - - - + + - - + - 3 3 - - - - + - - + - 3 4 - - - + - - - + - 3 5 + - - + + + + + - T 9 6 0 14 8 5 9 12 3 T: total number of positives IFA: indirect fluorescence antibody test ?: inconclusive result 40 Figure II.1. Specificity test of the multiplex PCR. M, 50 bp DNA ladder; Lane 1, E. ictaluri; lane 2, E. tarda; lane 3: F. columnare; lane 4, F. psychrophilum; lane 5, F. johnsoniae,; lane 6, Y. ruckeri; lane 7, S. typhimurium; lane 8, F. hydatis; lane 9, Escherichia coli. Primers used for top gel (A) were universal forward primer with EicEta- 23S-R. Primers used for bottom gel (B) were universal forward primer with Fco-23S-R. M 1 2 3 4 5 6 7 8 9 negative Primer- dimer A B 41 III. IDENTIFICATION AND CHARACTERIZATION OF AN INTERVENING SEQUENCE WITHIN THE 23S RIBOSOMAL RNA GENE OF EDWARDSIELLA ICTALURI Abstract Comparison of the 23S rRNA gene sequences of Edwardsiella tarda and E. ictaluri confirmed a close phylogenetic relationship between these two fish pathogen species and a distant relation with the ?core? members of the Enterobacteriaceae family. Analysis of the rrl gene for the 23S rRNA in E. ictaluri revealed the presence of an intervening sequence (IVS) in helix-45. This new 98 bp IVS shared 97% nucleotide identity with Salmonella typhimurium helix-45 IVS. E. ictaluri helix-45 IVS was present in all E. ictaluri strains analyzed and in at least six rrl operons within each cell. Fragmentation of 23S rRNA due to IVS excision by RNase III was observed by methylene blue staining of ribosomal RNA extracted from E. ictaluri isolates. This is the first report of an IVS in the 23S rRNA gene of the genus Edwardsiella. Keywords: Edwardsiella, intervening sequences (IVS), rrl genes 42 Introduction The genus Edwardsiella was first described in 1965 by Ewing et al. to harbor a collection of 37 strains biochemically distinct from other taxa within the family Enterobacteriaceae. The proposed type species for the genus was Edwardsiella tarda and at that time comprised isolates of fecal origin from the USA. At about the same time, a Japanese group studying bacterial isolates from reptiles identified a distinct yet homogeneous group, referred to as the Asakusa group, within the Enterobacteriaceae family (Sakazaki and Murata 1962) which appeared to be very similar to the newly described E. tarda. It was not until 1980 that a second species, E. hoshinae, was added to the genus by Grimont et al. (1980) who isolated it from reptiles and birds. Finally, Hawke et al. (1981) added the last species to the genus, E. ictaluri, described from isolates recovered from diseased channel catfish (Ictalurus punctatus). The phylogenetic position of the genus Edwardsiella within the family Enterobacteriaceae is somewhat unclear since few evolutionary studies including this genus have been attempted (Ahmad et al. 1990). Nevertheless, DNA-DNA hybridization data pointed to Serratia as the closest genus to Edwardsiella at 20% similarity (Krieg and Holt 1984). Other studies based on different signature sequences, such as tRNA Leu gene and 16S rRNA gene, confirmed a distant phylogenetic relationship between Edwardsiella and core members of the Enterobacteriaceae such as Escherichia coli or Salmonella (Francino et al. 2003). Phylogenetic comparisons between the three Edwardsiella species are also scarce. Panangala et al. (2005) studied the interspecies and intraspecies differences between E. tarda and E. ictaluri based on the 16S-23S rRNA intergenic spacer region and concluded both species were highly similar (>96% identity). To our 43 knowledge, no formal ribosomal genes-based trees including E. tarda and E. ictaluri sequences have been published. At the intraspecific level, it is widely accepted that E. tarda presents a higher degree of intrinsic diversity than E. ictaluri (Klesius et al. 2003; Panangala et al. 2005). Only by using the high resolution fingerprinting technique AFLP (Amplified Fragment Length Polymorphism) has it been possible to discriminate among E. ictaluri isolates (Klesius et al. 2003). E. ictaluri is the causal agent of enteric septicemia of catfish (ESC) and is responsible for great economic losses to the catfish industry across the world (Plumb 1999; Wagner et al. 2002). Only in the USA, where channel catfish production accounts for more than half of total aquaculture products, ESC costs the catfish industry millions of dollars yearly in direct fish losses (Plumb and Shoemaker 2006). Although a few non- ictalurid species (e.g. tilapia Oreochromis niloticus, Chinook salmon Oncorhynchus tshawytscha, and rainbow trout Oncorhynchus mykiss) are susceptible to experimental infection, E. ictaluri natural host range is mainly restricted to members of the Ictaluridae family, particularly channel catfish. On the contrary, E. tarda has shown a much broader host spectrum with the ability to infect fish from different families as well as reptiles and even humans, though is a minor source of disease in aquaculture. The pathogenic factors that made E. ictaluri a specialized catfish pathogen while E. tarda retained its broader spectrum are not known, nor are the phylogenetic relationships between these two pathogens. The aim of this work was to expand our knowledge on the phylogenetic relationships between E. tarda and E. ictaluri by comparing their 23S rRNA gene sequence and to find new signature sequences for ESC detection. As a result of this study, 44 a new intervening sequence (IVS) in the 23S rRNA of E. ictaluri was identified for the first time. Materials and Methods Bacterial strains and cultivation conditions Twenty nine E. ictaluri strains, including the type strain CECT 885, along with five E. tarda strains and Escherichia coli K-12 strain were used in this study (Table III.1). All strains were grown on brain heart infusion (BHI) (Becton, Dickinson and Company, Sparks, MD) broth or BHI agar (supplemented with 1.5% agar) at 28?C, except for Escherichia coli which was incubated at 37?C. Stock cultures were maintained in 15% glycerol at -70?C, and a single colony was isolated prior to use. 23S rRNA gene amplification, cloning and sequencing Total bacterial DNA was extracted using the Qiagen DNeasy Tissue kit (Qiagen, Valencia, CA) following manufacturer?s instructions. Two sets of universal primers 118V 5?-CCGAATGGGGAAACCCA-3? (positions 112 to 130 in Escherichia coli), 1037R 5?- CGACAAGGAATTTCGCTAC-3?(positions 1930 to 1948), 23_2F (5?- GGCGGCCGTAACTATAACG-3?) (positions 1901 to 1919), and 23_2R (5?- AGCCTCACGGTTCATTAGTACC-3?) (positions 2874-2895), complementary to highly conserved regions of eubacterial 23S rRNA genes were used to amplify the 23S rRNA gene from E. ictaluri CECT 885 and EILO and E. tarda CECT 849. PCR conditions were performed according to Arias et al. (1995). The expected size amplicons (both primer pair combinations, 118V-1037R and 23_2F-23_2R, yielded similar products ~1,8 kb) were resolved through standard agarose gel electrophoresis. Amplified products were purified using the High Pure PCR Product Purification Kit (Roche Diagnostic 45 Corporation, Indianapolis, IN) and cloned into pGEMTEasy (Promega, Madison, WI). To ensure sequence accuracy, two clones from each strain were sequenced at the Auburn University Sequencing Core (Auburn, AL). Internal primers were designed to obtain the full length sequence for each clone. Sequences were trimmed and their similarity to 23S rRNA genes confirmed by BLAST (Basic Local Alignment Search Tool) search . A previously unknown intervening sequence (IVS) in the 23S rDNA of E. ictaluri was identified by comparison with other Enterobacteriaceae 23S rRNA genes. Intervening sequences PCR amplification Primers flanking the newly identified IVS were designed, and all strains used in the study were tested for IVS presence. The IVS primers used were: Eic-IVF-5?- CTAAACCATGCACCGAAGC-3? and Eic-IVR-5?-TTACAGAACGCTCCCCTACC-3?. Fifty nanograms of total DNA was used as template in a PCR reaction containing 0.5 ?M primer Eic-IVF, 0.5 ?M primer Eic-IVR, 10 ?M deoxynucleotides, 1.5 mM MgCl 2 , one unit of Taq DNA polymerase and 1x buffer (Eppendorf, Hamburg, Germany). PCR conditions were as follows: initial denaturing at 94?C for 5 min; 35 cycles of denaturing at 94?C for 30 sec, annealing at 50?C for 30 sec and extension at 72?C for 45 sec, with a final extension step of 10 min at 72?C. Amplified products were electrophoresed, purified, cloned, and sequenced as described above. Ribotyping Analysis of the ribosomal operons present in E. ictaluri was performed accordingly to Arias et al. (1997a). Briefly, 5 ?g of total DNA were digested with BamHI and KpnI, transferred onto a nylon membrane, and hybridized with a digoxigenin (DIG) labeled probe. Samples were run in duplicate and hybridized with two ribosomal probes. 46 1037-DIG is a universal probe complementary to a conserved sequence in the 23S rRNA gene of eubacteria . The entire IVS sequence discovered in E. ictaluri was PCR amplified using Eic-IVF and Eic-IVR, purified, DIG-labeled with the PCR DIG Probe Synthesis Kit (Roche), and used as probe (IVS-DIG) for ribotyping. While the 1037-DIG-labelled probe highlighted the number of ribosomal operons present in both E. ictaluri and E. tarda, hybridization with IVS-DIG indicated presence of IVS in a given operon. RNA isolation, Northern blotting, and methylene blue staining Cells were grown in BHI broth to mid-log phase and RNA extraction was done as described by Mattatall and Sanderson (1996). Five micrograms of total RNA were electrophoresed under the same denaturing conditions, transferred to a nylon membrane and stained with methylene blue. Data analysis The GenBank nr/nt database was searched for similar sequences using the BLASTn algorithm service . Fifty 23S rDNA complete (or near complete) sequences from other Gammaproteobacteria were downloaded from GenBank and from the Comparative RNA Web (http://www.rna.icmb.utexas.edu) for comparison with E. ictaluri and E. tarda 23S rRNA gene sequences. An initial phylogenetic analysis was performed to verify the phylogenetic position of the genus Edwardsiella within the Gammaproteobacteria class. Sequences used were (with accession numbers in brackets): Acinetobacter calcoaceticus (X87280), Aeromonas hydrophila (X67946), A. sobria (AF508060), Arsenophonus spp. (AY587142), Buchnera aphidicola (NC_004545), Citrobacter freundii (U77928), E. ictaluri (DQ211094, DQ314205), E. tarda (DQ211093), Enterobacter aerogenes (AY116918), Erwinia carotovora (BX950851), 47 Escherichia coli (AJ278710, J01695, AB035926, AB035922, U00096), Francisella tularensis (AY146950), Haemophilus influenza (U32742), Halomonas halmophila (AJ306879), H. venusta (AJ306887), Hamiltonella defense (AY296733), Klebsiella pneumoniae (X87284), Leucothrix mucor (X87285), Morganella morganii (AY116909), Photobacterium damselae (Y18520), Photorhabdus luminescens (NC_005126), Plesiomonas shigeloides (X65487), Proteus mirabilis (AY116929), P. vulgaris (AY116927), Pseudomonas aeruginosa (Y00432), P. perfectomarina (L03788), P. stutzeri (X87289), Regiella insecticolla (AY296734), Salmonella bongori (U77927), S. enterica (U77919, U77922, U77926, CP000026), S. typhimurium (AE008895, AE008893), Serratia symbiotica (AY296732), Shewanella oneidensis (AE015456), Shigella boydii (CP000036), S. flexnerii (NC_004741), Sodalis glossinidius (NC_007712), unclassified aphid secondary endosymbiont (AF263561), Vibrio cholerae (AE004157), V. splendidus (AJ294425), V. vulnificus (X87293), Yersinia enterocolitica (U77925), Y. pestis (AE013984, AE017127), and Y. pseudotuberculosis (NC_006155). Results of distance analysis in BioNumerics (see below) indicated that Edwardsiella spp. formed a strongly supported clade (100% bootstrap value) within the family Enterobacteriaceae. All Enterobacteriaceae sequences clustered together presenting a robust clade, consistent with previous phylogenetic studies. Further analysis, aimed at elucidating relationships of E. ictaluri and E. tarda to other lineages of Enterobacteriaceae, were conducted using 34 sequences representing 25 taxa. In this analysis we excluded species outside the Enterobacteriaceae family except for Vibrio choleare, which was designated as the outgroup. Sequences were aligned using BioNumerics v. 4.0 (Applied Maths, Kortrij, Belgium). Unknown bases were discarded 48 for the analysis and the multiple alignment was corrected manually. Similarity matrices, distances (using the Jukes and Cantor correction (Jukes and Cantor 1969)), neighbor- joining (NJ) (Saitou and Nei 1987), and maximum parsimony trees were generated. Bootstrapping analysis was undertaken to test the statistical reliability of the topology of the trees using 500 bootstrap resamplings of the data. Prediction of E. ictaluri 23S rRNA secondary structure was completed by free-energy minimization using online software at http://www.bioinfo.rpi.edu/applications/mfold/old/rna . Results 23S rRNA gene sequence and analysis Essentially complete 23S rRNA gene sequences (>2,700 bp) were determined for two Edwardsiella species (two E. ictaluri isolates and one E. tarda). A parsimony tree based on 23S rRNA gene sequences analysis is shown in Figure III.1. Species names and GenBank accession codes are included in the tree. The 23S rRNA gene consensus tree revealed a distant phylogenetic relationship of the genus Edwardsiella from the ?core? members (Salmonella and Escherichia) of the family Enterobacteriaceae. The two E. ictaluri isolates had very similar (>99%) but not identical 23S rRNA gene sequences. Both Edwardsiella species shared a high degree of nucleotide identity (97%) and formed a tight cluster. The phylogenetically closest species to Edwardsiella included in the study was Erwinia carotovora (94% nucleotide identity). Identification of IVS in E. ictaluri The alignment of E. tarda and E. ictaluri 23S rRNA gene sequences revealed an insertion of 98 bp in the 23S rRNA gene of E. ictaluri. BLAST analysis of the insert showed high similarity with an IVS present in Salmonella typhimurium. The newly found 49 IVS was present in both E. ictaluri isolates but absent in E. tarda. E. ictaluri IVS was located in helix-45 of the 23S rRNA gene. Available helix-45 IVS sequences from other Enterobacteriaceae were downloaded from GenBank and used to construct a neighbor- joining tree (data not shown) that revealed a high degree of sequence diversity between Enterobacteriaceae IVSs. S. typhimurium IVS was most similar to E. ictaluri followed by Yersinia enterocolitica. Specific nucleotide identities are displayed in Table III.2. The highest percent of nucleotide identity was found between E. ictaluri and S. typhimurium at 97%, suggesting that both sequences should be grouped under the same IVS family (Family M) according to Miller et al. (2000). Yersinia bercovieri and Proteus mirabilis shared 88% nucleotide similarity and could be included under a unique IVS family as well. Sequence similarities among remaining IVSs ranged from 42% to 86% nucleotide identity. When the secondary structure of both E. ictaluri and S. typhimurium was generated (Figure III.2) only two differences were found between the two structures. E. ictaluri lacked a small loop present in S. typhimurium due to a single transversion on position 77, which resulted in a longer stem in this region. The second difference was a transition on position 35 of the IVS that resulted in no structural change. Characterization of E. ictaluri IVS In order to assess the prevalence of IVS within the species E. ictaluri, a collection of E. ictaluri isolates was screened for IVS presence. Primers flanking the helix-45 IVS region were designed and used in a PCR assay. Amplified regions containing IVS were 152 bp in length, while 54 bp bands indicated an absence of IVS in helix-45. All twenty nine E. ictaluri yielded the expected 152 bp amplified product (Figure III.3) while all E. 50 tarda strains lacked the IVS band and only exhibited the 54 bp amplicon. However, all E. ictaluri strains tested also presented a weak band at 54 bp indicating that at least one rrn operon in E. ictaluri did not contain IVS. All amplified IVSs were cloned and sequenced. No differences in nucleotide sequences between the different E. ictaluri isolates tested were observed (data not shown). Ribotyping analysis was performed in order to further characterize the rrl operons in E. ictaluri. Restriction analysis by KpnI and BamHI followed by hybridization with the universal probe 1037-DIG revealed at least six rrl operons in the E. ictaluri genome (data not shown). When the IVS-specific probe was used, the same bands were displayed indicating that all rrl displayed by ribotyping contained the IVS. This result contradicts the PCR analysis by which at least one rrl operon did not contain IVS. However, bands highlighted by ribotyping were large in size, ranged from 8 kb to 25 kb, and could contain more than one rrl operon. Fragmentation of 23S rRNA due to IVS excision was revealed by methylene blue staining of ribosomal RNA extracted from E. ictaluri isolates as well as from controls. If an rrl gene contains an IVS in helix-45 and that IVS is removed during posttranscriptional processing (without religation of the rRNA), two smaller fragments of 1.2 kb and 1.7 kb result. Figure III. 4 displays the processing of rrn genes in E. ictaluri. 16S and 23S rRNAs were represented by a 2.9 kb and a 1.5 kb band respectively, that were displayed by all isolates tested including IVS-negative controls. E. ictaluri rRNAs pattern showed two additional bands of 1.2 kb and 1.7 kb demonstrating fragmentation of the 23S rRNA in this species. 51 Discussion Intervening sequences were first described by Burgin et al. (1990) as responsible for 23S rRNA fragmentation in Salmonella. These intervening sequences, or internal transcribed spacers, are removed (without splicing) after transcription of ribosomal genes by the action of RNaseIII, and the resulting 23S rRNA fragments are held together by the compact structure of ribosomes which remain functional. Since their discovery, IVS have been described in numerous bacteria including Alphaproteobacteria, Gammaproteobacteria, and Epsilonproteobacteria as well as in Spirochaeta (Evguenieva-Hackenberg and Klug 2000). Two characteristics of IVS have yet to be explained by microbiologists. First, a seemly random and sporadic distribution of IVS through bacteria has been observed. Second, their function is not known. Presence of IVSs in ribosomal genes seems not to interfere with cell viability nor appear to confer any evolutionary advantage. These unique sequences have been well studied within the family Enterobacteriaceae, where six out of 14 tested genera harbored isolates with IVS in their 23S rRNA genes (Pronk and Sanderson 2001). Intervening sequences in Enterobacteriaceae appear at two different sites, helix-25 and helix-45, according to the proposed secondary structure of the 23S rRNA, and can be grouped in different families based on nucleotide identity. However, these families (or superfamilies) cannot be correlated to other hierarchal classifications of Enterobacteriaceae based on taxonomy or phylogenetics. Horizontal and vertical gene transfer episodes had to be combined in order to justify IVSs distribution in Enterobacteriaceae (Pabbaraju et al. 2000). In this work, we have sequenced and characterized the 23S rRNA genes from two Enterobacteriaceae whose genus, Edwardsiella, has never been the target of ribosomal 52 gene-based phylogenetics studies and is rarely included or compared with other members of the family Enterobacteriaceae. The analysis of the 23S rRNA of E. ictaluri and E. tarda confirmed the close relationship between both species which was previously determined by 16S rRNA analysis (data not shown). The 23S rRNA gene derived parsimony tree showed the genus Edwardsiella is phylogenetically distant from the ?core? members of the family Enterobacteriaceae, such as Salmonella and Escherichia, as previously reported from DNA:DNA hybridization studies (Krieg and Holt 1984). Interestingly, E. ictaluri contains an insertion of 98 bp in helix-45 that has high similarity with IVS described in other Enterobacteriaceae. Sequence analysis of this new IVS revealed a 97% nucleotide sequence identity with S. typhimurium helix-45 IVS. In fact, only two nucleotide differences in the IVSs of these two species were found (one transition and one transversion). Based on this high similarity between E. ictaluri and S. typhimurium IVSs, both sequences should be included under the same 45 helix-IVS family M (Miller et al. 2000). Oddly, S. typhimurium IVS has a higher similarity (97%) with E. ictaluri IVS than with any other helix-45 IVS described to date. S. typhimurium has seven rrl genes, six of which present helix-45 IVSs of identical sequence (Mattatall and Sanderson 1996). The number of rrl genes in E. ictaluri has not been completely defined but, according to our ribotyping data, at least six rrl genes containing helix-45 IVSs are harbored in this species. When the IVSs from all thirty E. ictaluri isolates were cloned and sequenced, no differences in sequence between isolates or clones were noted, suggesting all helix-45 IVSs are identical. However, PCR results suggest at least one rrl gene in E. ictaluri lacks the 45-helix IVS. These results 53 could be explained by the existence of seven rrl genes in the chromosome, six of which would contain helix-45 IVS, as in S. typhimurium (Mattatall and Sanderson 1996). IVSs in members of the seven genera (Salmonella, Yersinia, Proteus, Providencia, Klebsiella, Citrobacter, and Edwardsiella) of Enterobacteriaceae show some degree of sequence similarity with each other, but no significant nucleotide identity between Enterobacteriaceae IVSs and any other sequences, including IVSs from other bacteria has been detected by BLAST searches (Pronk and Sanderson 2001). However, the possession of IVS is a rather random phenomenon in this family, since some genera (i.e. Escherichia and Enterobacter) do not harbor IVSs in their rrl genes. Earlier studies on nucleotide sequences supported the idea that IVSs must be exchanged between different genera by lateral transfer. IVSs have terminal inverted repeats flanking the central region, which is characteristic of a mobile genetic element. These mobile elements could be transferred from one species to another by phage-mediated transduction, transformation, or conjugation. However, alternative hypothesis suggests that the progenitor of these species obtained the IVS from a common ancestor but that subsequent variation between different species was simplely due to the loss of IVSs by some strains or species. Edwardsiella could have obtained the IVS prior to species divergence and, while E. ictaluri still maintains the insertion (in at least six rrn operons), E. tarda could have lost it. From an applied point of view, the availability of 16S rRNA gene sequences has facilitated the development of PCR-based methods in molecular diagnosis. However, cases of very closely related species have highlighted the need of using other conserved molecules such as the 23S rDNA (Arias et al. 1995) or the intergenic spacer region (ISR) 54 (Welker et al. 2005) as targets in a PCR detection methods since the 16S rRNA gene might not provide sufficient genetic variability. This is the case for E. tarda and E. ictaluri, as both species shared more than 99% identities in their 16S rDNA so that finding species-specific PCR primers has been an obstacle in developing PCR detection methods. To date, the only PCR detection method for E. ictaluri (Bilodeau et al. 2003) targets a mobile element present in the E. ictaluri genome since non specific signature sequences cannot be found in the 16S rRNA gene. Because of the negative economic impact caused by these two pathogens in the catfish industry, rapid and specific detection methods need to be implemented in fish diagnostic laboratories in order to hasten treatments and control outbreaks. The discovery of a helix-45 IVS sequence in the 23S rRNA gene of E. ictaluri, which is not present in E. tarda, opens possibilities for using this newly identified sequence in molecular diagnosis of enteric septicemia of catfish (ESC). The potential of the E. ictaluri IVS as a PCR-based detection target was described in chapter II. 55 Table III.1. Strains used in the IVS study. Species Strain Origin E. ictaluri CECT 885 T Catfish, Georgia, USA 195, 196, 151, 218, 219, AL-93-75, AL- 02-27, AL-93-58, ALG-03-278, ALG-03- 190, ALG-03-161, ALG-03-278, ALG-03-275, ALG-03-277, ALG-03-192 Catfish, Alabama, USA S94-728, S94-1034, S94-715, S94-703, S94-827, S94-862, S94-872, S94-1051, S94-707,1696, 1760, 1963 Catfish, Mississippi, USA EILO Catfish, Thailand E. tarda CECT 849 T Human feces 1909, Flb, Flb2, 172, Catfish, Alabama, USA Escherichia coli K-12 Human feces T, type strain CECT, Spanish Type Culture Collection Table III.2. Comparison of E. ictaluri helix-45 IVS with other Enterobacteriaceae helix-45 IVSs. Nucleotide sequence identities over 80% are shown in bold. % Nucleotide sequence identity with helix-45 IVS of Strains AF 176788 AF 176788 AF 176792 M 35802 AF 177051 U 49925 DQ 314205 M 35814 P. vulgaris AF176788 100 Pr. rettgeri AF176788 86 100 P. vulgaris AF176792 51 57 100 Y. bercovieri M35802 42 48 35 100 P. mirabilis AF177051 36 48 38 88 100 S. typhimurium U49925 80 79 52 50 49 100 E. ictaluri DQ314205 72 72 52 46 49 97 100 Y. enterocolitica M35814 80 82 52 44 51 85 83 100 5 6 57 Figure III.1. Parsimony tree based on 23S rRNA gene sequences. The tree shown is a consensus bootstrap tree based on 500 resampled parsimony trees. Significant (>50) bootstrap values for each branch are indicated. Species names and GenBank accession numbers are noted as well. Bar shows 5% sequence divergence. 100 100 93 86 84 53 100 100 93 60 100 99 100 100 100 51 100 80 94 100 100 100 Vibrio Unclassified aphid secondary endosymbiont Arsenophonus spp. endosymbiont Regiella Hamiltonella Photorhabdus Sodalis Shigella Escherichia Escherichia Escherichia Escherichia Shigella Salmonella Salmonella Salmonella Salmonella Salmonella Salmonella Salmonella Citrobacter Klebsiella Edwardsiella Edwardsiella Edwardsiella Yersinia Yersinia Yersinia Serratia Yersinia Erwinia Morganella Enterobacter Proteus Proteus cholerae insecticola defensa luminescens glossinidius flexnerii coli coli coli coli boydii enterica typhimurium typhimurium enterica enterica bongori enterica freundii pneumoniae ictaluri ictaluri tarda pseudotuberculosis pestis pestis symbiotica enterocolitica carotovora morganii aerogenes vulgaris mirabilis AE004157 AF263561 AY296734 AY296733 NC_005126 NC_007712 NC_004741 AB035922 AB035926 U00096 J01695 CP000036 U77919 AE008893 AE008895 CP000026 U77926 U77927 U77922 U77928 X87284 DQ211094 DQ314205 DQ211093 NC_006155 AE017127 AE013984 AY296732 U77925 BX950851 % 5 AY587142 AY116909 AY116918 AY116927 AY116929 Figure III.2. Proposed secondary structure of IVSs from E. ictaluri (A) and S. typhimurium (B). Secondary 58 structure predictions based on free-energy minimization. Nucleotide differences in E. ictaluri IVS compared with S. typhimurium IVS are marked by asterisks. The proposed conserved site at which helix-45 was replaced by the IVS is indicated by a horizontal line. A B 59 Figure III.3. Agarose gel containing the PCR products from amplification of 23S rRNA gene containing the IVS. Lane M, 50 bp molecular marker. Lane C, no template DNA; lane 1, Escherichia coli; lane 2, E. tarda CECT 849; lane 3 to lane 9, E. ictaluri isolates: CECT 885, EILO, 195, 196, 151, 218 and 219, respectively. rrl operons containing IVS in helix-45 yielded a 152 bp amplified product. Operons lacking IVS generated a smaller amplicon of 54 bp. This 54 bp amplified should not be confused with the primer-dimer band observed in the no-template control. 60 Figure III.4. rRNAs of E. ictaluri CECT 885 (1), EILO (2), E. tarda CECT 849 (3) and Escherichia coli (4). Total RNA was electrophoresed, blotted, and stained with methylene blue as described in Material and Methods. 1 2 3 4 61 IV. COMPARISON OF LIPOPOLYSACCHARIDE AND PROTEIN PROFILES BETWEEN FLAVOBACTERIUM COLUMNARE STRAINS FROM DIFFERENT GENOMOVARS Abstract Lipopolysaccharide (LPS) and total protein profiles from four Flavobacterium columnare isolates were compared. These strains belonged to genetically different groups and/or presented distinct virulence properties. F. columnare isolates ALG-00-530 and ARS-1 are highly virulent strains that belonged to different genomovars while F. columnare FC-RR is an attenuated mutant used as a live vaccine against F. columnare. Strain ALG-03-063 is included in the same genomovar group as FC-RR and presents a similar genomic fingerprint. Electrophoresis of LPS showed qualitative differences among the four strains. Further analysis of LPS by immunoblotting revealed that the avirulent mutant lacks the higher molecular bands in the LPS. Total protein analysis displayed by immunoblotting showed differences between the strains analyzed although common bands were present in all the isolates. FC-RR lacked two distinct common bands (34 kDa and 33 kDa) shared by the other three isolates. Based on the difference of LPS and total protein profiles, it is possible to discriminate the attenuated mutant FC-RR from other F. columnare strains. Keywords: LPS, whole protein profiles, F. columnare, modified live vaccine 62 Introduction Flavobacterium columnare is the causal agent of columnaris disease, one of the most important bacterial diseases of freshwater fish species. This bacterium is distributed world wide in aquatic environments, affecting wild and cultured fish as well as ornamental fish in aquaria (Austin and Austin 1999). F. columnare is considered the second most important bacterial pathogen in commercial cultured channel catfish, Ictalurus punctatus, in the southeastern USA, second only to Edwardsiella ictaluri (Wagner et al. 2002). Direct losses due to F. columnare are estimated in excess of millions of dollars per year. Mortality rates of catfish populations in ponds can reach 50% to 60% and can be as high as 90% in tank-held catfish fingerlings (USDA 2003a, 2003b). Columnaris disease usually begins as an external infection of fins, body surface, or gills. The fins become necrotic with grayish to white margins, and initial skin lesions appear as discrete bluish areas that evolve into depigmented necrotic lesions. Skin lesions might have yellowish mucoid material accompanied by mild inflammation. Lesions can develop exclusively on the gills, which usually results in subacute disease and mortality, as is typically the case in young fish (Plumb 1999). Due to the ubiquitous presence of F. columnare in aquatic environments, eradication of the disease in fish farms will not likely occur. Control and treatment of columnaris have primarily been directed toward the use of improved water-management practices to reduce physiological and environmental stress in the fish. Recently, a modified live F. columnare vaccine has been developed using a rifampicin-resistant strategy (Shoemaker et al. 2005b). This method had been previously used by Montaraz and Winter (1986) to generate a rough Brucella abortus strain, currently employed as the 63 official vaccine for cattle brucellosis in the USA. The same strategy was used by Klesius and Shoemaker (1999) to create an E. ictaluri rifampicin-resistant mutant patented as a modified live vaccine against enteric septicemia of catfish (ESC) (AQUAVAC-ESC, Intervet, Millsboro, DE). Characterization of B. abortus and E. ictaluri rifampicin- resistant mutants revealed that lipopolysaccharide (LPS), a main virulence factor for Gram-negative bacteria, lacked the high molecular bands observed in virulent isolates (Vemulapalli et al. 1999; Arias et al. 2003). Flavobacterium columnare belongs to the Cytophaga-Flavobacterium- Bacteroides (CFB) group and is phylogenetically distant from the better studied gammaproteobacteria subclass which contains major human and animal pathogens (including Brucella and Edwardsiella). To date, virulence factors in F. columnare are poorly characterized. Specifically, the role of LPS in columnaris pathogenicity has not been explored. The objective of this study was to investigate if the attenuated F. columnare mutant originated through a rifampicin-resistance strategy presented a modified LPS and a different total protein profile by comparison with virulent strains. Materials and Methods Bacterial isolates and culturing Four F. columnare strains, ALG-00-530, FC-RR, ARS-1, and ALG-03-063 were used in this study. Isolate ALG-00-530 was obtained from diseased channel catfish at the Alabama Fish Farming Center, Greensboro, Alabama. The ARS-1 isolate was recovered from diseased channel catfish at the Aquatic Animal Health Research Unit, USDA-ARS, Auburn, Alabama. These two isolates have been demonstrated to be virulent for channel catfish (Shoemaker et al. 2003a). F. columnare ALG-03-063 was also isolated from 64 diseased channel catfish. FC-RR strain is a rifampicin-resistant mutant, avirulent for catfish. This attenuated mutant has been licensed as a modified live vaccine by Intervet, Inc. and it is now undergoing field trials (Shoemaker et al. 2005b) (Patent No.: US 6,881,412 B1). Genetic fingerprinting A previous study by Arias et al. (2004) divided F. columnare species into four well-defined genetic groups. All catfish isolates were grouped into three different genogroups (I through III; genogroup IV included only tilapia isolates). Strains ALG-00- 530 and ARS-1 belonged to genogroups I and II, respectively; while ALG-03-063 belonged to genogroup III. The avirulent mutant FC-RR was not included in that former study but was fingerprinted in the present work. Amplified fragment length polymorphism (AFLP) patterns from FC-RR mutant were obtained as described by Arias et al. (2004), added to the existing database and analyzed. Production of polyclonal anti-serum against F. columnare Twelve channel catfish were randomly divided into three groups of four fish each. Two groups were used to produce antibodies (Ab) against F. columnare strains (Ab- ALG-00-530 and Ab-FC-RR) while the third group served as negative control. Antigen preparation was from cells grown overnight in modified Shieh broth (5.0 g tryptone, 2.0 g yeast extract, 1.0 mL of 45.6 mM CaCl 2 2H 2 O, 3 mL of 0.36 mM KH 2 PO 4 , 10 mL of 0.12 M MgSO 4 7H 2 O, 10 mL of 0.41 mM FeSO 4 7H 2 O, 1 L of distilled water, pH 7.2- 7.4) and quantified by plate count (10 X 10 8 CFU/mL). Bacterial cells were sonicated for 30 seconds at speed 4 using a Virsornic 600 sonicator (SP Industries Company, Gardiner, NY). One milliliter of the sonicated suspension was mixed with 1 mL complete Freund?s 65 adjuvant (Sigma, St. Louis, MO). Two hundred microliters of the mixture were intraperitoneally injected into each fish. Control fish were injected with 200 ?L of Shieh broth. A booster injection was administered 27 days after initial immunization using 200 ?L antigen prepared as above (Freund?s incomplete adjuvant (Sigma) was used this time). Prior to the booster immunization, about 1.5 mL of blood was obtained from the caudal vasculature of each fish for titer determination (Shoemaker et al. 2003b). Three weeks after booster immunization, fish were bled again, and this antiserum was used to probe the western blots (see below). Blood was allowed to clot at room temperature for 1 h and serum was carefully removed after centrifugation for 5 min at 1,000 g. Sera were stored at -20 ?C until use. In order to determine the relative titer of the sera, enzyme linked immunosorbent assays (ELISA) were conducted according to Shoemaker et al. (2003b). Sonicated cell suspensions were prepared as described above and serial ten-fold dilutions were used to coat the plates; proper positive and negative controls were included. Serum from each individual fish was tested against each antigen dilution. Lipopolysaccharides (LPS) and total protein extraction The four isolates were cultured in modified Shieh broth for 24 h at 28 ?C. Three milliliters of broth were centrifuged at 3,000 g for 15 min. Pelleted cells were resuspended in lysis buffer and proteins were extracted according to Arias et al. (1997b). Crude LPS was extracted following the phenol-water protocol described by Westphal and Jann (1965). After lyophilization, the LPS extract was diluted in sample buffer at a final concentration of 1mg/mL. Aqueous and phenol LPS phases were stored at -80?C until use. 66 LPS and total protein analysis Electrophoresis ? LPS from both phases were resuspended in sample treatment buffer and electrophoresed by discontinuous SDS-PAGE on a 4% stacking gel and a 20% separating gel (Laemmli 1970). Gels were run at 15 mAmp for 90 min. Gels were silver stained following the instruction of Bio-Rad Silver Stain Kit (Bio-Rad, Hercules, CA). Total protein electrophoresis followed the procedure described by Arias et al. (1997b). Coomassie staining was used to visualize the protein bands following standard methods (Sambrook and Russell 2001). Western blotting ? Protein and LPS samples were analyzed in duplicate. First, samples were resolved on 12% SDS-PAGE gel, and then transferred to polyvinylidene difluoride (PVDF) membrane (Bio-Rad) at 100 volts for 1 h. After blotting, one membrane was incubated with anti-ALG-00-530 serum while the second membrane was incubated with anti-FC-RR serum. After blocking for 30 min, 1:500 diluted polyclonal catfish serum was incubated with the membrane overnight followed by a one-hour incubation with a monoclonal antibody (E-8) specific for channel catfish IgM (1:10) (Klesius 1990) , and labeled with conjugated goat anti-mouse Ig (1:5,000), followed by detection with Opti-4CN T M (Bio-Rad). Results AFLP analysis Figure IV.1 shows results of the cluster analysis of the AFLP patterns from the strains used in the study. Attenuated mutant FC-RR displayed a unique AFLP fingerprint although clustered among other genogroup III strains. AFLP profile of strain ALG-03- 063 presented the highest percent of similarity with the one displayed by the attenuated 67 mutant. Based on this result, ALG-03-063 was selected for further comparisons as the most genetically similar strain to FC-RR present in our collection (that currently includes more than 50 F. columnare isolates; data not shown). ELISA analysis All fish immunized with F. columnare yielded positive titers. Control fish presented a very low cross-reactivity with the sonicated cells. All immunized fish produced strong immunity against both antigens. Although antisera from immunized fish reacted with both ALG-00-530 and FC-RR antigens, the corresponding antigen provided the highest titer (data not shown). Individual fish sera showing the highest titer (?1/800) were chosen for immunoblotting analysis (fish no. 3 anti-ALG-00-530 and fish no.1 anti- FC-RR respectively). LPS characterization The results of SDS-PAGE of LPS from F. columnare are shown in Figure IV.2. The characteristic ladder-like pattern typical of Gram-negative pathogens was not observed. Only a few bands ranging from 3.5 kDa to 17 kDa were present in all four strains. Phenol phase extracts contained higher amounts of LPS than the aqueous- extracted samples. Strain ALG-00-530 showed a few bands in the phenol phase extract while only one weak band was visualized in the aqueous phase. A similar pattern was observed with strain ARS-1 although band intensities were stronger. Attenuated mutant FC-RR shared a similar LPS pattern with ALG-03-063 although FC-RR bands were slightly smaller. Bands from F. columnare LPS were visualized much more effectively in immunoblots. When catfish polyclonal antiserum was used for immunoblotting, 68 important differences between the four strains were revealed (Figure III.3). While all strains isolated from diseased channel catfish (ALG-00-530, ARS-1, and ALG-03-063) presented high molecular bands (above 21 kDa) in both phenol and aqueous phases, the FC-RR mutant exhibited bands below that range. ALG-00-530 and ARS-1 again showed a similar LPS pattern regardless of antisera or extraction phase used. ALG-03-063 presented the same size bands in the phenol phase as FC-RR but also exhibited higher molecular weight bands that were absent in the mutant. Interestingly, these results were consistent regardless of antisera used (anti-ALG-00-530 or anti-FC-RR). However, significant differences between aqueous and phenol phase samples were found in both immunoblots. Total protein analysis Western blots of total proteins were also analyzed with both sera (Figure IV.4). The immunoblots indicated that the four strains did contain different antigenic proteins but also shared some common bands. ARS-1 showed more distinct antigenic protein bands than any other strain analyzed (Figure IV.4). The strains ALG-00-530, ARS-1 and ALG-03-063 contained two very distinct bands around 35 kDa. FC-RR mutant lacked these bands but instead presented two additional bands in the 30 kDa range. Each strain had very similar banding pattern with anti-ALG-00-530 and anti-FC-RR sera. When ALG-00-530 serum was used, reactivity against low molecular weight bands present in the three clinical strains was observed. FC-RR low molecular bands did not react with anti-ALG-00-530 serum. On the contrary, when anti-FC-RR was used, low molecular weight proteins in FC-RR were revealed. 69 Discussion Lipopolysaccharide is a major superstructure of Gram-negative bacteria, contributes greatly to the structural integrity of the bacteria, and protects them from host immune defenses. The LPS has been described and well characterized as a virulence factor in many bacterial species (Salmonella sp. (Makela et al. 1973), Escherichia coli (Medearis et al. 1968), Shigella flexneri (Rajakumar et al. 1994), Brucella abortus (Vemulapalli et al. 2000) including fish pathogens such as Vibrio vulnifucus (Amaro et al. 1997), E. ictaluri (Arias et al. 2003), and F. psychrophilum (MacLean et al. 2001). Attempts to characterize the structure of the LPS in F. columnare were made by MacLean et al. (2003) and by Vinogradov et al. (2003). Unfortunately, the strain used by these authors (ATCC 43622) had been originally misidentified as F. columnare but was actually F. johnsoniae (Darwish et al. 2004; Shoemaker et al. 2005a). Therefore, no information on F. columnare LPS composition or immunogenic role was available. In this study, catfish polyclonal antisera against two strains, ALG-00-530 and FC- RR (attenuated mutant), were generated and used to determine LPS immunogenic properties between genetically different F. columnare strains. Differences in LPS between four strains of F. columnare were evident after electrophoresis and silver staining but even more obvious after western blot analysis. Although a typical LPS ladder has been reported in other Flavobacterium species (MacLean et al. 2001), our silver staining of F. columnare LPS analysis revealed only a few bands under 7 kDa. Immunogenic bands present in the LPS were revealed by immunoblots. However, there was a major difference in band molecular weight between strains isolated from diseased channel catfish and the attenuated strain FC-RR. The three strains from diseased 70 fish (ALG-00-530, ARS-1, and ALG-03-063) presented LPS bands between 21 and 29 kDa while FC-RR exhibited a unique band under 21 kDa. Our results suggest that the induction of rifampicin-resistant mutation(s) in F. columnare results in loss of the high molecular weight bands displayed by virulent isolates. Similar results were reported in B. abortus and E. ictaluri (Arias et al. 2003) when LPS from rifampicin-resistant mutants was analyzed. Rifampicin is a potent and broad-spectrum antibiotic well known as DNA- directed RNA polymerase inhibitor. Mutations in the RNA polymerase gene that conferred resistance against rifampicin have been widely documented (Jin and Gross 1988). On the other hand, the mechanism by which rifampicin induces mutations that affect LPS structure are relatively unknown. It is remarkable how such different pathogens (Brucella, Edwardsiella and Flavobacterium) behaved similarly under the same stressor suggesting that short O-chain LPS will favor cell survival. It has been postulated that rifampicin, like other hydrophobic molecules, enters the cell via simple diffusion through the outer membrane with LPS being the main hurdle (Kirschbaum and Gotte 1993). A structural change in the LPS might enhance the barrier effect protecting the cell against rifampicin. Vemulapalli et al. (1999) suggested that multiple genes involved in LPS biosynthesis are disrupted in B. abortus and this results in the attenuated phenotype. Our results indicate that the LPS in F. columnare may play an important role in columnaris pathogenesis because the lack of LPS high molecular weight bands seems to be correlated with total lack of virulence in other bacterial species (Kimura and Hansen 1986; Rajakumar et al. 1994; Amaro et al. 1997; Arias et al. 2003). Although a thorough antigenic characterization in F. columnare has not yet been attempted, some serological 71 diversity has been reported between isolates (Anacker and Ordal 1959; Sanders et al. 1976). This antigenic variation can be justified by the genetic variability observed in this species (Arias et al. 2003; Thomas-Jinu and Goodwin 2004) . In fact, we observed some serological differences in LPS within the virulent strains. While ALG-00-530 and ARS-1 both showed a very similar pattern denoting common antigens, ALG-03-063 presented at least one additional LPS band regardless of extraction phase and antisera used. Interestingly, these three strains belong to different genogroups and present very distinct fingerprinting profiles; however, ALG-00-530 and ARS-1 resulted very similar pattern by LPS immunoblots. These two strains have been used by our group in virulence studies and although both are able to cause columnaris disease in catfish by immersion, ALG-00- 530 has consistently exhibited a higher mortality rate than ARS-1 (unpublished information). When we compared the antigenic variability exhibited by the total protein profiles, all four strains showed a different profile. Differences in protein composition between isolates were expected as they presented differences in their genome. However, there was a large set of common bands shared by all strains and revealed by both antisera. This indicates that the FC-RR mutant presents a number of common epitopes with other F. columnare isolates and could explain why the FC-RR strain used as a modified live vaccine is able to confer protection against different F. columnare isolates. In conclusion, we showed that the F. columnare FC-RR strain lacks the high molecular weight components of LPS present in other virulent strains. Differences in LPS and immunogenic proteins were observed between genetically different F. columnare strains. A secondary outcome of this study was the genetic fingerprinting of FC-RR. The 72 attenuated mutant presents a unique AFLP profile, not present in our existing database, which can be used to track the modified live vaccine strain (FC-RR) in the fish farm environment. 73 Figure IV.1. Dendrogram based on AFLP patterns of four strains of F. columnare (ALG- 00-530, FC-RR, ARS-1, and ALG-03-063). The tree was derived by UPGMA cluster analysis of the AFLP profiles. FC-RR ALG-03-063 ARS-1 ALG-00-530 10 0 9590858075 Genogroup III Genogroup II Genogroup I % similarity 74 Figure IV.2. Silver stained sodium dodecyl sulfate-polyacrylamide gel electrophoresis profiles of the LPS preparations from F. columnare strains used in this study. Lane 1,molecular standard; lanes 2 through 5, LPS aqueous phase from ALG-00-530 (2), FC- RR (3), ARS-1 aqueous (4), and ALG-03-063 (5); lanes 6 through 9 show LPS phenol phase from ALG-00-530 (6), FC-RR (7), ARS-1 (8) and ALG-03-063 (9). 1 2 3 4 5 6 7 8 9 6.9 kDa 21 kDa 29 kDa 75 Figure IV.3. Immunoblot using anti-ALG-00-530 serum (A) and anti-FC-RR serum (B) with LPS preparation. Lane 1, molecular standard; lanes 2 through 5, LPS aqueous phase from ALG-00-530 (2), FC-RR (3), ARS-1 aqueous (4), and ALG-03-063 (5); lanes 6 through 9 show LPS phenol phase from ALG-00-530 (6), FC-RR (7), ARS-1 (8) and ALG-03-063 (9). 21 kDa 6.9 kDa 21 kDa 6.9 kDa A B 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 35.6 kDa 35.6 kDa 76 Figure IV.4. Western blot analysis of total protein with anti-ALG-00-530 serum (A) and anti-FC-RR serum (B). Lane 1, molecular standard; lane 2, ALG-00-530; lane 3, FC-RR; lane 4, ARS-1; lane 5, ALG-03-063. A B 1 2 3 4 5 1 2 3 4 5 35.6 kDa 21 kDa 6.9 kDa 56 kDa 101 kDa 125 kDa 210 kDa 35.6 kDa 21 kDa 6.9 kDa 56 kDa 101 kDa 125 kDa 210 kDa 77 V. IDENTIFICATION, CHARACTERIZATION AND EXPRESSION PATTERNS OF GTF, HEMH, NORB, AND TRX GENES IN FLAVOBACTERIUM COLUMNARE Abstract Four genes from a shotgun genomic library of F. columnare virulent strain ALG- 00-530 were identified, characterized, and selected for differential expression analysis based on sequence similarity to putative virulence genes from related species. These genes were: glycosyltransferase (gtf), ferrochelatase (hemH), nitric oxide reductase (norB) and thioredoxin (trx). A collection of 30 F. columnare strains, including strains from genomovars I and II, were tested for the presence of these genes. Distribution patterns of gtf, hemH, norB, and trx across the species were not uniform. Nucleotide sequence variation was observed between genomovars for each gene; however, strains within the same genomovar shared identical gene sequences. Nine strains of F. columnare (from both genomovars I and II) were chosen for gene expression analysis. The expression profile of these genes varied when selected strains were grown in vitro under identical conditions. ALG-00-530 was chosen for gene expression comparison under standard growing conditions, iron-limited conditions and in the presence of skin explants from channel catfish (Ictalurus punctatus). gtf, hemH, norB, and trx gene expression levels varied when ALG-00-530 was incubated under different conditions. 78 Keywords: Flavobacterium columnare, gtf, hemH, norB, and trx gene expression, iron- limited, catfish skin explants, pathogenicity 79 Introduction Flavobacterium columnare, the etiologic agent of columnaris disease, is a Gram- negative bacterium that can infect most freshwater fish species. Differences in virulence are known to exist among F. columnare isolates, resulting in variable mortality of fish (Pacha and Ordal 1970; Suomalainen et al. 2006). Understanding the pathogenesis of F. columnare is critical for fish health, but our knowledge of F. columnare virulence factors is far from being complete. Only a few genes from F. columnare have been described to date, and most known sequences from this pathogen correspond to ribosomal genes. Among the non- ribosomal sequences identified, the chondroitin AC lyase gene in F. columnare has been previously sequenced and characterized (Xie et al. 2005). The enzyme codified by this gene is able to degrade acidic polysaccharides, such as hyaluronic acid and chondroitin sulfates (Griffin 1991; Teska 1993; Stringer-Roth et al. 2002). Differences in chondroitin AC lyase activity have been observed between F. columnare isolates (Stringer-Roth et al. 2002). Relation of chondroitin AC lyase activity to virulence in F. columnare has been recently reported (Suomalainen et al. 2006). Genes encoding outer membrane proteins, such as zinc metalloprotease and prolyl oligopeptidase, have also been characterized in a virulent F. columnare isolate (Xie et al. 2004). However, association of these two genes with F. columnare pathogenicity is unclear. In addition, several proteases have been identified in F. columnare, but their specific roles in columnaris pathogenicity are still unknown (Bertolini and Rohovec 1992; Newton et al. 1997). The ability of F. columnare to attach to fish tissues is thought to be a factor correlated to virulence (Decostere et al. 1999b). Bader et al. (2005) selected for an 80 adhesive-defective F. columnare strain that exhibited reduced virulence to channel catfish. Unfortunately, no adhesin gene has been identified in F. columnare. Genetic variability of F. columnare species has been characterized using different molecular markers (Triyanto and Wakabayashi 1999; Thomas-Jinu and Goodwin 2004; Arias et al. 2004; Darwish and Ismaiel 2005). Restriction fragment length polymorphism analysis of the 16S rRNA gene divided F. columnare species into three genomovars (Triyanto and Wakabayashi 1999). The coexistence of three main genomic groups within the species (genomovars I, II, and III) has been corroborated by intergenic spacer region (ISR) sequencing (Darwish and Ismaiel 2005), and single-strand conformation polymorphism (SSCP) analysis (Olivares-Fuster et al. 2007). To date, most of the genetic markers used for F. columnare strain typing have relied on ribosomal gene analysis. The aim of this work was to identify and characterize putative virulence genes in this species and to further investigate the genetic diversity of F. columnare using non- ribosomal markers. To achieve this objective, a partially sequenced shotgun genomic library constructed from a virulent F. columnare strain was used. Four putative virulence genes were selected for further analysis. Gene presence and expression patterns were analyzed across the species. Two in vitro experiments aimed to mimic some of the environmental conditions F. columnare encounters during infection (limited iron and fish skin) were used for gene expression analysis. 81 Materials and Methods Bacterial strains and culture conditions Thirty strains of F. columnare, representing genomovars I and II, were used in this study (Table V.1). Cells were cultured in modified Shieh broth (Shoemaker et al. 2005a) at 28 ?C. Putative virulence gene identification A shotgun genomic library of the F. columnare virulent strain ALG-00-530 was constructed. Briefly, F. columnare DNA was extracted and purified following standard protocols (Sambrook and Russell 2001). Total DNA was partially digested with Sau3A I. Digested DNA ranging from 1 to 1.25 kb was purified by double size selection and ligated to the digested pBluescript II (KS-) vector (Stratagene, Cedar Creek, TX) at a molar ratio of 1:2. The cloning site was BamH I. Library titer was estimated at 35 recombinant clones per microliter. Quality control tests showed 1% or less empty clones and an expected genomic coverage of 20X based on approximate genome size of 4.1 mb. Three thousand clones of this library have been sequenced to date at the USDA-ARS MSA Genomics Laboratory (Stoneville, MS) following standard procedures. Sequences from the shotgun genomic library were compared with sequences in the GenBank database available at NCBI (National Center for Biotechnology Information http://www.ncbi.nlm.nih.gov/) using the BLASTx algorithm (Altschul et al. 1990). Sequences with more than 35% identity were recorded. Among the identified genes, four were chosen for further analysis due to their high identity (?65%) with known genes described as virulence factors in other bacteria. Open reading frames (ORF) of the four sequences were identified using Vector NTI ? Suite 8 software package (Invitrogen, 82 Carlsbad, CA). ORF were translated into amino acid sequences by using Vector NTI ? Suite 8 and the GenBank database was searched for protein sequences using the BLASTp tool (Altschul et al. 1990). Protein-protein identity percentage was recorded for each chosen ORF. Specific primers for each gene were designed using Vector NTI ? Suite 8. Primer sequences for each gene were shown in Table V.2. These primers were tested on 30 strains of F. columnare by PCR. Unless otherwise stated, all PCR reagents were purchased from Promega (Madison, WI). Each 50 ?L PCR reaction included 2.5 ?M MgCl 2 , 1X buffer, 0.2 ?M of both primers, 0.2 ?M of dNTPs, 1.7 unit of Taq polymerase, and 60 ng of DNA template. The PCR amplification profile was 5 min hot start at 95 ?C; 35 cycles of 30 s at 94 ?C, 45 s at 58 ?C, and 1.5 min at 72 ?C; and 10 min at 72 ?C. DNA, RNA extraction and Reverse transcription PCR (RT-PCR) Total bacterial DNA was extracted using the Qiagen DNeasy Tissue kit (Qiagen, Valencia, CA) following the manufacturer?s instructions. Total RNA was extracted from 1.5 mL of bacterial culture using RNeasy plus Mini Kit (Qiagen). Turbo DNA-free kit (Ambion, Austin, TX) was used to eliminate DNA contamination in RNA samples. cDNA synthesis and RT-PCR were performed using the Reverse Transcription System (Promega). One microgram of total RNA was used to synthesize the cDNA in 20 ?L of reaction. cDNA synthesis reaction was diluted to 100 ?L, and 10 ?L were used as template for PCR using gene specific primers, as described above. Cloning and sequencing analysis PCR products were resolved through standard agarose gel electrophoresis. Amplified products were purified using the High Pure PCR Product Purification Kit (Roche Diagnostic Corporation, Indianapolis, IN) and cloned into pGEMTEasy 83 (Promega). To ensure sequence accuracy, two clones from each strain were sequenced at the Auburn University Sequencing Core (Auburn, AL). Nucleotide sequence alignments and deduced amino acid sequence alignments were performed by CLUSTAL W multiple sequence alignment algorithm (Chenna et al. 2003). Gene expression Expression of gtf, hemH, norB, trx was tested under standard conditions, iron- limited conditions, and in the presence of catfish skin explants. Following RNA extraction and cDNA synthesis, genes were amplified from cDNA by the designed primers as mentioned above. A 16S rRNA cDNA internal control was amplified with universal primers UFUL (5?-GCCTAACACATGCAAGTCGA-3?) and URUL (5?- CGTATTACCGCGGCTGCTGG-3?) (Chen et al. 2004). Positive (genomic DNA) and negative (total RNA) controls were included for each gene amplification. Amplified products were electrophoresed at 100 V for 30 min and visualized under UV light with ethidium bromide added to a 1% gel to a final concentration of 0.5 ?g/mL. Gene expression under standard conditions ? Nine strains of F. columnare, representing genomovars I and II, were cultured in modified Shieh broth at 28 ?C for 20 h. Bacterial cells were harvested from one and a half milliliters broth culture followed by RNA extraction and RT-PCR. ALG-00-530 gene expression under iron-limited conditions ? Two iron chelators, 2, 2-dipyridyl and transferrin, were used to remove free iron from the growth medium. 2, 2- dipyridyl was added to 50 mL of Shieh broth to a final concentration of 50 ?M. Preparation of the transferrin solution followed the protocol described by Biosca et al. (1993). Briefly, human apo-transferrin (Sigma, St. Louis, MO) was dissolved at a 84 concentration of 1 ?M in a solution containing 100 ?M Tris, 150 ?M NaCl, and 50 ?M NaHCO 3 (pH 8.0). This solution was sterilized by filtration. Prior to inoculation with F. columnare, transferrin was added to 50 mL Shieh medium to achieve a final concentration of 10 ?M and incubated at 37?C for 60 min. Cultured ALG-00-530 bacterial cells were collected by centrifugation at 2,000 g for 20 min. Pellets were washed once in PBS and centrifuged again. Cells were resuspended in approximately 1 mL Shieh broth and then 1/3 mL (of this broth) was transferred into 50 mL Shieh broth containing 2, 2-dipyridyl, or apo-transferrin. Fifty milliliters of Shieh broth was inoculated as control. One and a half milliliters of bacterial cells were collected at time 0 h, 2 h, 4 h, 6 h, 8 h, and 24 h after inoculation. Cells were centrifuged and pellets were resuspended in 100 ?L of RNA later (Qiagen) and frozen at -80 ?C until used. Reverse transcription PCR was performed as explained above. ALG-00-530 gene expression in the presence of catfish skin explants ? Skin explants from channel catfish that tested negative for F. columnare antibody presence were obtained according to the procedure by Xu and Klesius (2002). One milliliter of Medium 199 (Sigma) supplemented with fetal bovine serum at 2.5% was added into each well of a 24-well culture plate (Costar, Cambridge, MA). One cm 2 of fish skin explant was placed onto the bottom of 12 wells. One and a half milliliters of an overnight culture of ALG-00-530 bacterial cells was collected for the baseline control (C 0 ); one and half milliliters of the same bacterial culture was added into each of the 24 wells. Wells without fish skin explants were regarded as controls, while wells containing the fish explants were referred to as treatments. Bacterial cells were collected from each 85 individual well at 30 min, 1 h, 3 h, 6 h, 10 h, and 24 h from both control and treatment wells. After centrifugation, bacterial pellets were suspended in RNA later (Ambion, Austin, TX) and frozen at -80 ?C until used. Total RNA extraction and RT-PCR were performed as described above. Results Putative virulence gene identification Out of 3,000 clones sequenced from the ALG-00-530 shotgun library, 73 sequences displayed a high nucleotide similarity with known genes present in GenBank (based on BLASTx search, data not shown). Within this subset, four sequences were identified from published literature as putative virulence genes and therefore selected for this study: glycosyltransferase gene (gtf) (GenBank accession number DQ911521), ferrochelatase gene (hemH) (GenBank accession number DQ911518), nitric oxide reductase gene (norB) (GenBank accession number DQ911522), and thioredoxin gene (trx) (GenBank accession number DQ911523). ORFs were translated into amino acid sequences which were then compared to GenBank sequences using BLASTp. Identity percentages of ORF based on protein-protein alignment were as follows: Gtf was 81% identical to F. johnsoniae sequence accession number ZP_01245614, HemH was 63% identical to F. johnsoniae sequence accession number ZP_01246582, NorB was 77% identical to an unidentified Flavobacterium species sequence accession number ZP_01106280, while Trx was 83% identical to F. johnsoniae sequence accession number ZP_01243853. Based on the PCR results, all strains contained norB and trx genes (Table V.1). However, no PCR amplification for hemH was observed in eight strains (GA-02-14, 86 ARS-1, BM, HS, GZ, BZ-1, BZ-4, BZ-5), while three strains (ARS-1, BM, HS) failed to amplify the gtf gene. Cloning and sequencing analysis Two F. columnare strains from genomovar I (ATCC 49512 and ALG-03-57) and genomovar II (ALG-00-530 and MS-02-475) were selected for multi locus sequence analysis using gtf, hemH, norB, and trx genes. Nearly complete sequences were amplified from gtf, norB, and trx and used for comparison. Only partial sequence (250 bp out of 626) of hemH was obtained. Alignment of these sequences showed two distinct populations. Both genomovar I strains shared an identical sequence for all three genes as did ALG-00-530 and MS-02-475 genomovar II strains (data not shown). However, nucleotide sequences differed between genomovars (Figure V.1). The variation percentage observed was 6.9% for gtf, 4.9% for hemH, 3.8% for norB, and 4.3% for trx. Translation of the nucleotide sequences showed the distinction between genomovars was carried out at the protein level. Amino acid sequence variation derived from each gene was not identical between genomovars except for Trx. Gft displayed a 2.2% variation between both genomovars. HemH protein sequence was 5.1% variable, while NorB presented only 0.88% sequence variation. Gene expression Gene expression under standard conditions ? The 16S rRNA gene, used as internal control, was expressed at any given time in all experiments. Gene expression patterns of nine strains of F. columnare varied at 20 h post inoculation under standard growing conditions and are summarized in Table V.3. Gene expression differed not only between genomovars but also within each genomovar. Genomovar I displayed three 87 expression patterns. Strains GA-02-14 and HS shared an identical expression profile, showing no expression of both norB and trx gene. Strains BM and ARS-1 also shared a similar expression profile with no expression of norB but positive expression of trx; however, the expression of trx gene in BM was stronger than in ARS-1. FC-RR (an avirulent mutant) displayed a distinct pattern with expression of norB but no expression of the other three genes. Genomovar II strains showed differences in expression patterns, as well. Stronger expression of norB and trx was observed in ALG-00-530 compared to other strains. LSU strain showed weak expression of trx gene, while BZ-1-02 and ALG- 00-527 failed to express any of the tested genes. ALG-00-530 gene expression under iron-limited conditions and in the presence of skin explants ? Iron-limited conditions did not interfere with F. columnare housekeeping gene expression (Table V.4). The 16S rRNA gene was uniformly expressed throughout the experiment. By contrast, gtf did not express at all in the experiment. hemH expression was only detected when cells were transferred to standard growing conditions for the first 2 h of incubation and after 2 h of incubation in the presence of 2, 2-dipyridyl (Table V.4, Figure V.2). Overall, transferrin seemed to have little effect on norB and trx gene expression when compared to standard conditions, although trx transcript levels seemed lowered when transferrin was incorporated into the medium. 2, 2-dipyridyl had an immediate repression effect on norB and trx expression, while the 16S rRNA gene was not affected by this chemical. However, transcript levels for norB and trx went back to high levels 2 h after transfer to iron-limited conditions, remaining stable until 8 h and 24 h for norB and trx, respectively. 88 When F. columnare cells were incubated in cell culture medium with or without skin explants, a distinct expression pattern was observed. Little expression of trx was detected in both controls and skin explants samples. The norB transcript was detected in both samples at different sampling times, although expression levels never reached those observed under standard growing conditions. However, expression levels of the 16S rRNA gene were comparable to those observed under standard conditions. Prior exposure to the culture medium, F. columnare cells cultivated overnight displayed the same pattern of expression previously observed under normal growing conditions. Discussion F. columnare has been reported to be a morphological (Song et al. 1988; Thomas- Jinu and Goodwin 2004), serological (Anacker and Ordal 1959), and genomically (Toyama et al. 1996; Triyanto and Wakabayashi 1999; Thomas-Jinu and Goodwin 2004; Arias et al. 2004; Darwish and Ismaiel 2005; Schneck and Caslake 2006) diverse species. According to Bernardet and Grimont (1989), only 78% similarity was found among F. columnare strains using DNA-DNA hybridization. Triyanto and Wakabayashi (1999) described three different genomovars among the species based on the analysis of the 16S rDNA-RFLP. Genomovar ascription has become routine for F. columnare strain characterization (Triyanto and Wakabayashi 1999; Arias et al. 2004; Olivares-Fuster et al. 2007). However, beyond ribosomal variability, no study has been conducted to investigate the intra-species variation of F. columnare at the single gene level, mainly due to the lack of genetic information available for this species. In the current study, the glycosyltransferase gene (gtf), ferrochelatase gene (hemH), nitric oxide reductase gene (norB), and thioredoxin gene (trx) have, for the first 89 time, been identified and characterized in F. columnare. These genes presented a high similarity to homologous genes within the Flavobacterium genus, showing the highest similarity with F. johnsoniae gene sequences. Although strains within the same genomovar showed identical gene sequences, there was a 4 to 7% nucleotide sequence variation observed between genomovars for each gene. Flavobacterium columnare strains were clearly divided into two groups, and such division was even expressed at the amino acid level. Lee et al. (1998) suggested that single amino acid substitutions could change the biological activity of proteins. We found some non-conserved amino acid substitutions between the genomovar I ALG-00-530 and the genomovar II ATCC 49512 in Gtf, HemH, and NorB protein sequences. Therefore, the activity of these proteins could differ between F. columnare genomovars. Our data agreed with the genomovar segregation previously reported for F. columnare. Genomovars could be identified by sequencing either norB or trx genes. gtf and hemH cannot be considered good markers, because they were not present in all F. columnare strains. Some examples have been published of such correlation between ribosomal gene-based variability and non-ribosomal gene-based variability i.e. F. psychrophilum (Soule et al. 2005). Non-ribosomal variation divided F. psychrophilum into two lineages that are associated with different host species. This is not the case for F. columnare, since channel catfish is susceptible to both F. columnare genomovars. However, preliminary studies showed genomovar II strains were more virulent to channel catfish than genomovar I strains (Craig A. Shoemaker, USDA-ARS, Auburn, AL, personal communication). 90 gtf, hemH, norB, and trx genes characterized in the present study have been described as virulence factors in other bacterial species (Almiron et al. 2001; Narimatsu et al. 2004; Bjur et al. 2006; Loisel-Meyer et al. 2006). Our expression data showed that hemH, norB, and trx genes are weakly or not expressed when cells reached the end of the log phase approximately 24 h growth; however, they were strongly expressed during the first few hours following inoculation into fresh Shieh broth culture medium. Gene expression was weak when cells were transferred to cell culture medium, regardless of the presence of catfish skin explants. This may be the result of the relative high salinity in this medium. During the catfish skin explants experiment, bacterial cells were cultivated in cell culture medium 199 diluted 1:1 with bacterial broth (final salinity was about 0.45%). This salinity was required to keep the skin cells alive. It has been reported that increased salinity significantly reduced growth and adhesion ability of F. columnare (Altinok and Grizzle 2001). These authors reported that the adhesion of F. columnare was reduced as salinity increased from 0.1% to 0.3%, and from 0.3% to 0.9%. The general trend observed was a reduction in expression of the four genes analyzed. This might be due to the elevated salinity levels; however, expression levels of the 16S rRNA gene seemed to be unaffected. Both norB and trx gene expression in F. columnare was not inhibited under iron- limited conditions. Holmes et al. (2005) reported that iron-limited conditions actually increased the transcript level of thioredoxin in Campylobacter jejuni. Iron-limited conditions also lead to oxidative stress in Anabaena sp.(Latifi et al. 2005). Additionally, both norB and trx can defend against oxidative and/or NO-mediated stress (Wieles et al. 1997; Jaegera et al. 2004; Philippot 2005; McGee et al. 2006). As a result, it is possible 91 that F. columnare continues to express both norB and trx in order to reduce oxidative stress due to iron deficiency or limitation. In conclusion, our data confirmed the genomic diversity of F. columnare at the single gene level. Nucleotide sequences of gtf, hemH, norB, and trx of F. columnare differed between genomovars I and II. Multiple gene expression patterns existed both between and within genomovars. Although the relation between gene expression pattern and virulence is unclear, this study is the first to address in vitro gene expression in F. columnare. Further studies are ongoing in order to confirm the difference in virulence between genomovars I and II. Nevertheless, and due to the clear genetic division between genomovars, we strongly recommend the inclusion of more than one genomovar in future F. columnare studies. 92 Table V.1. Strains of F. columnare used in the study showing genomovar (G) ascription and standard PCR amplification results of gtf (glycosyltransferase gene), hemH (ferrochelatase gene), norB (nitric oxide reductase gene), and trx (thioredoxin gene). ?+? represents positive PCR amplification of the gene, ?-? represents no amplification when DNA was used as template. Isolates Fish species Geographic origin G gtf hemH norB trx MS 467 Channel catfish Mississippi, USA I + + + + MS-02-465 Channel catfish Mississippi, USA I + + + + 27 Channel catfish Alabama, USA I + + + + IR Common carp Israel I + + + + MS 463 Channel catfish Mississippi, USA I + + + + FC-RR Channel catfish Alabama, USA I + + + + ALG-03-063 Channel catfish Alabama, USA I + + + + ALG-03-57 Channel catfish Alabama, USA I + + + + ATCC 23463 Chinook salmon Washington, USA I + + + + ATCC 49512 Brown trout France I + + + + ALG-03-069 Channel catfish Alabama, USA I + + + + GA-02-14 Rainbow trout Georgia, USA I - - + + GZ Channel catfish Alabama, USA I + - + + ARS-1 Channel catfish Alabama, USA I - - + + BM Channel catfish Alabama, USA I - - + + HS Channel catfish Alabama, USA I - - + + MO-02-23 Largemouth bass Missouri, USA II + + + + ALG-00-530 Channel catfish Alabama, USA II + + + + LSU Channel catfish Louisiana, USA II + + + + ALG-00-513 Channel catfish Alabama, USA II + + + + PT-14-151 Channel catfish Mississippi, USA II + + + + ALG-02-36 Largemouth bass Alabama, USA II + + + + ALG-00-515 Channel catfish Alabama, USA II + + + + ALG-00-527 Channel catfish Alabama, USA II + + + + MS 475 Channel catfish Mississippi, USA II + + + + ALG-00-522 Channel catfish Alabama, USA II + + + + ALG-00-521 Channel catfish Alabama, USA II + + + + BZ-1-02 Nile tilapia Brazil II + - + + BZ-4-02 Nile tilapia Brazil II + - + + BZ-5-02 Nile tilapia Brazil II + - + + 93 Table V.2. Primers for gene detection and gene expression of F. columnare. Genes Primers gtf Glytrans-F: 5?-CCAACATTTGGAGGTAGCGG-3? Glytrans-R: 5?-CCACCATCATCAGTTTAGCAGGT-3? hemH Ferrochelatase-F: 5?-AAGGAATCATATTGAATACGCGACC-3? Ferrochelatase-R: 5?-GCCAAAGCTTCCTTTATTGATGTTC-3? norB Nitric-F: 5?-TGCACGGACATCTAGCATTTTG-3? Nitric-R: 5?-CCTCATCGTTCGTTTTACCATGTTT-3? trx Thioredoxin-F: 5?-TGGCATTAGCAATTACAGATGCTAC-3? Thioredoxin-R: 5?-CCCACTACTTCCCCATTTTGAA-3? 94 Table V.3. RT-PCR results from nine strains of F. columnare. G represents genomovar ascription. ???represents no expression, ?w? represents weak expression, ?+? represents expression, ?++? represents strong expression. NA, no amplified product was obtained when genomic DNA was used as template. G strains gtf hemH norB trx 16SrRNA I GA-02-14 NA NA - - +++ I ARS-1 NA NA - w +++ I BM NA NA - + +++ I HS NA NA - - +++ I FC-RR - - + - +++ II BZ-1-02 - NA - - +++ II LSU - - - w +++ II ALG-00-530 - - ++ ++ +++ II ALG-00-527 - - - - +++ 95 Table V.4. F. columnare ALG-00-530 gene expression under standard conditions, iron- limited conditions, and in the presence of catfish skin explants. C 0 is the baseline control for catfish skin explant experiment. ?-? represents no expression, relative intensity of the expressed genes is expressed by the number of ?+? symbols. Time (h) hemH norB trx 16S rRNA gene 0 + ++++ ++++ +++ 2 ++ ++++ ++++ +++ 4 - +++ + +++ 6 - +++ ++ ++ 8 - + - +++ In Shieh broth (normal condition) 24 - +++ +++ +++ 0 - + + +++ 2 + ++++ +++ +++ 4 - +++ ++ ++ 6 - +++ + ++ 8 - +++ +++ +++ In the presence of 2,2-dipirydyl (iron-limited condition) 24 - - ++++ +++ 0 - ++++ + +++ 2 - ++++ ++ +++ 4 - +++ ++ +++ 6 - - - ++ 8 - + + +++ In the presence of transferrin (iron-limited condition) 24 - + + +++ C 0 0 - + + +++ 0.5 - ++ + +++ 1 - ++ + +++ 3 - +++ - +++ 6 - + - +++ 10 - + + +++ In the presence of fish skin explants 24 - - - +++ 0.5 - - - ++ 1 - +++ - +++ 3 - ++ - +++ 6 - + + +++ 10 - + + +++ Controls for the fish skin explants 24 - - + +++ 96 Figure V.1. Nucleotide sequence alignment of gtf, norB, and trx of F. columnare from genomovars I and II (GenBank assession numbers in brackets). 530 stands for ALG-00- 530 (genomovar II) and 49512 is the abbreviation of ATCC 49512 (genomovar I). Nucleotides that are different between genomovars are indicated by shaded/unshaded box. * indicates putative start condon. ?.? indicates missing nucleotides. Y=C+T, R=A+G, and W=A+T gtf * 530 (DQ911521) ATGAAAATAG CAATAGTTTG TTATCCAACA TTTGGAGGTA GCGGTGTAGT 50 49512(EF547540) .......... .......... ....CCAACA TTTGGAGGTA GCGGTGTAGT 50 530 (DQ911521) AGCAACAGAA TTAGGTTTAG AGTTGGCTCG CCGAGGTCAT GAAATTCATT 100 49512(EF547540) AGCAACAGAA TTAGGTTTAG AGCTGGCTCG TCGAGGTCAT GAAATTCATT 100 530 (DQ911521) TTATTACCTA TCGTCAACCT GTGCGTTTGG CACTTTTGAA TCATAATGTA 150 49512(EF547540) TTATTACCTA TCGTCAACCC GTGCGTTTAG CTCTTTTGAG TCACAATGTA 150 530 (DQ911521) CATTATCATG AAGTAAATGT GCCTGAATAT CCATTATTTC ATTACCAACC 200 49512(EF547540) CATTATCATG AAGTAAACGT TCCTGAATAT CCATTATTTC ATTACCAACC 200 530 (DQ911521) CTATGAATTA GCTCTTTCGA GTAAACTAGT AGATATGGTA AAACTTTATA 250 49512(EF547540) CTATGAATTA GCTCTTTCGA GTAAATTAGT AGATATGGTA AAACTCTATA 250 530 (DQ911521) AAATAGATAT ATTACATGTA CATTATGCTA TACCTCATGC TTATGCGGGT 300 49512(EF547540) AAATAGATGT ATTGCATGTG CATTATGCCA TACCTCATGC TTATGCAGGT 300 530 (DQ911521) TATATGGCCA AACAGATCTT AAAAGAAGAA GGAATTAATC TGCCAATGGT 350 49512(EF547540) TATATGGCTA AACAAATTTT AAAAGAAGAG GGAATTAATC TACCAATGGT 350 530 (DQ911521) TACCACATTG CACGGTACAG ATATAACCTT AGTAGGAAAC CATCCTTACT 400 49512(EF547540) TACTACATTG CATGGTACAG ATATAACGTT AGTTGGAAAT CATCCTTATT 400 530 (DQ911521) ATAAACCAGC TGTTACTTTT AGTATCAATA AGTCAGATGT AGTAACATCT 450 49512(EF547540) ATAAACCAGC AGTTACTTTT AGTATCAATA AGTCAGATGT AGTAACATCT 450 530 (DQ911521) GTCTCAAAAA GCTTAAAAGA AGATACGTTG AAATATTTTG ATGTAACAAA 500 49512(EF547540) GTTTCAAAAA GCTTAAAAGA AGATACACTA AAATATTTTG ATGTAACTAA 500 530 (DQ911521) AGAAATAAAA GTAGTCCCTA ATTTTATAGA AATWGAAGAA AYAGAAGCTC 550 49512(EF547540) GGAAATAAAA GTAGTTCCTA ATTTTATAGA AATAGAAGAA ACAGAGGCTC 550 530 (DQ911521) AGAATCTAAT TTGTAAGAGG TCTGTAATGG CTACACCAGA AGAGAAAATT 600 49512(EF547540) ACAATCTAAT TTGTAAAAGG TCTGTAATGG CTACGCCAGA AGAGAAAATA 600 530 (DQ911521) ATTACCCATA TTAGTAATTT TAGAAAAGTT AAAAAAATAC CTGATGTTAT 650 49512(EF547540) ATTACCCATA TTAGTAATTT TAGAAAGGTT AAAAAGATAC CAGATGTAGT 650 530 (DQ911521) TAAAGTTTTT TATAAAATAC AAGAAAAAAT ACCTGCTAAA CTGATGATGG 700 49512(EF547540) TAAAGTCTTT TTTAAGATAC AAGAAAAAAT ACCTGCTAAA CTGATGATGG 700 97 530 (DQ911521) TGGGTGATGG ACCTGAAAAA GAAAAGGCAG AATGGCTTTG TAGAGAATTA 750 49512(EF547540) TGG....... .......... .......... .......... .......... 750 530 (DQ911521) GGTATTGAAG AAAAAGTTAT TTTTTTTGGG AACCAGTCAT GA 792 49512(EF547540) .......... .......... .......... .......... .. 792 norB * 530 (EQ911522) ATGCACGGAC ATCTAGCATT TTGGGGAGCA TACGCCATGA TTGTATTAGC 50 49512(EF547542) .TGCACGGAC ATCTAGCATT TTGGGGAGCA TACGCCATGA TTGTATTAGC 50 530 (EQ911522) AATTATTAGT TATGCTATAC CTAATCTTAC AGGAAGAAAA AGATATGATT 100 49512(EF547542) AATTATTAGT TATGCAATAC CTAATCTTAC AGGAAGAAAA AGATACGATT 100 530 (EQ911522) CAGTTACAGG ACGTATGGCA TTYTGGTTAT CAAATATTGG TATGTTAGGT 150 49512(EF547542) CAGTTACAGG ACGTATGGCC TTTTGGTTAT CAAATATTGG TATGTTAGGT 150 530 (EQ911522) ATGACCACCG CTTTTGGGGT AGCTGGAGTA GCTCAAGTAT ATCTTGAAAG 200 49512(EF547542) ATGACCACCG CTTTTGGGGT AGCTGGAGTC GCTCAAGTAT ACCTTGAAAG 200 530 (EQ911522) AAAGTTTAAA ATGGAGTTCA TGACCGTTCA GAATGAAATA GCTATTCACT 250 49512(EF547542) AAAGTTTAAA ATGGAGTTTA TGACCGTTCA AARTGAAATA GCTATTCATT 250 530 (EQ911522) TTGTAGTATT ACTACTGTGT GCCACATTAT TTACTCTAGG AATYTCTTTA 300 49512(EF547542) TTGTAGTACT GCTACTATGT GCCACATTAT TCACTATAGG AATTTCTTTA 300 530 (EQ911522) TACATATACG ATTTTATTAA ACATGGTAAA ACGAACGATG AGGCCATAAT 350 49512(EF547542) TACATATACG ATTTTATTAA ACATGGTAAA ACGAACGATG AGG....... 350 530 (EQ911522) TGGCTAA 357 49512(EF547542) ....... 357 trx * 530 (DQ911523) ATGGCATTAG CAATTACAGA TGCTACATTT GATGAGGTAG TATTACAATC 50 49512(EF547543) .TGGCATTAG CAATTACAGA TGCTACATTT GATGAGGTAG TATTACAATC 50 530 (DQ911523) AGACAAACCT GTTTTAGTGG ATTTTTGGGC GGCTTGGTGT GGACCTTGTC 100 49512(EF547543) AGACAAACCA GTTTTAGTGG ATTTTTGGGC AGCTTGGTGC GGGCCTTGTC 100 530 (DQ911523) GTATGGTAGG TCCCGTTATT GATGAAATTG CAACAGAATA TGAAGGAAAA 150 49512(EF547543) GTATGGTAGG TCCCGTTATT GACGAAATTG CAACGGAATA TGAAGGAAAA 150 530 (DQ911523) GCAGTAATAG GTAAGGTAGA TGTCGATGCC AATCAAGAAT TTGCTGCAAA 200 49512(EF547543) GCAGTAATTG GTAAGGTAGA TGTCGATGCC AATCAAGAGT TTGCTGCAAA 200 530 (DQ911523) GTACGGTGTT CGCAATATTC CTACTGTTTT GATTTTTCAA AATGGGGAAG 250 49512(EF547543) GTATGGCGTA CGTAATATTC CTACTGTTTT GATTTTTCAA AATGGGGAAG 250 530 (DQ911523) TAGTGGGACG TCAGGTAGGT GTAGCTCCTA AAGATACTTA TGCTAAAGCA 300 49512(EF547543) TAGTG..... .......... .......... .......... .......... 300 530 (DQ911523) ATTGATGCCT TATTGTAA 318 49512(EF547543) .......... ........ 318 98 Figure V. 2. hemH, norB, and trx gene expression of F. columnare strain ALG-00-530 by reverse transcription PCR (RT-PCR). Expression patterns are shown under iron-limited conditions in the presence of 2, 2-dipyridyl and in Shieh broth at 0 h, 2 h, 4 h, 6 h, 8 h, and 24 h. 16S rRNA gene (16S) was used as an internal control. 16S hemH trx norB 0 2 4 6 8 24 0 2 4 6 8 24 Shieh broth Iron-limited Time (h) 99 VI. CONCLUSIONS This dissertation focused on molecular characterization of the two main fish pathogens, E. ictaluri and F. columnare, affecting Alabama aquaculture. The main outcome of my work is the new genetic information that has been identified, characterized and made available to the scientific community. My first objective was to increase the number of known ribosomal sequences from these two pathogens in order to identify new signature sequences for a PCR-based multiplex protocol. Although this objective could not be fulfilled due to technical and practical reasons (see Chapter II), I was able to identify a new IVS in the 23S rRNA gene of E. ictaluri. This is the first report (Chapter III) of this type of sequence in the genus Edwardsiella. In relation to F. columnare, an intraspecies characterization allowed me to analyze the avirulent mutant currently used as a vaccine against columnaris disease. Not only can this mutant now be tracked using fingerprinting methods, but I correlated its lack of virulence to changes in the LPS structure (Chapter IV). In addition, by analyzing sequence data from a shotgun library of F. columnare, I identified and characterized 4 putative virulence genes in this species. I showed there is a deep division between genomovars I and II at the genomic level. These results strongly encourage the use of both genomovars in future F. columnare studies. In brief, some of the main results from my dissertation can be summarized as follows: 100 1) An intervening sequence (IVS) is present in the 23S rRNA gene of E. ictaluri but not in E. tarda 2) E. ictaluri IVS was present in the majority of the rrl operons 3) E. ictaluri IVS shared 97% nucleotide identity with S. typhimurium helix-45 IVS 4) E. ictaluri is closely related to E. tarda based on 23S rDNA analysis; however, both species are phylogenetically distant from the ?core? members within the Enterobacteriaceae family 5) The modified F. columnare live vaccine FC-RR lacks the higher molecular weight bands in the lipopolysaccharide (LPS) 6) gtf, hemH, norB, and trx gene sequences divide F. columnare strains into two groups that correspond with genomovar ascription 7) Expression profiles of genes gtf, hemH, norB, and trx varied among different F. columnare strains under identical culture conditions Contribution The work described in this dissertation opens several venues for future studies. New phenotypic and genotypic data were generated for E. ictaluri and F. columnare. Effectively utilizing the data presented in the dissertation may be relevant to E. ictaluri IVS evolution, LPS structure definition in F. columnare, iron acquisition system in F. columnare, expression of putative virulence factors in mutant strains of F. columnare, new virulence genes identification, or evidence collection for defining subspecies in F. columnare. ? E. ictaluri habors an IVS in 23S rRNA gene and such IVS exhibits 97% similarity to Salmonella typhimurium. As is known, S. typhimurium can infect many 101 mammalian species. Did this IVS get transferred between Salmonella to E. ictaluri? Why does E. tarda not harbor this IVS, although E. tarda and E. ictaluri belong to the same genus, and both can infect fish and display similar/identical sequences in ribosomal gene? The discovery of IVS raises questions about Edwardsiella spp. evolution. ? Lipopolysaccharide has been well characterized as a virulence factor in a variety of bacterial pathogens. In this work, LPS profiles are different between virulent and avirulent strains of F. columnare. Additionally, F. columnare does not display the typical ladder-like pattern of LPS. Further work is needed to show where the difference is located in the genome and the importance of LPS in F. columnare pathogenesis. A functional study of LPS will also improve our understanding of F. columnare pathogenicity. ? An iron acquisition system has not been identified in F. columnare. The results of the dissertation suggested that ferrochelatase, an important iron importing enzyme, was not involved in iron absorption in F. columnare. Since iron is such an important element in bacterial growth, survival, and virulence, it is logical that F. columnare expresses uptake systems that mediate iron acquisition from the environment and from a catfish. ? Putative virulence gene expression in the mutant strain of FC-RR could contribute to a future research project. FC-RR is a modified live vaccine against F. columnare. In the current study, we showed that this strain is phenotypically different from the other virulent strains; however, we did not include this strain in the putative virulence gene expression study. It may be interesting to characterize 102 gene expression patterns in FC-RR in response to iron-limitation and presence of fish skin. ? Whole genome sequencing of F. columnare is ongoing at Mississippi State University. However, the strain selected for this project is a genomovar I isolate. The dissertation discovered that F. columnare genomovar I and II differed in gene sequences. It seems necessary to sequence strains from both genomovars. Comparison of whole genomes between the two genomovars should provide some answers for columnaris pathogenesis and guidelines for vaccine development. ? Since a shotgun genomic library of F. columnare is already available, further data mining could lead to the identification of new virulence genes. ? The data presented in the dissertation showed that gtf, hemH, norB, and trx sequences divided F. columnare strains into two populations which correspond to genomovars. Based on my data and other information generated by our group, we see a deep division within the species at the genetic level that is also correlated to host-species-specificity and virulence properties. These findings point towards a division of F. columnare into two subspecies. We need more data to support this hypothesis including DNA:DNA hybridization studies. Publications Zhang, Y., and Covadonga R. Arias. 2007. Identification and characterization of an intervening sequence within the 23S ribosomal RNA genes of Edwardsiella ictaluri. Systematic and Applied Microbiology 30:93-101. Zhang, Y., Covadonga R. Arias, Craig A. Shoemaker and Phillip H. Klesius. 2006. Comparison of lipopolysaccharide and protein profiles between 103 Flavobacterium columnare strains from different genomovars. Journal of Fish Diseases 29:657-663. Zhang, Y., Craig A. Shoemaker, Phillip H. Klesius, and Covadonga R. Arias. Identification and characterization of gtf, norB, and trx gene in Flavobacterium columnare. Journal of Fish Diseases (under review). 104 CUMULATIVE BIBLIOGRAPHY Abbas, A. K., and A. H. Lichtman 2000. Cellular and molecular immunology. Saunders, Philadelphia, PA. Ahmad, S., W. G. Weisburg, and M. A. Jensen. 1990. Evolution of aromatic amino acid biosynthesis and application to the fine-tune phylogenetic positioning of enteric bacteria. Journal of Bacteriology 172:1051-1061. Ainsworth, A. J., G. Capley, P. Waterstreet, and D. Munson. 1986. Use of monoclonal antibodies in the indirect fluorescent antibody technique for the diagnosis of Edwardsiella ictaluri. Journal of Fish Diseases 9:439-444. Almiron, M., M. Martinez, N. Sanjuan, and R. A. Ugalde. 2001. Ferrochelatase is present in Brucella abortus and is critical for its intracellular survival and virulence. Infection and Immunity 69:6225-6230. Altinok, I., and J. M. Grizzle. 2001. Effects of low salinities on Flavobacterium columnare infection of euryhaline and freshwater stenohaline fish. Journal of Fish Diseases 24:361-367. Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. Journal of Molecular Biology 215:403-410. 105 Amaro, C., B. Fouz, E. Biosca, E. Marco-Noales, and R. Collado. 1997. The lipopolysaccharide O side chain of Vibrio vulnificus serogroup E is a virulence determinant for eels. Infection and Immunity 65:2475-2479. Anacker, R. L., and E. J. Ordal. 1959. Studies on the myxobacterium Chondrococcus columnaris I. serological typing. Journal of Bacteriology 78:25-32. Aoki, T., T. Sakaguchi, and T. Kitao. 1987. Multiple drug-resistant plasmids from Edwardsiella tarda in eel culture ponds. Nippon Suisan Gakkaish 53:1821-1825. Aoki, T., and I. Hirono. 1995. Detection of the fish-pathogenic bacteria Edwardsiella tarda by polymerase chain reaction. Proceedings of the International Symposium on Biotechnology Applications in Aquaculture, Asian Fisheries Society Special Publication No.10:135-146. Arias, C. R., E. Garay, and R. Aznar. 1995. Nested PCR method for rapid and sensitive detection of Vibrio vulnificus in fish, sediments, and water. Applied and Environmental Microbiology 61:3476-3478. Arias, C. R., L. Verdonck, J. Swings, E. Garay, and R. Aznar. 1997a. Intraspecific differentiation of Vibrio vulnificus biotypes by amplified fragment length polymorphism and ribotyping. Applied and Environmental Microbiology 63:2600-2606. 106 Arias, C. R., L. Verdonck, J. Swings, R. Aznar, and E. Garay. 1997b. A polyphasic approach to study the intraspecific diversity amongst Vibrio vulnificus isolates. Systematic and Applied Microbiology 20:622-633. Arias, C. R., C. A. Shoemaker, J. J. Evans, and P. H. Klesius. 2003. A comparative study of Edwardsiella ictaluri parent (EILO) and E. ictaluri rifampicin-mutant (RE-33) isolates using lipopolysaccharides, outer membrane proteins, fatty acids, Biolog, API 20E and genomic analyses. Journal of Fish Diseases 26:415-21. Arias, C. R., T. L. Welker, C. A. Shoemaker, J. W. Abernathy, and P. H. Klesius. 2004. Genetic fingerprinting of Flavobacterium columnare isolates from cultured fish. Journal of Applied Microbiology 97:421-428. Austin, B., and D. A. Austin. 1999. Bacterial fish pathogens: disease of farmed and wild fish. Springer, New York. Bader, J. A., and E. J. Shotts. 1998. Identification of Flavobacterium and Flexibacter species by species-specific polymerase chain reaction primers to the 16S ribosomal RNA gene. Journal of Aquatic Animal Health 10:311-319. Bader, J. A., K. E. Nusbaum, and C. A. Shoemaker. 2003a. Comparative challenge model of Flavobacterium columnare using abraded and unabraded channel catfish, Ictalurus punctatus (Rafinesque). Journal of Fish Diseases 26:461-7. 107 Bader, J. A., C. A. Shoemaker, and P. H. Klesius. 2003b. Rapid detection of columnaris disease in channel catfish (Ictalurus punctatus) with a new species-specific 16S rRNA gene-based PCR primer for Flavobacterium columnare. Journal of Microbiological Methods 52:209-220. Bader, J. A., C. A. Shoemaker, and P. H. Klesius. 2005. Production, characterization and evaluation of virulence of an adhesion defective mutant of Flavobacterium columnare produced by beta-lactam selection. Letters in Applied Microbiology 40:123-127. Baird, K. D., H. M. Chikarmane, R. Smolowitz, and K. R.Uhlinger. 2003. Detection of Edwardsiella infections in Opsanus tau by polymerase chain reaction. Biological Bulletin 205:235-236. Bernardet, J. F., and P. A. D. Grimont. 1989. Deoxyribonucleic acid relatedness and phenotypic characterization of Flexibacter columnaris sp. nov., nom. rev., Flexibacter psychrophilus sp. nov., nom. rev., and Flexibacter maritimus Wakabayashi, Hikida, and Masumura 1986. International Journal of Systematic Bacteriology 39:346-354. 108 Bernardet, J. F., P. Segers, M. Vancanneyt, F. Berthe, K. Kersters, and P. Vandamme. 1996. Cutting a gordian knot: Emended classification and description of the genus Flavobacterium, emended description of the family Flavobacteriaceae, and proposal of Flavobacterium hydatis nom. nov. (Basonym, Cytophaga aquatilis Strohl and Tait 1978). International Journal of Systematic Bacteriology 46:128- 148. Bertolini, J. M., and J. S. Rohovec. 1992. Electrophoretic detection of proteases from different Flexibacter columnaris strains and assessment of their variability. Diseases of Aquatic Organisms 12:121-128. Bilodeau, A. L., G. C. Waldbieser, J. S. Terhune, D. J. Wise, and W. R. Wolters. 2003. A real-time polymerase chain reaction assay of the bacterium Edwardsiella ictaluri in channel catfish. Journal of Aquatic Animal Health 15:80-86. Biosca, I. G., H. Llorens, E. Garay, and C. Amaro. 1993. Presence of a capsule in Vibrio vulnificus biotype 2 and its relationship to virulence for eels. Infection and Immunity 61:1611-1618. Bjur, E., S. Eriksson-Ygberg, F. Aslund, and M. Rhen. 2006. Thioredoxin 1 promotes intracellular replication and virulence of Salmonella enterica serovar typhimurium. Infection and Immunity 74:5140-5151. Booth, N. J. 2005. The role of urease in the pathogenesis of Edwardsiella ictaluri. Doctoral dissertation. Louisiana State University, Baton Rounge, Louisiana, USA. 109 Bullock, G. L., T. C. Hsu, and E. B. Shotts. 1986. Columnaris disease of fishes. Fish disease leaflet 72. U. S. Fish and Wildlife Service, Washington, D. C. Burgin, A. B., K. Parodos, D. J. Lane, and N. R. Pace. 1990. The excision of intervening sequences from Salmonella 23S rRNA. Cell 60:405-414. Chen, C. R., H. Y. Chung, and G. H. Kuo. 1982. Studies on the pathogenicity of Flexibacter columnaris I. effect of dissolved oxygen and ammonia on the pathogenicity of Flexibacter columnaris to eel (Anguilla japonica). CAPD fisheries serials no. 8 Reports on Fish Disease Research 4:57-61. Chen, J. D., and S. Y. Lai. 1998. PCR for direct detection of Edwardsiella tarda from infected fish and environmental water by application of the hemolysin gene. Zoological Studies 37:169-176. Chen, Y. C., M. C. Chang, Y. C. Chuang, and C. L. Jeang. 2004. Characterization and virulence of hemolysin III from Vibrio vulnificus. Current Microbiology 49:175-9. Chenna, R., H. Sugawara, T. Koike, R. Lopez, T. J. Gibson, D. G. Higgins, and J. D. Thompson. 2003. Multiple sequence alignment with the clustal series of programs. Nucleic Acids Research 31:3497-3500. Chowdhury, B. R., and H. Wakabayashi. 1990. Detection of Flexibacter columnaris from fish and water using indirect fluorescent antibody technique. Bangladesh Journal of Microbiology 7:25-35. Crumlish, M., T. T. Dung, J. F. Turnbull, N. T. N. Ngoc, and H. W. Ferguson. 2002. Identification of Edwardsiella ictaluri from diseased freshwater catfish, 110 Pangasius hypophthalmus (Sauvage), cultured in the Mekong Delta, Vietnam. Journal of Fish Diseases 25:733-736. Darwish, A. M., A. A. Ismaiel, J. C. Newton, and J. Tang. 2004. Identification of Flavobacterium columnare by a species-specific polymerase chain reaction and renaming of ATCC43622 strain to Flavobacterium johnsoniae. Molecular and Cellular Probes 18:421-427. Darwish, A. M., and A. A. Ismaiel. 2005. Genetic diversity of Flavobacterium columnare examined by restriction fragment length polymorphism RNA gene and sequencing of the 16S ribosomal RNA gene and the 16S-23S rDNA spacer. Molecular and Cellular Probes 19:267-274. Davis, H. S. 1922. A new bacterial disease of fresh-water fishes. Bulletin of the U. S. Bureau of Fisheries 38:261-280. Decostere, A., F. Haesebrouck, and L. A. Devriese. 1997. Shieh medium supplemented with tobramycin for selective isolation of Flavobacterium columnare (Flexibacter columnaris) from diseased fish. Journal of Clinical Microbiology 35:322-324. Decostere, A., F. Haesebrouck, G. Charlier, and R. Ducatelle. 1999a. The association of Flavobacterium columnare strains of high and low virulence with gill tissue of black mollies (Poecilia sphenops). Veterinary Microbiology 67:287-298. Decostere, A., F. Haesebrouck, J. Turnbull, and G. Charlier. 1999b. Influence of water quality and temperature on adhesion of high and low virulence Flavobacterium columnare strains to isolated gill arches. Journal of Fish Diseases 22:1-11. 111 Dietmaier, W., and F. Hofstadter. 2001. Detection of microsatellite instability by real time PCR and hybridization probe melting point analysis. Laboratory Investigation 81:1453-1456. Evguenieva-Hackenberg, E., and G. Klug. 2000. RNase III processing of intervening sequences found in helix 9 of 23S rRNA in the alpha subclass of Proteobacteria. Journal of Bacteriology 182:4719-4729. Ewing, W. H., A. C. McWhorter, M. R. Escobar, and A. H. Lubin. 1965. Edwardsiella, a new genus of Enterobacteriaceae based on a new species, E. tarda. International Bulletin of Bacteriological Nomenclature and Taxonomy 15:33-38. Francino, M. P., S. R. Santos, and H. Ochman. 2003. Phylogenetic relationships of bacteria with special reference to endosymbionts and enteric species. In M. Dworkin, editor. Prokaryotes: A handbook on the biology of Bacteria: Ecophysiology, isolation, identification, applications, 3rd edition. Springer- Verlag, New York. Garnjobst, L. 1945. Cytophaga columnaris (David) in pure culture: a myxobacterium pathogenic to fish. Journal of Bacteriology 49:113-128. Griffin, B. R. 1991. Characteristics of a chondroitin AC lyase produced by Cytophaga columnaris. Transactions of the American Fisheries Society 120:391-395. Griffin, B. R. 1992. A simple procedure for identification of Cytophaga columnaris. Journal of Aquatic Animal Health 4:63-66. 112 Grimont, P. A. D., F. Grimont, C. Richard, and R. Sakazaki. 1980. Edwardsiella hoshinae, a new species of Enterobacteriaceae. Current Microbiology 4:347-351. Hanson, L. A., and J. M. Grizzle. 1985. Nitrite induced predisposition of channel catfish Ictalurus punctatus to bacterial diseases. Progressive Fish Culturist 47:98-101. Hanson, L. A., and W. A. Rogers. 1989. Enzyme immunoassay identification of Edwardsiella ictaluri in decomposing channel catfish. Journal of the World Aquaculture Society 20:279-280. Hawke, J. P. 1979. A bacterium associated with disease of pond cultured channel catfish, Ictalurus punctatus. Journal of the Fisheries Research Board of Canada 36:1508- 1512. Hawke, J. P., A. C. McWhorter, A. G. Steigerwalt, and D. J. Brenner.1981. Edwardsiella ictaluri sp. nov., the causative agent of enteric septicemia of catfish. International Journal of Systematic Bacteriology 31:396-400. Hawke, J. P., and R. L. Thune. 1992. Systemic isolation and antimicrobial susceptibility of Cytophaga columnaris from commercially reared channel catfish. Journal of Aquatic Animal Health 4:109-113. Hawke, J. P., R. M. Durborow, R. L. Thune, and A. C. Camus. 1998. ESC-enteric septicemia of catfish. Publication No. 477, Southern Regional Aquaculture Center, Stoneville, Mississippi. 113 Hawke, J. P. 2003. Edwardsiella tarda septicemia. In AFS-FHS (American Fisheries Society-Fish Health Section). FHS blue book: suggested procedures for the detection and identification of certain finfish and shellfish pathogens 2006 edition. AFS-FHS, Bethesda, Maryland. Heid, C. A., J. Stevens, K. J. Livak, and P. M. Williams. 1996. Real time quantitative PCR. Genome Research 6:986-994. Hiney, M. P., and P. R. Smith. 1998. Validation of polymerase chain reaction-based techniques for proxy detection of bacterial fish pathogens: framework, problems and possible solutions for environmental applications. Aquaculture 162:41-68. Hirono, I., N. Tange, and T. Aoki. 1997. Iron-regulated hemolysin gene from Edwardsiella tarda. Molecular Microbiology 24:851-856. Holmes, K., F. Mulholland, B. M. Pearson, C. Pin, J. McNicholl-Kennedy, J. M. Ketley, and J. M. Wells. 2005. Campylobacter jejuni gene expression in response to iron limitation and the role of Fur. Microbiology 151:243-257. Hoshina, T. 1962. On a new bacterium, Paracolobactrum anguillimortiferum sp. Bulletin of the Japanese Society of Scientific Fisheries 28:162-164. Jaegera, T., H. Buddea, L. Flohe, U. Mengeb, M. Singhb, M. Trujilloc, and R. Radic. 2004. Multiple thioredoxin-mediated routes to detoxify hydroperoxides in Mycobacterium tuberculosis. Archives of Biochemistry and Biophysics 423:182- 191. Janda, J. M., S. L. Abbott, S. Kroske-Bystrom, W. K. Cheung, C. Powers, R. P. Kokka, and K. Tamura. 1991. Pathogenic properties of Edwardsiella species. Journal of Clinical Microbiology 29:1997-2001. 114 Jin, D. J., and C. A. Gross. 1988. Mapping and sequencing of mutations in the Escherichia coli rpoB gene that lead to rifampicin resistance. Journal of Molecular Biology 202:45-58. Jukes, T. H., and C. R. Cantor. 1969. Evolution of protein molecules. Pages 21-132 in H. N. Munro, editor. Mamalian protein metabolism. Academic Press, New York. Kasornchandra, J., W. A. Rogers, and J. A. Plumb. 1987. Edwardsiella ictaluri from walking catfish, Clarias batrachus L., in Thailand. Journal of Fish Diseases 10:137-138. Kent, M. L., and J. M. Lyons. 1982. Edwardsiella ictaluri in the green knife fish, Eigemannia (sic) virescens. Fish Health News: a service to the field of fish health research 11:ii. Keskin, O., S. Secer, M. Izgur, S. Turkyilmaz, and R. Mkakosya. 2004. Edwardsiella ictaluri infection in rainbow trout (Oncorhynchus mykiss). Turkish Journal of Veterinary and Animal Sciences 28:649-653. Kimura, A., and E. J. Hansen. 1986. Antigenic and phenotypic variations of Haemophilus influenzae type b lipopolysaccharide and their relationship to virulence. Infection and Immunity 51:69-79. Kirschbaum, T. M., and R. F. Gotte. 1993. Rifampicin resistance of Mu lysogenic strains by rpoB mutations. Biological Chemistry Hoppe-Seyler 374:657-64. Klesius, P. H. 1990. Effect of size and temperature on the quantity of immunoglobulin in channel catfish, Ictalurus punctatus. Veterinary Immunology and Immunopathology 24:187-95. 115 Klesius, P. H., K. Johnson, R. Durborow, and S. Vinitnantharat. 1991. Development and evaluation of an ELISA for catfish serum antibody to Edwardsiella ictaluri. Journal of Aquatic Animal Health 3:94-99. Klesius, P. H., and C. A. Shoemaker. 1997. Heterologous isolates challenge of channel catfish, Ictalurus punctatus, immune to Edwardsiella ictaluri. Aquaculture 157:147-155. Klesius, P. H., and C. A. Shoemaker. 1999. Development and use of modified live Edwardsiella ictaluri vaccine against enteric septicemia of catfish. Advances in Veterinary Medicine 41:523-37. Klesius, P. H., J. Lovy, J. J. Evans, E. Washuta, and C. R. Arias. 2003. Isolation of Edwardsiella ictaluri from tadpole madtom, Noturus gyrinus in a southwestern New Jersey River. Journal of Aquatic Animal Health 15:295-301. Krieg, N. R., and J. G. Holt. 1984. Bergey's Manual of Systematic Bacteriology. Williams and Wilkins, Baltimore, Maryland. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685. Latifi, A., R. Jeanjean, S. Lemeille, M. Havaux, and C. C. Zhang. 2005. Iron starvation leads to oxidative stress in Anabaena sp. strain PCC 7120. Journal of Bacteriology 187:6596?6598. Lawrence, M. L., and M. M. Banes. 2005. Tissue persistence and vaccine efficacy of an O polysaccharide mutant strain of Edwardsiella ictaluri. Journal of Aquatic Animal Health 17:228-232. 116 Leadbetter, E. R. 1974. Genus II. Flexibacter Soriano 1945, 92, Lewin 1969, 192 emend. mut. char. Pages 105-107 in R. E. Buchanan and N. E. Gibbons, editors. Bergey's manual of determinative bacteriology, 8th edition. Williams and Wilkins, Baltimore, Maryland. Lee, C. C., S. P. Craig, and A. E. Eakin. 1998. A single amino acid substitution in the human and a bacterial hypoxanthine phosphoribosyltransferase modulates specificity for the binding of guanine. Biochemistry 37:3491-3498. Loeb, M. 1995. Ferrochelatase activity and protoporphyrin IX utilization in Haemophilus influenzae. Journal of Bacteriology 177:3613-3615. Loisel-Meyer, S., M. P. Jimenez de Bagues, E. Basseres, J. Dornand, S. Kohler, J. P. Liautard, and V. Jubier-Maurin. 2006. Requirement of norD for Brucella suis virulence in a murine model of in vitro and in vivo infection. Infection and Immunity 74:1973-1976. MacLean, L. L., E. Vinogradov, E. M. Crump, M. B. Perry, and W. W. Kay. 2001. The structure of the lipopolysaccharide O-antigen produced by Flavobacterium psychrophilum (259-93). European Journal of Biochemistry 268:2710-2716. MacLean, L. L., M. B. Perry, E. M. Crump, and W. W. Kay. 2003. Structural characterization of the lipopolysaccharide O-polysaccharide antigen produced by Flavobacterium columnare ATCC 43622. European Journal of Biochemistry 270:3440-3446. Madigan, M. T., J. M. Martinko, and J. Parker 2003. Brock biology of microorganisms. Prentice Hall, Upper Saddle River, New Jersey. 117 Makela, P., V. Valtonen, and M. Valtonen. 1973. Role of O-antigen (lipopolysaccharide) factors in the virulence of Salmonella. The Journal of Infectious Diseases 128:81- 85. Mattatall, N. R., and K. E. Sanderson. 1996. Salmonella typhimurium LT2 possesses three distinct 23S rRNA intervening sequences. Journal of Bacteriology 178:2272-2278. McCarthy, D. H. 1975. Columnaris disease. Journal of Institute Fisheries Management 6:44-47. McGee, D. J., S. Kumar, R. J. Viator, J. R. Bolland, J. Ruiz, D. Spadafora, T. L. Testerman, D. J. Kelly, L. K. Pannell, and H. J. Windle. 2006. Helicobacter pylori thioredoxin is an arginase chaperone and guardian against oxidative and nitrosative stresses. The Journal of Biological Chemistry 281:3290-3296. Medearis, D. N., Jr., B. M. Camitta, and E. C. Heath. 1968. Cell wall composition and virulence in Escherichia coli. Journal of Experimental Medicine 128:399-414. Meyer, F. P., and G. L. Bullock. 1973. Edwardsiella tarda a new pathogen of channel catfish Ictalurus punctatus. Applied Microbiology 25:155-156. Miller, W., K. Pabbaraju, and K. E. Sanderson. 2000. Fragmentation of 23S rRNA in strains of Proteus and Providencia results from intervening sequences in the rrn (rRNA) genes. Journal of Bacteriology 182:1109-1147. Miyazaki, T. 1985. Comparative histopathology of Edwardsiellosis in fishes. Fish Pathology 23:219-227. Montaraz, J. A., and A. J. Winter. 1986. Comparison of living and nonliving vaccines for Brucella abortus in BALB/c mice. Infection and Immunity 53:245-51. 118 Morrison, E. E., and J. A. Plumb. 1994. Olfactory organ of channel catfish as a site of experimental Edwardsiella ictaluri infection. Journal of Aquatic Animal Health 6:101-109. Mullis, K., and F. Faloona. 1987. Specific synthesis of DNA in vitro via a polymerase- catalyzed chain reaction. Methods in Enzymology 155:335-350. Narimatsu, M., Y. Noiri, S. Itoh, N. Noguchi, T. Kawahara, and S. Ebisu. 2004. Essential role for the gtfA gene encoding a putative glycosyltransferase in the adherence of Porphyromonas gingivalis. Infection and Immunity 72:2698-2702. Newton, J. C., L. G. Wolfe, J. M. Grizzle, and J. A. Plumb. 1989. Pathology of experimental enteric septicemia in channel catfish Ictalurus punctatus Rafinesque following immersion-exposure to Edwardsiella ictaluri. Journal of Fish Diseases 12:335-348. Newton, J. C., T. M. Wood, and M. M. Hartley. 1997. Isolation and partial characterization of extracellular proteases produced by isolates of Flavobacterium columnare derived from channel catfish. Journal of Aquatic Animal Health 9:75- Olivares-Fuster, O., C. R. Arias, C. A. Shoemaker, and P. H. Klesius 2007 Molecular typing of Flavobacterium columnare isolates by single-strand conformation polymorphism analysis (SSCP). FEMS Microbiology Letters. (in press) Ordal, E. J., and R. R. Rucker. 1944. Pathogenic myxobacteria. Proceedings of the Society for Experimental Biology and Medicine 56:15-18. Owens, D. R., S. L. Nelson, and J. B. Addison. 1974. Isolation of Edwardsiella tarda from swine. Applied and Environmental Microbiology 27:703-705. 119 Pabbaraju, K., W. Miller, and K. E. Sanderson. 2000. Distribution of intervening sequences in the genes for 23S rRNA and rRNA fragmentation among strains of the Salmonella reference collection B (SARB) and SARC sets. Journal of Bacteriology 182:1923-1929. Pacha, R. E., and E. J. Ordal. 1970. Myxobacterial diseases of salmonids. Pages 243-257 in S. F. Snieszko, editor. A Symposium on Diseases of Fishes and Shellfishes. American Fisheries Society, Washington, D. C. Panangala, V. S., V. L. van Santen, C. A. Shoemaker, and P. H. Klesius. 2005. Analysis of 16S-23S intergenic spacer region of the rRNA operons in Edwardsiella ictaluri and Edwardsiella tarda isolates from fish. Journal of Applied Microbiology 99:657-669. Panangala, V. S., R. A. Shelby, C. A. Shoemaker, P. H. Klesius, A. Mitra, and E. E. Morrison. 2006. Immunofluorescent test for simultaneous detection of Edwardsiella ictaluri and Flavobacterium columnare. Diseases of Aquatic Organisms 68:197-207. Panangala, V., C. A. Shoemaker, V. Van Santan, K. Dybvig, and P. H. Klesius. 2007. Multiplex-PCR for simultaneous detection of three fish-pathogenic bacteria, Edwardsiella Ictaluri, Flavobacterium columnare and Aeromonas hydrophila. Diseases of Aquatic Organisms 74:199-203. Philippot, L. 2005. Denitrification in pathogenic bacteria: for better or worst? Trends in Microbiology 13:191-192. Plumb, J. A., and D. J. Sanchez. 1983. Susceptibility of five species of fish to Edwardsiella ictaluri. Journal of Fish Diseases 6:261-266. 120 Plumb, J. A. 1993. Edwardsiella septicemia. Pages 60-79 in V. Inglis, R. J. Roberts, and N. R. Bromage, editors. Bacterial diseases of fish. Halsted Press, New York. Plumb, J. A. 1999. Health maintenance and principal microbial diseases of cultured fish. Iowa State University Press, Ames, Iowa. Plumb, J. A., and C. A. Shoemaker. 2006. Enteric septicemia of catfish in aquaculture compendium. CAB international, Wallingford, U. K. Pronk, L. M., and K. E. Sanderson. 2001. Intervening sequences in rrl genes and fragmentation of 23S rRNA in genera of the family Enterobacteriaceae. Journal of Bacteriology 183:5782-5787. Rajakumar, K., B. Jost, C. Sasakawa, N. Okada, M. Yoshikawa, and B. Adler. 1994. Nucleotide sequence of the rhamnose biosynthetic operon of Shigella flexneri 2a and role of lipopolysaccharide in virulence. Journal of Bacteriology 176:2362- 2373. Reichenbach, H. 1989. Genus I. Cytophaga Winogradsky 1929, 577, AL emend. Pages 2015-2046 in J. T. Staley, M. P. Bryant, N. Pfennig, and J. G. Holt, editors. Bergey's manual of systematic bacteriology. Williams and Wilkins, Baltimore, Maryland. Reid, W. S., and J. A. Boyle. 1989. Plasmid homologies in Edwardsiella ictaluri. Applied and Environmental Microbiology 55:3253-3255. Rucker, R. R., B. J. Earp, and E. J. Ordal. 1953. Infectious diseases of Pacific salmon. Transactions of the American Fisheries Society 83:297-312. Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Molecular Biology and Evolution 4:406-425. 121 Sakazaki, R., and Y. Murata. 1962. The new group of the Enterobacteriaceae, the Asakusa group. Japanese Journal of Bacteriology 17:616-620. Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, New York. Sanders, J. E., R. A. Holt, and J. L. Fryer. 1976. Serological comparison of Flexibacter columnaris isolates using rabbit and rainbow trout (Salmo gairdneri) antisera. Journal of the Fisheries Research Board of Canada 33:1386-1388. Schneck, J. L., and L. F. Caslake. 2006. Genetic diversity of Flavobacterium columnare isolated from fish collected from warm and cold water. Journal of Fish Diseases 29:245-248. Schurig, G. G., R. M. Roop, T. Bagchi, S. Boyle, D. Buhrman, and N. Sriranganathan. 1991. Biological properties of RB51; a stable rough strain of Brucella abortus. Veterinary Microbiology 28:171-188. Sechter, I., M. Shmilovitz, G. Altmann, R. Sellgmann, B. Kretzer, I. Braunstein, and B. Gericheter. 1983. Edwardsiella tarda isolated in Israel between 1961 and 1980. Journal of Clinical Microbiology 17:669-67. Sharma, V. K., Y. K. Kaura, and I. P. Singh. 1974. Frogs as carriers of Salmonella and Edwardsiella. Antonie Van Leeuwenhoek 40:171-175. Shoemaker, C. A., P. H. Klesius, and J. M. Bricker. 1999. Efficacy of a modified live Edwardsiella ictaluri vaccine in channel catfish as young as seven days post hatch. Aquaculture 176:189-193. Shoemaker, C. A., P. H. Klesius, C. Lim, and M. Yildirim. 2003a. Feed deprivation of channel catfish, Ictalurus punctatus (Rafinesque), influences organosomatic 122 indices, chemical composition and susceptibility to Flavobacterium columnare. Journal of Fish Diseases 26:553-61. Shoemaker, C. A., R. Shelby, and P. H. Klesius. 2003b. Development of an indirect ELISA to detect humoral response to Flavobacterium columnare infection of channel catfish, Ictalurus punctatus (Rafinesque). Journal of Applied Aquaculture 14:43-52. Shoemaker, C. A., C. R. Arias, P. H. Klesius, and T. L. Welker. 2005a. Technique for identifying Flavobacterium columnare using whole-cell fatty acid profiles. Journal of Aquatic Animal Health 17:267-274. Shoemaker, C. A., P. H. Klesius, and J. Evans. 2005b. Modified live Flavobacterium columnare against columnaris disease in fish. U. S. patent application 6881412 B1. 2005-04-19. Shotts, E., and W. Waltman. 1990. A medium for the selective isolation of Edwardsiella ictaluri. Journal of Wildlife Diseases 26:214-218. Song, Y. L., J. L. Fryer, and J. S. Rohovec. 1988. Comparison of gliding bacteria isolated from fish in North America and other areas of the Pacific rim. Fish Pathology 23:197-202. Soule, M., K. Cain, S. LaFrentz, and D. R. Call. 2005. Combining suppression subtractive hybridization and microarrays to map the intraspecies phylogeny of Flavobacterium psychrophilum. Infection and Immunity 73:3799-3802. Speyerer, P. D., and J. A. Boyle. 1987. The plasmid profile of Edwardsiella ictaluri. Journal of Fish Diseases 10:461-470. 123 Srinivasa-Rao, P. S., T. M. Lim, and K. Y. Leung. 2003. Functional genomics approach to the identification of virulence genes involved in Edwardsiella tarda pathogenesis. Infection and Immunity 71:1343-1351. Stanley, L. A., J. S. Hudson, T. E. Schwedler, and S. S. Hayasaka. 1994. Extracellular products associated with virulent and avirulent strains of Edwardsiella ictaluri from channel catfish. Journal of Aquatic Animal Health 6:36?43. Stringer-Roth, K. M., W. Yunghans, and L. F. Caslake. 2002. Differences in chondroitin AC lyase activity of Flavobacterium columnare isolates. Journal of Fish Diseases 25:687-691. Suomalainen, L. R., M. A. Tiirola, and E. T. Valtonen. 2005. Influence of rearing conditions on Flavobacterium columnare infection of rainbow trout, Oncorhynchus mykiss (Walbaum). Journal of Fish Diseases 28:271?277. Suomalainen, L. R., M. Tiirola, and E. T. Valtonen. 2006. Chondroitin AC lyase activity is related to virulence of fish pathogenic Flavobacterium columnare. Journal of Fish Diseases 29:757-763. Teska, J. D. 1993. Assay to evaluate the reaction kinetics of chondroitin AC lyase produced by Cytophaga columnaris. Journal of Aquatic Animal Health 5:259-264. Thomas-Jinu, S., and A. E. Goodwin. 2004. Morphological and genetic characteristics of Flavobacterium columnare isolates: correlations with virulence in fish. Journal of Fish Diseases 27:29-35. Thune, R. L., J. P. Hawke, and M. C. Johnson. 1994. Studies on vaccination of channel catfish, Ictalurus punctatus, against Edwardsiella ictaluri. Journal of Applied Aquaculture 3:11-23. 124 Toyama, T., K. Kita-Tsukamoto, and H. Wakabayashi. 1996. Identification of Flexibacter maritimus, Flavobacterium branchiophilum and Cytophaga columnaris by PCR targeted 16S ribosomal DNA. Fish Pathology 31:25-31. Triyanto, and H. Wakabayashi. 1999. Genotypic diversity of strains of Flavobacterium columnare from diseased fishes. Fish Pathology 34:65-71. Tucker, C. S. 1985. Channel catfish culture. Elsevier Science Publisher, Amsterdam, the Netherlands. Tyagi, S., and F. R. Kramer. 1996. Molecular beacons: probes that fluoresce upon hybridization. Nature Biotechnology 14:303-308. Ullah, A., and T. Arai. 1983. Pathological activities of the naturally occurring strains of Edwardsiella tarda. Fish Pathology 18:65-70. USDA (United States Department of Agriculture). 2003a. Catfish 2003 Part II: Reference of foodsize catfish health and production practices in the United States, 2003. USDA:APHIS:VS,CEAH, National Animal Health Monitoring System, Fort Collins, Colorado. USDA (United States Department of Agriculture). 2003b. Catfish 2003 Part I: Reference of fingerling catfish health and production practices in the United States, 2003. USDA:APHIS:VS, CEAH, National Animal Health Monitoring System, Fort Collins, Colorado. USDA (United States Department of Agriculture). 2005. Census of aquaculture (2005). Volume 3, USDA: National Agricultural Statistics Service (NASS). Vemulapalli, R., J. R. McQuiston, G. G. Schurig, N. Sriranganathan, S. M. Halling, and S. M. Boyle. 1999. Identification of an IS711 element interrupting the wboA gene of 125 Brucella abortus vaccine strain RB51 and a PCR assay to distinguish strain RB51 from other Brucella species and strains. Clinical and Diagnostic Laboratory Immunology 6:760-764. Vemulapalli, R., Y. He, L. Buccolo, S. Boyle, N. Sriranganathan, and G. Schurig. 2000. Complementation of Brucella abortus RB51 with a functional wboA gene results in O-antigen synthesis and enhanced vaccine efficacy but no change in rough phenotype and attenuation. Infection and Immunity 68:3927-3932. Vinogradov, E., M. B. Perry, and W. W. Kay. 2003. The structure of the glycopeptides from the fish pathogen Flavobacterium columnare. Carbohydrate Research 338:2653-2658. Wagner, B. A., D. J. Wise, L. H. Khoo, and J. S. Terhune. 2002. The epidemiology of bacterial diseases in food-size channel catfish. Journal of Aquatic Animal Health 14:263-272. Waltman, W., E. B. Shotts, and V. S. Blazer. 1985. Recovery of Edwardsiella ictaluri from danio (Danio devario). Aquaculture 46:63-66. Waltman, W., E. B. Shotts, and T. C. Hsu. 1986. Biochemical characteristics of Edwardsiella ictaluri. Applied and Environmental Microbiology 51:101-104. Waterstrat, P., J. Ainsworth, and G. Capley. 1989. Use of an indirect enzyme-linked immunosorbent assay ELISA in the detection of channel catfish Ictalurus punctatus Rafinesque antibodies to Edwardsiella ictaluri. Journal of Fish Diseases 12:87-94. 126 Welker, T. L., C. A. Shoemaker, C. R. Arias, and P. H. Klesius. 2005. Transmission and detection of Flavobacterium columnare in channel catfish Ictalurus punctatus. Diseases of Aquatic Organisms 63:129-138. Wellborn, T. 1988. Channel catfish life history and biology. Publication No. 180, Southern Regional Aquaculture Center, Stoneville, Mississippi. Westphal, O., and K. Jann. 1965. Bacterial lipopolysaccharide extraction with phenol- water and further application of the procedure. Methods in Carbohydrate Chemistry 5:83-96. Wieles, B., T. Ottenhoff, T. Steenwijk, K. Franken, R. de Vries, and J. Langermans. 1997. Increased intracellular survival of Mycobacterium smegmatis containing the Mycobacterium leprae thioredoxin-thioredoxin reductase gene. Infection and Immunity 65:2537-2541. Williams, M. L., P. Azadi, and M. L. Lawrence. 2003. Comparison of cellular and extracellular products expressed by virulent and attenuated strains of Edwardsiella ictaluri. Journal of Aquatic Animal Health 15:264-273. Williams, M. L., and M. L. Lawrence. 2005. Identification and characterization of a two- component hemolysin from Edwardsiella ictaluri. Veterinary Microbiology 108:281-289. Wolters, W. R., D. J. Wise, and P. H. Klesius. 1996. Survival and antibody response of channel catfish, blue catfish, and channel catfish female X blue catfish male hybrids after exposure to Edwardsiella ictaluri. Journal of Aquatic Animal Health 8:249-254. 127 Xie, H. X., P. Nie, and B. J. Sun. 2004. Characterization of two membrane-associated protease genes obtained from screening out-membrane protein genes of Flavobacterium columnare G4. Journal of Fish Diseases 27:719-729. Xie, H. X., P. Nie, M. X. Chang, Y. Liu, and W. J. Yao. 2005. Gene cloning and functional analysis of glycosaminoglycan-degrading enzyme chondroitin AC lyase from Flavobacterium columnare G(4). Archives of Microbiology 184:49-55. Xu, D. H., and P. H. Klesius. 2002. Antibody mediated immune response against Ichthyophthirius multifiliis using excised skin from channel catfish, Ictalurus punctatus (Rafinesque), immune to Ichthyophthirius. Journal of Fish Diseases 25:299-306. Yeh, H. Y., C. A. Shoemaker, and P. H. Klesius. 2005. Evaluation of a loop-mediated isothermal amplification method for rapid detection of channel catfish Ictalurus punctatus important bacterial pathogen Edwardsiella ictaluri. Journal of Microbiological Methods 63:36-44. Yeh, H.-Y., C. A. Shoemaker, and P. H. Klesius. 2006. Sensitive and rapid detection of Flavobacterium columnare in channel catfish Ictalurus punctatus by a loop- mediated isothermal amplification method. Journal of Applied Microbiology 100:919-925. Yin, J. L., N. A. Shackel, A. Zekry, P. H. McGuinness, C. Richards, K. V. Putten, G. W. McCaughan, J. M. Eris, and G. A. Bishop. 2001. Real-time reverse transcriptase- polymerase chain reaction (RT-PCR) for measurement of cytokine and growth factor mRNA expression with fluorogenic probes or SYBR Green I. Immunology and Cell Biology 79:213-21. 128 Zaldivar, M. 1985. Attachment of the pathogen Flexibacter columnaris to fish cells. Master's Thesis, Oregon State University, Corvallis, Oregon.