|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 THE ROLE OF TOXINS AND SECRETED FACTORS IN THE PATHOGENESIS OF Vibrio vulnificus By JENNIFER LEE JOSEPH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREME NTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009
2 2009 Jennifer Lee Joseph
3 To my parents, William and Susan Joseph
4 ACKNOWLEDGMENTS I thank my mentor, Paul Gulig, who has guided me with patience through out my graduate career. With his wisdom and advice he has served as a mentor not only in the la boratory, but in life as well I acknowledge my committee members, Henry Baker, Shouguang Jin, and Anita Wright. They have brought new perspectives to my research and have k ept me constantly thinking and questioning everything My mentor and committee members have provided for me the foundation to have a successful career I would also like to express my appreciation to all of my colleagues in the Gulig lab (past and present ) for making my time in lab enjoyable. I am especially grateful for all of the assistance provided by Roslyn Brown, Julio Martin, Rupam Sharma, and Jessica Ascencio Of all of my colleagues, I am most grateful to Patrick Thiaville With his constant enc ouragement, assistance, and friendship, he has gone above and beyond his role as a coworker, and has become a cherished friend. Above all, I could not have done this without my friends and family especially m y parents They have pushed me harder than any one else and have never lost faith in me or my abilities. They e xpress unconditional love and encouragement, and I realize everyday how blessed I am to have them by my side.
5 TABLE OF CONTENTS ACKNOWLEDGMENTS .................................................................................................. 4 page LIST OF TABLES ............................................................................................................ 9 LIST OF FIGURES ........................................................................................................ 10 ABSTRACT ................................................................................................................... 12 CHAPTER 1 INTRODUCTION .................................................................................................... 14 Vibrio vulnificus ....................................................................................................... 14 Disease Caused by V. vulnificus ............................................................................. 14 Classification of V. vulnificus ................................................................................... 15 Virulence Factors of V. vulnificus ............................................................................ 18 The Polysaccharide Capsule ............................................................................ 18 Acquisition of Iron from the Host ...................................................................... 19 Flagella ............................................................................................................. 21 Pili and Attachment .......................................................................................... 21 Extracellular Toxins .......................................................................................... 22 Hemolysin/cytolysin ................................................................................... 23 Other hemolysins ....................................................................................... 23 Metalloprotease ......................................................................................... 24 Phospholipases .......................................................................................... 25 Focus of Investigation ............................................................................................. 26 Specific Aim 1: Examine Role of the RtxA1 Toxin in Pathogenesis of Vibrio vulnificus ....................................................................................................... 26 Specific Aim 2: Examine the Other RTX Loci in Virulence of V. vulnificus and the Importance of Activation of the RtxA Toxins ..................................... 27 Specific Aim 3: Examine the Type VI Secretion System o f V. vulnificus ......... 28 2 MATERIALS AND METHODS ................................................................................ 30 Standard Microbiological and Animal Infection Protocols ....................................... 30 Bacterial Cultures, Chemicals, and Media ........................................................ 30 Infection of Mice ............................................................................................... 31 Histological Analysis ......................................................................................... 32 Tissue Culture .................................................................................................. 32 Infection of monolayers .............................................................................. 32 Crystal violet a ssay for detachment and destruction of monolayers .......... 34 Lactate dehydrogenase assay for measuring lysis of INT 407 Cells .......... 34 Apoptos is assay ......................................................................................... 36 Statistical Analysis ............................................................................................ 37
6 Molecular Genetics and Mutagenesis ..................................................................... 37 Southern Blots .................................................................................................. 37 Extraction and digestion of genomic DNA .................................................. 37 Preparation of digoxigenin labeled probes for Southern bl ot analysis ........ 38 Transfer of DNA ......................................................................................... 38 Hybridization and detection of the labeled DNA probe ............................... 39 Mutagenesis of V. vulnificus ............................................................................. 39 USER friendly cloning ................................................................................ 39 Three way USER friendly cloning .............................................................. 40 Construction of mutations by conjugation of plasmid DNA into V. vulnificus ................................................................................................. 41 Chitin induced natural transformation ........................................................ 42 3 ROLE OF THE RtxA1 TOXIN IN PATHOGENESIS OF Vibrio vulnificus ................ 50 Rationale for Study ................................................................................................. 50 Introduction ............................................................................................................. 50 Results .................................................................................................................... 54 Disruption of rtxA1 by aph Insertion ................................................................. 54 Cytotoxicity of rtxA1 ::aph mutant ............................................................... 55 Virulence of rtxA1 ::aph mutant in mice ...................................................... 56 Deletion of rtxA1 ............................................................................................... 57 rtxA1 :: aph mutants ........................................................... 59 rtxA1 :: aph mutants in mice .................................................. 59 rtxA1 :: aph Mutation by ChitinInduced Natural Transformation .............................................................................................. 60 rtxA1 :: aph mutants ........................................ 61 rtxA1 :: aph with the wild type rtxA1 allele ....................................................................................................... 61 rtxA1 :: aph infected mice ....................... 63 Insertion of aph at 5 End of rtxA1 ................................................................. 64 In vitro characterization of rtxA1 :: aph ...................................................... 64 Virulence of rtxA1 :: aph in mice ................................................................ 64 Verification of the virulence defect of the rtxA1 :: aph mutant ................... 65 RtxA1 Causes Apoptosis .................................................................................. 66 A Combination of RtxA1 and VvhA Contributes to Cytotoxicity ........................ 67 Construction of a double mutation of rtxA1 :: aph vvhA ................... 67 Cytotoxicity of rtxA1 :: vvhA ............................................................. 67 Virulence of rtxA1 :: vvhA ................................................................ 68 Prevalence of rtxA1 in V. vulnificus .................................................................. 69 Discussion .............................................................................................................. 70 RtxA1 is the Major Cytotoxic Factor of V. vulnificus ......................................... 71 RtxA1 has a Role in Virulence of V. vulnificus .................................................. 72 Are Other Factors Involved in Cytotoxicity and Tissue Damage? ..................... 74 Presence of the rtxA1 Gene is Widespread Among V. vulnificus Strains ......... 75 What is the Function of RtxA1 in Virulence? .................................................... 76
7 4 THE ROLE OF THE OTHER RTX LOCI IN VIRULENCE OF V. vulnificus AND THE IMPORTANCE OF ACTIVATION OF THE RtxA TOXINS .............................. 96 Rationale for Study ................................................................................................. 96 Introduction ............................................................................................................. 96 Results .................................................................................................................... 98 Identification and Examination of RtxA2 ........................................................... 98 Mutation of rtxA2 to Examine a Role in Virulence ............................................ 99 Construction of rtxA2 :: ............................................................................. 99 Cytotoxicity of the rtxa2 :: ........................................................... 100 Virulence of the rtxA2 :: mutant in irontreated mice .............................. 100 Double Mutation of rtxA1 and rtxA2 ................................................................ 101 Virulence of double rtxA1 / rtxA2 mutant in mice ....................................... 102 rtxA1 ::aph rtxA2 :: ...................................................... 103 Cytotoxicity of ........................................................ 104 rtxA1 :: aph rtxA2 :: ........................................................... 104 Identification and Examination of RtxA3 ......................................................... 105 Deletion of the rtxA3 Gene to Examine a Role in Virulence ........................... 106 Construction of the rtxA3 mutant ........................................................... 1 06 rtxA3 mutant ............................................................. 106 rtxA3 mutant ................................................................ 106 Future Direction: Construction of a Triple rtxA Mutant ................................... 107 Examination of RtxC1 in Virulence ................................................................. 107 Deletion of VV2_0480 encoding RtxC1 .................................................... 108 rtxC1 mutant ...................................................... 109 Virulence of rtxC1 in irontreated mice .................................................. 109 Discussion ............................................................................................................ 110 RtxA2 is Not Essential for Virulence ............................................................... 111 RtxA3 is Not Essential for Virulence ............................................................... 113 RtxC is Not Required for Virulence Caused by RtxA1 .................................... 114 5 EXAMINING THE ROLE OF THE TYPE VI SECRETION SYSTEM IN PATHOGENESIS OF Vibrio vulnificus .................................................................. 125 Rationale for Study ............................................................................................... 125 Introduction ........................................................................................................... 125 Results .................................................................................................................. 129 Identification and Deletion of V. vulnificus vgrG ............................................. 129 vgrG ::aph mutant ............................................................ 130 vgrG ::aph in mice ............................................................... 132 Identification and Deletion of V. vulnificus hcp ............................................... 132 hcp ::cat ........................................................................... 134 hcp ::cat .............................................................................. 134 Deletion of the T6SS Factors in an rtxA1 Background ................................... 135 Cytotoxicity of T6SS, rtxA1 double mutants ............................................. 136 Virulence of T6SS, rtxA1 double m utants ................................................ 137 Discussion ............................................................................................................ 138
8 VgrG Causes Apoptosis; However, it is Not Essential for Virulence ............... 138 Hcp is Not Necessary for Cytotoxicity or Virulence ........................................ 140 What is the Function of T6SS? ....................................................................... 141 6 DISCUSSION ....................................................................................................... 153 What Role Do the RTX Toxins Play? .................................................................... 154 Type VI Secreted Factors are Not Essential for Virulence .................................... 155 If Not RTX Toxins or T6SS, What is Causing Damage? ....................................... 156 V. vulnificus is an Accidental Human Pathogen .................................................... 159 Final Remarks ....................................................................................................... 160 REFERENCE LIST ...................................................................................................... 162 BIOGRAPHICAL SKETCH .......................................................................................... 173
9 LIST OF TA BLES Table page 2 1 Bacterial strains used in this study ..................................................................... 44 2 2 Plasmids used in this study. ............................................................................... 46 2 2 Oligonucleotides used in this study. ................................................................... 48 3 1 Southern blot to detect rtxA1 in V. vulnificus isolates ......................................... 95
10 LIST OF FI GURES Figure page 3 1 Schematic of rtx gene clusters of Vibrio vulnificus strain CMCP6. ...................... 78 3 2 Detachment/destruction of INT 407 monolayers by rtxA1 ::aph mutant.. ............ 79 3 3 Virulence of rtxA1 ::aph mutant in iron dextran treated mice.. ............................. 80 3 4 Skin le sions of mice infected with rtxA ::aph ....................................................... 81 3 5 Detachment/destruction of INT rtxA1 ::aph mutant.. .... 82 3 6 Vi rtxA1 :: aph mutants in mice.. ......................................................... 83 3 7 Detachment/destruction of INT 407 cell monolayers by chitinrecreated rtxA1 mutants.. ................................................................................................. 84 3 8 rtxA1 :: aph mutants recreated by chitin transformation.. ............... 85 3 9 Cytotoxicity to INT rtxA1 :: aph complemented with wild type rtxA1 in trans.. .................................................................................................... 86 3 10 rtxA1 :: aph by expressing wildtype rtxA1 in trans on pGTR1227. ........................................................................................... 87 3 11 Histopathology of s.c. lesi rtxA1 infected mice. ...................................... 88 3 12 Detachment/destruction of INT 407 monolayers by rtxA1 :: mutant. ........... 89 3 13 Virulence of the rtxA1 :: aph mutant at increasing inocula. ................................ 90 3 14 Apoptosis of J774 cells infected with rtxA1 mutants.. ......................................... 91 3 15 Detachment/des truction of INT 407 monolayers infected with rtxA1 vvhA or rtxA1/vvhA mutants.. .......................................................................................... 92 3 16 Virulence of double rtxA1 ::vvhA ::tet mutant ........................................... 93 3 17 vvhA and rtxA1 :: aph mutants.. ............................................................................................................. 94 4 1 Detachment/destruction of INT 407 monolayers infected with rtxA2 :: ......... 116 4 2 Virulence of rtxA2 :: ........................................................................... 117 4 3 Cytotoxicity to INT 407 cells caused by the double mutant ............................................................................................................ 118
11 4 4 Virulence of rtxA1 ::aph, rtxA2 :: ............................................. 119 4 5 Lysis of INT rtxA1 ::aph rtxA1 :: aph rtxA2 :: mutants.. ........................................................................................................... 120 4 6 rtxA1 :: :: ............................................ 121 4 7 rtxA3 mutant in mice. ............................................................ 122 4 8 Cytotoxicity to INT rtxC mutant. ............................... 123 4 9 rtxC1 mutant in mice .................................................................. 124 5 1 vgrG :: aph. ............................. 143 5 2 vgrG ::aph ................................................................... 144 5 3 vgrG ::aph in s.c. inoculated mice ............................................... 145 5 4 hcp :: cat .. ............................... 146 5 5 hcp ::cat .. .............................................. 147 5 6 hcp ::cat in s.c. inoculated mice. .................................................. 148 5 7 rtxA1 :: tetA vgrG :: aph double mutant ................................... 149 5 8 Cytotoxicity of rtxA1 and T6SS double mutants to J774 cells. .......................... 150 5 9 Virulence of rtxA1 T6SS double mutants in mice.. ........................................... 151 5 10 Subcutaneous lesions of mice infected with T6SS, rtxA1 mutants. ................. 152
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE ROLE OF TOXINS AND SECRETED FACTORS IN THE PATHOGENESIS OF Vibrio vulnificus By Jennifer Lee Joseph December 2009 Chair: Paul A. Gulig Major: Medical Sciences Immunology and Microbiology Vibrio vulnificus is a gram negative bacterium capable of causing serious infections after ingestion of contaminated seafood or contact of wounds with conta minated water or objects. The bacteria are highly cytotoxic to host cells and cause extensive tissue damage during infection. The factors involved in this damage remain unknown. The focus of this investigation was to examine the role of toxins and other secreted factors of V. vulnificus in virulence and cytotoxicity We examined the R epeats in T o x in (RTX) toxins and proteins secreted via the type VI secretion system (T6SS) by constructing mutations in V. vulnificus and evaluating their effects on virule nce in mice and cytotoxicity in cell culture models. V. vulnificus encodes three RTX toxins, RtxA1, RtxA2, and RtxA3. We examined all three toxins and their activation by RtxC. rtxA1 mutants were defective in cytotoxicity to intestinal epithelial INT 407 cells and the ability to induce apoptosis in J774 murine macrophagelike cells. rtxA1 mutants were also attenuated for virulence in subcutaneously inoculated, iron dextrantreated mice; however, they were still able to cause subcutaneous lesions similar t o the wild type, suggesting that other virulence
13 factors cause these lesions. Additionally, RtxA1 does not require activation by RtxC for virulence or cytotoxicity. Deletion of either rtxA2 or rtxA3 had no significant effect on cytotoxicity in cell cultu re or virulence in mice. The recently discovered T6SS was examined for its role in cytotoxicity and virulence. The hemolysin coregulated protein (HCP) and Valineglycine repeat protein (VgrG) are suggested to be secreted effectors and components of the T6SS apparatus. Deletion of each of these genes in V. vulnificus had no effect on cytotoxicity in cell culture or virulence in mice, suggesting that the T6SS is not essential in virulence and tissue damage caused by V. vuln i ficus These studies evaluated several toxins and secreted factors of V. vulnificus for cytotoxicity and virulence. We identified RtxA1 as the major cytotoxic factor; however, other accessory toxins contribute to cytotoxicity. Despite examining several toxins and identifying the major c ytotoxic factor, the key factor(s) involved in tissue damage remains elusive.
14 CHAPTER 1 INTRODUCTION Vibrio vulnificus Vibrio vulnificus is a gram negative, motile, curved rod shaped bacterium found commonly in estuarine waters such as the coastal regions of the Gulf of Mexico. V. vulnificus can be free living in the water, as well as in association with filter feeding shellfish, such as oysters and clams ( 1) It resides in areas with temperate climates and thrives when water temperatures are above 18C (1 2) During t he warm summer months, nearly all of the oysters harvested from Gulf of Mexico coastal waters are contaminated with V. vulnificus (1). Disease Caused by V. vulnificus V. vulnificus is the leading cause of reported seafoodrela ted deaths in the United States (1). Infection occurs by two routes : ingestion of r aw, contaminated seafood and contact of open wounds with contaminated water or objects. Ingestion of contaminated seafood, such as raw oysters, results in primary septicemia, characterized by fever, chills, and the formation of secondary bullous lesions on the lower extremities (3 5) The disease progresses rapidly, with the onset of symptoms as soon as 24 hours after ingestion, and death can occur within 48 hours (1 2) Contamination of open wounds after contact with V. vulnificus results in wound infection. In severe cases, wound infection may progres s to necrotizing fasciitis, requiring debridement of the tissue or amputation of the infected limb, and in extreme cases will lead to secondary sepsis and death (2 6) Individuals with elevated serum iron levels due to conditions such as hemochromatosis and those with other predisposing conditions, such as hepatic disease, are at highest risk for systemic infection due to either ingestion or wound
15 infection (1 5 7) Both septicemia and wound infection are characterized by extremely rapid replication of the bacteria in the host and extensive tissue damage to the skin. The rapid na ture of the disease makes treatment difficult, resulting in mortality rates as high as 50% for septicemia and 30% for wound infection (1). Classification o f V. vulnificus V. vulnificus can be classified based on biotypes, lipo polysaccharide (LPS) antigens, and genetic sequences There are three biotypes of V. vulnificus Biotype 1 is associated with oysters and predominantly causes disease in humans, bioty pe 2 primarily affects eels and fish, and biotype 3 is an emerging biotype affecting people handling fish in Israel (8). Strains can also be classified by serotyping the LPS O antigen. Biotype 2 strains possess a single serotype; however, there is much more heterogeneity among the biotype 1 strains (9 10) More recently, various forms of molecular typing have been used differentiate strains of V. vulnificus Not all strains of V. vulnificus have the same virulence potential, and many researchers are interested in identifying genetic markers or patterns that are able to distinguish those which are more adept at c ausing disease from those which are less adept. Many of these studies involve molecular typing of clinical isolates from human patients and environmental isolates from water and oysters. Nilsson, et al (11) examined the small subunit (16S) rRNA of 67 clinical and nonclinical isolates, and determined there are 17nucleotide differences throughout the sequence. These differences in the 16S rRNA divide V. vulnificus strains into two major groups, designated types A and B. The majority of nonclinical isolates (31 of 33) were of the A type, while the majority of the clinical isolates (26 of 34) were type B (11)
16 Development of a real time PCRbased assay to determine the 16S rRNA type revealed a third type, designated AB, which correlated with nonclinical isolates (12) Warner and Oliver (13) used random amplification of polymorphic DNA (RAPD) to differentiate V. vulnificus isolates. Using this method, they found ex tensive heterogeneity among the samples. However, they identified a band that correlated with the clinical V vulnificus isolates and was occasionally present in the environmental iso lates This group subsequently developed a PCR based assay to identify the presence of this band, designated as virulence correlated gene ( vcgC or vcgE ), in clinical (C) and environmental (E) type strains, respectively. Using this PCR method, they could distinguish between the C type and E type strains, and their clas sific ation matched the A and B ribotyping of Nilsson, et al for strains that overlapped in the two studies (14) Chatzidaki Livanis et al (15) used an alternative typing method called R epetitive Extragenic Palindromic DNA PCR (rep PCR) to classify isolates of V. vulnificus Rep PCR is a genetic typing method that targets conserved repetitive elements at multiple loci throughout the genome, also distinguished between strains of clinical versus environmental origin (15) As opposed to the other typing methods, which classified V. vulnificus into two major groups, this method showed greater genetic diversity among the strains of clinical origin 68 strains were divided into seven groups based on patterns of PCR amplicons. Groups I IV and VII corresponded to clinical isolates and group III corresponded to environmental strains These r epPCR groups matched the A and B ribotyping of Nilsson, et al and the E and C RAPD results of Warner and Oliver. Chatzidaki Livanis et al. also examined polymorphism s in group 1
17 capsule polysaccharide (CPS) gene sequences by PCR and identified 2 allele groups: CPS allele 1 corresp onded to clinical isolates that classified as mostly r epPCR groups I IV, and VII and CPS allele 2 corresponded to environmental isolates in repPCR group III (15) Multi Locus Seque nce Typing (MLST) examining DNA sequence polymorphisms of housekeeping genes has also been used to examine genetic relationships of V. vulnificus strains. Bisharat et al (16 17) used MLST of a set of housekeeping genes to type 159 V. vulnificus strains and determined that V. vulni ficus strains could be placed into two clusters C luster 1 was comprised mainly of environmental isolates and cluster 2 was comprised primarily of clinical isolates. Cohen, et al (18) subsequently examined 6 genes to type 67 V. vulnificus strains by MLST. This st udy developed different designations opposite that of Bisharat and coworkers. Almost all lineage 1 strains were of clinical origin, while lineage 2 comprised most environmental isolates. Additionally, Cohen, et al identified a genomic island that was pr esent in most lineage 1 strains, but was absent from most l ineage 2 strains examined. The different typing methods have successfully related genotype to source of isolation of a strain, implying that strains associated with clinical isolation would be of i ncreased virulence. V. vulnificus strains fall primarily into two clades based on the different typing methods. Clade 1 is comprised of strains primarily isolated from the environment that are ribotypes A and AB vcgE MLST group 1, and lineage 2. Clade 2 consists of strains primarily isolated from clinical samples that are ribotype B, vcgC MLST group 2, and lineage 1. Recently, our laboratory performed a detailed analysis of the virulence of 71 strains from clinical and environmental sources in the
18 su bcutaneously (s.c.) inoculated, iron dextran treated mouse model of infection to determine the relationship of genotype to virulence ( Thiaville, et al ., in preparation). Almost all of the strains caused severe skin infection. However, only a subset of st rains had potential to cause systemic infection and death. This higher virulence potential significantly correlated with the genotypes classified into clade 2 strains ( vcgC ; B type 16S rRNA; MLST type 2; lineage 1) although there were several highly virul ent clade 1 strains, as well as attenuated clade 2 strains. From these studies, it is evident that the clade 2 V. vulnificus strains are more adept at causing systemic infection and death in our mouse model than the clade 1 strains, although there are exc eptions for in both directions Virulence Factors of V. vulnificus The major hallmarks of V. vulnificus disease are the rapid growth of the organism, extensive tissue damage, and resistance to host innate defenses. Although studies in pathogenesis of V. vulnificus have revealed factors necessary for virulence, the means by which V. vulnificus causes such a rapid and destructive infection remain uncertain (2 6 19) Several virulence factors have been proposed, including a polysaccharide capsule, mechanisms for iron acquisition, LPS, flagella, pili, hemolysin/cytolysin, metalloprotease and several exoenzymes (19) The Polysaccharid e Capsule The polysaccharide capsule is the most important virulence factor identified to date. The presence of the capsule relates to colony morphology, as encapsulated bacteria form opaque colonies and unencapsulated bacteria form translucent colonies (20 21) All virulent strains of V. vulnificus are encapsulated, and unencapsulated isolates occurring naturally in the environment are attenuated in mouse models of
19 infection (20, 22, 23) Wright, et al (22) demonstrated the importance of the capsule by examining translucent, acapsular transposon mutants of a virulent encapsulated strain of V. vulnificus These acapsular mutants were more sensitive to normal human sera than the encapsulated parent strain, and they were highly attenuated in int raperitoneal ( i.p.) inoculated mice (22). V. vulnificus is an extracellular pathogen, so to cause a successful infection, it must have the ability to resist phagocytic activ ities of host defense cells and complement mediated lysis. T he capsule appears to be the main mechanism for the resistance to these host defenses demonstrated by the serum sensitivity and the sensitivity to phagocytosis by translucent, acapsular isolates ( 22, 24, 25) Acquisition of Iron from the Host The importance of iron for growth of microorganisms has long been recognized (26) Free iron is limited in the host environment due to sequestration by ironbinding proteins, such as transferrin, lactoferrin, and hemoglobin. Pathogens have developed various mechanisms for sensing low iron levels and acquiring iron from the host during infection, in cluding the producti on of low molecular weight siderophores for binding iron and receptors for host ironcontaining proteins Iron overload in the host is one of the most important susceptibility factors for V. vulnificus disease, suggesting that the ability to sequester iron from the host is critical for virulence of the organism Wright et al. (27) experimentally demonstrated the importance of excess iron for pathogenesis of V. vulnificus infection Iron treatment of mice reduced the i.p. LD50 from 106 CFU to 1 CFU. Starks, et al. (28) emphasized the importance of iron in the host by demonstrating that during subcutaneous (s.c.) inocul ation of virulent strains the inocula must be increased by 105fold in non iron
20 dextrantreated mice to achieve similar levels of bacteria in tissues as in irontreated mice Additionally, the mechanisms of iron sensing and iron acquisition have proven t o be important virulence factors of V. vulnificus V. vulnificus produces a catechol type siderophore, vulnibactin, which enables V. vulnificus to utilize iron from transferrin, a host ironbinding protein (29) Litwin, et al. (29) characterized a mutant that was unable to utilize transferrinbound iron and found it was defective in production of the catechol type siderophore. This mutant was attenuated for virulence in an infant mouse model of infection Different mechanisms for the enhanced virulence of V. vulnificus in ironoverloaded hosts have been proposed. Elevated iron levels are speculated to increase host susceptibility by inhibiting certain innate immune defenses. For example, excess iron in the host has been demonstrated to have an inhibitory effect on phagocyte function (30) Hor (31) demonstrated th at the activity of neutrophils during infection with V. vulnificus was decreased in irontreated mice compared to noniron treated mice. Excess iron also contributes to increased virulence of V. vulnificus by increasing the growth rate of the bacteria. I n the same study as above, Hor, et al (31) also d emonstrated that irontreatment results in increased growth of V. vulnificus during infection. Using a marker plasmid system, Starks, et al. (32) demonstrated that treating s.c. inoculated mice with iron dextran significantly decreased the replication time of the V. vulnificus during infection. While iron contributes to both host susceptibility and virulence of V. vulnificus it appears that the main role of iron scavenging by V. vulnificus is to facilitate growth of the bacteria in the host
21 Flagella V. vulnific us possesses a single polar flagellum and is motile. Kim and Rhee (33) investigated an insertion mutation of the flgC gene, encoding a flagellar basal body rod protein, of a vvhA/vvpE mutant strain Their results revealed a defect in motility, a decrease in adherence and cytotoxicity to HeLa cells, and attenuat ion in mice (33) Lee, et al (34) examined a flagellum deficient mutant by a knockout of the flgE gene. This mutant had an increase in the LD50 over the wild type in s.c. inoculated, iron dextrantreated mice. Similar to the flgC mutants, flgE mutants also showed a severe defect in adherence to epithelial cells (34) Our laboratory has investigated the role of flagella in virulence in mice. Signaturetagged mutagenesis screens identified an attenuated mutant containing an insertion in the fliP gene involved in flagellar biosynthesis. This mutant was nonmotile and unable to cause systemic infection in the mice ; however it maintained the ability to cause severe skin infection. Further investigation of the flagellin gene clusters, flaCDE and revealed that the flaCDE flagellin genes are necessary for motility and for systemic disease in our s.c. inoculated, iron dextrantreated mouse model of infection, but that the flaFBA genes are dispensable for motility and virulence (Tucker, et al ., in preparation). Pili and Attachment To initiate infection at host surfaces, bacteria must be able to attach and colonize these surfaces. Attachment is mediated by production of adhesions, including pili. Gander and LaRocco (35) identified the presence of pili on the surface of V. vulnificus strains, particularly on clinical isolates. The authors determined the number of adherent
22 bacteria on HEp2 cells and concluded that clinical strains were more adherent than were environmental strains (3 5) Paranjpye, et al. (36) examined a pilD mutat ion that abolished expression of surface pili and resulted in a 100fold increase in LD50 in an iron treated, i.p. inoculated mouse model of infection. This study revealed a potential role of pili in virulence. The pilD gene encodes the type IV pilus leader peptidase/N methyltransferase that is involved in pilus formation, as well as processing of other type 2 secretion system secreted proteins. As expected, t he pilD mutation had other pleiotropic effects, including defective secretion of the hemolysin/ cytolysin protease, and chitinase, so the defect in virulence could not be definitively attributed to the absence of pili. To examine pili further, Paranjpye and Strom (37) examined a pilA mutant to assess the effect on adherence and virulence. A lthough pili were still present on the surface, the pilA mutant was defective in adhesion to HEp2 cells and had a 10fold increase LD50 in i.p. inoculated, irontreated mice (37) Extracellular Toxins V. vulnificus produces and secretes many extracellular proteins and toxins. It is hypothesized that some of these secreted proteins contribute to the significant tissue damage observed during infection. The most well studied extracellular proteins have been the metalloprotease and the hemolysin/cytolysin; however their roles in pathogenesis have remained questionable. Other putative hemolysins have been suggested, such as the Hemolysin III (HlyIII) (38) and a homolog of Legiolysin (VllY) (39) Phospholipase activity has also been implicated in virulence, although no isogenic mutant has been studied to date (40 41) Finally, the three RTX toxins, examined during this investigation, have been speculated to play a role in virulence. As discussed
23 later, only the Rt xA1 toxin has been confirmed to be a cytotoxic factor with a role in virulence. Hemolysin/ c ytolysin Kreger and Lockwood (42) first identified the hemolytic and cytotoxic ability of V. vulnificus The hemolysin/cytolysin protein, VvhA, was later purified and shown to have hemolytic activity (43) Injection of purified VvhA into animals produced skin damage very similar to that seen after infection with bacteria (44) suggesting that VvhA could be responsible for the damage caused by the bacteria during infection. Subsequent studies using purified toxin preparations have dissected various mechanisms of cytotoxicity such as apoptosis and poreformation (45 46) Despite the e vidence that the purified VvhA hemolysin causes tissue damage, Wright and Morris (47) demonstrated that an isogenic mutant strain remained virulent in mouse models of infection. These results cast doubt on the role of VvhA in virulence and tissue damage during infection. Other hemolysins In addition to VvhA, V. vulnificus pr oduces other proteins with putative hemolytic activities Chen, et al (38) iden tified a gene with similarity to the hemolysin III ( hlyIII ) of Bacillus cereus and it is highly similar to other putative hemolysin, including ones in V ch ol erae Y ersinia pestis and S almonella enterica. E. coli expressing the V. vulnificus HlyIII from a plasmid was hemolytic towards human erythrocytes. H owever, a V. vulnificus hlyIII mutant was still hemolytic on sheep blood agar plates indicating that the presence of other hemolysins compensate for the loss of hlyIII or that V. vulnificus HlyIII is not hemolytic to sheep erythrocytes. While it is not certain if HlyIII is a true hemolysin of V. vulnificus it does have a role in virulence. The h lyIII mutant was
24 attenuated by 16fold in i. p. inoculated, non irontreated mice (38) We have also examined a V. vulnificus hlyIII mutant and observed approximately a 3fold attenuation in s.c. inoculated, iron dextrantreated mice (unpublished data) Another, more ambiguous hemolysin is the VllY protein. Chang, et al (39) ident ified a clone from a V. vulnificus genomic library that confers hemolysis and pigment production to transformed E. coli The gene identified had similarity to the lly gene encoding the legiolysin of L egionella pneumophila. The VllY and Lly are also relat ed to the family of 4 hydroxyphenylpyruvate dioxygenase (H pp D) proteins involved in the catabolism of tyrosine. The exact function of H ppD in bacteria is uncertain, although it is involved in the production of pyomelanin, which has been implicated in stre ss survival and colonization for other bacteria (48) Duri ng our studies into the role of toxins in V. vulnificus we deleted vllY and determined that it was not essential for hemolysis on blood agar plates or cytotoxicity in cell culture. The vll Y mutant did have a slight attenuation in virulence in s.c. inoculated, iron dextrantreated mice; however, the exact function during infection remains unknown (unpublished data) Metalloprotease The zinc dependent metalloprotease produced by VvpE degrades elastin and collagen(49 50) As with VvhA hemolysin, injection of purified VvpE into mice caused dermal necrosis, similar to what is seen during infection (49) T wo groups constructed vvpE mutations and observed no attenuation of the mutants in i.p. or s.c. inoculated mice (51 52) Our laboratory showed that the vvpE mutation had no effect on skin damage or liver (systemic) inf ection in s.c. inoculated mice (52) indicating that VvpE is not likely involved in virulence or tissue damage. These results were similar to
25 what was seen with VvhA, emphasizing the importance of studying isogenic mutants to evaluate virulence, as opposed to studying the effects of a purified protein. It remained a possibility that the hemolysin/cytolysin and metalloprotease may be redundant virulence factors, so that mutation of one of them could be compensated for by the other. Fan et al (53) reported that a strain with constructed mutations in both of these genes retains some cytotoxicity in cell culture and is virulent in mice. This suggests that there are other cytotoxins being produced that are contributing to tissue damage. Phospholipases The phospholipase activity of V. vulnificus also has been proposed to have a role in virulence. Nearly 25 years ago, Testa, et al (40) demonstrated that V. vulnificus possesses phospholipase A1/A2 and lysophospholipase activities, but not phospholipase C activity. In 2007, Koo, et al (41) proposed that this phospholipase A (PLA) activity was important for virulence in a mouse model. The researchers inhibited phospholipase activity during infection of mice by treatment with tetracycline and determined that the attenuation that they observed was due to the inhibition of PLA activity (41) T here were many gaps in the study by Koo, et al (41) including the lack of construction and analysis of a mutant deficient in PLA activity. We followed up on the report by Koo, et al ., to determine if PLA activity of V. vulnificus was important for virulence. We deleted two genes encoding p hospholipase/lecithinase/hemolysin ( tlh ) and an outer membrane phospholipase A ( ompla). Both mutants were hemolytic on rabbit or sheep blood agar plates and had phospholipase activity on egg yolk agar plates Both mutants were also as virulent as the wild type in s.c. inoculated, irontreated mice (unpublished data) Without a PLA
26 deficient mutant, the claims made by Koo cannot be completely disregarded. Phospholipase activity may still contribute the virulence of V. vulnificus. Focus of Investiga tion The goal of this investigation was to examine the role of toxins and extracellular proteins of V. vulnificus in pathogenesis. Despite years of research, the factors causing the extensive tissue damage during infection remain unknown. We speculated t hat secreted toxins are causing destruction of host cells, leading to the tissue damage observed during infection. We followed the molecular version of Kochs postulates (54 55) to evaluate a role in virulence for the three RTX toxins (RtxA1, RtxA2, and RtxA3) and the type VI secretion system (T6SS). We analyzed these mutants for cytotoxicity in cell culture and virulence in the s.c. inoculated, iron dextran treated mouse model (28) This model allows us to examine both the ability of V. vulnificus to cause a local, skin infection and the ability to cause systemic infection and death. Specific Aim 1 : Examine Role of the RtxA1 Toxin in Pathogenesis of Vibrio vulnificus RtxA1 of V. vulnificus belongs to the Multifunctional A utoprocessing RTX (MARTX) family of RTX toxins (56) It is a very large protein (>500kDa), consisting of conserved amino acid repeat regions N terminus and C termi nus, flanking less conserved central domains with enzymatic activities. RTX toxins are cytotoxic in cell culture by a variety of mechanisms, and many of them contribute to the virulence. Previous studies in our laboratory identified the RtxA1 as a major cytotoxic factor of V. vulnificus strain MO6 24/O. These studies also showed it had a minor role in virulence. During this investigation, we analyzed the role of the RtxA1 in V. vulnificus strain CMCP6. We constructed mutations in the rtxA1 gene and analyzed the mutant strains for cytotoxicity
27 in cell culture and virulence in our mouse model of infection. We concluded that RtxA1 is the major cytotoxic factor of V. vulnificus CMCP6 and that it has a role in virulence in the mouse model. It is interesti ng that, although RtxA1 is the major cytotoxic factor of V. vulnificus rtxA1 mutants are still able to cause significant amounts of gross tissue damage, as well as systemic infection and death in the mouse model. RtxA1 may contribute to tissue damage, but it is clear that there are other factors involved, possibly VvhA hemolysin/cytolysin. We examined strain with mutations in the rtxA1 gene and the vvhA gene. Deletion of vvhA in the rtxA1 mutant background eliminated residual cytotoxicity; however, the double mutant was as virulent as the rtxA1 mutant. The factor causing tissue damage remains unknown. Specific Aim 2: Examine the O ther RTX Loci in Virulence of V. vulnificus and the Importance of Activation of the RtxA Toxins In addition to RtxA1, CMCP6 encodes two other large, putative RTX toxins, RtxA2 and RtxA3. The residual cytotoxicity and virulence phenotypes of rtxA1 mutants hinted that there could be an interesting story with the RTX toxins and pathogenesis. We speculated the RTX toxins could a ct in concert and that mutation of any one of them could affect cytotoxicity or virulence. To determine if these other RTX proteins contributed to tissue damage and virulence, we constructed mutations in the rtxA2 and rtxA3 genes. Interestingly, the rtxA 2 and rtxA3 genes proved to be nonessential for cytotoxicity in cell culture and for virulence in mice. A double rtxA1 / rtxA2 mutant was examined for virulence in mice, and, despite a reduction in systemic infection, it was still able to cause skin infecti on and tissue damage in mice.
28 RTX toxins typically require activation by an acyltransferase, RtxC. CMCP6 only encodes one RtxC in its genome. The gene rtxC1 is located directly upstream of rtxA1 We investigated the requirement of RtxC1 for activation of the RtxA toxins by deleting the rtxC1 gene and examining the mutant strain for cytotoxicity and virulence. We expected the phenotype of an rtxC1 mutant to be similar to the rtxA1 mutant examined in Specific Aim 1. Interestingly, the rtxC1 mutant was c ytotoxic in cell culture, and virulent in the mouse model. We concluded that activation by RtxC1 is not required for RtxA activity. While this differs from what is known about typical RTX toxins, it must be remembered that RtxA1 belongs to the MARTX fami ly of toxins, and the requirement for activation of this subfamily of RTX toxins has not yet been shown. Specific Aim 3 : Examine the Type VI Secretion System of V. vulnificus We identified the presence of a putative type VI secretion system (T6SS) in V. vulnificus by BLAST searches with known T6SS genes. The T6SS is newly discovered, but it has already been implicated to be involved in virulence of several gram negative pathogens, including Vibrio cholerae We examined T6SS by deleting the genes encodi ng Hcp and VgrG. These two proteins are suggested to be necessary for secretion, as well as being secreted substrates themselves (57) Deletion of these two genes had no effect on cytotoxicity in cell culture and no effect on virulence in mice, indicating that T6SS is not important for virulence. Because all of the mutants that we constructed and tested, including combinations of mutations, retained significant virulence and pathology in infected skin tissues, we speculate that an unidentified toxin or combination of toxins and secreted factors, including those studied here, contribute to damage. If the latter is true, then it is
29 possible that deletion of just one or two toxins at time will not have an effect on virulence, because any o f them could be sufficient for virulence.
30 CHAPTER 2 MATERIALS AND METHODS Standard Microbiological and Animal Infection Protocols Bacterial Cultures, Chemicals, and Media Bacterial strains used in this study are listed in Table 21. The virulent clinical strain of V. vulnificus CMCP6, was used as the background strain for all mutations discussed. Some mutations were originally constructed in FLA399, a spontaneous rifampicin resistant mutant of CMCP6; however, we halted the use of mutants in this backgr ound due to concerns over the virulence of FLA399. E. coli Top10 (Invitrogen) and E. coli EC100D (Epicentre) were used for routine cloning E. coli S17 (58) was used as the donor strain for plasmid conjugations with V. vulnifi cus. V. vulnificus and E. coli strains were grown in LuriaBertani broth containing 0.85% (w/v) NaCl (LB N) or LB N plates containing 1.5% (w/v) ag ar. Strains were stored at 80C in LB N with 35% (v/v) glycerol. Strains were grown on LB N plates containing 6% (w/v) sucrose for counterselection for loss of suicide plasmids. When required, antibiotics were included in the media at the following concentrations for E. coli : ampicillin (100 g/mL), tetracycline (12.5 g/mL), chloramphenicol (30 g/mL), kanam ycin (40 g/mL). Antibiotics were used at the following concentrations for V. vulnificus : ampicillin (10 g/mL), tetracycline (6.25 g/mL), chloramphenicol (5 g/mL), kanamycin (100 g/mL), rifampicin (50 g/ml), and colistin (102 U/mL ). To select for V. vulnificus and against donor E. coli during filter mating conjugations, VVM agar (59) or LB N agar containing 102 U/mL colistin and appropriate antibiotics was used.
31 Unless noted otherwise, components for media were from Difco (Franklin Lakes, NJ), chemicals were from Sigma (St. Louis, MO), DNA extraction and purification kits were from Qiagen (Valencia, CA), molecular genetics enzymes were from New England Biolabs (Ipswich, MA), and oligonucleotides w ere from IDT (Coralville, IA). Infection of Mice For mouse infections, a static overnight cultu re of the bacteria was grown in LB N at room temperature. Prior to infection, the starter cultures were diluted 1:20 into LB N and shaken at 37C until the bacteria reached exponential growth phase, when the optical density at 600 nm (OD600) reached 0.4 t o 0.6. The cultures were diluted in phosphatebuffered saline (PBS) to an appropriate concentration for infection. CFU/ml was determined by diluting and plating. All mouse infections used the s.c. inoculation model described by Starks et al (28) Seven to tenweek old female ICR mice (Harlan SpragueDawley, Indianapolis, Ind.) housed under specific pathogenfree conditions were used for all experiments. At least 1 hr prior to inoculation, mice were injected i.p. St. Louis, MO) per gram body weight. Mice were injected s.c. into the right lower flank with bacteria suspended in 0.1 mL PBS. When the mice became moribund, as indicated by a rectal temperature below 33C, they were euthanized by carbon dioxide asphyxiation. If the mice did not become moribund, they were euthanized at 20 hr postinoculation. After the mice were euthanized, their skin was peeled back to reveal the s.c. lesion at the i njection site. For quantitative analysis of the bacteria in the tissues, samples of the s.c. lesion and the liver were aseptically removed from mice, homogenized in 5 mL of PBS using glass tissue homogenizers, diluted, and plated on LB N agar. Strains
32 carrying plasmids were plated both nonselectively (LB N agar) and selectively (LB N agar containing appropriate antibiotics). Samples were not taken from mice with no visible lesion. When no CFU were recovered from a skin lesion or liver sample, a minimum detectable CFU/g was used for these mice for statistical analysis. Histological Analysis Sam ples of the subcutaneous lesions resulting at inoculation sites w ere collected immediately after sacrifice of mice and fixed by immersion in 10% (v/v) buffered formalin (60) Formalin fixed tis the University of Florida Department of Pathology, Immunology, and Laboratory Medicine Diagnostic Referral Laboratory Histological sections were s tained with hematoxylineosin. Tissue Culture T wo cell lines were used in this study, the human intestinal epithelial cell line, INT 407, and the murine macrophagelike cell line, J774, both obtained from American Type Culture Collection (ATCC, Manassas, Virginia). All tissue culture was maintained in Dulbeccos modified Eagle medium (DMEM) containing 10% (w/v) fetal bovine serum (FBS) and an antibiotic antimycotic mix with a final concentration of 100 U/ml penicillin, 100 g/ml streptomycin, and 25 ng/ml Amphotericin B. All tissue culture media compo nents were supplied by Invitrogen (Carlsbad, California). All tissue cultures were incubated at 37C in a humidified atmosphere with 5% CO2. Infection of monolayers For measurement of monolayer destruction/detachment and cell lysis, INT 407 cells were was hed with Hanks balanced salt solution (HBSS), suspended in DMEM with FBS and antibiotics at a concentration of 1x105 cells/mL, and seeded in 24well
33 tissue culture plates (Corning, Cambridge, Massachusetts). The plates were incubated for 2 days until the monolayers reached 8090% confluency. Two hours prior to infection, antibiotic containing medium was removed from each well and replaced with 1 ml antibiotic free DMEM with FBS. All V. vulnificus strains, grown to logarithmic phase as detailed above, were pelleted, suspended in 3 ml antibiotic free DMEM with FBS and diluted to 2 x 107 CFU/ ml. 0.5 ml of appropriate bacterial suspension was added to each well, in triplicate for a multiplicity of infection (MOI) of 10. After 1 hour of incubation at 37C, of gentamicin was added to each well to kill the bacteria. Following antibiotic treatment, plates were incubated at 37C up to 24 hours. It should be noted that cell cultures were not washed after addition of gentamicin, so that any extracellular products and toxins produced by the bacteria during the initial 1 hour infection period would remain in the cell culture. 24 h later, infected cell cultures were assayed for destruction or detachment of INT 407 monolayers or lysis of INT 407 cells. Fo r measurement of apoptosis, J774 cells were washed as above, diluted to a concen tration of 1 x 105 cells/ml, and 0.1 ml of the J774 cells were added to each well of a black Nunc F96 MicroWell plate (Nunc). The plates were incubated for 2 days, to allow the monolayers to reach 8090% confluency. Two hours prior to infection, antibiotic containing medium was removed from each well and replaced with 0.1 ml antibiotic free medium. All V. vulnificus strains, grown to logarithmic phase as detailed above, were pelleted, suspended in 5 ml antibiotic free DMEM with FBS and diluted to 2 x 107 CFU/ml. 0.1 ml (2 x 106 CFU) of appropriate bacterial suspension was added in triplicate to the 96 well plates. After 1 hour of incubation at 37C, extracellular bacteria were killed by addition of gentamicin to each well at a final concent
34 Following antibiotic treatment, plates were incubated at 37C for 3 hours, at which time infected cell cultures were assayed for apoptosis. Crystal v iolet a ssay for d etachment and d estruction of monolayers A crystal violet staining as say, adapted from Ruff and Gifford (61) was used to assess destruction and detachment of infected INT 407 monolayers. 24 hr after infection, supernatants were aspirated, and wells were washed two times with 1 ml PBS to remove dead or damaged cells. The cells remaining attached to the culture well were stained with 1 ml of 0.05% (w/v) crystal violet diluted in PBS and incubated at room temperature for 10 mi n. The wells were washed four times with 1 ml PBS and 1 ml 95% ethanol was added to solubilize the remaining crystal violet. Uninfected INT 407 cells and blank wells containing media only were used as positive and negative controls, respectively, for pr esence of epithelial cells. 150 l of the crystal violet/ethanol solution from each well was transferred to a 96well plate, and absorbance was measured at 490 nm (A490) using an ELx800uv microplate reader (BioTek Instruments, Inc.) Percent monolayer detachment was calculated as follows: 1 Raw A490 (from each infected and uninfected well) A490 media only (average of all three wells) = Normalized A490 2 Normalized A490 (for each infected well) / Normalized A490 of uninfected cells (average of all three wells) = ratio of attached infected cells to attached uninfected cells 3 100 (ratio x 100) = % detachment /destruction (cells detached from monolayer) Lactate d ehydrogenase as say for me asuring l ysis of INT 407 Cells The Cytotoxicity Detection Kit (Roche Appli ed Science, Indianapolis, IN) was used to detect lysis of the infected INT 407 cells. This is a colorimetric assay that measures lactate dehydrogenase ( LDH) a cytoplasmic protein released by eukaryotic cells as a
35 result of lysis. LDH interacts with tetr azolium salt (yellow) in the LDH assay reagent supplied in the kit, resulting in the formation of formazen salt (red). The color formed, measured by absorbance at 490 nm, is proportional to the number of lysed cells. 24 hr after infection of INT 407 monol removed and transferred to a microcentrifuge tube. The supernatant was centrifuged at 200 x g at 4C for 10 minutes to remove cells and debris 96well plate. This represented To measure total LDH activity of either infected or uninfected cells, TritonX 100 (Tx) was added to each well for a final concentration of 1% and was mixed vigorously to removed from each well, centrifuged, and 25 l was transferred to a 96well plate and diluted with PBS. 100 well plate, according to the manufacturers instructions, and the plate was incubated in the and the amount of LDH released was measured using an ELx800uv microplate reader (Bio Tek Instruments, Inc.). Percent lysis was calculated as follows: 1 Raw A490 (from each infected and uninfected well, both TrtionX treated and untreated) A490 media only (average of all three wells) = Normalized A490 2 [100 x (Normalized A490 of infected well (no Tx treatment) x 0.525 mL)]/ [(Normalized A490 (no Tx treatment) x 0.15 mL)+(Normalized A490 of maximum release of infected well (with Tx treatment) x 0.425 mL)] = % lysis For each well, we computed the LDH activity in the supernatant divided by the total LDH activity in the well. Th e percent LDH release by uninfected cells, approximately 15%, was subtracted from the percent LDH release in V. vulnificus -
36 infected wells. Triplicate wells were run for each sample, and each experiment was performed at least twice. Apoptosis assay The ApoOne Homogeneous Caspase3/7 Assay (Promega, Madison, Wisconsin) was used to measure the ability of V. vulnificus strains to cause apoptosis to J774 cells. This assay measures the activities of caspase3 and caspase7, members of the cysteine aspartic aci d specific protease family which are key effectors in the process of apoptosis in mammalian cells. Measurements of these caspases are used as indicators of apoptosis in cell culture models. The ApoOne caspase3/7 substrate rhodamine 110, bis (N CBZ L as partyl L glutamyl L valyl L aspartic acid amide) (Z DEVD R110), exists as a profluorescent substrate prior to the assay. Upon sequential cleavage and removal of the DEVD peptides by caspase3/7 activity, the rhodamine 110 leaving group becomes intensely f luorescent allowing apoptosis to be quantified. J774 cells were infected with V. vulnificus as described above and incubated for three hours at 37C post addition of gentamicin. Following incubation, 100 l of the Homogeneous Caspase3/7 Reagent was added to each well and then incubated shaking gently at room temperature in the dark for 30 minutes. Fluorescence of each well was measured in the FLx800 Microplate Fluorescence Reader (BioTek Instruments, Inc., Winooski, Vermont) at an excitation wavelength of 485 20 nm and an emission wavelength of 530 25 nm. The apoptotic agent gliotoxin was used as a positive apoptotic control, wells containing media only were used as negative controls and uninfected J774 cells were used to establish a background level of apoptosis. The following formula was used to calculate apoptosis:
37 1 Raw wavelength reading (from each infected and uninfected well) Wavelength reading of media only (average of all three wells) = Normalized wavelength reading 2 For % apoptosis: [100 x (Normalized wavelength reading (from each infected well) / Normalized wavelength reading (from gliotoxin treated wells, average of three wells)]= % apoptosis 3 For apoptosis over background: Normalized wavelength reading (from each infected well) / Normalize d wavelength reading (from uninfected wells, average of three wells) = ratio of apoptosis over cell background Means and standard deviations of the percent apopotsis or wavelength ratios obtained for each set of triplicate wells w ere calculated for each st rain. Statistical Analysis The Student s t test was used to examine for significant differences between means of two sample groups. For experiments with more than two sample groups, an ANOVA was performed to determine if a significant difference was present in the group ( P ) If a significant difference was detected, a Fishers Least Significant Difference (LSD) 2 tests were used in mouse experiments to determine if the number of mice with detectable CFU was significantly changed in mutant versus wildtype infections. Statistical analyses were performed using Excel and XLSTAT. The statistical test used is delineated in the text for each experiment. Values were considered statistically significant for P Molecular Genetics and Mutagenesis Southern Blots Extracti on and digestion of genomic DNA For S outhern b lot analysis, genomic DNA was extracted from a selection of clinical and environmental strains representing different genotypes with varying cytotoxicities and virulence ( Thiaville et al ., in prep aration) Genomic DNA was isolated from each
38 strain using the Qiagen DNeasy Blood and Tissue Kit The genomic DNA was digested with the restriction endonuclease Eco R I for 3 h r resolved in a 0.8% (w/v) agarose gel. The digoxigeninlabeled DNA Molecular Weight Marker II (Roche Applied Science, Indianapolis, IN) was used as a size standard. The thidium bromide and photographed. Preparation of d igoxigenin l abel e d p robes for S outhern b lot a nalysis Four probes (rtxA1A through rtxA1 D) spanning the rtxA 1 gene of CMCP6 were used for detection of rtxA genes in multiple strains of V. vulnificus Each probe (approximately 1.5 kb in length) was PCR amplified. The PCR amplicons were labeled with digoxigenin by random primed labeling using the DIG High Prime DNA Labeling and Detection Starter Kit II (Roche, Mannheim, Germany). Each labeled probe was pur ified using the QIAquick PCR Purification Kit (Qiagen) to remove excess contaminating proteins prior to use. Efficiency of probe labeling was detect ed by spotting serial dilutions of labeled probe onto a positively charged nylon membrane and proceeding wi th the detection protocol described in the kit technical manual. Transfer of DNA After the agarose gel was stained with ethidium bromide and photographed, the DNA was depurinated by incubating the gel in 250 mM HCl with shaking using an orbital shaker a t room temperature for 15 minutes. The gel was rinsed briefly two times with filtered ddH2O. The DNA in the gel was then denatured in denaturing solution (0.5 M NaOH; 1.5 M NaCl) for 15 min shaking at room temperature. After incubation in the denaturing solution, the gel was briefly rinsed with ddH2O two times. Next, the gel was incubated in a neutralization solution (0.5 M Tris HCl, pH 7.5; 1.5 M NaCl) for 15 min shaking at room temperature. The gel was briefly washed two times with ddH2O. The
39 gel wa s equilibrated for 10 min in 20X SSC (3M NaCl; 0.3M sodium citrate; pH 7.0) prior to transfer of the DNA to a positively charged nylon membrane ( Roche). The transfer of the DNA from the gel to the positively charged nylon membrane was accomplished by over night capillary transfer. After overnight transfer, the DNA was fixed to the membrane by UV crosslinking using a GS Gene Linker UV Chamber (BioR ad, Hercules, CA). Hybridization and d etection of the labeled DNA probe The membrane was prehybridized with 10 ml prewarmed DIG Easy Hyb (Roche) in a hybridization bag at 42C for 30 min with gentle rocking. 350 ng of DIG labeled probe was denatured by boiling and added to 5 ml prewarmed DIG Easy Hyb. The prehybridization solution was discarded the hybridizat ion solution containing the probe was added to the membrane, and the bag was sealed. The hybridization proceeded overnight at 42C with gentle rocking. Hybridization solutions containing DIG labeled probes were saved, stored at 20C, and reused. To denature probes stored in DIG Easy Hyb, the hybridization solution was heated at 68C for 10 min. Chemiluminescent detection with CSPD was used to visualize the hybridized probe. The DIG Wash and Block Buffer Set (Roche) was used in the detection process. D etection of the hybridized probe was carried out as described in the Roche kit t echnical m anual Mutagenesis of V. vulnificus USER f riendly c loning USER (U racil S pecific E xcision R eagent ) Friendly cloning (New England Biolabs, Ipswich, MA) is a method for easily capturing PCR products into a USER compatible vector. Target DNA was PCR amplified using oligonucleotide primers designed with
40 eight additional nucleotides, either GGAGACAU or GGGAAAGU complementary to overhangs on the linearized USER vector. PCR products were treated with the USER enzyme, a mixture of uracil DNA glycosylase which excises the uracil residue leaving an abasic site, and the DNA glycosylaselyase Endo VIII, which breaks the phosphodiester backbone flanking the abasic site, releasing the terminal seven nucleotides. The PCR product, flanked by the 8 nucleotides, was ligated with the USER vector which had been digested with the nicking enzyme Nt.BbvC I and the restriction enzyme Xba I, to create 8nucleotide overhangs complementary to the 8nucleotide 3 extensions on each end of the PCR product, allowing for easy capture of the product. Our laboratory has constructed USER friendly versions of cloning and allelic exchange vectors commonly used in V. vulnificus described in detail in Guli g, et al (62) An allelic exchange vector, pCVD442 (63) was modifie d by insertion of lacZ with a USER cloning site to yield pGTR1113. The cat gene, conferring chloramphenicol resistance, was inserted into pGTR1113, to yield pGTR1129. Both of these vectors were subsequently used to construct plasmids for mutation of selected genes. 10 l of PCR product was incubated with 1 unit USER enzyme and 20 ng USER vector at 37C for 15 minutes. T4 DNA ligase and ligase buffer were added to the mixture and incubated for 15 minutes at room temperature. The reaction was cleaned usi ng the DNA Clean and ConcentratorTM Kit (Zymo Research Corporation, Orange, CA), electroporated into E. coli EC100D pir+ and plated LB N agar plates containing 40 g/ml X gal and either 30g/ml chloramphenicol or 100 g/ml ampicillin Three way USER friendl y c loning We modified the original USER cloning method to clone upstream and downstream flanking DNA together in a single step. 500to 1000 bp upstream and
41 downstream DNA sequences flanking a target gene were amplified using oligonucleotides designed for capture into a USER vector. The oligonucleotides used to amplify the outside ends of the upstream and downstream DNA included the USER sequences complementary to the vector. The inside ends of the two fragments to be joined were amplified with oligonucl eotides designed with compatible USER ends ( ACCCGGGU), containing a Sma I site. Equal amounts of the upstream and downstream PCR products were mixed together, cleaned using the DNA Clean and ConcentratorTM Kit and eluted in 10 l water. The cleaned PCR products were mixed with 20 ng USER vector and incubated at 37 C for 15 min with 1 unit USER enzyme. The remainder of the USER cloning procedure was followed as described above. To allow for selection of the deletion, an antibiotic resistance cassette was inserted between the upstream and downstream cloned sequence. The deletion plasmid was digested with Sma I at the site engineered between the upstream and downstream sequences. A blunt end antibiotic resistance cassette ( aph encoding kanamycin resistance ; cat encoding chloramphenicol resistance; or tet encoding tetracycline resistance) was ligated into the Sma I site with T4 DNA ligase and incubated overnight at room temperature. The insertion of the antibiotic resistance cassette was selected on the appropriate antibiotic, the correct size of the plasmid was confirmed on an agarose gel, and the presence of the insert was confirmed by PCR amplification across the upstream and downstream sequences. Construction of mutations by c onjugation of p lasmid DNA into V. vulnificus Allelic exchange plasmids were electroporated into E. coli S17 pir for conjugation into V vulnificus via filter mating. Static overnight cultures of V. vulnificus and E. coli were grown in LB N at room temperature and 37C, respectively. The
42 cultures were diluted in fresh LB N and grown shaking at 37C until they reached exponential phase. The two cultures were mixed in a ratio of 3 to 1, donor E. c oli cells to recipient V. vulnificus cells, filter mated, and transconjugants were selected on either VVM or LB N containing 102 U /mL colistin and appropriate ant ibiotics. Once the correct insertion of the allelic exchange vector into the chromosome was confirmed by PCR, the single crossover isolates were grown on LB N agar containing 6% (w/v) sucrose at room temperature to enrich for cells in which a second crossover event occurred, resulting in loss of the plasmid. Sucroseresistant colonies were passaged onto LB N containing appropriate antibiotics to select for the mutation and screened for loss of antibiotic resistance encoded by the vector. Allelic exchange of the mutant DNA for the wild type DNA was confirmed by PCR. Chitin induced natural t ransformation Chitin induced natural transformation of V. vulnificus was performed as described in Gulig et al (62) Sterile pieces of crab shell from blue crabs were used as a source of chitin. Static cultures of V. vulnificus were grown overnight in LB N at room temperature. The overnight cultures were dilu ted 1:20 into fresh LB N and grown shaking at 37C until the cultures reached exponential phase. The bacteria were centrifuged, washed with filter sterilized seawater (University of Florida Whitney Laboratory for Marine Bioscience) diluted to a salinity o f 25 ppt centrifuged again, and suspended in a volume of 25ppt seawater twice the starting culture volume. 2 mL of the bacteria in seawater were placed in a well of a 12 well tissue culture plate (Costar, Corning, NY) with a piece of sterile crab shell, and the plate was incubated overnight at 30C. The following day the supernatant was removed and replaced with 2 mL of fresh 25ppt seawater and o f genomic DNA was
43 added to the culture. Plasmid DNA prepared using the Qiagen QIAprep Spin Miniprep Kit was linearized by digestion with an enzyme that cu t the vector opposite the insert sequences Genomic DNA wa s prepared using Qiagen DNeasy Blood and Ti ssue Kit. The plates were incubated overnight at 30C. The following day the supernatants were removed and plated on LB N with appropriate antibiotics. The c rab shells were placed in a 50mL conical tube conta ining 2 mL of PBS, vortexed to release bacteria, and the supernatant was plated on LB N with appropriate antibiotics. Mutations were confirmed by PCR to verify correct allelic exchange of the mutant DNA for the wild type DNA.
44 Table 2 1 Bacterial strains used in this study Strains Relevant Characteristics Reference or Source E. coli TOP10 F mcrA mrr hsdRMS mcrBC lacZ lacX74 recA1 araD ara leu )7697 galU galK rpsL (Strr) endA1 nupG Invitrogen S17 pir pir lysogen; thi pro hsdR h sdM + recA RP4 2 Tc::Mu Km::Tn7(Tpr Smr) (58) EC100D pir + F mcrA mrr hsdRMS mcrBC ) lacZ lacX74 recA1 endA1 araD139 )7697 galU galK rpsL nupG pir +(DHFR) Epicentre V. vulnificus CMCP6 Clinical Isolate (64 65) FLA399 Spontaneous Rif r derivative of CMCP6 This study FLA413 FLA399 rtxA1 :: aph; Km r This study FLA439 FLA399 rtxA1 :: Km r This study FLA441 FLA399 rtxA2 This study FLA554 CMCP6 rtxA1 :: aph ; Km r This study FLA558 CMCP6 rtxA1 :: Km r This study FLA590 CMCP6 rtxA1 :: aph reversion This study FLA591 rtxC This study FLA899 rtxA3 This study FLA90 0 rtxA1 :: aph isolate # 1; Km r This study FLA901 rtxA1 :: aph isolate # 2; Km r This study FLA904 CMCP6 rtxA1 aph ; Km r This study FLA912 vvhA :: tetAR ; Tc r This study FLA917 vvhA :: tetAR rtxA1 aph ; Tc r Km r This study FLA923 rtx A1 :: aph (via chitin transformation with FLA901 genomic DNA); Kmr This study FLA943 rtxA1 :: aph (via chitin transformation with FLA900 genomic DNA); Kmr This study FLA943(pGTR1228) rtxA1 :: aph ( rtxA1 + ); Km r Cm r This study FLA943(pGTR1204) CMCP rtxA1 :: aph (vector); Km r Cm r This study FLA947 rtxA1 :: aph Km r This study FLA954 rtxA1 :: tetA ; Tc r This study
45 Table 21. Continued Strains Relevant Characteristics Reference or Source FLA960 rtxA1 :: tetA vgrG :: aph ; Tc r Km r This study FLA965 rtxA1 :: tetA hcp :: cat ; Tc r Cm r This study FLA1030 hcp :: cat ; Cm r This study FLA1035 vgrG :: aph ; Km r This study
46 Table 22. Plasmids used in this study. Plasmid Description Source or Reference pBR aph pBR322 aph cloned at EcoR aph element containing transcription and translation termination signals; Ap r Km r Tc r (66 67) pCOS5 Cloning vector containing cat oriT ColE1, and ori V ; Ap r Cm r (68) pCR2.1 TOPO TOPO cloning vector for capturing PCR products using TA cloning, Apr Invitrogen pCVD442 R6K ori mob sacB bla ; suicide vector for allelic exchange; Ap r (63) pNEB206A USER friendly cloning vector; Ap r NEB pUC19 Clon ing vector; lacZ ; Ap r (69) pGTR256 1 kb region of rtxA2 amplified from CMCP6 with oligos vv rtxA2 A and vv rtxA2 B captured into pCR2.1TOPO This study pGTR257 pGTR256 digested with Hind III, filled in with Klenow fragment, and religat ed,for loss of Hind III site This study pGTR259 pGTR257 digested with Nar I and regligated; deletion of portion of aph gene This study pGTR260 aph cassette at Ssp I site in rtxA2 fragement of pGTR259 This study pGTR261 aph from pGTR26 0 fragment subcloned into pUC19 with Hind III site deleted This study PGTR262 Digestion of pGTR261 with Hind III and religation to delete aph aph This study pGTR263 ; rtxA2 Ssp I and Xba I of pGTR1122 This study pGTR265 rtxC; rtxC upstream and downstream PCR amplicons digested with Not I and USER cloned into pGTR1113; Apr This study pGTR267 rtxA3; rtxA3 upstream and downstream amplicons captured into pGTR1113 by threeway USER cloning; Apr This study pGTR268 rtxA3; rtxA3 upstream and downstream amplicons captured into pGTR1129 by three way USER cloning; Ap r Cm r This study pGTR272 rtxA1 upstream and downstream PCR amplicons captured into pGRT1129 by threeway USER cloning (isolate 1); Apr, Cm r This study pGTR273 rtxA1 upstream and downstream PCR amplicons captured into pGRT1129 by threeway USER cloning (isolate 2); Apr, Cm r This study pGTR274 rtxA1 :: aph; aph inserted at Sma I site between the rtxA1 upstream and downstream sequences of pGTR272; Ap r Cm r This study pGTR276 500 bp upstream and downstream flanking site in 5' end of rtxA1 gene captured into pGTR1129 by three way USER cloning; Ap r Cm r Km r This study pGTR278 rtxA1 aph aph cassette inserted at Sma I site between cloned rtxA1 sequenc es of pGTR276 for insertion into 5' end of rtxA1 ; Ap r Cm r Km r This study
47 Table 22. Continued pGTR285 rtxA1 :: tet ; tetA gene subcloned into the Sma I wite between rtxA1 upstream and downstream sequences of pGTR273; Apr, Tcr This study pGTR288 rtxA3 : : cat; cat gene subcloned into Sma I site between rtxA3 upstream and downstream sequences in pGTR267; Apr, Cmr This study pGTR652 3.8 kb internal rtxA1 fragment from MO6 24/O sequencing clone inserted into Sma I site of pCVD442; Apr This study pGTR653 rtxA1 :: aph ; aph cassette from pUC4K inserted in rtxA1 fragment in pGTR652; Apr, Kmr This study pGTR1113 pCVD442:: lacZ lacZ USER Friendly cloning oligonucleotide linker incorporated (62) pGTR1119 rtxA1 :: aph reversion ; tetAR gene i nserted at blunt ended XbaI site of pGTR652; Apr, Tcr This study pGTR1122 pCVD442:: cloned at blunt ended Nde I of pCVD442:: lacZ r Cmr (62) pGTR1129 pCVD442:: ::USER cloned into EcoR V site of pGTR1113; Apr, Cmr (62) pGTR1204 pCOS5:: ::USER; USER friendly cloning site from pGTR1113 incorporated into pCOS5:: Apr, Cmr (62) pGTR1208 500 bp vvhA upstream and downstrea m PCR amplicons captured into pGTR1129 by three way USER cloning; Apr, Cmr This study pGTR1221 vvhA :: tetAR; tetAR cassette subcloned into Sma I site between vvhA upstream and downstream sequences in pGTR1208; Apr, Cmr, Tcr This study pGTR1227 rtxA1 comp lementation; 15.6 kb rtxA1 gene PCR amplified from CMCP6 and cloned into pGTR1204 at Sac I and Xba I sites; Apr, Cmr This study pGTR1303 500 bp hcp upstream and downstream PCR amplicons captured into pGTR1113 by three way USER cloning; Apr This study pGTR1 304 500 bp vgrG up s tream and downstream PCR amplicons captured into pGTR1113 by three way USER cloning; Apr This study pGTR1305 hcp :: cat ; cat cassette subcloned into Sma I site between hcp upstream and downstream sequences of pGTR1303; Apr, Cmr This study pGTR1307 vgrG :: aph; aph cassette subcloned into Sma I site between vgrG upstream and downstream sequences of pGTR1304; Apr, Cmr, Kmr This study
48 Table 22 Oligonucleotides used in this study. Oligonucleotide primer Sequence 5 3 a rtxA1test down ATCGGTGTAGCGTCAAACACAGG rtxA1 test up GGAACTTGATGCTCCGGGGC rtxA1 up 5' A CCCGGG rtxA1 up 3' U TAAGCCAAACTCTTCTTTAGGAG GGAGACA UGAG CTTGCAGGCGGAGAGTGA rtxA1 down 3' GGGAAAGU CGCTTATGGCAACGGAATTCG rtxA1 down 5' A CCCGGG rtxA1 insert up 5' U CAACGAGGCACGAGTATAGAG A CCCGGG rtxA1 insert up 3' UACAGTGACATGATCATCACCAC GGAGACAU GTTGCAGAATGGCAGCAGCG rtxA1 insert down 5' A CCCGGG rtxA1 insert down 3' U CGGATCG ATTGGTGCA ACGG GGGAAAGU GGACCAAATGGTTGTCAGCAC rtxA1 omega down 5' GGTAATATCGCCCACTTTGCC rtxA1 omega CGGTGGATGACCTTTTGAATG rtxA1 omega up 3' CCCCTTCATCGTCGCTACG rtxA1 clone 3' GCAT GAGCTC rtxA1 clone 5' GTGAATATTACACCGTTATACCCT TCGA TCTAGA rtxA1 A 5' CTCCT AAAGAAGAGTTTGGCTTATG CACCGTTATACCCTTTTTATGGA rtxA1 A 3' GAGATGTTGGAGGTCATTGGTG rtxA1 B 5' CCATGTTCAGATCGGACTTGCC rtxA1 B 3' CACGGTTCTGGTTCTTCTGCTG rtxA1 C 5' GTCTGGGCATTGTCGAGAACG rtxA1 C 3' GTCACAAAGTCAGTTGGAGTCGAC rtxA1 D 5' CGGTCAGGTCC TTCTCGACAG rtxA1 D 3' CGCCGACGATGGAAACAACAG rtxA2 a G TCTAGA rtxA2 b GGTGGCCAATGTTGAAGATGG C GAGCTC rtxA3 down 3 GTGATCGCCGCGCTGTCAACC GGGAAAGU GCTCAATGGCGCATGTCGTG rtxA3 down 5SmaI USER ACCCGGGU TGCTGTCTGCAATTTGGTCTAAG rtxA3 up 3 SmaI USER ACCCGGGU AG ATGACTGTAGTACTTTCGTC rtxA3 up 5 GGAGACAU GTAACCAATACCAGCGTCTAAG rtxC down 3 GGGAAAGU A GTACGGTCTAGTTTATTGCCC rtxC down 5 GATATAGGGCAGA GCGGCCGC CCTAAAGAAGAGTTT GGCTTATG
49 Table 22 Continued. Oligonucleotide primer Sequence 5 3 a rtxC up 3 CTCTTCT TTAGG GCGGCCGC rtxC up 5 TCTGCCCTATATCAACC AATATG GGAGACAU TTGAGCTTGCAGGCGGAGAG vvha up 5' USER GGAGACAU CGCCTTACCGTACTCTGCTG vvhA up 3' USER A CCCGGG vvha down 5' USER U CATCTTATTTTCCCTCAGATTGG A CCCGGG vvha down 3' USER U GCTCTGTTGCCTTAGCGATAAC GGGAAAGU CAATTG CCAGCGTAGACGTGAG vgrG down 5' A CCCGGG vgrG up 3' U GAATTGAGAATAATCGCATCTTTC GGGAAAGU GAAGCGGTGATCAACGATCG hcp up 5' GGAGACAU CCACCATGACCTACTTGGTTG hcp up 3' A CCCGGG T6SS up 5' U GAAAGATGCGATTATTCTCAATTC GGAGACAU CCAAGGCGAAAAACGGTGTAG T6SS up 3' A CCCGGG T6SS down 5' U GT AAACCTGCATCTGTCATTGG A CCCGGG T6SS down 3' U TCCTTATACAATATGCTCAGGAG GGGAAAGU GACTATTATCTCTGTGGCGGTC vgrG hcp down 5' ACCCGGG vgrG hcp down 3' U CTGAGATGCTCGCTATTTTGTAG GGGAAAGU CATCTGCAAGAGCGAGTGGTC T6SS int hcp 5' GCACTTTCGAACCCTCCAAATG T6SS int hcp 3' CTACAAAATAGCGAGCATCTCAG vgrG outside 5' GTTCCAGCACCGAGTTGCCC vgrG outside 3' CACCGCGCTGAAACTCTCCA hcp outside 5' CACAAAGTTAACCGCCGGACG hcp outside 3' GAAAGCCAATCAGCTGCAATCC lacZ test 5: CAAGGCGATTAAGTTGGGTAAC lacZ test 3: GACCATGATTA CGCCAAGCTC sacB5' 2 AAGTTCCTGAATTCGATTCGTCC sacB3' 2 CCTTTCGCTTGAGGTACAGCG a Restriction sites are underlined and USER cloning sequences are bolded.
50 CHAPTER 3 ROLE OF THE RtxA1 TOXIN IN PATHOGENESIS OF V ibrio vulnificus Rationale for Study During a genomic sequencing project of the V. vulnificus clinical isolate MO6 24/0 in our laboratory, a clone was identified which contained a 3.8kb insert with homology to the V. cholerae rtxA structural gene. Initial studies in the laboratory, performed by Ang ela Starks, Ph.D., indicated that RtxA1 had a major role in the cytotoxicity of V. vulnificus MO6 24/O in vitro and that it played a detectable, yet minor role in virulence. We were interested in conducting a more in depth analysis of the RtxA1 toxin in t he virulent, clinical strain CMCP6, whose genome has been completely sequenced. There was some discrepancy in the virulence the original rtxA1 ::aph mutation in strain MO6 24/O, studied previously, and other mutations constructed in strain CMCP6, discussed here. These results demonstrated that RtxA1 is not only a major cytotoxic factor, but also plays a role in the virulence of V. vulnificus CMCP6. However, the residual virulence of rtxA1 mutants demonstrated that other, as of yet unidentified, virulence factors contribute to the damage to host tissues caused by infection with V. vulnificus Introduction The RTX ( r epeats in t o x in) toxins are produced by many gram negative bacteria, and include such toxins as E. coli HlyA (70) Bordetella pertussis CyaA (71) Neisseria meningitidis FrpA (72) and V. cholerae RtxA (73) Each member of the family is characterized by a glycinerich repeated motif at the C terminal end of the protein. Generally, the RTX toxins, encoded by rtxA are secreted by a type I secretion system (T1SS), encoded by rtxBD in the same operon and tolC encoded elsewhere (74 75) An
51 acyltransferase, RtxC, is encoded upstream of rtxA and has been shown in several of RTX encoding bacteria to be essential for toxin activity (76 79) The amino acid sequence identified during the MO624/O genomic sequencing project shared homology with the V. cholerae RtxA, the proto typical member multifunctional autoprocessing RTX toxins (MARTX) (56) The genomic sequences of two other clinical isolates, CMCP6 and YJ016 (6 4) revealed that each strain encodes two rtx gene clusters on chromosome 2 and one rtx gene on chromosome 1. The first gene cluster contains five genes in a similar order to the rtx locus of V. cholerae ( 73) ( F igure 3 1 shows gene organization) The genes are grouped into two divergent operons. The first operon encodes a 15.62kb gene encoding the toxin structural protein; rtxC1 encoding an acyltransferase putatively essential for activity of the RTX toxin; and VV2_0481, encoding an uncharacterized hypothetical protein. The second operon encodes the T1SS proteins, RtxB1, RtxD1, and RtxE1. RtxB1 and RtxE1 are ABC transport proteins, and RtxD is a membrane fusion protein (80 81) Together along with tolC, located elsewhere on the chromosome, these genes encode a T1SS for export of RtxA. The second rtx gene cluster located on chromosome 2 includes a 13.96kb rtxA 2 gene and the genes rtxB 2 and rtxD 2 putatively encoding parts of the secretion system for RtxA2. This cluster lacks rtxC that would normally encode the toxin activator, suggesting that the product of rtxA2 may not become an active RTX toxin. The third locus on chromosome 1 contains only an 8.8kb gene encoding an RtxA toxin, rtxA3 and lacks the accessory secretion and activation genes.
52 MARTX toxin proteins have four important components (56) : N terminal repeats, internal enz ymatic domains, a cysteine protease domain, and the C terminal RTX repeats. The N terminal and C terminal repeats are highly conserved among the MARTX toxins ; however the internal regions are variable and are predicted to encode the cyt otoxic functions o f the toxins. Typical RTX toxins are poreforming toxins and not multifunctional, with the exception of B. pertussis CyaA, which has both poreforming ability and adenylate cyclase activity (71) The RtxA of V. cholerae the best characterized MARTX toxin, does not form pores or cause cell lysis (73 82) Instead, the RtxAvc causes rounding of cells and loss of the integrity of tight junctions of polarized cells by covalently crosslinking actin monomers into multimers (73 82 83) The actin crosslinking domain (ACD) is one of the central enzymatic domains of RtxAvc. While ACD is essential for actin crosslinking, deletion of the domain did not ablate the ability to cause cell rounding (84) This cell rounding is mediated through inactivation of small Rho GTPases leading to the downstream effect of disassembly of the actin cytoskeleton (85) The domain important for this activity is the Rho GTPase inactivation domain (RID) (85) These enzymatic domains carry out their function inside the host cell. It is predict ed that the repeats at the N terminus and C terminus play a role in translocation of the internal domains into the host cell, and the domains are released by autocleavage of itself at a cysteine protease domain (CPD) located just upstream of the C terminal repeats (56 86) Bioinformatics analyses of RtxA1 of V. vulnificus revealed that it does not have the ACD domain, a nd functional studies have shown that V. vulnificus does not induce actin
53 crosslinking (84 87) RtxA1 of V. vulnificus does encode the RID domain, essential for the inactivation of Rho GTPases, which may account for observed cell rounding that occurs prior lysis of cells (87 88) V. vulnificus RtxA1 has three additional domains of putative enzymatic activity; however, the function of these remains unknown (56) The RtxA1 toxin of V. vulnificus has become a topic of interest in recent years. Liu, et al (89) searched for CMCP6 genes controlled by the in vivoexpressed regulator HlyU and identified rtxA1 Upon further examination, they demonstrated that the CMCP6 rtxA1 mutant was noncytotoxic and was attenuated in i.p. inoculated, iron dextrantreated mice. They observed a 103fold increase in LD50 (50% lethal dose) in this mouse model and concluded that RtxA1 was a major virulence factor. Lee, et al (88) and Kim, et al (87) independently identified transposon mutants of V. vulnificus MO6 24/O with reduced cytotoxicity. These mutants had insertio ns in genes within the rtx1 locus. Both groups examined their respective MO624/O rtxA1 mutants for virulence and cytotoxicity. The rtxA1 mutant constructed by Lee, et al had a 103fold increase in i.p. LD50 in irontreated mice (88) Kim, et al (87) observed an LD50 increase of 102fold in i.p. or intragast ric (i.g.) inoculated, noniron treated mice. Each group also concluded that RtxA1 is major cytotoxic factor, although the rtxA1 mutants exhibited some residual cytotoxicity with increasing MOIs and with longer infection times (87 88) and one group determined that the residual cytotoxicity was due to the hemolysin/cytolysin, VvhA (87) Lee, et al later demonstrated that RtxA1 caused apoptosis in INT 407 cells (90) Subsequent studies provided evidence that rtxA1 is secreted by a T1SS encoded by rtxBDE at th e same locus (81)
54 All of the previous studies were performed in either i.p. or i.g. inoculated mice. These mouse models examine the ability to cause systemic disease, but do not produce a skin infection. From these studies, one proposed function in virulence is that RtxA1 was to assist the bacteria in invading through the intestinal barrier into the bloodstream (87) ; however, a function for this cytotoxic factor in tissue damage during skin infection has not yet been studied. We investigated the role of RtxA1 after s.c. inoculation in mice, which is allows us to examine the ability to cause skin infection as well as systemic infection. Our examination of RtxA1 yielded cytotoxicity results similar to those of others. However, our rtxA1 mutants were not as attenuated in the s.c. inoculated, iron dextrantreated mouse model as were the rtxA1 mutants of the other groups in the i.p inoculated mouse model. We initially thought that this discrepancy was due to the location of our original mutation within t he rtxA1 gene, and so we deleted the entire gene. However, deletion of the entire rtxA1 gene still yielded less attenuation in our mouse model than that observed by others in their mouse models. Despite their noncytotoxic phenotype, the rtxA1 mutants wer e still able to cause extensive tissue damage and lethality in s.c. inoculated mice. Results Disruption of rtxA1 by aph Insertion To examine the role of RtxA1 in the pathogenesis of V. vulnificus an insertion mutation had previously been constructed in the in the rtxA1 gene of the clinical virulent strain, MO6 24/O. The 3.8 kb rtxA1 DNA of the MO6 24/O sequencing clone described above was cloned into pCVD442 and subsequently disrupted by insertion of an aph cassette, encoding kanamycin resistance. The resulting plasmid, pGTR653, was conjugated into the sequenced, clinical strain, CMCP6, for insertion of aph into the rtxA1
55 gene. The mutants were selected on LB N containing kanamycin and verified to have undergone allelic exchange by PCR with the primers r txA test a and rtxA test b. The resulting CMCP6 mutant with the aph insertion approximately 10.6kb downstream in the rtxA1 gene is called FLA554. The rtxA1 gene is very large, making complementation in trans, and therefore fulfillment of the molecular K ochs postulates (54 55) difficult. Instead of complementation with the wildtype allele on a plasmid, the aph mutation of FLA554 was reverted back to wildtype by allelic exchange with the MO6 24/O rtxA1 se quence carried on the allelic exchange plasmid pGTR1119. The reversion of the rtxA1 :: aph mutati on in FLA554 yielded strain FLA590. Cytotoxicity of rtxA1 :: aph m utant Previous studies in the laboratory demonstrated that the RtxA1 toxin is a major cytotoxic factor of MO6 24/O. To determine the role of RtxA1 in cytotoxicity of CMCP6, in vitro characterization was carried out in tissue culture using the intestinal epithelial cell line INT 407. Two in vitro assays were used to measure the ability of the rtxA1 mutant to cause monolayer detachment/destruction and lysis of epithelial cells. A crystal violet staining assay (61) was used to assess ability of the strai ns to cause monolayer detachment/destruction of INT 407 cells. Confluent monolayers of INT 407 cells were infected with V. vulnificus at an MOI of 10 and stained as described in the Materials and Methods. Similar to the observations with MO6 24/O, there was a significant decre ase in the ability of the CMCP6 rtxA1 :: FLA554, to cause monolayer detachment/destruction as compared to the parental strain (5.6% vs. 76. 6%, P =0. 0001) (F igure 32). Reversion of the mutation back to wild type restored cytotoxi city for CMCP6 mutants ( P = 0.09)
56 The crystal violet assay detects general detachment of the monolayers, but does not indicate the method of cell death. RTX toxins are typically poreforming toxins, resulting in lysis of the host cells. Members of the M ARTX subfamily have different methods for cytotoxicity due the variability in the activity domains. The toxins insert into the membrane of host cells and translocate the internal enzymatic domains. A method used to assess lysis of host cells is by measur ement of lactate dehydrogenase (LDH) release. Infection of INT 407 monolayers was carried out exactly as for the crystal violet assay. LDH release was measured using the Cytotoxicity Detection Kit (Roche Boehringer Manneheim, Indianapolis, IN). Upon inf ection of INT 407 cells, the rtxA1 ::aph mutant, FLA554, caused a significantly lower amount of LDH to be released by the cells than the parental strain, CMPC6 (17.8% lysis vs. 71.3% lysis, P = 0.002) (data not shown). These defects in lysis were restored by reversion of the mutation back to wild type. This indicates that the RtxA1 toxin causes lysis of epithelial cells in vitro, likely due to poreformation and disruption of the membrane, as with many other RTX toxins. Due to the significant reduction of cytotoxicity caused by the mutant strain, RtxA1 is considered to be one of the major cytotoxic factors produced by V. vulnificus. Virulence of rtxA1 :: aph m utant in m ice The attenuated virulence of the rtxA1 :: aph mutant of MO6 24/O seen in previous studies in the laboratory gave indications that RtxA1 has a minor role virulence. The rtxA1 ::aph mutant of CMCP6 was examined for virulence using the iron dextrantreated, s.c. inoculated mouse model. This infection model allows a quantitative assessment of the ability of a strain to cause both localized and systemic infection, as measured by enumeration of bacteria recovered from skin lesion and liver tissues, respectively. Mice are treated with iron dextran at least 1 hour prior to infection to mimic the predisposition
57 for elevated serum iron levels in the susceptible host. Mice are then inoculated by s.c. injection of bacteria into the lower right flank Infection with FLA554, CMCP6 rtxA1 :: resulted in decreased virulence at the minimum lethal dose of 300 CFU (Figure 3 3) Five out of five mice had visible skin lesions that looked similar to wild type skin lesions (Figure 3 4) ; however only four out of five mice yielded detectable CFU from the skin, resulting in a mean of 107.4 CFU/g skin tissue s. Of five mice infected, only three had detectable CFU isolated from the liver, yielding a significantly lower level of bacteria compared to the wild typeinfected mice (102.9 CFU/g vs. 105.7 CFU/g, P = 0.0005). At a 10fold higher inoculum, 3,000 CFU, the rtx A1 mutant was able to cause wildtype levels of skin infection in all five mice infected. All of the mice had systemic infection, indicated by bacteria isolated from the liver, albeit at a lower level than the wild type infected at 300 CFU (103.7 CFU/g vs 105.7 CFU/g, P = 0.00 7 ). Infection with the reversion strain restored virulence to levels comparable to the wild type at the minimum lethal dose of 300 CFU/mouse. These results are in agreement with what has been observed for the rtxA1 mutation in MO624/O, suggesting that rtxA1 does have a minor role in virulence, because increasing the inoculum by only 10fold overcomes the attenuation of the rtxA1 mutants. Deletion of rtxA1 Our results from the rtxA1 ::aph mutant strains indicated that RtxA1 is a major cytotoxic factor of V. vulnificu s; however, it is not a major virulence factor. While our cytotoxicity data agree with other reports in the literature for V vulnificus rtxA1 other laboratories observed a larger attenuation of virulence using other rtx A1 mutations in V. vulnificus (87 89) These groups reported a 102 to 103fold increase in i.p. LD50, whereas we observed full virulence after increasing the inoculum by 10fold in our s.c.
58 injected mouse model. The aph insertion described above is approximately 10 kb into the rtxA1 open reading frame, leaving over half of the gene undisrupted. We considered the possibility that a portion of the RtxA1 protein was being produced and secreted, resulting in less attenuation than observed with the rtxA1 mutations of others. To resolve this discrepancy, we constructed a mutant strain in which the entire rtxA1 open reading frame was deleted and replaced with an aph cassette. The rtxA1 :: aph mutant, FLA943, was expected to have similar phenotypes as the rtxA1 ::aph mutant in both cytotoxicity assays and in the mouse model for virulence. Three way USER cloning was used to construct the allelic exchange plasmid to del ete the 15.6kb rtxA1 gene. The oligonucleotide pairs rtxA1 up3/ rtxA1 up5 and rtxA1 down3/ rtxA1 down5 were used to amplify 1.0kb upstream and downstream sequences of rtxA1 with the USER friendly cloning sequences complementary to the USER vector on the outside ends and the sequence ACCCGGGU on the inside ends where the two fragments were to be joined to create a common Sma I restriction site. These upstream and downstream rtxA1 fragments were cloned into the USER friendly allelic exchange vector pGTR1129 exactly as described for threeway USER cloning in the Materials and Methods, yielding pGTR272 ( rtxA1 ). To enable the selection of rtxA1 deletion mutants of V. vulnificus a blunt ended aph kanamycin resistance cassette was cloned into the Sma I site between the upstream and downstream rtxA1 ::aph. This allelic exchange plasmid was moved into V. vulnificus CMCP6 by conjugation and used to recombine the mutation into CMCP6 in the twostep sac B assisted allelic exchange process, resulting in two independent isolates, FLA900 and FLA901. The correct
59 recombination events were confirmed by PCR using the oligonucleotides RtxA1deletion up and RtxA1deletion down and Southern blot analyses. Cytotoxi city of rtxA1 :: aph m utants As a confirmation of the mutant phenotype, the rtxA1 ::aph isolates FLA900 and FLA901 were tested for cytotoxicity to INT 407 cells in vitro. The mutants had a great reduction in the ability to cause detachment/destruction of the monol ayers of INT 407 cells (1.4% for FLA900 and 11.3% FLA901; P = 0.00 01 and 0.0003, compared to wildtype, respectively) (Figure 35 ). This reduction in cytotoxicity was the equivalent to that observed for the original rtxA1 :: aph mutant. Virulence of rtxA1 : : aph m utants in m ice To examine if deletion of the entire rtxA1 gene caused greater attenuation than the rtxA1 ::aph mutation, the first rtxA1 ::aph isolate, FLA900, was inoculated into mice at the wild type minimum lethal dose of 300 CFU. At this inoculum FLA900 caused a detectable skin lesion in three of the five mice, and the mean CFU recovered from the skin tissues was 105 CFU/g skin lesion. Bacteria were not recovered from the livers of any of the mice, indicating that FLA900 was unable to cause syst emic infection at this dose. Interestingly, this experiment showed that this rtxA1 ::aph isolate was more attenuated than the rtxA1 ::aph mutant, which caused liver infection in three of five mice at this inoculum. To substantiate this result, we infect ed mice with FLA900 and a second independent rtxA1 ::aph isolate, FLA901, each a t an inoculum of 5,000 CFU. At this inoculum, we would expect that the mice would succumb to the infection with either of these mutants, similar to the infection with the original rtxA1 :: aph mutation. Surprisingly, the two rtxA1 ::aph isolates had differ ent virulence p henotypes (Figure 36 ). The first isolate, FLA900, remained attenuated at this dose, causing skin lesions in
60 three of five mice and liver infections in only two of five mice. The amount of bacteria isolated from the skin lesions caused by FLA900 was approximately 100fold lower than the amount of bacteria isolated from the skin lesions of mice infected with only 300 CFU of the wild type (105.7 CFU/g vs. 107.9 CFU/g, respectively, P = 0.0 09 ). Systemic infection by FLA900 was nearly undetect able, with a bacterial load of 102.0 CFU/g in the liver ( P = 106) Interestingly, the second rtxA1 ::aph isolate, FLA901, was virulent when inoculated at the same inoculum (5,000 CFU) at which the first isolate was attenuated. The amount of bacteria rec overed from the skin and liver was not significantly different from wildtype levels (108.4 CFU/g skin lesion and 105.5 CFU/g liver) ; however, it was significantly higher than the amount o f bacteria recovered from FLA900 ( P = 0.002 for skin; P = 106 for l iver ) Based on these results, the level of attenuation caused by the rtxA1 :: aph mutations remained questionable. Reconstruction of the rtxA1 :: aph Mutation by Chitin Induced Natural Transformation It is possible that one of the two rtxA1 :: aph isolates, most likely the more attenuated FLA900, had a secondary mutation els ewhere in the chromosome causing an additional defect in virulence. T o resolve this issue, each mutation was recreated in the wild type CMCP6 background. We took advantage of the chitininduced natural transformation of V. vulnificus to move the mutation from each rtxA1 ::aph isolate into CMCP6. Genomic DNA was extracted from each mutant and incubated with CMCP6 growing in seawater in the presence of crabshell. Transformed bacteria were selected by plating on LB N agar containing kanamycin to select for allelic ex change of aph for the rtxA1 The rtxA1 :: aph mutant recreated from more attenuated FLA900 was named FLA943, and the mutant recreated from more virulent FLA901 was named FLA923.
61 Deletion of the rtxA1 gene was confirmed by PCR, and the noncytotoxic phenoty pe of each isolate was verified (Figure 37 ). Virulence of r econstructed rtxA1 :: aph m utants Each reconstructed rtxA1 ::aph mutant was examined for virulence in mice. If the cause for the attenuation in the original FLA900 was due to a secondary mutation, then both of the mutants recreated from the original rtxA1 :: aph isolates would have the same virulence phenotype as the more virulent of the two original isolates, FLA901. Iron dextrantreated mice were inoculated with 6,000 CFU of each mutant, and FLA943 and FLA923 were similarly attenuated (Figure 3 8) FLA943 caused a visible skin lesion in four of the five mice, with a bacterial load of 107.4 CFU/g. FLA923 caused a visible infection in five out of five mice; however the bacterial yield was slightl y lower 106.9 CFU/g tissue. Each mutant caused systemic infections in three of five mice, yielding 104.4 CFU/g and 104.3 CFU/g from the livers of mice infected with FLA943 and FLA923, respectively. This was lower than what was observed for the wildtype C MCP6 at a 10fold lower inoculum. The virulence phenotypes of these two reconstructed mutants lie between the two original rtxA1 :: aph mutants, for which one was avirulent and the other was fully virulent at 5,000 CFU. Because the two mutants behaved sim ilarly, we chose to use FLA943 rtxA1 :: aph for the remainder of the studies. Attempted c omplementation of rtxA1 :: aph with the w ild t ype rtxA1 al lele The two rtxA1 mutants, rtxA1 :: aph FLA554 and rtxA1 :: aph FLA943 differed slightly in their virulence in mi ce. The rtxA1 ::aph infected every mouse at a 10fold higher inoculum than the minimum lethal dose for wildtype CMCP6. The rtxA1 ::aph mutant, FLA943, did not consistently infect every mouse when the inoculum was as high as 20fold above the wildtype mi nimum lethal dose. While two and sometimes three of five
62 mice succumbed to infection by rtxA1 :: aph with wild type levels of bacteria in the skin and liver, each time there was one or two mice without any detectable bacterial load in the liver To fulfil l the molecular version of Kochs postulates (54 55) and confirm that this defect in virulence is indeed due to the mutation, the wildtype allele must be expressed in trans to complement the mutat ion. The rtxA1 open reading frame was PCR amplified using iProof HighFidelity DNA polymerase (Bio Rad, Hercules, CA) with the oligonucleotides rtxA1 clone 5 and rtxA1 clone 3, containing the Sac I and Xba I restriction sites at the 5 ends of the oligonucleotide, respectively. The 15.6kb rtxA1 amplicon was digested with the restriction endonucleases Sac I and Xba I and cloned into the plasmid pGTR1204 (pCOS5:: lacZ USER) which had been digested with Sac I and Xba I. The rtxA1 gene was captured in a directional manner, so as to be expressed by the lacZ promoter. The resulting rtxA1 complementation plasmid, pGTR1227, was moved into the rtxA1 :: aph mutant FLA943 via conjugation. Cytotoxicity was restored to FLA943 by expressing rtxA1 in trans on pGTR1227. Cyt otoxicity to INT 407 cells was verified by both crystal violet stain to measure detachment/destruction of the monolayers and by measuring lysis by detecting LDH release (Figure 39 ). This result confirmed that the rtxA1 gene was being expressed from the plasmid and was able to form a cytotoxic RtxA1 toxin capable of killing INT 407 cells in vitro. Next, it was imperative to test the complementation strain for virulence in the iron dextrantreated mouse. Unfortunately, the complementation was not as straig htforward in vivo as it was in vitro. An initial infection with FLA943 and its complementation counterpart carrying pGTR1227 showed that the complementing plasmid was able to
63 restore virulence to the mu tant (Figure 310(a)). The complemented mutant infec ted the skin and livers of five out of five mice, and the bacterial load in the liver was restored to wild type levels (105.1 CFU/g liver). Unfortunately, this result was not reproduced in subsequent infections; sometimes virulence was restored and someti mes it was not (Figure 3 10(b)). Histopathology of s.c. lesions of rtxA1 :: aph infected mice The gross appearance of the s.c. skin lesions formed in mice infected with the rtxA1 mutants resembled those of mice infected with CMCP6. We were interested to se e if there was a difference in tissue damage at the histological level. Samples of the s.c. lesion from infected mice were fixed in 10% (v/v) buffered formalin, embedded in par affin, sections Histological sections were stained with hem atoxylin eosin. FLA943 ( rtxA1 ::aph ) caused similar damage as CMCP6 at the histological level (Figure 3 11). Bacteria were present throughout the s.c. layer, and the s.c. muscle was fragmented similar to a wild type infection. Damage extending into the dermis was also observed. It has been suggest ed that RtxA1 may have a role in evasion/killing of immune defense cells and while we observed an influx of neutrophils or polymorphonuclear cells ( PMNs ) in some sections t he majority of the PMNs observed were damaged or dead. There was also evidence of perivascular infection in the rtxA1 :: aph lesions ; however, this does not rule out the possibility of a defect in breaching the vasculature to invade into the bloodstream. Overall, d espite the abolishment of cytoto xicity in cell culture, deletion of rtxA1 had little effect on tissue d amage in mice. These observations indicate that perhaps RtxA1 has some other role
64 in virulence other than damage or killing of leukocytes, and that there are other cytotoxic factors involved in the tissue damage. Insertion of aph at 5 End of rtxA1 A sl ight disagreement remained between results obtained with the original rtxA1 ::aph mutant FLA554 and the rtxA1 ::aph mutant FLA943. Unfortunately, the failure of the complementation to fully restore virulence some of the time left this disagreement unsettle d. To substantiate the virulence results, we decided to construct yet another mutation in the rtxA1 gene. Three way USER cloning was used to construct a plasmid to insert the aph cassette 100bp into the rtxA1 open reading frame. The aph cassette (67) consists of tran scriptional termination signals in both directions and translational stop codons in every reading frame flanking an aph kanamycin resistance gene. By inserting the aph into the 5 end of rtxA1 gene, transcription and translation would be blocked and the RtxA1 protein would not be made. The resulting mutant CMCP6 rtxA1 :: named FLA904, should have behaved similarly to rtxA1 FLA943. In v itro c haracterization of rtxA1 :: aph FLA904 was examined for cytotoxicity for INT 407 cells by assessing detachment/destruction of the monolayers and LDH release, as described above. Similar to the previous rtxA1 mutants, FLA904 ( r txA1 :: aph ) was significantly reduced in its ability to cause destruction/detachment of the monolayer compared to the parent CMCP6 (11% vs. 89%, P = 0.001 ) (F igure 31 2 ). Virulence of rtxA1 :: aph in m ice FLA904 was examined for virulence using the iron dex tran treated, s.c. injected mouse model. The rtxA1 :: aph mutant was examined at the minimum lethal dose for
65 wild type CMCP6, 300 CFU, and in 10fold increme nts, 3,000 CFU and 30,000 CFU (F igure 313). At the lowest inoculum, FLA904 was attenuated compared to CMCP6. Three of five mice inoculated with the mutant at this dose had detectable skin infections, giving a mean of 105.9 CFU/g isolated from the skin lesion. Of the three mice with detectable skin infection, only one had a temperature drop below 33 C, indicatin g the mouse was moribund. This was the only mouse with bacteria isolated from liver tissue, and at 106.8 CFU/g liver, this amount of bacteria signified the potential of this mutant to cause systemic disease. We took into consideration the mi ce with no detectable CFU in the liver and assigned them a minimum detectable level, resulting in a mean liver CFU of 103.2 CFU/g liver tissue. At a 10 fold higher inoculum (3,000 CFU), FLA904 was able to cause skin lesions in all of the mice, with a mean bacterial load of 107.8 CFU/g in the skin. Two of the five mice remained healthy, with no detectable liver infection. The other three mice became moribund, resulting in a mean of 104.7 CFU/g of liver, not significantly different than wildtype CMCP6. A n increase in the inoculum to 30,000 CFU resulted in nearly full virulence, where four of the five mice became moribund and had a mean of 105.0 CFU/g in liver. Verification of the v irulence d efect of the rtxA1 :: aph m utant To fulfill the molecular version of Kochs postulates and confirm that the attenuation and decreased cytotoxicity was due to the aph insertion in rtxA1 we attempted to revert the mutation by allelic exchange with the wild type rtxA1 5 sequence. This proved to be technically challenging, and was unsuccessful. Taking into account that this mutant behaved similarly in vitro and in vivo as the rtxA1 :: aph mutant, we believe that the defect seen in each of these mutants was, in fact, due to mutation of rtxA1 mutations, and not a secondary mutation elsewhere.
66 RtxA1 Causes Apoptosis V. vulnificus induces apoptosis in host cells both in vitro and during infection of mice (90 92) Until rece ntly, no specific factors have been implicated in causing apoptosis. RtxA1 and VvhA can induce apoptosis in INT 407 cells and in HUVEC cells, respectively (45 90) The rtxA 1 mutants described above were tested for their ability to cause apoptosis in J774 murine macrophagelike cells. We measured apoptosis using the ApoONE Caspase 3/7 assay (Promega). Caspase3 and caspase7 are effector caspases activated during the process of apoptosis. The ApoONE kit utilizes a profluorescent caspase3/7 substrate, rhodamine 110 bis (N CBZ L aspartyl L glutamyl L valyl aspartic acid amide) (Z DEVD R110). Upon cleavage of t he DEVD substrate by caspase3 or caspase7, the rhodamine 110 can be detected at an excitation wavelength 498 nm and emission wavelength 521 nm. The amount of fluorescence represents the amount of caspase activity in the sample. Monolayers of J774 cells established in black 96well tissue culture plates were infected with V. vulnificus at an MOI of 10. Each strain was infected in triplicate, and cells treated with gliotoxin were used as a positive apoptotic control. Each rtxA 1 mutant induced significant ly less apoptosis in J774 cells than CMCP6 and gliotoxin (Figure 31 4 ). Whereas wildtype CMCP6 induced 76% apoptosis in J774 cells, rtxA1 :: aph FLA554 caused 39% apoptosis ( P = 0.0001), FLA904 caused 40% apoptosis ( P = 105), and rtxA1 :: aph FLA943 caused 27% apoptosis ( P = 106). Reversion of the rtxA1 :: aph mutation and complementation of the rtxA1 ::aph mutation restored apoptotic ability to these mutants. It is noteworthy that rtxA1 mutants retained the ability to cause approximately 30% apoptosis, suggesting that V. vulnificus has additional means of causing apoptosis in host cells.
67 A Combination of RtxA1 and VvhA Contributes to Cytotoxicity Based on the above results, Rtx A1 is considered to be one of the major cytotoxic factors. However, it is not the only factor contributing to cytotoxicity. Cytotoxicity of rtxA1 mutants increases with increasing MOI and increased infection times (87 88) At a MOI of 10, the rtxA1 :: aph mutant FLA904 is cytotoxic to INT 407 cells at infection times longer than one hour (Figure 315). It is possible that other cytotoxic factors are produced by V. vulnificus during infection of INT 407 cells. One intriguing protein to examine is the hemolysin/cytolysin VvhA. The role of VvhA in virulence has been a source of contention for years. Originally, it was considered to be a major hemolysin and cytotoxin. Purified VvhA is highly cytotoxic in vitro, and when injected into mice, VvhA causes skin damage similar to infection with V. vulnificus (43 44) Despite the activity of the purified toxin, a vvhA mutant of V. vulnificus is as virulent in mice as is the wild type strain (47) Construction of a d ouble m utation of rtxA1 :: aph vvhA To determine if VvhA accounts for the residual cyto to xicity observed in the rtxA1 mutants, a dou ble mutant in vvhA and rtxA 1 was created. The vvhA gene was deleted by threeway USER cloning combined with chitin based transformation to replace the vvhA gene with tetAR vvhA mutant was named FLA912. The genomic DNA from rtxA1 :: FLA904 was added to FLA912 growing in the presence of crabshell. The chitintransformed rtxA1 :: vvhA mutant was called FLA916. Cytotoxicity of rtxA1 :: vvhA FLA912, FLA904, and FLA916 were tested for cytotoxicity to INT 407 cells. The cells were infected as before, except gentamicin was added either at 1 hr or 3 hr
6 8 postinfection to allow for a longer infection times. The cells were washed and stained with crystal violet at 4 hr postinfection to assess the detachment/destruction of the monolayers (Figure 31 5 ). The rtxA1 mutant FLA904 was cytotoxic when the infection proceeded for longer than 1 hr without the addition of gentamicin (17.2% destruction at 1 hr vs. 103.8% destruction at 3 hr, P=0.0008). Deletion of vvhA had no effect on cytotoxicity compared to CMCP6 (55.8% vs. 53.9% at 1 hr, P = 0.9; 82.5% vs. 101.7% destruction at 3 hr, P = 0. 1 ). Interestingly, the double mutant FLA916 maintained reduced cytotoxicity up to 3 hr postinfection and was significantly less cytot oxic than the rtxA1 mutant alone (27.5% destruction by FLA916 vs. 103.8% destruction by FLA904, P=0.0 001 ). Infection times longer than three hours have not been examined. Therefore, VvhA contributes to the residual cytotoxicity caused by the rtxA 1 mutant s during long term infections. The cytotoxicity observed when vvhA is deleted by itself is probably due to the presence of the potent cytotoxin RtxA1. Virulence of rtxA1 :: vvhA The greater reduction in cytotoxicity of the double rtxA1 vvhA mutant over the rtxA1 mutant raised the question if the double mutant would be more attenuated in the mouse model. The rtxA1 mutants are still able to cause lethal infection at 3,000 CFU. Additionally, skin lesions caused by these mutants are as severe as ones caused by wild type infection, suggesting that, despite its cytotoxicity, RtxA1 alone does not account for the tissue damage observed during infection. Most likely a combination of toxic factors contributes to this damage; therefore, it is relevant to examine the double rtxA1 / vvhA for virulence in the mouse model. FLA904, FLA912, and FLA916 were examined for virulence in mice and the ability to cause tissue damage in the skin lesion (Figure 31 6 ). FLA912 was as virulent as
69 wild type at the low inoculum of 300 CFU (107.8 CFU/g skin and 105.5 CFU/g liver). FLA904 was slightly attenuated at 3,000 CFU, causing skin infections in all five mice (107.9 CFU/g skin) and lethal systemic infection in two of five mice (103.1 CFU/g liver). The double mutant FLA916 was not more attenuated than the FLA904 at the same inoculum of 3,000 CFU. FLA916 caused visible skin lesions in four of five mice (106.6 CFU/g skin) and systemic infections in three of five mice (104 CFU/g liver). The histopathology of skin lesions caused by these mutants indicated that there was still extensive tissue damage (Figure 317). FLA916 caused less damage in the dermis than FLA904 or FLA912; however, it was st ill able to cause tissue edema and necrosis in the subcutis, with few live PMNs present. These results indicate the presence of other factors involved in tissue damage. Prevalence of rtxA1 in V. vulnificus Previous studies in the laboratory have demonstrated that cytotoxicity does not necessarily correlate with virulence. We have a col lection of environmental and clinical V. vulnificus strains who se genotypes, virulence in mice, and cytotoxicity phenotypes have been determined. There are several strains with little or no cytotoxicity in cell culture, and m any of these strains are still virulent in mice. On the opposite end of the spectrum, we have several attenuated or avirulent strains that are as cytotoxic as CMCP6 or MO6 24/O in cell culture. Since RtxA1 is the major cytotoxic factor, we wanted to determine if the presence of RtxA1 correlated with the cytotoxicity potential of the strain. Southern blot analysis was used to determine the prevalence of rtxA1 in our collection of strains. Four digoxigeninlabeled probes (A, B, C, D) spanning the rtxA1 gene were generated. A selecti on of 41 strains encompassing clade 1 and clade 2,
70 with varying virulence and cytotoxicity phenotypes were analyzed. Surprisingly, the rtxA1 probes hybridized to all strains regardless of genetic clade, virulence, or cytotoxicity (Table 31) There was some variation in the sizes of the bands to which the probes A, B, and D hybridized, but with the exception of probe B, the sizes did not correlate to any specific group. Probe s A and D are at each end of the gene and the hybridization to different siz ed bands may be attributed to sequence variation flanking the gene. Probe B, designed to detect an internal portion of the gene upstream of the C terminal repeats, hybridized to two different sized bands in clade 1 (environmental type) and clade 2 (clinic al type) strains. In 15 of 20 clade 1 strains, probe B hybridized to a smaller band. In 13 of 14 clade 2 strains probe B hybridized to a larger band. The band sizes did not correlate with cytotoxicity or virulence potential, only with genotype. The di fference in size could be due to a few nucleotide changes resulting in an additional restriction site, or it could be indicative of different domains present in the rtxA1 gene among the V. vulnificus clades These Southern blot analyses indicate that rtxA 1 is widespread among V. vulnificus strains; however, there is some sequence variation that may influence the potential for cytotoxicity and virulence. We have simply analyzed the presence of the rtxA1 gene, and it is possible that RtxA1 is not produced or secreted at all in the less cytotoxic strains. Discussion V. vulnificus is highly cytotoxic in cell culture, and studies to identify the key cytotoxic factor(s) have been ongoing for more than twenty years. Initial reports of the cytotoxic activity of culture supernatants (42) initiated the interest in the VvhA cytolysin (43 44 4 7) and the VvpE metalloprotease (49 50, 93) Treatment of cell cult ure or
71 injection of the purified proteins in mice indicated that these were cytotoxic factors that were able to cause tissue damage similar to what is observed during infection (43 44 49 93) In contradiction to studies with the purified toxins, mutation of either one or both of the genes encoding these proteins in V. vulnificus has no effect on vi rulence in mice (47 52 53) These studies left an unanswered question: What is causing the extensive tissue damage during infection? RtxA1 is the Major Cytotoxic Factor of V. vulnificus Our laboratory first identified a gene, now known as rtxA1 during a pilot genomic sequencing project of MO624/O. Initial studies concluded that RtxA1 is one of the major cytotoxic factors of V. vulnificus MO6 24/O. We continued examining RtxA1 in the V. vulnificus strain CMCP6. An insertion of the aph cassette about twothirds into the rtxA1 gene nearly abolished cytotoxicity as it did for V. vulnificus MO6 24/O. Subsequent mutations, including an insertion in the 5 end of the rtxA1 gene and a complete deletion of the gene, emphasized the importance of RtxA1 for cytotoxicity to the intesti nal epithelial cell line INT 407 V. vulnificus causes rapid detachment/destruction of the cell monolayer within one hour postinfection, which is nearly abolished in the rtxA1 mutants. Most typical RTX toxins disrupt cell membranes and cause lysis of the host cells. The V. cholerae RtxA toxin does not form pores or cause lysis. Instead it induces cell rounding via actin crosslinking which leads to cell death ( 84) The RtxA1 of V vulnificus does not contain the ACD causing actin crosslinking but contains other domains such as the RID, which may target other signaling pathways in the host cell. Each of the rtxA1 mutants had a reduced ability to cause lysis of INT 407 cells, and these results were similar to what we observed for monolayer detachment/destruction. These results
72 indicated that RtxA1 ultimately is able to disrupt membrane permeability resulting in lysis of the cells. While the exact mechanism of c ell death by RtxA1 remains unknown, we have demonstrated that RtxA1 induces apoptosis in J774 murine macrophagelike cells (Figure 3 14) The rtxA1 mutants were defective in activating the caspase 3 and caspase 7, leading to decreased apoptosis. Lee, et al (90) also observed the ability of RtxA1 to induce apoptosis in INT 407 cells. Interestingly, RtxA1 contains a domain that with similarity to a portion of the MCF toxin of Photorhabdus luminescens The MCF toxin causes apoptosis in mammalian cells; however, it is uncertain if the portion of the toxin with similarity to the RtxA1 toxin i s necessary for the apoptotic activity (94 95) It is important to note that RtxA1 is not the only apoptotic factor of V. vulnificus Infection with rtxA1 mutants resulted in decreased apoptos is, but there was still residual apoptosis above background levels. Further examination into the other apoptotic factors is needed. RtxA1 has a Role in Virulence of V. vulnificus While we and others have definitively demonstrated that RtxA1 is the major cytotoxic factor of V. vulnificus its precise role in virulence is still uncertain. Our initial mutation by insertion of the aph cassette into rtxA1 of CMCP6 resulted in attenuation of virulence in iron dextrantreated mice at a low inoculum ; however i ncreasing the inoculum 10fold resulted in all of the mice developing systemic infection (Figure 3 3) Despite, the noncytotoxic phenotype of the mutants (Figure 3 2) the rtxA1 mutants were still able to cause visible tissue damage in the skin lesion. A t the same time, other groups reported that rtxA1 mutants caused a 102to 103fold increase in LD50 (87 89) While this difference could be due to the use of two different mouse models of infection,
73 i.p. vs. s.c., we also considered that we may not have had a complete knockout mutant in rtxA1 The insertion disrupted rtxA1 approximately 10kb downst ream in the gene, just upstream of the CPD and C terminal repeats. This meant that nearly twothirds of the protein could have been produced. If the 5 end region of rtxA1 carried domains essential for virulence, we would see less attenuation with our mutant than if the whole gene was inactivated. Therefore, we constructed two other mutations, a complete deletion ( rtxA1 ::aph ) and a 5 insertion of the aph element ( rtxA1 :: aph ). These mutants were slightly more attenuated in mice than the rtxA1 :: aph mu tant. At 3,000 CFU and 5,000 CFU (10fold to 17fold higher than the minimum lethal dose), the new rtxA1 mutants were able to cause skin infection, with lesions similar to wild type. However, the mutants were defective at causing lethal, systemic infecti on, only causing systemic infection in some of the mice (Figure s 3 8 3 10 and 31 3 ) These results indicated that RtxA1 was involved in virulence, more so than we originally thought. We attempted to fulfill the molecular Kochs postulates, by cloning t he wild type rtxA1 gene into a plasmid and expressing it in the rtxA1 ::aph mutant to complement the mutation. The complementing plasmid was able to restore cytotoxicity (Figure 3 9) but this result was not consistent The fact that it did not consistently restore virulence does not negate the previous results (Figure 3 10) We cloned only the rtxA1 gene, but not the surrounding accessory genes. It is possible that the RtxA1 expressed in trans on a plasmid may not be properly exported by the T1SS encod ed on the chromosome. In fact, not one of the other groups publishing about RtxA1 showed complementation of their mutation by expressing rtxA1 in trans during infection of mice. Kim, et al (87) were able to complement their mutant in cytotoxicity assays by
74 expressing the rtxA1 rtxC1 and rtxB1 genes in a cosmid vector, but this group did not show comple mentation of virulence in mice. What is most interesting is that the deletion of rtxA1 has no effect on tissue damage. Histopathology of the skin lesions of mice infected with the rtxA1 :: aph showed similar damage in the subcutis and dermis as that caused by wild type (Figure 3 11). We were able to observe a reduction in tissue damage in lesions of rtxA1 ::aph infected mice only when the amount of bacteria isolated from the lesion was low ( approximately 105 CFU/g skin) (not shown). This decrease in damage may be due to the reduced level of bacteria and a lower production of other factors, rather than being an effect of the absence of RtxA1. Unfortunately, we have not examined histopathology of wild type lesions yielding low levels of bacteria to compare the damage. It is clear, however, that when the infection produces high levels of bacteria in the skin the rtxA1 :: aph mutants are capable of causing tissue damage similar to wild type. Are Other Factors Involved in Cytotoxicity and Tissue Damage? While we can conclude that RtxA1 is the major cytotoxin of V. vulnificus this does not rule out presence of other accessory cytotoxins. Infection of cell culture for longer than one hour with rtxA1 mutants causes increasing cytotoxicity (Figure 3 15) (87 88) The bacteria are likely growing rapidly during the infection of cell culture so the cytotoxicity could be due to the increased number of bacteria in the well. Supporting this hypothesis is the fact that at increasing MOIs, the rtxA1 mutants become more cytotoxic (87 88) It is possible that the ex pression of other cytotoxins will eventually lead to cytotoxicity. VvhA is the hemolysin/cytolysin previously identified to be a major cytotoxic factor of V. vulnificus ; however deletion of vvhA alone did not attenuate virulence in mice or cytotoxicity i n our cell culture model, possibly because the presence
75 of RtxA1 causes such rapid cytotoxicity. Evidence supporting the role of vvhA as an accessory cytotoxin was provided by a vvhA / rtxA1 double mutant. The double mutant was noncytotoxic up to 3 hours p ostinfection in our cell culture assay (Figure 3 1 5 ) and Kim, et al (87) re ported a reduction in cytotoxicity up to 6 hours postinfection. These cytotoxicity results were intriguing, suggesting that VvhA may be contributing to the tissue damage observed in the r txA1 mutant. Upon s.c. inoculation of irontreated mice, the double mutant was able to cause systemic infection similar to the rtxA1 mutant (Figure 3 16) Despite its noncytotoxic phenotype, the double mutant was able to cause skin lesions with gross tissue damage that resembled lesions caused by the wild type. The hist opathology of the lesions revealed that the double mutant causes slight less damage in the dermis, although it is by far not at causing damage throughout the dermis (Figure 3 17). Presence of the rtxA1 Gene is Widespread Among V. vulnificus Strains The f act that Southern blot analysis revealed that the rtxA1 gene was present in all V. vulnificus strains examined, regardless of source of isolation or genotype, is very interesting. Some of these strains are less cytotoxic and some are even noncytotoxic co mpared to CMCP6 and MO624/O. We had expected that at least a portion of the rtxA1 gene would be absent in these strains ; however all four probes for rtxA1 hybri dized to all strains examined (Table 31) Additionally, not all of the strains encoding rtx A1 have the potential to cause skin infection and lethal infection in mice. This result indicates that RtxA1 is not sufficient to cause virulence that perhaps RtxA1 is not expressed in these strains during infection, or that these RtxA1 toxins are missing some critical domains necessary for virulence and/or cytotoxicity. It is also likely that
76 the presence of RtxA1 alone is not sufficient for virulence, and that these strains are missing other key virulence factors. What is the Function of RtxA1 in Vi rulence? What is interesting from these studies is that in our s.c. mouse model of infection, we see less attenuation than that observed during i.p. inoculation of mice. During s.c. inoculation, the bacteria must be able to establish a local skin infect ion, and then invade through the tissues to reach s ystemic sites. Our rtxA1 mutants are able to establish a local infection and cause tissue damage, indicating that there are other factors involved in damage. The defect of the rtxA1 mutants, both in our s.c. mouse model and in the i.p. mouse model, is that they are not able to cause lethal infection. Kim, et al. (87) demonstrated a defect in the rtxA1 mutants invading into the bloodstream after inoculation into ligated ileal loops. This result suggested that the RtxA1 toxin aids in invasion through the intestinal wall. This inv asion defect could explain our results in the s.c. mouse. While rtxA1 mutants are still able to cause establish a local infection capable of causing damage in the s.c. tissue, perhaps they cannot cause the vasodilation or damage the vasculature that may be necessary to invade into the bloodstream. Another possible role for RtxA1 may be in evading/killing host immune defenses. We have demonstrated that the presence RtxA1 is induces apoptosis in J774 macrophagelike cells (Figure 3 14) We have previously shown that very lit tle intact PMN response occurs during s.c inoculation of mice (28) and there are of RTX toxins targeting leukocytes (96) However, the histopathology of skin lesions of rtxA1 :: aph infected mice show destruction of the subcutaneous tissue and evidence of PMN killing occurring, resembling wildtype infections, suggesting that RtxA1 is not targeting PMNs. Finally, we must not disregard that RtxA1 may not have a role in tis sue
77 damage. The presence of other rtxA genes and other putative destructive toxins suggests that a combination of these factors most likely necessary to cause destruction, and deletion of one of the factors may not be sufficient to eliminate damage.
78 Figure 31. Schematic of rtx gene clusters of Vibrio vulnificus strain CMCP6. The prototypical rtx operon from E. coli hlyA is shown at the top, with a ruler for size comparison. In CMCP6, t here are two rtx gene clusters on chromosome II and one on chromosome I. rtxA encodes the RtxA toxin protein, rtxC encodes an acyltransferase, rtxB and rtxE encode ABC transport proteins and rtxD encodes a membrane fusion protein.
79 Figure 32. Detachment/d estruction of INT 407 monolayers by rtxA1 ::aph mutant Confluent INT 407 monolayers were infected with wildtype, the rtxA1 ::aph mutant FLA554, or the rtxA1 ::aph reversion FLA590. Each strain was infected in triplicate for one hour prior to treatment with gentamicin. After overnight incubation, the attached cells were stained with crystal violet, and the percent of destruction/ detachment caused by infection was calculated as described in Materials and Methods. FLA554 had significantly reduced cytotoxicity compared to wildtype. Cytotoxicity was restored by reversion with the wild type allele ( P = 0.09, reversion compared to wild type ) *, P = 0.0001; ** P =0.000 5 by Fishers LSD for difference in mean % destruction by rtxA1 ::aph compared to wild type or the reversion.
80 Figure 33. Virulence of rtxA1 ::ap h mutant in iron dextrantreated mice. Mice were inoculated subcutaneously with either 300 CFU of the wildtype CMCP6, 300 or 3,000 CFU of the rtxA1 ::aph mutant FLA554, or 300 CFU of the rtxA1 ::aph reversion FLA590. Mice were euthanized when temperatures dropped below 33C or at 23 hours postinoculation, and samples of the skin lesion and liver were homogenized for quantification of bacteria. All mice had skin lesions Fractions beneath the bars indicate the numbers of mice that yielded detectable numbers of bacteria from the skin or liver samples over the number of inoculated mice. Bars without fractions indicate that bacteria were recovered from 5/5 mice. FLA554 was attenuated for systemic infection in mice compared to the wild type. Virulence was r estored by reversion with the wild type allele ( P = 0.58, reversion vs. CMCP6) Asterisks indicate statistical significance of CFU/g liver tissue in mutant infections by Fishers LSD (*, P = 0.00 7 ; **, P = 0.002; ***, P = 0.0005; ****, P = 0.0002).
81 A B C Figure 34. Skin lesions of mice infected with rtxA ::aph After s.c. inoculated mice were euthanized, the skin was peeled back to reveal the s.c. lesions. Lesions of mice infected with 300 CFU of (A) CMCP6, (B) FLA554 ( rtxA1 :: aph ), or (C) FLA590 ( r txA1 :: aph reversion ) are shown. There was no visible difference in lesions caused by FLA554 and wild type CMCP6.
82 Figure 35 Detachment/d estruction of INT 407 monolayers by the rtxA1 :: aph mutant. Confluent INT 407 monolayers were infected with wildtype, the rtxA1 ::aph rtxA1 isolates, FLA900 (#1) or FLA901 (#2) After infection, the attached cells were stained with crystal violet and % destruction/detac hment was calculated. Deletion of rtxA1 caused a similar reduction in cytotoxicity as the mutation by aph insertion. Mutant strains were not significantly different from each other ( P 0.5). Asterisks in dicate statistical significance of mean % destruc tion by mutant strains compared to wild type by Fishers LSD ( *, P = 0.0 0 03; **, P = 0.00 02 ; ***, P = 0.0001).
83 Figure 36 Virulence of rtxA1 ::aph mutants in mice Iron dextran treated mice were inoculated s.c. with 300 CFU of wildtype CMCP6, or 5,000 CFU of either FLA900 rtxA1 #1) or FLA901 rtxA1 #2 ) The two rtxA1 isolates had different virulence phenotypes Fractions beneath the bars indicate the proportion of samples that yielded bacteria. Bars without fractions indicate that bacteria were recovered from 5/5 mice. Samples were not taken from mice with no visible skin lesion, and the minimum detectable CFU/g was used for these mice for statistical analysis (104 CFU/g skin and 102 .5 CFU/g liver). Asterisks indicate statistical significa nce of CFU/g tis sue or temperature rtxA1 infections compared to wildtype infections or each other by a Fishers LSD (*, P = 0.0 5 ; **, P = 0.0 1 ; ***, P = 0.002; *** *, P = 0.0004; *** P = 106). A dagger indicates statistical significance of number of samples yielding bacteria in the mutant infection compared to wild2 test (, P = 0.04).
84 Figure 37 Detachment/destruction of INT 407 cell monolayers by chitin recreated rtxA1 mutants Confluent INT 407 monolayers were infected with wildtype or the recrea rtxA1 mutants, FLA943 (#1) or FLA923 (#2) After infection, the attached cells were stained with crystal violet and % destruction/detachment was calculated. The chitin recreated rtxA1 mutants had reduced cytotoxicity *, P = 0.0 05; **, P = 0.00 3 b y Fishers LSD for difference in mean % destruction by mutant strains compared to wild type
85 Figure 38 Virulence of rtxA1 ::aph mutants re created by chitin transformation. Iron dextrantreated mice were s.c. rtx A1 #1 rtxA1 #2 chitin recreated mutant FLA923. The two chitinrecreated had similar virulence in mice ( P > 0.7) Fractions beneath the bars indicate the proportion of samples that yielded bacteria. Bars without fracti o ns indicate bacter ia were recovered from 5/5 mice. Samples were not taken from mice with no visible skin lesion, and the minimum detectable CFU/g was used for these mice for statistical analysis.
86 A B Figure 3 9 Cytotoxicity to INT 407 cells by rtxA1 ::aph complemented with wild type rtxA1 in trans INT 407 monolayers were infected with either wildtype rtxA1 mutant FLA943, or FLA943 carrying the complementation plasmid pGTR1227. Cytotoxicity was restored to FLA943 by expressing rtx A1 in trans on pGTR1227. (A) The amount of detachment/ destruction of the cell monolayers was determined by crystal violet assay (B) Lysis of INT 407 cells was measured by LDH release. Asterisks indicate statistical significance by Fishers LSD comparin g rtxA1 to CMCP6 or to complemented mutant (*, P = 0.003; **, P 6) The complemented mutant was not significant ly different from CMCP6 ( P > 0.7).
87 A B Figure 310. Complementation of virulence of rtxA1 :: aph by expressing wild type rtxA1 in tra ns on pGTR1227. Iron dextrantreated mice were inoculated with rtxA1 mutant FLA943, or FLA943 complemented with pGTR1227. Fractions beneath the bars indicate the proportion of samples that yielded bacteri a. (A) Virulence was restored in the initial test of complemented mutant. *, P = 0. 0 3 ; **, P = 0.0 2 ; ***, P = 0. 0 002; ****, P < 0. 0 001 by Fishers LSD rtxA1 to CMCP6 or to the complemented mutant The complemented mutant was not significantly different from CMCP6 (B ) Virulence was not restored in a separate experiment *, P = 0. 0 07; **, P = 0.005 ; *, P < 0.0002 by Fishers LSD comparing rtxA1 or complemented mutant to wildtype. Daggers indicate statistical significance of numbers of samples yielding bacteria in the mutant or complemented infection compared to wild2 test (, P = 0.04 ; P = 0.01).
88 Figure 3rtxA1 infected mic e. Mice were s.c. rtxA1 ). After infection, tissue samples were collected, fixed in buffered formalin, embedded in paraffin, and cut into 5m sections. Sections were stained with hematoxylin and eosin. Magnifica tions, x100 (A through C) and x400 (D through E). (A) Uninfected mouse skin. Epidermis (e), dermis (d), and subcutis (s) with adipocytes, blood vessels, and muscle layer. (B) Skin of mouse infected with CMCP6. Extensive edema throughout subcutis layer (*), and edema and necrosis extending into dermis (arrow). Fragmentation and destruction of muscle layer rtxA1 ). Severe infection and edema in subcutis (*), with damage extending into dermis (arrow). Muscle layer is fragmented, similar to the CMCP6 infecte d mouse (arrowhead). (D) CMCP6 infected mous e skin showing very few live neutrophils (PMNs) present (arrowhead). Most PMNs are necrotic or degenerated (arrows). (E) Skin of mouse infected with FLA943. More PMNs are present in subcutis (arrowheads); however, many of them are dying or necrotic (arr ows). (F) Perivascular infection caused by FLA943. Clot forming within blood vessel (arrow). Staining indicates suggests presence of bacteria surrounding the blood vessel (*).
89 Figure 31 2 Detachment/destruction of INT 407 monolayers by rtxA1 :: mutant. INT rtxA1 mutant FLA901, or the rtxA1 :: mutant FLA904. The amount of destruction of the cell monolayers was determined by staining with crystal violet. FLA904 had a similar reduction in cyto toxicity as FLA901, and the two strains were not significantly different from each other Asterisks indicate statistical significance by Fishers LSD comparing FLA901 or FLA904 to wild type (*, P = 0.00 2 ; **, P = 0.00 1 ).
90 Figure 31 3 Virulence of the rtxA1 :: aph mutant at increasing inocula. Iron dextrantreated mice were inoculated s.c. with wild type CMCP6 or rtxA1 :: aph FLA904 at inocula of 300 CFU, 3,000 CFU, or 30,000 CFU. Virulence of FLA904 increased at higher inocula. Fractions beneath the bars i ndicate the proportion of samples that yielded bacteria. B ars without fractions indicate that bacteria were recovered from 5/5 mice. Samples were not taken from mice with no visible skin lesion, and the minimum detectable CFU/g was used for these mice for statistical analysis *, P = 0.03 by Fishers LSD comparing FLA904 ( 300 CFU ) to CMCP6 or FLA904 (3,000 CFU) and **, P = 0.0 1 by Fishers LSD comparing FLA904 (300 CFU) to FLA904 (30,000 CFU). FLA904 inoculated at 3,000 CFU or 30,000 CFU caused no signi ficant difference in CFU/g of skin compared to wild type. There was no significant difference in liver or systemic infection caused by FLA904 compared to wild type.
91 Figure 31 4 Apoptosis of J774 cells infected with rtxA1 mutants. J774 monolayers seeded in black 96well tissue culture plates were infected with either wildtype CMCP6, FLA554 ( rtxA1 ::aph ), FLA590 ( rtxA1 :: aph reversion) FLA904 ( rtxA1 :: aph ) FLA943 ( rtxA1 :: aph) or FLA943(pGTR1227) ( complemented rtxA1 ). Caspase3/7 activity was measured using the ApoONE Homogeneous Caspase3/7 assay. % apoptosis was calculated by dividing the normalized relative fluorescence units (RFU) for each sa mple by the mean normalized RFU for gliotoxin treated well s (considered 100% apoptosis). The rtxA1 mutants induced less apoptosis than the wild type and gliotoxin (positive control). The ability to cause apoptosis was restored by reversion or complementation FLA590 caused more apoptosis than CMCP6 ( P 0.000 1 ); however, the complement ed FLA943 (pGTR1227) caused significantly lower apoptosis than CMCP6 ( P = 0.0 0 2 ) Asterisks indicate statistical significance by Fishers LSD comparing each strain to wild type or to the strain indicated by lines (*, P = 0.0 0 2 ; **, P 0.0001 ; ***, P < 105; ****, P < 106).
92 Figure 31 5 Detachment/destruction of INT 407 monolayers infected with rtxA1 vvhA or rtxA1/vvhA mutants. INT 407 monolayers were infected with wild type CMCP6, the rtxA 1 :: aph mutant FLA904, t vvhA mutant FLA912, or the double rtxA1 / vvhA mutant FLA916 at an MOI of 10. Each strain was infected in triplicate. Gentamicin was added either 1 hour or 3 hours postinfection, and monolayer detachment/ destruction was assessed by crystal violet stain ing at 4 hours postinfection. FLA904 wa s cytotoxic after 3 hours. This cytotoxicity wa s reduced upon inactivation of rtxA1 and vvhA Asterisks indicate statistical significance by Fishers LSD of mutant to either wildtype or the strain indicated by lines (*, P = 0. 0 2 ; ** P = 0. 001 ; ***, P = 0.0 001 ).
93 Figure 31 6 Virulence of double rtxA 1 ::vvhA ::tet mutant. Iron dextrantreated mice were inoculated s .c. with either 300 CFU of wild type CMCP6, 3,000 CFU of rtxA 1:: aph mutant FLA904, 3,000 CFU of rtxA 1:: aph vvhA mutant vvhA mutant FLA912. FLA916 wa s not more a ttenuated than FLA904, indicating that vvhA was not necessary for virulence. Fractions beneath the bars indicate the proportion of samples that yielded bacteria. Bars without fractions indicate that bacteria were recovered from 5/5 mice. Samples were not taken from mice with no visible skin lesion, and the minimum detectable CFU/g was used for these mice for statistical analysis. Statistical significance is calculated by a Fishers LSD (*, P = 0.0 1 for FLA904 to CMCP6 ; **, P = 0.02 for FLA916 to FLA912)
94 Figure 3vvhA and rtxA1 :: aph mutants. vvhA ), FLA904 ( rtxA1 :: aph ), or FLA916 ( rtxA1 :: aph vvhA ) After infection, tissue samples were col lected, fixed in buffered formalin, embedded in paraffin, and cut into 5m sections. Sections were stained with hematoxylin and eosin. Magnification x100 (A) Skin of mouse infected with FLA912 Edema and destruction of the tissues through the subcutis and extending into the dermis (*). Heavy staining of subcutis beneath muscle layer indicates massive bacterial infection in this region (arrow). (B) Skin of mouse infected with FLA904. Extensive destruction in the subcutis layer, including fragmentation of the muscle layer indicated by an arrows. Necrosis extending into dermis indicated by an arrow head. (C) Skin of mouse infected with FLA916. I nfection and edema in subcutis ( arrow ) with damage to the muscle (arrow). Dermis was mostly intact, with very little edema (arrowhead). The double rtxA1 :: aph vvhA mutant was still able to cause significant tissue damage, although damage was less severe in the dermis than the rtxA1 :: aph mutant.
95 Table 31. Southern blot to detect rtxA1 in V. vulnificus isolates Strain 16s rRNA RAPD MLST Skin CFU Live r CFU Cytotoxicity (%) Probe A Probe B Probe C Probe D CMCP6 B C 2 7.2 4.3 93.9 + 2 + + FLA101 AB E 1 8.0 4.9 62.9 + 1 + + FLA102 B C 2 8.3 7.3 59.1 + 2 + + FLA104 A E 1 6.8 3.4 100.0 + 2 + + FLA105 B C 2 5.9 2.6 96.0 + 2 + + FLA106 AB E 1 6.1 3.2 95 .7 + 1 + + FLA107 A E 1 8.3 4.7 43.6 + 2 + + FLA108 A E ND 6.8 2.5 100.0 + 1 + + FLA109 AB E 1 6.0 3.7 99.0 + 1 + + FLA111 A E 1 7.2 3.0 100.0 + 1 ND ND FLA112 B C 2 6.1 3.4 100.0 + 2 + + FLA113 B C 2 5.9 3.7 100.0 + ND + + FLA114 AB E 1 8.1 6.7 10 0.0 + 2 + ND FLA115 AB E 1 8.4 4.5 100.0 + 1 ND ND FLA116 B C 2 8.4 3.3 62.9 + 2 ND ND FLA117 B C 2 8.5 5.4 99.7 + 2 + + FLA118 A E 1 4.6 2.5 100.0 + ND ND ND FLA119 A E 1 6.0 2.5 38.1 ND 1 + + FLA120 B E 1 7.3 2.5 85.0 + 1 + + FLA121 A E 1 6.3 2.5 90.3 ND 1 + + FLA122 A E 1 7.0 2.6 2.0 ND 1 + + FLA125 A E 1 8.3 5.4 100.0 + ND ND ND FLA128 A E 1 7.7 2.5 87.9 + 1 + + FLA136 B C 2 8.1 3.5 66.7 + 2 + + FLA137 AB E 1 7.9 2.5 100.0 + 2 + + FLA139 B C 2 4.9 2.8 99.9 ND 2 + + FLA140 B C 2 5.9 2.5 91. 4 ND ND ND ND FLA141 AB E 1 6.9 2.7 20.8 + 1 + + FLA142 A E 1 7.9 2.8 82.3 ND 2 ND ND FLA143 AB E 1 6.5 2.5 98.5 ND ND ND ND FLA144 AB E 1 7.8 5.7 39.8 + 1 + + FLA145 A E 1 7.7 3.6 41.4 + 1 ND ND FLA146 B C 2 6.1 2.5 30.6 + 2 + + FLA147 B C 2 7.9 5. 3 26.4 + 2 + + FLA148 A E ND 6.7 4.8 22.8 + 1 ND ND MO6 24/O B C 2 98.0 + 2 + + LL728 B C 7.5 4.8 37.3 + 2 + + MLT367 B C 3.5 2.5 84.4 ND 1 ND ND YJ016 B C 2 7.7 5.7 ND + 2 + + 302/99 AB E ND 7.8 5.7 91.3 ND 2 ND ND 313/98 AB E ND 7.0 2. 8 3.1 ND 2 ND ND +, hybridization; ND, No data/ probe not tested; 1, hybridization to small band; 2, hybridization to large band.
96 CHAPTER 4 THE ROLE OF THE OTHER RTX LOCI IN VIRULENCE OF V. vulnificus AND THE IMPORTANCE OF ACTIVATION OF THE RtxA TOXINS Rationale for Study Deletion of rtxA1 nearly abolishes cytotoxicity; however, the RtxA1 toxin does not appear to be as important for virulence in the s.c. model of infection as it is for virulence after i.p. inoculation studied by other laboratories (87 88) I n the s.c. injected mouse model, t he rtxA1 mutants are still able to cause skin lesions comp arable to the wildtype and they are able to cause systemic infection with an inoculum of 3,000 CFU ( Figure s 3 7 and 311 ) To provide a more complete analysis of the RTX toxins of V. v ulnificus we wanted to examine the other two RTX like proteins, RtxA2 and RtxA3, as well as the necessity of RtxC 1 for activity of the RtxA toxins. There was a possibility that the RtxA toxins encoded at these other loci could account for residual virulence in the absence of RtxA1. We investigated the role of RtxC1 and the putative RTX toxins through mutagenesis, followed by cell culture analysis and in vivo infections. Introduction V. vulnificus encodes three rtx gene loci. The RtxA1 toxin proved to be the major cytotoxic factor in vitro, and it contributes to virulence i n s.c. inoculated, iron dextrantreated mice. This locus encodes a complete MARTX system (56) (Figure 3 1). Upstream in the rtxA1 operon is rtxC1 annotated as a h emolysin acyltransferase. RtxC1 has 96% identity to the V cholerae RtxC and 32% identity to the HlyC of E. coli HlyC activates the E. coli HlyA protein by posttranslational fatty acylation on lysine residues of the HlyA (77) Other previously studied RTX toxins also require activation by acylation of the RtxA protein (75 76 78) but this has not been shown for V. cholerae RtxA.
97 The other two rtx loci are incomplete by consensus RTX standards. The second locus encompasses the 14kb gene VV2_151 4 encoding RtxA2, a putative RTX toxin. Included in this locus are genes encoding RtxB2 and RtxD2, which are annotated to be involved in secretion of RtxA2. The third rtx3 locus includes only the 8.8kb gene VV1_2715 encoding an RTX like protein, RtxA3, which at 2,937 amino acids is smaller than RtxA1 and RtxA2. To address if there is redundancy in the functions of the RtxA toxins, we constructed an insertional mutation in the rtxA 2 gene and deleted the rtxA3 gene. The mutations either were combined or are in the process of being combined with rtxA1 mutations to produce double and triple rtxA mutants. Given that rtxA1 mutants alone had a vast reduction in cytotoxicity to INT 407 cells we did not expect mutations in rtxA 2 and rtxA 3 to have an effect on c ytotoxicity. However, there remained the possibility that the RTX toxins act in concert and that mutation of any one of them could affect cytotoxicity or virulence. If the RtxA1 toxin is not the only RTX toxin contributing to virulence, we would observe attenuation of the rtxA 2 and rtxA 3 mutants in the s.c. inoculated mouse. We also investigated the role of the RtxC1 acyltransferase in activation of the RtxA toxins by deletion of the rtxC1 gene. Typical RtxA toxins are activated by the RtxC enzyme encoded in the same operon, so we anticipated that RtxC1 would be necessary for RtxA1 activity and probably not involved in activation of the other two RtxA toxins. However, because there are no rtxC homologs in the regions surrounding rtxA2 and rtxA3 we consi dered the possibilities that these RTX toxins were either not active RTX toxins, or that the RtxC1 in the rtx1 locus could also activate these RTX toxins.
98 Results Identification and Examination of RtxA2 Upon examining the published genomic sequence of CMCP 6, we were immediately interested in the gene VV2_1514 due to its large size comparable to the VV2_0479 rtxA1 gene. The VV2_1514 gene is 14kb, about 1.6 kb smaller than the rtxA1 gene, which is the largest known gene of V. vulnificus According to the N CBI Conserved Domain search, glycinerich peptide repeats similar to RTX repeats are located in the C terminus of the RtxA2 protein. The repeat region and the large size of the protein indicated that it may be a member the RTX family. Just downstream of rtxA2 are the putative secretion genes VV2_1515, annotated as a toxin secretion ATP binding protein, and VV2_1516, annotated as an HlyD family protein. BLAST analysis revealed that the amino acid sequences of the two proteins have approximately 25% identi ty to the RtxB1 and RtxD1 proteins, respectively. The similarities with these and other T1SS proteins provide support that the products of these genes should be called RtxB2 and RtxD2, and that these proteins are components of a secretion system for RtxA2 Despite the evidence that the rtx2 locus encodes a putative RtxA2 toxin secreted by a T1SS, BLAST analysis revealed very low sequence similarity between the amino acid sequence of RtxA2 and either the V. vulnificus RtxA1 or V. cholerae RtxA. While RtxA2 contains RTX like repeats in the C terminus, the N terminal repeats of the MARTX family have not yet been identified in RtxA2. The lack of similarity our inability to identify the conserved N terminal repeats, and the absence of nearby accessory genes r txC and rtxH support that the RtxA2 may not be a true member of the MARTX
99 sub family. However, RtxA2 could still be a member of the RTX family, and therefore warranted study as a potential virulence factor. Mutation of rtxA2 to Examine a Role in Virulence Construction of rtxA2 :: A mutation in gene VV2_1514, rtxA2 was constructed by insertion of the translational stop sequences of the element (67) For insertion of the into the 5 end of the rtxA2 gene a 1 kb fragment was PCR amplified using the oligonucleotides rtxA 2 A and rtxA2 B and cloned into pCR2.1 by TOPOTA Cloning (Invitrogen) to create pGTR256 aph element was excised from pBR322: aph by BamH I digestion and subcloned at a unique Ssp I restriction site within the rtxA2 fragment. The rtxA 2 aph was exci sed using restriction endonucleases Xba I and Sst I and was subcloned into pGTR260 (pUC19 with the Hind III restriction site destroyed). The resulting plasmid was named pGTR261. Because one of the ultimate goals was to combine the rtxA2 mutation with the rt xA1 mutation selected by kanamycin resistance, the kanamycin resistance cassette of the aph needed to be removed. The transcriptional stop signal and the kanamycin resistance cartridge were removed by Hind III digestion, gel extraction, and religation of the plasmid to yield pGTR262, carrying the 1kb rtxA 2 region interrupted by 35bp of the reading frames. The rtxA2 fragment was excised by Xba I and Sst I and inserted into the allel ic exchange vector pGTR1122 (pCVD442:: ) yielding pGTR263. The allelic exchange vector was moved into V. vulnificus FLA399 ( a spontaneous rifampicin resistant isolate of CMCP6) via conjugation. rtxA2 mutants were selected by the sacB assisted allelic exchange proces s, and insertion of the 35 bp sequence was
100 confirmed by PCR using the primers rtxA2A and rtxA2 B. The FLA399rtxA2 mutant was named FLA 441 Cytotoxicity of the rtxa2 :: mutant The RtxA1 toxin of V. vulnificus has an obvious role in cytotoxicity, and the large reduction observed during infection of cells with rtxA1 mutants implied that the other RTX toxins produced by V. vulnificus do not have the same level of cytotoxicity as Rtxa1 in this cell culture model. However, if all of the RTX toxins are essential for cytotoxicity, then deleting any one of them would abolish cytotoxicity. Therefore, it was possible that the rtxA2 mutant would have a reduction in cytotoxicity to INT 407 cells. INT 407 monolayers were seeded in 24well plates and infected with either wild type FLA399 or the rtxA2 mutant FLA441. Each st rain was infected in triplicate, and uninfected wells and wells containing media alone served as positive and negative controls respectively FLA399 was used as a control for a less cytotoxic strain. It should be noted that FLA399rtxA1 :: aph behaves similar ly to FLA554 (C MCP6 rtxA1 ::aph ) during infection of cell culture and mice. Interestingly mutation of the rtxA2 gene caused no reduction in the detachment or destruction of the monolayers compared to FLA399 ( 77.8% vs. 82.6 %, respectively, P =0. 6 ) whereas the was less cytotoxic (16.4%, P = 0.0002 compared to FLA399) (Figure 4 1) These results implied that RtxA1 was the main RTX toxin causing cytotoxicity in INT 407. Virulence of the rtxA2 :: mutant in irontreated mice The iron dextrantreated mouse model was used to assess the virulence of the rtxA2 mutant. S.c. inoculations of 300 CFU of FLA441 caused all mice to become moribund. The levels of bacteria isolated from the skin lesion (108 CFU/g) and the liver (105 CFU/g) were comparable to what is observed with the wild type (Figure42).
101 Therefore, RtxA2 is not necessary for virulence, but may be a redundant factor to RtxA1. If this is the case, construction a double mutation in rtxA1 and rtxA2 would be more attenuated than the single rtxA mutants. Double Mutation of rtx A 1 and rtxA2 To investigate the role of RtxA2 as an accessory toxin involved in virulence, a double mutant with the rtxA1 mutation combined with the r txA2 :: constructed. We constructed a double mutant by moving the pGTR263 ( r allelic exchange plasmid) into FLA399rtxA1 :: aph by conjugation, followed by sacB assisted allelic exchang e The resulting rtxA1 :: aph double, named FLA439, was attenuated in cytotoxicity similar to the rtxA1 :: aph mutant (data not shown), confirming that RtxA2 was not involved in cytotoxicity. The mutant was similarly attenuated in virulence, similar ly to the rtxA1 ::aph mutant; however, there was experimental variation in virulence. Due to concerns over the vir ulence of the FLA399 background in relation to other mutant strains in the laboratory, we discontinued the us e of mutants created in FLA399 and reconstructed all of the mutations in the wildtype CMCP6 background. The CMCP6 named FLA554, is discussed in detail in Chapter 3. The rtxA2 :: mutant of CMCP6, named FLA561, behaved similar ly to FLA441 (FLA399rtxA2 :: in cell culture The do uble mutant was constructed as before and the resulting strain, FLA558, was examined. S imilar to results from the mutants in the FLA399 background, RtxA2 was not necessary for cytotoxicity (Figure 43) The double mutant FLA558 caused 13% destruction of INT 407 monolayers, which was not significantly different from the 9%
102 destruction caused by the rtxA1 mutant FLA554. As expected, CMCP6 and the rtxa2 :: FLA561 were highly cytotoxic to INT 407 cells. Virulence of double rtxA1 / rtxA2 mutant in mice FLA554 ( rtxA :: aph) or FLA558 ( rtxA2 :: ) were s.c. inoculated into iron dextrantreated mice to determine if the double mutant was more attenuated than the single mutant (Figure 4 4). When inoculated at 300 CFU, FLA554 caused skin lesions in four of four mice and systemic infection in two of four mice. Means of 107.4 CFU/g and 103.3 CFU/g of bacteria were isolated from the skin and liver, respecti vely. The double mutant, FLA558, was more attenuated than the FLA554 mutant at this inoculum. Of the five mice infected, three had a detectable skin infection, but only one of these mice had a wildtype skin lesion with 108 CFU/g. The other two mice had small skin lesions with very low levels of bacteria in the skin lesion (103.8 CFU/g and 103.3 CFU/g). Not one of the five mice infected with FLA558 had a detectable infection of the liver, indicating that the double mutant was not able to cause systemic infection at an inoculum of 300 CFU (Figure 44 (A)). Based on these results, RtxA2 appeared to be contributing to the initial local infection in the skin lesion, since only one of the mice had a wild type skin infection. If the double mutant is less abl e to cause a local infection, then it would also be unable to cause a systemic infection, which is the result we observed in this experiment. We increased the inoculum of FLA554 and FLA558 to 3,000 CFU (Figure 44 (B)) to further test the level of attenuation of these mutants. The results of this higher inoculum reflected the inconsistency in the virulence of the rtxA1 mutant. Four of the five FLA554infected mice developed skin lesions similar to the wildtype ( 107.1 CFU/g of skin); however, only one o f these mice developed a detectable systemic infection,
103 although much lower (102.6 CFU/g of liver) than what is observed during a typical wildtype infection (approximately 105 CFU/g). In contrast to results from infection with 300 CFU, infection with 3,0 00 CFU caused the double mutant FLA558 to be slightly more virulent than FLA554. Every mouse infected with FLA558 developed a skin lesion containing a high level of bacteria (107.9 CFU/g of skin). Two of the five mice became moribund, with temperatures below 33C and a high amount of bacteria in their livers (104.5 CFU/g and 103.5 CFU/g). The other three mice had no detectable liver infection. The mean liver infection of FLA558infected mice was not significantly different from the liver infection in FL A554 infected mice (103.0 CFU/g vs. 102.3 CFU/g, P = 0.15). Therefore, i noculation of a higher dose of the mutants gave the impression that RtxA2 is disposable for virulence in mice. Construction of rtxA1 :: aph, rtxA2 :: The inconsistency in the virulence of the rtxA 1 ::aph mutant at low and high inocula could be a reason for the attenuation observed for the double mutant at 300 CFU. The attenuation may not be due to the rtxA2 mutation; instead, it is more likely due to experimental fluctuation observed with rtxA1 :: aph. A fter examining the other rtxA1 mutant constructs, discussed in Chapter 3, we felt that the deletion of rtxA1 provided a more consistent virulence phenotype in mice. Therefore, the del etion was a better rtxA1 mutant background in which to study the effect of the rtxA2 mutation. Genomic DNA of the rtxA1 :: aph mutant FLA943 was used to move the rtxA1 ::aph mutation into FLA561 via chitininduced natural transformation. Transformants were selected on LB N with kanamycin and verified by PCR. The :: mutant strain was named FLA947.
104 Cytotoxicity of The mutant was first examined for its ability to cause lysis in INT 407 cells (Figure 4 5 ). Infection with FLA947 caused 12.4% LDH release, which was not significantly lower than 16.4% LDH release by cells infected w ith the rtxA1 mutant FLA943 ( P = 0.3 ). Therefore, as we expected from the previous results RtxA2 was not contributing to cytotoxicity to INT 407 cells. Virulence of rtxA1 ::aph, rtxA2 :: We next examined the double mutant for virulence in mice. FLA943 t ypically caused infection in some of the mice at an inoculum of 3,000 CFU, and so if FLA947 was more attenuated than the FLA943, we would observe complete attenuation at this inoculum. In agreement with the double mutant FLA558 described above, the FLA947 was not more attenuated than the rtxA1 ::aph FLA943 (Figure 46 ). At an inoculum of 3,000 CFU, FLA943 caused detectable skin and liver infection in three of five mice, with mean yields of 106.6 CFU/g from skin and 103.1 CFU/g from liver. FLA947 was able to cause detectable infection in five of five mice, although the amount of bacteria recovered was not significantly higher than the amount recovered from FLA943 (107.7 CFU/g skin, P = 0.3 ; a nd 104.3 CFU/g liver, P = 0.1). Although the double mutant FLA947 caused skin lesions resembling t hose of wild type infections it would be interesting to examine the histopathology of the skin lesions of mice infected with the rtxA1 :: aph rtxA2 :: mutant and compare it to the lesions of mice infected with or mutants. This would reveal if ther e are any differences in damage caused in the skin lesion.
105 Identification and Examination of RtxA 3 In addition to th e rtx1 and rtx2 loci, CMCP6 encodes a third, putative rtxA gene, VV1_2715, which we have designated r txA3 Unlike the other two rtx loci, it appears that r txA3 is the only rtx gene in the region. The typical type I secretion proteins, rtxB and rtxD are not found in the surrounding region as they are in the rtx1 and rtx2 gene clusters. The size of the rtxA3 gene is 8.8 kb, which is much shorter than the other two rtxA genes, yet still larger than most V. vulnificus genes. There is very little sequence si milarity between the CMCP6 RtxA toxins at the amino acid level. The C terminal region of RtxA3 has 36% similarity with the C terminal repeat region of RtxA1, but these sequences are not as conserved in RtxA2. The internal portion of the RtxA3 sequences h as less than 40% similarity with the RtxA2 sequence. The RtxA3 sequence was used in a BLAST search of the E. coli sequences in the NCBI database, and although there was low similarity, every hit was a n RTX protein including the prototypical HlyA and large r putative RTX toxins of E. coli The most conserved region of these proteins is at the C terminal end. A Prosite scan t h rough ExPASy identified a series of Hemolysin Ca+Binding repeats, also known as the RTX repeats, in the C terminal end of RtxA3. Th ese repeats are also used to predict type I secretion of proteins. Evidence of part of a T1SS is found just downstream of rtxA3 Gene VV1_2752 encodes an outer membrane protein with homology to the TolC family of proteins involved in TISS. As mentioned above, the other TISS proteins, ABC transport protein and membrane fusion protein, are not encoded near to the rtxA3 gene. The presence of the RTX glycinerich repeats indicates RtxA3 is accurately labeled as an RTX toxin. The absence of the rtxCBD genes suggests that RtxA3 may not be activated or secreted; therefore, it may not be a functional RTX toxin. To provide
106 a more complete analysis of RTX toxins of V. vulnificus we examined an rtxA3 mutant for its role in pathogenesis. Deletion of the rtxA3 Gen e to Examine a Role in Virulence Construction of the rtxA3 mutant An allelic exchange vector for deletion of rtxA3 was constructed by threeway USER cloning. Upstream and downstream sequences of rtxA3 were PCR amplified using oligonucleotides with USER friendly ends. A USER Sma I site was used to join the upstream and downstream fragments. The PCR products were captured in pGTR1129, resulting in the plasmid pGTR268. The plasmid was conjugated into CMCP6, and deletion of rtxA3 was selected by the sacB a ssisted allelic exchange proces s. Deletion of rtxA3 was verified by PCR, and the resulting mutant was named FLA899. There was no antibiotic resistance cassette used for allelic exchange, so that rtxA3 mutation could be used in combination with any antibiotic resistancelabeled mutation. rtxA3 mutant Because our hypothesis was that RtxA1 was the major cytotoxin, we did not expect that FLA899 would exhibit a significant reduction in cytotoxicity compared to wild type CMCP6. FLA899 caused 95.8 2.6 % destruction/detachment of the INT 407 cell monolayer, which was comparable to the 92.5 3.8 % destruction/detachment of the monolayer caused by CMCP6 ( P = 0.28). These resul ts substantiated our belief that RtxA3 is not a cytotoxic factor of CMCP6. Virulence of the rtxA3 mutant Although deletion of rtxA3 did not have an effect on cytotoxicity to INT 407 cells, the possibility of a role for RtxA3 in virulence was not ruled out. FLA899 was examined
107 for virulence at the minimum lethal inoculum of 300 CFU. All of the mice infected with this mutant became moribund within 16 hours postinoculation. 107.8 CFU/g was harvested from the skin lesions, and 106.4 CFU/g was harvested from the livers of mice infected with FLA899 (Figure 47 ) We concluded that RtxA3 is not neces sary for virulence of V. vulnificus in mice or for cytotoxicity to INT 407 cells. Future Direction: Construction of a Triple rtxA Mutant A triple rtxA1 / rtxA2 / rtxA3 mutant will be the key construct to confirm that the only RTX protein that has role in vir ulence in mice is RtxA1. To take advantage of the chitininduced natural transformation of V. vulnificus to move mutations each mutation must be labeled with a selectable marker. Because rtxA3 construct did not have an antibiotic resistanc e marker, a new rtxA3 deletion plasmid construct was created by capturing the upstream and downstream PCR products in pGTR1113 by threeway USER cloning and inserting the cat gene (for chloramphenicol resistance) at the Sma I site between the upstream and d ownstream sequences. This plasmid, designated pGTR288, was moved into CMCP6 by conjugation, and the rtxA3 allele was replaced by cat The rtxA3 mutant was named FLA969, and genomic rtxA3 mutation into FLA947 ( rtxA1 ::aph rtxA2 :: to construct a triple rtxA mutant Chloramphenicol resistant mutants will be verified by PCR, and confirmed mutants will evaluated for cytotoxicity to INT 407 cells, ability to cause apoptosis J774, and virulence in mice. Examination of RtxC 1 in Virulence RTX toxins often require activation by an acyltransferase, typically named RtxC. Whil e most studies of this activation have been carried out using E. coli HlyA and HlyC (77) the importance of acylation for activation has also been demonstrated for other
108 RT X toxins, including B. pertussis CyaA (78) and Pasteurella haemolytica LktA (76) V. vulnificus encodes multiple annotated acyltransferases, but only one has similarity to the E. coli HlyC protein, and it is designated RtxC1. The rtxC 1 gene of V. vulnificus is located directly upstream of rtxA1 in the operon. We hypothesized that if RtxC1 is activating RtxA1, which is essential for cytotoxicity and full virulence, then a deletion of rtxC1 would have a similar phenotype to the rtxA1 mutants. Since there is no rtxC nearby in the clusters encoding rtxA2 and rtxA3 our first thought was they may be activated by the RtxC1. However, the results from the rtxA2 and rtxA3 mutants indica ted that they may not be active RTX toxins, may have a phenotype we have not yet identified, or may be activated by a different mechanism. Deletion of VV2_0480 encoding RtxC1 The rtxC1 gene is located directly upstream of rtxA1 in an operon. We designed a clean deletion of the rtxC1 open reading frame to ensure that the rtxC1 mutation would not have a polar effect on the downstream rtxA1 The allelic exchange plasmid for the rtxC1 deletion was designed to be constructed by crossover PCR (97) Approximately 500 bp upstream and downstream of rtxC1 were amplified by PCR. The i nside primers (rtxC up3 and rtxC down5 ) contained 33 bp of overlapping s equences including a Not I site. The outside primers (rtxC u p5 and rtxC down3 ) included the USER friendly cloning sites at the 5 ends that corresponded to the USER vector. In crossover PCR, a second PCR reaction is performed to anneal the upstream and downstream fragments to form a single joined fragment This step was unsuccessful for the rtxC1 upstream and downstream fragments. As an alternative, the upstream and downstream amplicons were digested with Not I to create compatible 5 overhangs. The two Not I digested PCR products were USER cloned into the alleli c exchange vector pGTR1113,
109 and the two Not I ends were annealed during the ligation step. The resulting plasmid, pGTR265, was used for two step sacB assisted allelic exchange mutagenesis to create the CMCP6 rtxC1 mutant, which was named FLA591. The mutation was confirmed by PCR using the outside primers rtxC up5 and rtxC down3. rtxC1 mutant We expected the rtxC1 mutant, FLA591, to be as attenuated in cytotoxicity assays as the rtxA1 mutants were. Surprisingly, the results were not as expected. Unlike the rtxA1 mutations, deletion of rtxC1 had no effect on monolayer detachment/destruction or lysis of INT 407 cells. Infection with FLA591 caused 58.6% detachment/destruction, whereas the rtxA1 mutant FLA554 caused only 5.6% det achment/destruction of the INT 407 cell monolayers ( P = 0.00 2 ) (Figure 4 8 (A)). The destruction caused by FLA591 was comparable to that caused by wild type (76.5% detachment/destruction, P = 0. 12 ). Because the rtxC1 mutation had no effect on detachment or destruction, we did not expect it to affect lysis. Indeed, CMCP6 and FLA591 caused high levels of lysis of INT 407 cells (71.3% and 81.4% lysis, respectively) compared to 17.6% lysis caused by the FLA554 ( P < 0.001) (Figure 4 7 (B)). Virulence of rt xC1 in irontreated mice rtxC1 mutant suggested that RtxA1 activity was not dependent on fatty acylation by RtxC1 However, it remained possible that activation of RtxA1 by RtxC1 was required for virulence in m ice. We inoculated iron dextranrtxC1 mutant to determine if RtxC1 was required for RtxA1 activity and virulence. We inoculated the mice with 300 CFU of FLA591 or CMCP6, an inoculum at which rtxA1 mutants are consistently attenuate d. Similar to the cytotoxicity results, the deletion of rtxC1 had no effect on virulence. All mice infected
110 with FLA591 developed skin lesions and became moribund. The amounts of bacteria recovered from FLA591infected mice (108.1 CFU/g of skin lesion a nd 104.8 CFU/g of liver) were comparable to the amounts recovered after infection with CMCP6 ( 108.0 CFU/g of skin, P = 0.4; 104.9 CFU/g of liver, P = 0.9) (Figure 4 9 ) It is evident from the mousevirulence data and the cytotoxicity in cell culture that the RtxA1 toxin does not require activation by RtxC1 to cause lethality in mice. Discussion The RtxA1 toxin of V. vulnificus has an obvious role in cytotoxicity and the reduction in cytotoxicity observed during infection of INT 407 cells with rtxA1 mutants implies that the other RTX toxins produced by V. vulnificus either do not have similar functions or are not active during these growth conditions. Another possibility was that the RTX toxins act in concert, and that deletion of any one of them would af fect the cytotoxicity phenotype. In mice, there is an even more intriguing story. The potent cytotoxicity of the RtxA1 in vitro suggested that, if expressed, it would be contributing to tissue damage during infection in vivo. However, as illustrated in Chapter 3, mutation of the rtxA1 gene did attenuate virulence of CMCP6, but there was no visible effect on tissue damage. The skin lesions looked very similar to those caused by the wild type, both at the macroscopic level and the histological level. It was possible that either RtxA2 or RtxA3 by themselves could be factor s in virulence or that the presence of any of the three toxins, RtxA1, RtxA2 and/or RtxA3, was suffi cient to contribute to virulence and perhaps tissue damage. For this reason, we examined RtxA2 and RtxA3 through mutagenesis and analysis of virulence and cytotoxicity
111 RtxA2 is Not Essential for Virulence The size of the rtxA2 gene and the genes in the surrounding region drew our attention to rtxA2 At 14 kb, rtxA2 is one of the largest g enes of V vulnificus The C terminal region of the RtxA2 protein contains a series of glycinerich repeats, similar to the conserved RTX repeats. The genes directly downstream of rtxA2 encode an ABC transport protein and a membrane fusion protein. This gene arrangement is similar to other RTX toxin loci, in which the genes for their T1SS are encoded downstream of the toxin. The C terminal RTX repeats and the evidence for T1SS suggests that RtxA2 is a secreted RTX toxin. Other than the large size (over 400 kDa), there is no indication that RtxA2 is a member of the MARTX subfamily of RTX toxins. MARTX toxins, such as RtxA1 of V. vulnificus and RtxA of V. cholerae, are characterized by their large size, conserved repeats in the N terminus, RTX repeats in the C terminus, central enzymatic domains, and a cysteine protease domain. RtxA2 does not follow this pattern. Despite numerous bioinformatics searches and scanning by eye, N terminal MARTX repeats could not be identified in RtxA2. The enzymatic domains are a bit more ambiguous, as there is very little conservation within the MARTX family. One domain identified in RtxA2 was a Von Willebrand Factor (VWF) domain. VWFs are eukaryotic glycoproteins that mediate adhesion of platelets and binding of the blo od clotting factor VIIIA. While VWF domains are present in bacteria, there have been no reports identifying a function for them. We constructed a mutation in rtxA2 by inserting the translational stop codons of the (67) Interestingly, mutation of rtxA2 had no effect on the cytotoxicity of V. vulnificus towards INT 407 cells or virulence in mice (Figures 41 and 4 2). The
112 rtxA2 mutants were able to cause severe skin lesions in the s. c. inoculated mice, indicating that RtxA2 is not essential for tissue damage. It is possible that RtxA1 and RtxA2 could have redundant functions during infection of mice. If RtxA1 is the more potent cytotoxin, then its presence in the rtxA2 mutant could c ompensate for the lack of RtxA2. Hence, we constructed a double mutation of rtxA1 :: aph and rtxA2 :: noncytotoxic phenotype as the single rtxA1 mutant, FLA554 (Figure 4 3) When inoculated into mice, FLA558 appeared more attenuated than FLA554 at an inoculum of 300 CFU (Figure 44 (A)). At this inoculum, FLA558 was defective at causing a wildtype skin infection in four of the five mice, and was therefore unable to cause lethal infection in the mice. This was an interesting result because it signified that both RtxA1 and RtxA2 were necessary for skin infection, and perhaps tissue damage. However, when we increased the inoculum to 3,000 CFU, FLA558 was able to cause skin lesions in all mice and was not more atten uated than the rtxA1 mutant (Figure 44 (B)). Therefore, it is still possible that rtxA2 is contributing to virulence. However, it is more likely that the discrepancy in the results at the two different inocula could be a reflection of the experimental f luctuation observed with the rtxA1 mutant. Deletion of the rtxA1 gene causes more consistent attenuation than the rtxA1 ::aph mutation; therefore, a double mutant was constructed by deletion of rtxA1 and rtxA2 It had a similar reduced cytotoxicity phenotype as the previous double mutant (Figure 4 5) At an inoculum of 3,000 CFU, it was more virulent than the rtxA1 :: aph mutant (Figure 4 6) and so it confirms our initial conclusion t hat RtxA2 is not essential for virulence in the s.c. inoculated mouse model.
113 Very recently, Chou, et al ., (98) characterized the rtxA2 gene of V. vulnificus YJ016 and reported that an rtxA2 mutant yielded similar results to our own mutant. In addition, expression of the rtxA2 gene was decreased at 37C compared to 30C. Iron also seemed to affect expression, with higher expression levels in ironlimiting conditions than during growth in media supplemented with iron. These data argue against a pivotal role of rtxA2 i n virulence. The only domain identified in RtxA2 is the VWF domain, which could mediate adhesion of the toxin to cells When Chou, et al. examined the adherence of a YJ016 rtxA2 mutant for adherence to human laryngeal epithelial cell (HEp2) monolayers, there was no significant difference in adherence to HEp2 cells compared to the wild type (98) RtxA3 is Not Essential for Virulence The third rtxA locus in CMCP6 encodes only one gene, rtxA3. The RtxA3 protein is predicted to be 2,937 amino acids, much smaller than the other two RtxA proteins of CMCP6. Analysis of the RtxA3 protein sequence revealed the presence of the C terminal Hemolysin Ca+Binding repeats, but not the N terminal MARTX repeats. RtxA3, like RtxA2, is not a MARTX toxin, but still may be classified a s a member of the RTX family. We deleted the rtxA3 gene and evaluated its role in virulence and cytotoxicity. Similar to results that we obtained with RtxA2, deletion of rtxA3 had no effect on virulence mice or cytotoxicity in cell culture (Figure 4 7 ) BLAST analysis revealed little similarity to well studied proteins. Similar proteins are encoded by other Vibrio species, as well as E. coli Burkholderia amifaria, Pectobacterium species, Shewanella species, and Acinetobacter baumanii to name a
114 few M ost of the proteins identified from alignments are annotated as putative RTX toxins, calcium binding hemolysin proteins, or outer membrane adhesinlike proteins. The presence of a VWF domain in the C terminal end of the protein, indicates that RtxA3 may b e involved in adhesion. The fact that most of the similar proteins identified by BLAST alignments are present in environmental bacteria, as opposed to known human or animal pathogens, suggests that RtxA3 may have a role in survival and adherence in the environment and may not be important for virulence. It should be noted that V. vulnificus has evolved to be marine organism and not a human pathogen, and most likely any factor we identify as involved in virulence also has a function for survival or growth in its ecological niche. RtxC is Not Required for Virulence Caused by RtxA1 We also examined possible activation the RtxA toxins by an RtxC protein. RtxC1, encoded by a gene upstream of rtxA1 is the only acyltransferase with homology to E. coli HlyC. Th erefore, we proposed that if RtxC mediated acylation of the RtxA toxins is required for activity of the toxin, then deletion of rtxC1 should have the same effect on virulence as we observed for the rtxA mutants. Of the three RTX toxins, only RtxA1 appeared to be necessary for cytotoxicity in cell culture or virulence in mice. Therefore, we expected to see a similar phenotype of rtxC in our cell culture assays and mouse infection as we saw with the rtxA1 mutants. Instead, t rtxC mutant was similar to t he wild type for cytotoxicity (Figure 4 8) and virulence in mice (Figure 4 9) This result agreed with cytotoxicity results by Liu, et al (89) who noted that th eir rtxC mutant was fully cytotoxic These researchers observed a 10fold increase in i.p. LD50 compared to the wild type; however, we observed no attenuation in rtxC in s.c. inoculated mice. The rtx C mutant constructed by Liu, et
115 al was more virulent than the rtxA1 mutant they studied, which had a 102fold increase in LD50. These results were surprising because it had always been assumed that the RTX toxin, RtxA, requires activation by RtxC (75) Although this requirement has not been confirmed in all RTX producing bacteria, it has been demonstrated for well studied RTX toxins, such as E. coli HlyA, B. pertussis CyaA and P. haemolytica LktA (76 77 99) Our study and that of Liu, et al. are the only known reports of the function of rtxC in the MARTX family of toxins. Activation by RtxC should not be discounted, though. It is possible that RtxC has a role in activating RtxA1 to produce an uncharacterized phenotype. Along the same thought, it is poss ible that RtxA2 and RtxA3 may still require activation by RtxC, but we have not yet determined a measurable phenotype to use in analysis.
116 Figure 41. Detachment/destruction of INT 407 monolayers infected with rtxA2 :: Confluent INT 407 monolayers were infected with wildtype FLA399 rtxA2 mutant FLA441, or the rtxA1 ::aph mutant FLA413. The cells remaining attached after incubation were stained with crystal violet. The rtxA2 :: mutant caused similar levels of det achment/destruction as the wildtype did ( P = 0.6) *, P < 0.0003 by Fishers LSD comparing % destruction by rtxA1 ::aph to FLA399 or the rtxA2 ::
117 Figure 42. Virulence of rtxA2 :: in mice. Iron dextrantreated mice were s.c. inoculated with 300 CFU of either FLA399 or the rtxA2 :: The rtxA2 :: mutant was not significantly attenuated c ompared to the wild t ype Skin lesions were observed for all mice. Fractions below bars indicate proportion of mouse samples that yielded detectable CFU of b acteria. Bars without fractions indicate that bacteria were recovered from 5/5 mice.
118 Figure 43 Cytotoxicity to INT 407 cells caused by the double mutant INT 407 monolayers were infected with CMCP6, FLA554 ( ), FLA5 58 ( rtxA1 :: :: ) or FLA 561 ( rtxA2 :: 10. Cells remaining attached to the well after infection were stained with crystal violet to assess destruction of the monolayers. Inactivation of rtxA2 had no additional effect on cytotoxicity of the rtxA1 mutant ( P = 0.2) Asterisks indicate significance of the mutants to wildtype or the parent mutant strain determined by Fishers LSD (*, P = 106; *, P = 107).
119 A B Figure 44. Virulence of rtxA1 ::aph rtxA2 :: Iro n dextrantreated mice were s.c. inoculated with either FLA554 ( rtxA1 ::aph ) or FLA558 ( rtxA1 ::aph rtxA 2 ::(A) 300 CFU or (B) 3,000 CFU Fractions beneath the bars indicate the proportion of mouse samples that yielded bacteria. (A) F L A558 was more attenuated than FLA554 at 300 CFU. *, P = 0.05 by Students t t est comparing FLA554 to FLA558. P = 0.0 02, comparing proportion mice infected with FLA558 to FLA554 by 2 test (B) H owever, FLA558 was not more attenuated thanFLA554 at 3, 000 CFU ( P = 0.1 comparing liver CFU).
120 Figure 45 Lysis of INT 407 cells infected with rtxA1 ::aph or rtxA1 ::aph rtxA2 :: mutants. INT rtxA1 rtxA1 Lysis was measured by detection of LDH release. The double mutant FLA947 and the single mutant FLA943 have reduced lytic ability compared to wildtype ; however, they were not different from each other ( P = 0.3) Asterisks indicate statistical significanc e compared to wild type CMCP6 by a Fishers LSD (*, P = 105).
121 Figure 46 Virulence of rtxA1 ::aph rtxA2 :: treated mice were s.c. inoculated with either FLA943 ( rtxA1 ) or FLA947 ( rtxA 2 : : of 3,000 CFU. Fractions beneath the bars indicate the proportion of mouse samples that yielded bacteria. FLA943 was slightly more attenuated at this inoculum, indicating rtxA2 is not essential for virulence. (P = 0.3, comparing skin CFU/g; P = 0.1, comparing liver CFU/g; determined by Students t test)
122 Figure 47 rtxA3 mutant in mice. Iron dextrantreated mice were rtxA3 ). Deletion of rtxA3 had no effect on the ability to cause systemic infection ( P = 0.1) *, P = 0.003 by Students t test comparing FLA899 to CMCP6
123 A B Figure 48 Cytotoxicity to INT 407 cells caused by the rtxC mutant INT 407 cell rtxC1 mutant FLA591, or the rtxA1 ::aph mutant FLA554. (A) Detachment/destruction of the monolayers was assessed by crystal vi olet stain. FLA591 was not significantly different from CMCP6 (P = 0.12). Significance determined by Fishers LSD. ( P = 0.00 2 comparing FLA554 to FLA591; **, P = 0.00 04 comparing FLA554 to CMCP6) (B) Lysis was measured by LDH release. FLA591 was not significantly different from CMCP6 ( P = 0. 35 ) Significance determined by Fishers LSD ( *, P = 0.002; **, P = 0.001 comparing mutant FLA554 to either the wild type or FLA591).
124 Figure 49 rtxC1 mutant in mice Iron treated mice were s.c. inoculated rtxC1 mutant FLA591. Mice were euthanized when their temperatures dropped below 33C, a sign that they were moribund. FLA591 was as virulent as CMCP6 was in mice, indicating RtxC is not necessary for RtxA1mediated virulence. ( P = 0.43, comparing skin CFU/g; P = 0.9, comparing liver CFU/g)
125 CHAPTER 5 EXAMINING THE ROLE OF THE TYPE VI SECRETION SYSTEM IN PATHOGENESIS OF Vibrio vulnific us Rationale for Study Deletion of rtxA1 encoding one of the major cytotoxic factors of V. had little effect, if any, on tissue damage in the subcutaneous lesion of s.c. infected mice. This result indicated that V. vulnificus produces other c ytotoxic factors that are involved in the destruction of host tissues. Cytotoxic factors are often extracellular proteins that are exported by one of the secretion systems bacteria. Knocking out the secretion of the proteins can have the same, if not greater, effect as knocking out the cytotoxic effectors themselves. Genomic sequence analysis of V. vulnificus revealed the presence genes with homology to the type VI secretion system (T6SS) genes of V. cholerae. The T6SS is a recently characterized secret ion system in gram negative bacteria (57) and it is believed to play a role in the virulence. Two secreted substrates are conserved among the bacteria possessing a T6SS: a hemolysincoregulated protein (Hcp) and a valineglycine repeat protein (VgrG). Not only are these proteins secreted via the T6SS, there is evidence that they are required to form a functional T6SS. We hypothesized that the effectors secreted by the V. vulnificus T6SS could contribute to virulence and tissue damage, and thus we constructed deletions of these two genes, hcp and vgrG and examined them for cytotoxicity in vitro and virulence in the mouse model of disease. Introduction Gram negative bacteria have developed numerous ways to transport proteins across the inner membrane, through the periplasm, and across the outer membrane into the extracellular milieu or into nearby eukaryotic cells. These secretion systems
126 play an important role in pathogenesis of many bacterial pathogens. The recently identified T6SS i s suspected to contribute to virulence of pathogens including V. cholerae (100) Pseudomonas aeruginosa (101) Aeromonas hydrophila (102) Fransicella tularensis (103) (104) Burkholderia mallei (105) and Agrobacterium tumefaciens (106) In addition to pathogenic interactions, the T6SS also contributes to symbiotic interactions of bacteria with their eukaryotic hosts, exemplified by studies of the Rhizo bium legume symbiotic relationship (107) Due to the recent discovery of this secretion system and variability in the gene tic composition in the T6SS regions, little is known about the genomic organization, the structure of the apparatus, and the exact mechanism of secretion. In silico analysis identified a set of 13 conserved proteins defined as the T6SS "core components (108) Some of these components share homology with the type IV secretion system (T4SS) components, although the two secretion systems are distinct from one another. The core components included the h emolysin c oregulated p rotei n (Hcp), v aline g lycine r epeat protein (VgrG), ClpB ATPase homologs (ClpV), and the T4SS homologs IcmF and DotU (108) Boyer, et al (108) have identified one or more T6SS loci present in the genomes of over 90 gram negative bacteria, many of which are pathogens or symbionts. Despite attempts to identify components involved in this new secretion system, numerous questions remain as to the exact function and mechanism of secretion. Two known secrete d substrates of the T6SS, Hcp and VgrG, are encoded by all T6SS containing bacteria. There is some debate as to if these proteins are secreted effector proteins, structural components of the secretion apparatus itself, or both. In 1996, Hcp
127 was first exa mined in V. cholerae by Williams, et al. (109) At that ti me, the T6SS was unknown and Hcp was thought to be regulated and secreted in a manner similar to the hemolysin, HlyA. Deletion of the two hcp genes had no effect on virulence in the infant mouse model or on cytotoxicity in vitro (109) A decade later, Pukatzki, et al (100) used transposon mutagenesis of a V cholerae non01/non0139 strain to identify factors causing virulence towards the amoeba Dictyostelium discoideum Dictyostelium attenuated mutants contained transposon insertions in vas ( v irulence a ssociated s ecretion) genes, some of which had similari ty to the icm genes of the L. pneumophila type IV secretion system. Secretion of four proteins (Hcp, VgrG 1, VgrG 2, VgrG 3, and VgrG 4) was affected by the transposon mutations. One of the mutations in vasH 54 activator) affected transcription of hc p 1 and hcp 2, encoding identical Hcp proteins. Additionally, an hcp double mutant was avirulent towards D. discoideum and was defective in secretion of the three VgrG proteins, indicating that Hcp has an integral role in the T6SS. Other studies have conf irmed that Hcp and VgrG are secreted factors and that their secretion is mutually dependent on each other. In E. tarda the three known T6SS secreted proteins are the Hcp homolog EvpC, a VgrG homolog EvpI, and an effector protein EvpP (not conserved among T6SSs). Inactivation of any of the 13 genes encoding T6SS core components, including evpC and disrupted secretion of the three secreted proteins EvpC, EvpI, and EvpP, and caused attenuation in virulence to fish (110) While EvpC (Hcp) and EvpI (VgrG) are required for a functional T6SS, EvpP is not required for secretion of the oth er proteins, indicating that Hcp and VgrG may be components of the secretion apparatus and that EvpP is a secreted effector protein
128 (110) Hcp is also a secreted substrate of P. aeruginosa. The crystal structure of the P. aeruginosa Hcp revealed the formation of hexameric rings (101) and these rings polymerize to form tubes up to 100 nm long (111) These nanotubes are proposed to form part of the secretion complex throug h which other effectors travel. All T6SS encoding bacteria have at least one VgrG, and many of them encode more than one. These proteins have domains sharing homology to structural features of the tail spike protein of the T4 bacteriophage (112) V. cholerae encodes three VgrG proteins, VgrG 1, VgrG 2, and VgrG 3, each containing domains with similarity to the gp44 protein of the bacteriophage Mu and gp5 protein of bac teriophage T4. The three proteins were predicted to form a trimeric complex, similar to the bacteriophage tail spike complexes used to puncture the bacterial membrane, and were demonstrated to interact with each other (112) The similarity to the bacteriophage tail spike proteins suggests that VgrG proteins could form a membranepuncturing device at the tip of the secretion apparatus. In addition to the gp44 and gp5 domains, V. cholerae VgrG 1 contains a C terminal extension with an actin crosslinking domain and VgrG 3 contains a peptidoglycanbinding domain. Deletion of VgrG 1 or VgrG 2 disrupts s ecretion of Hcp, VgrG 1, VgrG 2, and VgrG 3. VgrG 2 and its actincrosslinking domain are also cytotoxic for J774 and Raw264.7 cell lines and Dictyostelium amoeba, indicating that VgrG proteins could be effector proteins as well as a component of the secr etion apparatus (100 113) VgrG proteins containing C terminal extensions were identified in many, but not all, T6SS encoding bacteria and are classified as evolved VgrG proteins.
129 V vulnificus contains a genomic region encoding a T6SS on chromosome 2. This region, although not well defined, contains 18 genes homologous to T6SS genes of other bacteria and includes all of the 13 T6SS core components. Considering the importance of the T6SS in virulence of several organisms, including V. cholerae we speculated that the T6SS may have a role in virulence of V. vulnificus. Given that there is still little known about the T6SS and the specific proteins required for proper secretion, we planned to delete a 10kb portion of the chromosomal locus encoding the V. vulnificus T6SS. Seven genes were encoded in the 10kb region to be deleted, VV2_0428 through VV2_0434, including vgrG (VV2_0428) and hcp (VV2_0429). Other genes to be deleted in cluded VV2_0430 encoding a ClpB homolog, an ATPase hypothesized to provide the energy necessary for secretion (101 114) and other genes predicted to be core components of the T6SS. Deletion of this region would have knocked out secretion via this pathway, and we would have been able to assess the importance o f T6SS in pathogenesis of V. vulnificus Had we seen an effect, we would have deleted individual genes to study their functions. Unfortunately, we were unsuccessful at deleting the 10kb of this genetic locus. As an alternative plan, we decided to indiv idually delete the genes hcp and vgrG to disrupt secretion. Results Identification and Deletion of V. vulnificus vgrG As described above, all T6SS encoding bacteria have vgrG genes. V. cholerae and several other T6SS encoding bacteria have more than one v grG gene. V. cholerae encodes three VgrGs, two of which are evolved VgrGs, meaning they possess an extra domain at the C terminal end of the protein. BLAST analysis using the amino acid sequence of VgrG 1, VgrG 2, and VgrG 3 of V. cholerae identified o nly one protein in
130 V. vulnificus CMCP6 encoded by the gene VV2_0428. V. vulnificus VgrG has 31% identity with the conserved N terminal portion of the V. cholerae VgrGs; however, there is no additional C terminal extension as is present in VgrG 1 and VgrG 2 of V. cholerae VgrG encoded by V. vulnificus is predicted to be the type VI secretion protein VgrG according to the NCBI Conserved Domains database, and similarly to the other T6SS VgrGs, the VgrG sequence of CMCP6 has domains with similarity to the gp27/gp44 and the gp5 proteins of the bacteriophage tail spike proteins. The similarity of VgrG with other T6SS VgrGs indicates that VgrG should be essential for T6SS of effector proteins and may have cytotoxic activity itself. The vector to delete vgrG w as constructed by threeway USER cloning. 1 kb upstream and downstream of vgrG was PCR amplified with the primers vgrG down5/vgrG down3 and T6SS up5/T6SS up 3. The upstream and downstream fragments were USER cloned into the allelic exchange vector pGTR1113 (pCVD442:: lacZ USER), and the resulting plasmid was named pGTR1304. For selection of the deletion of vgrG the aph gene (encoding kanamycin resistance) was inserted at the Sma I site engineered between the upstream and downstream fragments. The resultin vgrG ::aph plasmid, pGTR1307, was linearized and added to CMCP6 growing in the presence of crabshell to induce chitinbased natural transformation. Chitin transformants with the correct mutation were selected on LB N containing kanamycin and were verifi ed by PCR to have undergone allelic exchange of the aph for vgrG vgrG ::aph mutant was named FLA1035. Cytotoxicity of vgrG :: aph mutant Since VgrG proteins are suggested to form a membranepuncturing device on the tip of the T6SS apparatus and potentially have cytotoxic activity, we tested FLA1035 for
131 cytotoxicity for both INT 407 cells and J774 cells. INT 407 monolayers were infected with FLA1035 for 1 hour, treated with gentamicin, and incubated overnight. The following day, the cells remaining attached to the wells were washed and stained with crystal violet to assess the level of detachment/destruction of the monolayer. FLA1035 did not have a decrease in cytotoxicity compared to CMCP6 (86.5% destruction vs. 88.7% destruction, respectively ; P = 0.63) (Figure 5 1). The T6SS of V. cholerae mediates cytotoxicity towards phagocytic cell types, including Raw264.7 and J774 cells (100 113) We therefore examined the T6SS mutants for cytotoxicity to J774 murinemacrophage like cells. Monolayer detachment/destruction was assessed for J774 cells in a similar manner to INT 407 cells. Monolayers were established and infected at an MOI of 10 for 1 hour. Detachment/destruction of the monolayers was measured as it was for INT 407 monolayers. Similar to the results from infection of the INT 407 cells, deletion of vgrG did not have a significant effect on cytotoxicity to J774 cells (79.2% destruction by FLA1035 vs. 88.2% destr uction by CMCP6 P = 0.16) (Figure 5 1 ). Although deletion of vgrG had no effect on the ability to detach/destroy the monolayers of J774 cells, there was a reduction in the ability of FLA1035 to induce apoptosis of J774 cells. The level of caspase 3/7 activity was measured in infected J774 cells and was lower in FLA1035infected cells than in CMCP6 infected cells. FLA1035 caused 38.2% apoptosis compared to 54% apoptosis caused by CMCP6 ( P = 0.0001) (Figure 5 2 ), indicating that VgrG or a protein secreted in a VgrG dependent manner is an apoptotic factor. The defect of the vgrG ::aph mutants was not as great as that observed with the rtxA1 :: aph mutants (38.2% compared to 20.5%, P = 105)
132 The low level of apoptosis induced by each of these mutants is above the bac kground level of uninfected cells, indicating that each protein contributes to apoptosis. Apparently, in the absence of either VgrG or RtxA1, the remaining factor can cause apoptosis. Virulence of vgrG ::aph in mice To determine the role of VgrG and pote ntially the T6SS in virulence, FLA1035 was s.c. inoculated in iron dextrantreated mice. The vgrG ::aph mutant was not attenuated at 1,000 CFU (three times the minimum lethal dose for wild type), although it did cause visible infection in only four of fiv e mice. Means of 106.7 CFU/g and 103.8 CFU/g were recovered from the skin lesions and livers of the FLA1035infected mice (Figure 53 ). These levels were not significantly lower than the level of bacteria in CMCP6infected mice (107.8 CFU/g skin lesion, P = 0.14 ; and 104.1 CFU/g liver, P = 0.69). In addition, the skin lesions caused by FLA1035 looked very similar at the gross level to the lesions caused by CMPC6. Therefore, VgrG and potentially the T6SS are not essential for virulence or tissue damage i n our s.c. model. Identification and Deletion of V. vulnificus hcp As an additional evaluation of the T6SS and its role in virulence of V. vulnificus we also examined the hemolysin coregulated protein, Hcp. Hcp is secreted in all T6SS encoding bacteria and is predicted to be a part of the T6SS machinery and therefore necessary for proper secretion of other substrates (100 111) While a BLAST search usin g the HCP sequences from V. cholerae and P aeruginosa did not reveal any hits in V. vulnificus searching the NCBI published genome of CMCP6 revealed a gene annotated as a hemolysincoregulated protein or hcp This gene, VV2_0429, is directly downstream of the vgrG gene discussed above and within the locus encoding the T6SS
133 genes of V. vulnificus V. vulnificus Hcp belongs to the DUF796 superfamily and the COG3157 according to the Conserved Domain database (NCBI), both of which are comprised of the T6SS Hcp proteins. The COG3157 group of proteins is specified by Boyer, et al (108) as one of the essential T6SS components. An InterProScan search (EBI database) for protein function also groups the V. vulnificus Hcp into the fam ily including the T6SS Hcp proteins. The Hcp of V. vulnificus shares 34% similarity and 62% identity to the A. tumefaciens Hcp protein. V. vulnificus Hcp also has 26% similarity and 40% identity to E. tarda EvpC (Hcp). Both of these proteins are secreted and are proposed to have similar functions to V. cholerae Hcp (100) although there is very little sequence similarity between them. Boyer (108) noted in their in silico analysis that there is a lot of sequence variation among the different T6SSs, especially in the secreted proteins. We therefore assumed that the protein encoded by VV2_0429 was indeed the T6SS Hcp protein, despite the lack of homology with V. cholerae. Since previous reports indicate that T6SS, i ncluding secretion of VgrG, is dependent on Hcp, we anticipated that inactivation of hcp would disrupt T6SS. The plasmid for deletion hcp pGTR1303, was constructed by threeway USER cloning of the PCR amplified upstream and downstream regions of hcp into the USER friendly allelic exchange vector, pGTR1113. The chloramphenicol acetyltransferase ( cat ) cassette from pCOS5 was inserted at the Sma I site engineered between the hcp upstream and downstream fragments. Once the deletion vector, pGTR1305, was confi rmed, it was linearized with Nde I and moved into CMCP6 via chitininduced transformation. Chitintransformants were selected on LB N agar plates containing
134 chloramphenicol and verified by PCR for deletion of the hcp hcp ::cat mutant strain was named FLA1030. Cytotoxicity of hcp:: cat FLA1030 was tested for cytotoxicity in vitro in J774 cells. Monolayers of J774 cells were infected with either CMCP6 or FLA1030 at MOI of 10. After one hour of infection, the bacteria were killed with gentamicin and the cells were incubated overnight. Monolayer detachment/destruction was assessed by staining with crystal violet. Deletion hcp of had no effect on cytotoxicity to J774 cells. FLA1030 caused a mean 89.6% destruction of J774 cell monolayer, compared to 88.2% destruction caused by CMCP6 ( P = 0.73) (Figure 54 ). As with mutant, FLA1030 was similarly cytotoxic to INT 407 cells (Figure 54). As detailed above, the vgrG ::aph mutant caused significantly less apoptosis in J774 cells than the wildtype parent. Therefore, we expected that if VgrG was secreted in an Hcp dependent manner, then the hcp ::cat mutant would have a similar reduction in ability to induce apoptosis. Interestingly, the hcp ::cat mutant, FLA1030, caused apoptosis in J774 cells similar to that caused by CMCP6 (Figure 55 ). This result indicate d that Hcp may not be necessary for T6SS mediated secretion, as opposed to the current model, or that VgrG secretion is independent of Hcp. Virulence of hcp :: cat We tested if Hcp was necessary for virulence by s.c. infection of iron dextranmice with 1, 000 CFU of FLA1030. We did not expect to see an effect on virulence, since the vgrG ::aph was virulent and we already speculated that the T6SS does not have an essential role in virulence. Confirming the previous results with VgrG, FLA1030 was as
135 virulent as the wild type in mice (Figure 56 ). Therefore, the T6SS does not appear to be essential for virulence. Deletion of the T6SS F actors in an rtxA1 B ackground It is possible that any cytotoxic defect of an hcp or vgrG mutant was covered up by the presenc e of the major cytotoxic factor, RtxA1. In many T6SS encoding bacteria, including V. cholerae and A. hydrophila (100 102, 113) the effects of T6SS effectors were examined in a parent strain lacking other major cytotoxic factors of these strains. For example, V. cholerae rtxA mutants we re used as background for all of the T6SS mutants in the studies identifying the role in cytotoxicity towards macrophagelike cells and D. discoideum (100, 113) To eliminate the cytot oxic effects of RtxA1, we decided to examine the T6SS mutations in an rtxA1 deficient strain. We constructed double rtxA1 ::tet A vgrG :: aph and rtxA1 :: tet A hcp ::cat mutants by chitininduced transformation. To construct the rtxA1 :: tet A vgrG :: aph mutant, genomic DNA from vgrG ::aph ) rtxA1 ::tetA ) growing on crabshell to ind uce natural transformation. Chitin transformants were selected on LB N containing kanamycin (to select for the deletion of vgrG ) and were verified to have tetracycline resistance (due to the rtxA1 ::tetA mutation). For construction of the rtxA1 ::tetA hcp ::cat rtxA 1 ::tetA ) was added to hcp ::cat ) growing on a crabshell. Chitintransformants were selected on LB N agar supplemented with tetracycline and verified to have resistance to chloramphenicol. The doubl e mutants were confirmed by PCR, and the resulting mutants were named FLA960 ( rtxA1 ::tetA vgrG ::aph ) and FLA965 ( rtxA1 :: tetA hcp ::cat )
136 Cytotoxicity of T6SS rtxA1 double mutants The rtxA1 ::tetA vgrG ::aph mutant, FLA960, was tested for the ability to cause destruction and detachment of INT 407 monolayers. As detailed above, the rtxA1 mutants are cytotoxic beyond one hour postinfection, so it was possible that VgrG or VgrG dependent factors could contribute to the residual cytotoxicity. INT 407 cells were infected with the various mutants for 1 hour and 2 hours prior to addition of gentamicin. The cells were incubated for a total of 4 hours after infection, and the cells remaining attached to the wells were stained with crystal violet. At 1 hr postinfection, the rtxA1 ::tetA mutant was reduced for cytotoxicity ( 13.1% destruction), as expected, and the double rtxA1 :: tetA vgrG :: aph mutant did not have a significant additional reduction in cytotoxicity (7.7% destruction) (Figure 57 ). At two hou rs postinfection, the rtxA1 ::tetA and the double rtxA1 ::tetA vgrG ::aph mutants caused similar levels of destruction of the monolayers (69.2% and 69.7% destruction, respectively) (Figure 57 ). These results indicate that VgrG is not essential for causi ng cytotoxicity to INT 407 cells, even in an rtxA1 background. We also tested the ability of the double rtxA1 and T6SS mutants to cause destruction/detachment of J774 cell monolayers. vgrG :: aph FLA1035 and hcp ::cat FLA1030 were cytotoxic towards J774 c ells and, as was expected, the rtxA1 ::tetA mutant, FLA954, had reduced cytotoxicity. Deletion of vgrG or hcp in the rtxA1 ::tetA mutant had no additional effect on cytotoxicity compared to the rtxA1 ::tetA (Figure 5 8 ). The apoptosis of the double rtxA1 ::tetA vgrG ::aph mutant, FLA960, needs to be examined. Both the rtxA1 ::tetA mutant and the vgrG :: aph mutant caused less apoptosis than the wild type in infected J774 cells, but each mutant caused some
137 apoptosis above the background levels. It will be interesting to see if FLA960 causes less apoptosis than either one of the two individual mutants. If RtxA1 and VgrG are the two major factors contributing to apoptosis in this cell culture model, we would expect to see very little apoptosis, similar to the background level in uninfected cells. Virulence of T6SS, rtxA1 double mutants We tested the virulence of the T6SS, rtxA1 double mutants by s.c. inoculation in iron dextrantreated mice. FLA960 ( rtxA1 ::tetA vgrG :: aph) and FLA965 ( rtxA1 ::tetA hcp ::cat ) each appeared to be slightly although not significantly, attenuated compared to FLA954 ( rtxA1 :: tetA ) (Figure 5 9). FLA960 caused visible lesions in five of five mice with a mean 107.5 CFU/g in the skin. This was similar to the bacteria recovered from the skin of FLA954 (108.1 CFU/g; P =0.2). FLA960 caused systemic infection in only two of five mice compared to five of five mice with systemic infection by FLA954 ( P = 0.4) Despite fewer mic e with systemic infections, the amount of bacteria recovered from the livers of FLA960 infected mice was not significantly different from the amount of bacteria recovered from FLA954 infected mice (103.6 CFU/g vs. 104.7 CFU/g; P = 0.3). Similar results we re observed for FLA965 ( rtxA1 ::tetA hcp ::cat ) (Figure 5 9). FLA965 caused detectable skin infections in three of five mice, and a mean of 106.3 CFU/g was recovered from the skins ( P = 0.01, compared to FLA954). The three mice with detectable skin infection had detectable liver infection, resulting in a mean of 103.5 CFU/g recovered from the livers ( P = 0.1, compared to FLA954). The similar decrease in bacteria recovered from mice infected with FLA960 and FLA965, though not significant from FLA954 inf ected mice may suggest a small role for T6SS in virulence. These mutants should be examined at higher inocula to investigate
138 the level of attenuation and at lower inocula to detect a significant difference between the rtxA1 :: tetA mutant and the double mutants. I nterestingly, the skin lesions caused by the double mutants resembled those caused by FLA954 and wild type (Figure 5 10) although the lesions have not yet been examined by histopathology. At the histological level, there may be a difference in tissue damage or PMN response in these mice, that was not observed when rtxA1 was knocked out alone. Discussion The discovery and characterization of the T6SS in several well known pathogens, as well as other gram negative bacteria, suggested that it might be a provocative target for our studies in pathogenesis of V. vulnificus Many reports suggest that the T6SS mediates cytotoxicity to host cells and contributes to virulence and survival during infection. The studies performed on the V. cholerae T6SS have proposed a role in cytotoxicity to phagocytic cells types, such as J774, Raw264.7, and Dictyostelium amoeba (100 113) The T6SS of E. F tularensis A. hydrophila, and P. aeruginosa have also been demonstrated or suggested to have a role in virulence. V. vulnificus CMCP6 contains a locus of 18 genes with similarity to or annotated domains of T6SS of other bacteria, including the 13 core components identified by (108) During this investigation, we examined the role of the V. vulnificus T6SS in virulence by construction of vgrG and hcp mutants and testing them in the cell culture and mouse models of infection. VgrG Causes Apoptosis; However, it is Not Essential for Virulence The VgrG proteins of the T6SS share structural similarity with the T4bacteriophage needle complex (gp27gp5). The three VgrG proteins of V. cholerae are hypothesized to form a trimeric co mplex, either at the tip of a secretion apparatus or
139 exposed somewhere on the bacterial surface, and this complex is speculated to act as a membrane puncturing device (112) While V. cholerae encodes three VgrG proteins, two of which have C terminal extensions, V. vulnificus encodes only one VgrG. The V. vulnificus VgrG has domains with homology to t he gp27/gp44 and the gp5 proteins; however, it does not have a C terminal extension. Possession of a single VgrG is not unusual, as there are other T6SS encoding bacteria with a single VgrG that lacks C terminal extensions (115) It is possible that the single VgrG is able to form into a homotrimer. We expected that deletion of vgrG would knock out the T6SS, as is the case for other bacteria. Deletion of the sing le vgrG gene of V vulnificus had no effect on destruction of INT 407 or J774 cell monolayers (Figure 5 1 ). We speculated that the lack of a cytotoxicity phenotype could be due to the activity of the major cytotoxic factor RtxA1, therefore we constructed a rtxA1 :: tetA vgrG :: aph mutant. The double mutant was not more attenuated in cell culture than the rtxA1 ::tetA mutant (Figure 5 7 and 59 ), indicating that VgrG does not contribute to the residual cytotoxicity caused by rtxA1 mutants. Interestingly, the vgrG :: aph mutant had reduced ability to cause apoptosis in J774 cells (Figure 5 2 ), although it has not yet been determined if apoptotic activity is completely abolished in the rtxA1 ::tetA vgrG ::aph mutant. Despite this small effect in cell cultur e, the vgrG ::aph mutant had no significant effect on virulence in mice when deleted by itself (Figure 5 3 ) H owever, when vgrG was deleted in the double mutant was slightly, though not significantly more attenuated than the rtxA1 ::tetA su ggesting that VgrG may have a role in virulence, but the presence of the cytotoxic factor RtxA1 may compensate for its absence.
140 Hcp is Not Necessary for Cytotoxicity or Virulence Hcp is another protein secreted by all known T6SSs, and it is proposed to f orm a tubular structure through which other T6SS secreted proteins pass. In well studied T6SSs, the secretion of the VgrG proteins and other T6SS effectors is dependent on Hcp, adding to the evidence that Hcp is a part of the secretion apparatus. We expected that if T6SS has a role in virulence, then a mutation in Hcp would cause attenuation in the mouse model. First we characterized the hcp ::cat mutant in vitro. As was expected based on the vgrG results, hcp ::cat had no observable decrease in cytotox icity towards J774 cells, even in the absence of the major cytotoxic fac tor, RtxA1 (Figure 5 7 ) Because we had observed a decrease in apoptosis of J774 cells infected with the vgrG :: aph mutants, we were surprised to discover that the hcp ::cat mutants w ere not defective at causing apoptosis (Figure 5 5 ). Additionally, we observed no difference in virulence of the hcp ::cat in mice; however, similar to the results from rtxA1 ::tetA vgrG :: aph mutant the rtxA1 :: tetA hcp ::cat mutant was slightly, though not significantly attenuated in mice compared to the rtxA1 :: tetA mutant. Given that both of the double mutants behaved similarly suggests that there may be a small role for T6SS in virulence of V. vulnificus I t is possible that RtxA1 and T6SS factor s have a similar function in pathogenesis, and although RtxA1 is more important in virulence, T6SS may have a smaller accessory role It is intriguing that the two T6SS mutants had different apoptosis phenotypes, since we expected Hcp to be necessary for secretion of VgrG. It is possible that V. vulnificus produces more than one Hcp, especially since V. cholerae encodes two hcp genes and deletion of both of them is necessary to disrupt secretion. However, close examination of the V. vulnificus genome confirmed that the hcp that was deleted
141 is the only identifiable hcp gene with similarity to any other hcp genes known. It is also notable that the Hcp sequence of V. vulnificus is quite different from the well studied Hcp proteins of V. cholerae and P. aeruginosa, but has similarity to Hcp proteins of other bacteria (e.g., A. tumefaciens Hcp, E. tarda EvpC, B. mallei Hcp, putative Hcp proteins of other Vibrio species). While domain searches, structural motifs, and protein function searches classify it as belonging to the same family of proteins as T6SS Hcp proteins, it is possible that the V vulnificus Hcp may not be a true T6SS Hcp protein, it might not have the same function, or may not be secreted. Another possibility that could explain the two different apoptosis phenotypes is that VgrG may be secreted in an Hcp independent manner and therefore a mutation in hcp does not affect secretion or activity of VgrG. What is the Function of T6SS? It is not entirely surprising that T6SS mutants are not attenuated in mice, and it is possible that the T6SS is not expressed in the presence of mammalian cells. T6SS has been demonstrated to be tightly regulated in other bacteria by different mechanisms (115) The environmental cues signaling T6SS expression vary among the bacteria, and there is still a lot unknown. Temperature can be a factor in regulation. For instance, the T6SS of Yersinia pestis is upregulated at 26C compared to 37C (116) Quorum sensing and growth phase influence expression of Hcp in V. cholerae (117) and these factors also mediate T6SS expression in A. hydrophila (11 8) Many of the T6SSs have a role in survival within or cytotoxicity to host cells, indicating that there are host factors that may mediate expression. For example, secretion of Hcp and VgrG proteins was increased in the plant pathogen Pectobacterium at rosepticum when the bacteria were grown in the presence of potato tuber extracts (119) Signals during
142 i nfection also regulate the T6SS secretion of as it is negatively regulated by RetS and positively regulated by LadR and is expressed during chronic infection (101) The T6SS of P. aeruginosa is also regulated by the serine/threonine kinase (PpkA) and phosphatase (PppA) encoded in the T6SS region, although the environmental cues for T6SS expression remain unknown (120) It would be advantageous to know what factors regulate expression of T6SS in V. vulnificus We did not evaluate the expression of the T6SS and do not know if it is active during growth in mammalian cell culture or in mice. V. vulnificus lives primarily in an estuarine environment, both in the water and cohabitating with shellfish, and it is possible that environmental cues such as temperature, salinity, or availability of nutrients may influence the expression of the T6SS. V. vulnificus can colonize and persist within the filter feeding mollusks, such as oysters. During the colonization of oysters, V. vulnificu s is able to resist being killed by the oyster defenses, such as the phagocytic hemocytes (121) Perhaps the T6SS has a role in persistence within the oyster and resistance to phagocytosis by hemocytes. This is not out of the realm of possibility given that many of the T6SSs of other bacteria influence survival and killing by phagocytic cell types. Since it appears that T6SS is not essential for virulence in mice, the role of the T6SS dur ing growth in oysters is a potential path worth following in the future.
143 Figure 51. Detachment /destruction of cell monolayers by vgrG ::aph INT 407 or J774 vgrG ) at an MOI of 10 for 1 hour. The cells remaining attached to the wells were stained with crystal violet and the percent destruction of the monolayer was calculated. Deletion of vgrG had no effect on the ability to destroy the monolayers of infected INT 407 or J774 cells.
144 Figure 52 Apoptosis induced by vgrG :: aph. J77 4 cells were infected with wildtype vgrG rtxA1 ::aph ) at an MOI of 10. Apoptosis was measured using the ApoONE kit to detect activity of caspase3 and caspase7. Gliotoxin was used as a positive apoptotic c ontrol and was considered to induce 100% apoptosis. % apoptosis for each infection was calculated by dividing the normalized RFU of each sample by the average normalized RFU from the gliotoxin treated wells and multiplying that number by 100. FLA1035 cau sed less apoptosis than wild type; howev er, it causes rtxA1 Asterisks indicate statistical significance by a Fishers LSD (*, P = 0.0 0 01; **, P = 105; *, P = 106)
145 Figure 53 Virulence of vgrG ::aph in s.c. inoculated mice. Iron dextrantreated mice were s.c. i noculated with 1,000 CFU of either wildtype CMCP6 or FLA1035 vgrG ). Mice were euthanized once their temperatures dropped below 33C. Samples of the skin lesion and liver were homogenized, diluted, and plated. Fractions below the bars indicate the pr oportion of mice with detectable bacteria in the samples over the total number of mice infected. While FLA1035 caused detectable infections in only 4 of 5 mice, it was as virulent as wild type in the remaining mice, indicating VgrG is not essential for vi rulence.
146 Figure 54 Detachment/destruction of cell monolayers by hcp :: cat J774 or INT 407 hcp ) at an MOI of 10 for 1 hour. The cells remaining attached to the wells were stained with crystal violet and the percent destruction of the monolayer was cal culated. Similar to the observations with VgrG, deletion of hcp had no effect on the ability to destroy the monolayers of infected J774.
147 Figure 55 Apoptosis of J774 cells infected with hcp ::cat J774 cells were infected with wild type CMCP6, FLA1 hcp vgrG ). Apoptosis was measured as before using the ApoONE kit. Apoptosis over background level was calculated by dividing the normalized RFU of each sample by the vgrG infection hcp caused an increase in apoptosis of J774 cells. *, P = 0.05; **, P = 0. 03; *, P = 0.003 determined by Fishers LSD
148 Figure 56 Virulence of hcp ::cat in s.c. inoculated mice. Iron dextrantreated mice were s.c. inoculated with 1,000 CFU of either wildtype CMCP6 or FLA1030 hcp ). Mice were euthanized once their temperatures dropped below 33C. FLA1030 was as virulent as wild type at this inoculum in all mice, although FLA1030 infected mice had higher temperatures than CMCP6infected mice, although the average temperature was still below 33 C (*, P = 0.02; **, P = 0.002 comparing FLA1030 to CMCP6 by Students t test)
149 Figure 57 Cytotoxicity of rtxA1 ::tetA vgrG ::aph double mutant. INT 407 monolayers were infected w ith either wild type CMCP6, FLA954 ( rtxA1 ), rtxA1 vgrG vgrG ) at an MOI of 10. Gentamicin was added either at 1 hour or 2 hours postinfection. At 4 hours postinfection, the attached cells were stained with crystal violet and the percent of detachment/destruction of the monolayer was calculated. Deletion of vgrG in rtxA1 mutant caused no additional reduction of cytotoxicity at 1 hour ( P = 0.08) or at 2 hours ( P = 0.9 4 ) indicating VgrG is not essential for cytotoxicity. Ast erisks indicate statistical significance by Fishers LSD (*, P = 0.03 comparing FLA9 54 or FLA960 to wildtype or FLA1035 ; **, P = 0.00 0 1 comparing FLA960 to FLA1035 or CMCP6; ***, P = 105 comparing FLA954 to CMCP6 )
150 Figure 58 Cytotoxicity of rtxA1 and T6SS double mutants to J774 cells. J774 cell monolayers w rtxA1 rtxA1 vgrG vgrG rtxA1 hcp hcp ) at an MOI of 10 for 1 hour. After overnight incubation, the cells remaining attached to the wells were stained with crystal violet and the level of detachment/destruction of the monolayer was calculated. As expected based on the results during INT 407 cell infections, deletion of either of the T6SS rtxA1 mutant ( P = 0.64 for rtx A1 vgrG ; P = 0.3 for rtxA1 hcp ) Asterisks indicate statistical significance by Fishers LSD (*, P = 0.00 1 ; **, P = 0.0 0 01; ***, P = 105)
151 Figure 59. Virulence of rtxA1 T6SS double mutants in mice Iron dextrantreated mice were s.c. inocul ate rtxA1 ), FLA960 rtxA1 vgrG ), or FLA9 6 rtxA1 hcp ). While it appear ed that the rtxA1 FLA954, d eletion of vgrG or hcp had no significant effect on amount of bacteria recovered from the skin compared to FLA954 by Students t test ( P = 0.2, FLA960; P = 0.1, FLA965). Similarly, there was no significant difference in bacteria recovered from the liver compared to FLA954 ( P = 0. 3 FLA960; P = 0.1, FLA965) Fractions below the bars indicate the proportion of mice with detectable bacteria in the samples over the total number of mice infected. FLA960 had a significant difference in the number of mice with systemic disease compared to FLA954 ( P = 0.04, FLA960 compared to FLA954 by 2 test)
152 Figure 510. Subcutaneous lesions of mice infected with T6SS rtxA1 mutants Iron dextrantreated mice were s.c. inoculated with 3,000 CFU of (A) FLA954 ( rtxA1 ), (B) FLA960 ( rtxA1 vgrG rtxA1 hcp ). After euthanization, the skin was peeled back to reveal the s.c. lesion at the site of injection.
153 CHAPTER 6 DISCUSSION Cutaneous lesions with extensive tissue damage occur during both primary septicemia and wound infection by V. vulnificus The wound infections and secondary bullous lesions that form during primary septicemia can become necrotic, often requiring surgical debridement of the tissue or amputation of the limb (2). V. vulnificus causes extensive damage and is highly invasive, reflected by its ability to invade through the intestinal wall or through the tissues at the site of the wound, causing a systemic infection and eventually death (2). What is causing the tissue damage observed during infection? This is a question that has remained unanswered, despite more than twenty years of research on the pathogenesis of V. vulnificus V. vulnificus is highly cytotoxic in cell culture, and studies to identify the key cytotoxic factor(s) are ongoing. Initial reports of the cytotoxic activity in culture supernatants initiated the interest in secreted toxic factors (42) V. vulnificus produces a spectrum of extracellular enzymes that could have potential cytotoxic activity, including the cytolysin (VvhA) (44) metalloprotease (VvpE) (50 93) the RTX toxins ( T his study, (87 89) type VI secreted factors (108) other hemolysins (38 39) and phospholipases (40 41) VvhA and VvpE are the best characterized of the putative toxins. Injection of purified VvhA or VvpE induces extensive dermonecrosis in the skin tissues mice comparable to infection by V. vulnificus (44 49) These studies concluded that VvhA and/or VvpE were the cause of the extensive tissue damage. However, strains with mutations constructed in either vvhA vvpE or both genes were still able to cause severe tissue damage and lethal infection, similar to the wild type (47,52, 53) It became
154 clear that while VvhA and VvpE may contribute tissue damage, they are not essential for it and other factors must be contributing t o it. What Role Do the RTX Toxins Play? The RTX toxins represented a promising direction in identifying the destructive factors of V. vulnificus For years, RTX toxins have been recognized for their hemolytic and leukotoxic activities in other bacterial p athogens (70) During this investigation, we have characterized the three putative RTX toxins of V. vulnificus and determined that RtxA1 is the major cytotoxic factor of V. vul nificus Despite the potent cytotoxicity of RtxA1 and attenuation of an rtxA1 mutant for virulence, rtxA1 mutants are still able to cause tissue damage at the site of the s.c. inoculation in mice. We speculated that the other RTX toxins, RtxA2 and RtxA3, may act in concert with RtxA1. Upon mutation of the rtxA2 and rtxA3 genes, we discovered that these toxins were not necessary for cytotoxicity or virulence. It still remains possible that the presence of any one of the toxins may cause damage and lethal infection. Therefore, we are in the process of constructing a triple rtxA1/rtxA2/rtxA3 mutant. Interestingly, an rtxA1 mutant had cytotoxic activity when incubated with INT 407 cells for longer time periods than the usual cytotoxicity assay (>1 hour). This result hinted that RtxA1 may not be the only important cytotoxic factor. Deletion of vvhA in combination with rtxA1 abolished this residual cytotoxicity, confirming previous reports (43) that vvhA does have a role in cytotoxicity in vitro. Studies in our laboratory demonstrated that cytotoxicity in vitro does not correlate with virulence in vivo, and the rtxA1/vvhA mutant strain provides supports this finding. Despite the noncytotoxic phenotype, the double mutant was not more attenuated than the rtxA1 mutant, and it
155 was still able to cause severe skin infections and lethality in mice at a 10fold higher inoculum than the minimum lethal dose for t he wild type parent. What, then, is the function of RtxA1 during infection? The defect of the bacteria to cause a lethal infection at low inocula and the increase in i.p. LD50 observed by other groups (87 89) indicate that RtxA1 may be involved in invasion into the bloodstream to cause systemic disease. Kim, et al (87) demonstrated that an rtxA1 mutant is defective at invading from the intestine into the bloodstream. RtxA1 may also aid in evasion of the host innate immune response. There are RTX toxins classified as leukotoxins (96) ; however, examination of the s.c. lesions in the mice suggests that the V. vulnificus rtxA1 mutant is still capable of killing PMNs. This does not rule out a role for RtxA1 in PMN killing, but it indicates that there are other important factors inv olved in the killing and evasion of immune defenses that are still active in the absence of RtxA1. Type VI Secreted Factors are Not Essential for Virulence The newly identified T6SS has been implicated in virulence of several pathogens, including V. cholerae (100) P aeuruginosa (101) F tularensis (103) and E tarda (104) We identified a region on chromosome II of CMCP6 that encodes the 13 conserved proteins defined as the T6SS "core components (108) Hcp and VgrG are believed to be the key factors for type VI secretion in other T6SS encoding bacteria (57) so we targeted these genes for deletion. Deletion of either gene had no effect on cytotoxicity to INT 407 cells or virulence in mice. The vgrG mutant induced less apoptosis than did the wild type, although this effect was not observed for the hcp mutant. This result indicates that hcp may not be necessary for vgrG secretion. Studies with V. cholerae suggest that phenotypes of T6SS mutants may not be observable in the presence of the RtxA toxin (113) so we constructed double mutants
156 rtxA1 ::tetA vgrG :: aph and rtxA1 ::tetA While the deletion of hcp or vgrG in the rtxA1 background has no effect on cytotoxicity compared to the rtxA1 there may be a small effect on virulence, although further studies are needed to confirm this. It is possible that RtxA1 compensates for the loss of T6SS in the single vgrG ::aph and hcp :: cat mutants, and that is why we were able to observe any effect in the absence of rtxA1 Despite these results, it remains c lear that T6SS is not essential for virulence in mice, and that it has a very small role, if any, in pathogene sis. If Not RTX Toxins or T6SS, What is Causing Damage? Other putative toxins such as the HlyIII, VllY, and phospholipases have been proposed to contribute virulence, although little experimental evidence exists. Chen, et al (38) i dentified the HlyIII as a putative hemolytic protein that confers hemolytic activity in E. coli. HlyIII has similarity to the hemolysin III ( hlyIII) of Bacillus cereus This V. vulnificus HlyIII is also highly similar to other putative hemolysins including those of V cholerae Y pestis and S enterica. HlyIII of B. cereus is a poreforming hemolysin; however, similar genes in other species have not been examined. V. vulnificus HlyIII is indicated to be a hemolysin, although the pore forming ability has not been observed (35). The V. vulnificus hlyIII mutant analyzed by Chen (38) remained hemolytic, although the mutant was attenuated in i.p. inoculated mice. During our analysis of toxins of V. vulnificus we also deleted hlyIII. Similar to observations made by Chen, et al ., our V. vulnificus hlyIII mutant was hemolytic on both rabbit blood agar plates and sheep blood agar plates and the hlyIII mutant was cytotoxic in cell culture hlyIII mutant was reduced for virulence at the minimum lethal dose of 300 CFU in s.c. inoculated mice, causing systemic infection in only three of five mice. However, the
157 mutant was able to cause wildtype levels of skin infection in the mice, suggesting HlyIII does not have a role in tissue damage. C hang, et al (39) characterized another putative hemolysin, VllY, with similarity to the legiolysin (Lly) of Legionella pneumophila Recombinant expression of VllY in E. coli caused the bacteria to becom e hemolytic and pigmented ; however, a vllY mutant in V. vulnificus was not constructed by Chang et al (39) During this study, we constructed a vllY mutant and observed that it was as hemolytic as the wild type on r abbit blood agar plates and sheep blood agar plates and was fully cytotoxic in cell culture. The vllY mutant was slightly attenuated, causing lethal infection in only three of five mice when inoculated at 300 CFU, yet it was able to cause wild type levels of skin infection VllY and Lly have similarity to the family of 4 hydroxyphenylpyruvate dioxygenase (HppD) proteins, which are eukaryotic and prokaryotic proteins involved in the catabolism of tyrosine (39 122) HppD is involved in the production of pyomelanin and results in pigmentation of the bacteria (48) An hppD mutant of Burkholderia cenocepacia is non pigmented and more susceptible to oxidative stress and killing by macrophages (123) L. pneum ophila Lly is involved in the production of melanin; however, a mutation in lly has no effect on hemolysis or on the survival of the intracellular survival of the bacteria in macrophages (124) I n V. cholerae, m elanin production is induced during osmotic and temperature stress, and the HppD/tyrosine catabolism pathway has recently been attributed t o expression of certain virulence facto rs (125) A V. cholerae mutant in homogentisate 1,2dioxygenase ( hmgA ) downstream of hppD was hyperpigmented due to overproduction of melanin. The
158 hmgA mutant had elevated expression of cholera toxin and the toxincoregulated pilus and had an increased ability to colonize mice compared to wild type (125) While differences in pigmentation of the V. vulnificus vllY have not been observed, this does not rule out a role in melanin production. V. vulnificus VllY may have a role in virulence, perhaps through regulation of other virulence factors, as observed for V. cholerae or by enhancing survival during oxidative stress and other environmental stresses. The phospholipase activity of V. vulnificus has also been proposed to have a role in virulence. T esta, et al (40) demonstrated that V. vulnificus possesses phospholipase A1/A2 and lysophospholipase activities, but not ph ospholipase C activity. Koo, et al (41) subsequently propose d that this phospholipase A (PLA) activity was important for virulence in a mouse model. The researchers used tetracycline to inhibit phospholipase activity during infection and claimed that it was the inhibition of PLA activity that caused attenuation in virulence (41) Unfortunately, there were many gaps in the study by Koo, et al (41) including the lack of construction and analysis of a phospholipase mutant. PLA has a role in virulence of other bacterial pathogens (126) A lecithin dependent hemolysin with PLA activity contributes to virulence of Yersinia enterocolitica by promoting colonization and by inducing inflammation and necrosis of Peyers patches and mesenteric lymph nodes (127) PLA enzymes can also be hemolytic/cytolytic by causing destabilization of cell membranes (126) For example, the lecithindependent hemolysin (LDH) V. parahaemolyticus with PLA activity causes lysis of erythrocytes (128) In light of this unresolved issue of the V. vulnificus PLA, we followed up on the report by Koo, et al ., to determine if PLA activity of V. vulnificus was important for
159 virulence. We examined the annotated genome of V. vulnificus CMCP6 and discovered that CMCP6 encodes two genes with annotated PLA activity, phospholipase/lecithinase/hemolysin ( tlh ) and an outer membrane phospholipase A ( ompla). We deleted these genes fr om CMCP6 individually and analyzed the mutants for hemolytic ability, crude phospholipase activity, and virulence. Both mutants were hemolytic on rabbit or sheep blood agar plates, despite the fact that the phospholipase/lecithinase/hemolysin protein has 73% identity to the thermolabile hemolysin TLH (also known as the lecithindependent hemolysin, LDH) of V. parahaemolyticus We observed lecithinase activity on tryptic soy agar plates supplemented with 5% (vol/vol) egg yolk. The tlh mutant had a very sl ight difference in lecithinase/phospholipase activity compared the wild type, but no difference in activity of the ompla mutant compared to wildtype. Both mutants were as virulent as the wild type in s.c. inoculated, irontreated mice. Until a mutant st rain lacking phospholipase activity is isolated, the claims made by Koo cannot be completely disregarded. Phospholipase activity may still contribute the invasiveness or the tissue damage caused by V. vulnificus. V. vulnificus is an Accidental Hum an Pathogen By simply looking at the annotated functions of the proteins described above, one would speculate that those proteins would be involved in virulence of V. vulnificus Hemolysins, proteases, RTX toxins, type VI secreted effectors, or phospholipases do contribute to virulence in many pathogens. V. vulnificus is first and foremost an environmental organism residing in shellfish and seawater that predominantly causes disease in people with underlying health conditions (19) The rapid nature of the disease and the fact that humans are a deadend host are reflective of the fact that V.
160 vulnificus has not evolved to be a human pathogen, and infection of hum ans is not a requirement for V. vulnificus to maintain its lifestyle Why, then, would V. vulnificus possess so many destructive enzymes? V. vulnificus lives in an environment in which it must compete for nutrients and colonize filter feeding shellfish, all the while avoiding predatory organisms which may feed on bacteria. The hemolysins and proteases, which we classify as toxins, may be important in acquiring and breaking down nutrients or avoiding predation by other organisms. The RtxA2 and RtxA3 toxi ns with the putative VWF domain may be involved in adherence and colonization of shellfish or other surfaces. The higher expression of rtxA2 at 20C and 30C, compared to 37C, provides evidence for a role of RtxA2 in the env ironment as opposed to human disease (98) The T6SS may also be an environmental survival factor. The T6SS was identified in V. cholerae by identifying mutants defective at killing the predatory amoeba D discoideum (100) Perhaps the T6SS of V. vulnificus has a similar role in avoiding predatory organisms. The V. vulnificus T6SS genes resemble the T6SS genes of symbiotic bacteria, more so than the genes of pathogenic bacteria, indicating T6SS effectors may have role in colonizing or survival within shellfish. In summary, each of the toxins described here most likely has an important role for survival and growth in the environment and by chance happen to be destructive to host cells and tissue. Final Remarks We have not been able to determine what is causing the destructive tissue damage dur ing infection with V. vulnificus although we have provided evidence that it is most likely multifactoral. V. vulnificus produces many extracellular enzymes with potential destructive activity, but deletion of any one of them individually has little or no effect on virulence. It is possible that the production of the other toxins will compensate
161 for the loss of one or that the key destructive factor has not yet been examined. Damage of the tissues may also be a side effect of the rapid growth of the bact eria and the inability of the host immune response to handle the infection. A fadR mutant exhibits slow growth in vitro and is highly attenuated in vivo (129) This mutant is defective at causing skin infections in mice at inocula as high as 105 CFU, despite evidence suggesting FadR has no role in cytotoxicity to host cells in vitro The most definitive virulence factors of V. vulnificus identified to date contribute mainly to growth and survival of the organism during infection (i.e., iron acquisition, capsule). Despite numerous attempts by many researchers, the factors causing destructive tissue damage and invasion during infection remain elusive.
162 REFERENCE LIST 1. Hlady, W.G. and Klontz, K.C. 1996. The epidemiology of Vibrio infections in Florida, 19811993. J. Infect. Dis. 173:11761183. 2. Strom, M.S. and Paranjpye, R. N. 2000. Epidemiology and pathogenesis of Vibrio vulnificus Microbes Infect. 2:177188. 3. Halow, K.D., Harner, R.C., and Fontenelle, L.J. 1996. Primary skin infections secondary to Vibrio vulnificus : the role of operative intervention. J. Am. Coll.Surg 183:329334. 4. Vollberg C.M. and Herrera, J.L. 1997. Vibrio vulnificus infection: an important cause of septicemia in patients with cirrhosis. South. Med. J. 90:10401042. 5. Kumamoto K.S. and Vukich, D.J. 1998. Clinical infections of Vibrio vulni ficus : a case report and review of the literature. J. Emerg. Med. 16:61 66. 6. Linkous D.A. and Oliver J.D. 1999. Pathogenesis of Vibrio vulnificus FEMS Microbiol. Lett. 174:207214. 7. Bullen J.J. Spalding P.B., Ward C.G., and Gutteridge J.M. 1991. Hemochromatosis, iron and septicemia caused by Vibrio vulnificus Arch. Intern Med. 151:16061609. 8. Bisharat N., Agmon V., Finkelstein R., Raz R., Ben Dror G., Lerner L., Soboh S., Colodner R., Cameron D.N., Wykstra D.L. et al 1999. C linical, epidemiological, and microbiological features of Vibrio vulnificus biogroup 3 causing outbreaks of wound infection and bacteraemia in Israel. Israel Vibrio Study Group. Lancet 354:14211424. 9. Martin S.J. and Siebeling R.J. 1991. Identificati on of Vibrio vulnificus O serovars with antilipopolysaccharide monoclonal antibody. J. Clin. Microbiol. 29:16841688. 10. Biosca E.G., Oliver J.D., and Amaro C. 1996. Phenotypic characterization of Vibrio vulnificus biotype 2, a lipopolysaccharidebas ed homogeneous O serogroup within Vibrio vulnificus Appl. Environ. Microbiol. 62:918927. 11. Nilsson W.B., Paranjpye, R.N., DePaola A., and Strom M.S. 2003. Sequence polymorphism of the 16S rRNA gene of Vibrio vulnificus is a possible indicator of strain virulence. J. Clin. Microbiol. 41:442 446. 12. Vickery M.C., Nilsson W.B., Strom M.S., Nordstrom J.L., and DePaola, A. 2007. A real time PCR assay for the rapid determination of 16S rRNA genotype in Vibrio vulnificus J. Microbiol. Methods 68:3 76384.
163 13. Warner J.M. and Oliver, J.D. 1999. Randomly amplified polymorphic DNA analysis of clinical and environmental isolates of Vibrio vulnificus and other vibrio species. Appl. Environ. Microbiol. 65:11411144. 14. Rosche T.M., Yano Y., and Ol iver J.D. 2005. A rapid and simple PCR analysis indicates there are two subgroups of Vibrio vulnificus which correlate with clinical or environmental isolation. Microbiol. Immunol. 49:381389. 15. Chatzidaki Livanis M., Hubbard, M.A., Gordon K., Harwo od, V.J., and Wright, A.C. 2006. Genetic distinctions among clinical and environmental strains of Vibrio vulnificus Appl. Environ. Microbiol. 72:61366141. 16. Bisharat N., Cohen D.I., Harding R.M., Falush D., Crook D.W., Peto T., and Maiden, M.C. 2005. Hybrid Vibrio vulnificus Emerg. Infect. Dis. 11:3035. 17. Bisharat, N., Cohen D.I., Maiden, M.C., Crook, D.W., Peto, T., and Harding, R.M. 2007. The evolution of genetic structure in the marine pathogen, Vibrio vulnificus Infect. Genet. Evol. 7:685 693. 18. Cohen, A.L., Oliver, J.D., DePaola, A., Feil, E.J., and Boyd, E.F. 2007. Emergence of a virulent clade of Vibrio vulnificus and correlation with the presence of a 33kilobase genomic island. Appl. Environ. Microbiol. 73:55535565. 19. Gu lig, P.A., Bourdage, K.L., and Starks, A.M. 2005. Molecular Pathogenesis of Vibrio vulnificus J. Microbiol. 43:118 131. 20. Yoshida, S., Ogawa, M., and Mizuguchi, Y. 1985. Relation of capsular materials and colony opacity to virulence of Vibrio vulnific us Infect. Immun. 47:446451. 21. Wright, A.C., Powell, J.L., Tanner, M.K., Ensor, L.A., Karpas, A.B., Morris, J.G., Jr., and Sztein, M.B. 1999. Differential expression of Vibrio vulnificus capsular polysaccharide. Infect. Immun. 67:22502257. 22. Wri ght, A.C., Simpson, L.M., Oliver, J.D., and Morris, J.G., Jr. 1990. Phenotypic evaluation of acapsular transposon mutants of Vibrio vulnificus Infect. Immun. 58:17691773. 23. Simpson, L.M., White, V.K., Zane, S.F., and Oliver, J.D. 1987. Correlation between virulence and colony morphology in Vibrio vulnificus Infect. Immun. 55:269272. 24. Kreger, A., DeChatelet, L., and Shirley, P. 1981. Interaction of Vibrio vulnificus with human polymorphonuclear leukocytes: association of virulence with resistanc e to phagocytosis. J. Infect. Dis. 144:244248.
1 64 25. Shinoda, S., Kobayashi, M., Yamada, H., Yoshida, S., Ogawa, M., and Mizuguchi, Y. 1987. Inhibitory effect of capsular antigen of Vibrio vulnificus on bactericidal activity of human serum. Microbiol. Imm unol. 31:393 401. 26. Weinberg, E.D. 2000. Microbial pathogens with impaired ability to acquire host iron. Biometals 13:85 89. 27. Wright, A.C., Simpson, L.M., and Oliver, J.D. 1981. Role of iron in the pathogenesis of Vibrio vulnificus infections. Inf ect. Immun. 34:503507. 28. Starks, A.M., Schoeb, T.R., Tamplin, M.L., Parveen, S., Doyle, T.J., Bomeisl, P.E., Escudero, G.M., and Gulig, P.A. 2000. Pathogenesis of infection by clinical and environmental strains of Vibrio vulnificus in iron dextrantre ated mice. Infect. Immun. 68:57855793. 29. Litwin, C.M., Rayback, T.W., and Skinner, J. 1996. Role of catechol siderophore synthesis in Vibrio vulnificus virulence. Infect. Immun. 64:28342838. 30. van Asbeck, B.S., Marx, J.J., Struyvenberg, A., and V erhoef, J. 1984. Functional defects in phagocytic cells from patients with iron overload. J. Infect. 8:232240. 31. Hor, L.I., Chang, Y.K., Chang, C.C., Lei, H.Y., and Ou, J.T. 2000. Mechanism of high susceptibility of ironoverloaded mouse to Vibrio vul nificus infection. Microbiol. Immunol. 44:871878. 32. Starks, A.M., Bourdage, K.L., Thiaville, P.C., and Gulig, P.A. 2006. Use of a marker plasmid to examine growth and death of Vibrio vulnificus in infected mice. Mol. Microbiol. 61:310 323. 33. Kim, Y.R. and Rhee, J.H. 2003. Flagellar basal body flg operon as a virulence determinant of Vibrio vulnificus Biochem. Biophys. Res. Commun. 304:405410. 34. Lee, J.H., Rho, J.B., Park, K.J., Kim, C.B., Han, Y.S., Choi, S.H., Lee, K.H., and Park, S.J. 2004. Role of flagellum and motility in pathogenesis of Vibrio vulnificus Infect. Immun. 72:49054910. 35. Gander, R.M. and LaRocco, M.T. 1989. Detection of piluslike structures on clinical and environmental isolates of Vibrio vulnificus J. Clin. Microbiol. 27:10151021. 36. Paranjpye, R.N., Lara, J.C., Pepe, J.C., Pepe, C.M., and Strom, M.S. 1998. The type IV leader peptidase/ N methyltransferase of Vibrio vulnificus controls factors required for adherence to HEp2 cells and virulence in ironoverloaded mi ce. Infect. Immun. 66:5659 5668. 37. Paranjpye, R.N. and Strom, M.S. 2005. A Vibrio vulnificus type IV pilin contributes to biofilm formation, adherence to epithelial cells, and virulence. Infect. Immun. 73:1411 1422.
165 38. Chen, Y.C., Chang, M.C., Chuan g, Y.C., and Jeang, C.L. 2004. Characterization and virulence of hemolysin III from Vibrio vulnificus Curr. Microbiol. 49:175179. 39. Chang, T.M., Chuang, Y.C., Su, J.H., and Chang, M.C. 1997. Cloning and sequence analysis of a novel hemolysin gene ( vl lY ) from Vibrio vulnificus Appl. Environ. Microbiol. 63:38513857. 40. Testa, J., Daniel, L.W., and Kreger, A.S. 1984. Extracellular phospholipase A2 and lysophospholipase produced by Vibrio vulnificus Infect.Immun. 45:458463. 41. Koo, B.S., Lee, J. H., Kim, S.C., Yoon, H.Y., Kim, K.A., Kwon, K.B., Kim, H.R., Park, J.W., and Park, B.H. 2007. Phospholipase A as a potent virulence factor of Vibrio vulnificus Int.J. Mol. Med. 20:913918. 42. Kreger, A. and Lockwood, D. 1981. Detection of extracellular toxin(s) produced by Vibrio vulnificus Infect. Immun. 33:583590. 43. Gray, L.D. and Kreger, A.S. 1985. Purification and characterization of an extracellular cytolysin produced by Vibrio vulnificus Infect.Immun. 48:6272. 44. Gray, L.D. and Kreger, A.S. 1987. Mouse skin damage caused by cytolysin from Vibrio vulnificus and by V. vulnificus infection. J. Infect. Dis. 155:236241. 45. Kwon, K.B., Yang, J.Y., Ryu, D.G., Rho, H.W., Kim, J.S., Park, J.W., Kim, H.R., and Park, B.H. 2001. Vibrio vulnificu s cytolysin induces superoxide anioninitiated apoptotic signaling pathway in human ECV304 cells. J. Biol. Chem. 276:4751847523. 46. Kim, H.R., Rho H.W., Jeong M.H., Park J.W., Kim J.S., Park B.H., Kim U.H., and Park S.D. 1993. Hemolytic mechanis m of cytolysin produced from V. vulnificus Life Sci. 53:571577. 47. Wright A.C. and Morris, J.G., Jr. 1991. The extracellular cytolysin of Vibrio vulnificus : inactivation and relationship to virulence in mice. Infect. Immun. 59:192197. 48. Nosanchuk, J.D. and Casadevall A. 2003. The contribution of melanin to microbial pathogenesis. Cell Microbiol. 5:203223. 49. Kothary M.H. and Kreger A.S. 1987. Purification and characterization of an elastolytic protease of Vibrio vulnificus J. Gen. Microbi ol. 133:17831791. 50. Miyoshi S., Hirata Y., Tomochika K., and Shinoda S. 1994. Vibrio vulnificus may produce a metalloprotease causing an edematous skin lesion in vivo. FEMS Microbiol. Lett. 121:321325. 51. Shao C.P. and Hor L.I. 2000. Metallo protease is not essential for Vibrio vulnificus virulence in mice. Infect. Immun. 68:35693573.
166 52. Jeong K.C., Jeong H.S., Rhee J.H., Lee S.E., Chung S.S., Starks A.M., Escudero, G.M., Gulig P.A., and Choi S.H. 2000. Construction and phenotypic evaluation of a Vibrio vulnificus vvpE mutant for elastolytic protease. Infect. Immun. 68:50965106. 53. Fan J.J., Shao C.P., Ho Y.C., Yu C.K., and Hor L.I. 2001. Isolation and characterization of a Vibrio vulnificus mutant deficient in both extracellular metalloprotease and cytolysin. Infect. Immun. 69:59435948. 54. Falkow S. 1988. Molecular Koch's postulates applied to microbial pathogenicity. Rev. Infect. Dis. 10:S274S276. 55. Gulig P.A. 1993. Use of isogenic mutants to study bacterial vir ulence factors. J. Microbiol. Methods 18:275287. 56. Satchell K.J. 2007. MARTX, multifunctional autoprocessing repeats in toxin toxins. Infect. Immun. 75:50795084. 57. Bingle L.E., Bailey C.M., and Pallen M.J. 2008. Type VI secretion: a beginner' s guide. Curr. Opin. Microbiol. 11:38. 58. Simon R., Priefer U., and Phler A. 1983. A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in gram negative bacteria. Bio/Technology 1:784791. 59. CerdaCuellar M., Jofre J., and Blanch, A.R. 2000. A selective medium and a specific probe for detection of Vibrio vulnificus Appl. Environ. Microbiol. 66:855 859. 60. Gulig P.A., Doyle T.J., Hughes J.A., and Matsui H. 1998. Analysis of host cells associated with the Spv mediated increased intracellular growth rate of Salmonella typhimurium in mice. Infect. Immun. 66:24712485. 61. Ruff M.R. and Gifford, G.E. 1980. Purification and physicochemical characterization of rabbit tumor necrosis factor. J. Immun ol. 125:16711676. 62. Gulig P.A., Tucker M.S., Thiaville P.C., Joseph J.L., and Brown, R.N. 2009. USERTM Friendly cloning coupled with chitinbased natural transformation enables rapid mutagenesis of Vibrio vulnificus Appl. Environ. Microbiol. 75. 63. Donnenberg M.S. and Kaper J.B. 1991. Construction of an eae deletion mutant of Enteropathogenic Escherichia coli by using a positiveselection suicide vector. Infect. Immun. 59:4310 4317. 64. Chen, C.Y., Wu K.M., Chang Y.C., Chang C.H., Tsai H.C., Liao T.L., Liu Y.M., Chen H.J., Shen A.B., Li J.C. et al 2003. Comparative genome analysis of Vibrio vulnificus a marine pathogen. Genome Res. 13:25772587.
167 65. Kim Y.R., Lee S.E., Kim C.M., Kim S.Y., Shin E.K., Shin D.H., Chung S.S., Choy H.E., Progulske Fox A., Hillman J.D. et al 2003. Characterization and pathogenic significance of Vibrio vulnificus antigens preferentially expressed in septicemic patients. Infect. Immun. 71:54615471. 66. Perez Casal J., Caparon, M.G., and Sc ott J.R. 1991. Mry, a trans acting positive regulator of the M protein gene of Streptococcus pyogenes with similarity to the receptor proteins of twocomponent systems. J. Bacteriol. 173:26172624. 67. Prentki P. and Krisch, H.M. 1984. In vitro inserti onal mutagenesis with a selectable DNA fragment. Gene 29:303 313. 68. Connell T.D., Martone A.J., and Holmes R.K. 1995. A new mobilizable cosmid vector for use in Vibrio cholerae and other gram negative bacteria. Gene 153:8587. 69. Yanisch Perron, C., Vieira, J., and Messing J. 1985. Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33:103119. 70. Welch R.A. 2001. RTX toxin structure and function: a story of numerous anomalies and f ew analogies in toxin biology. Curr. Top. Microbiol. Immunol. 257:85111. 71. Glaser P., Ladant D., Sezer O., Pichot, F., Ullmann A., and Danchin, A. 1988. The calmodulinsensitive adenylate cyclase of Bordetella pertussis : cloning and expression in Escherichia coli Mol. Microbiol. 2:19 30. 72. Thompson S.A., Wang L.L., West, A., and Sparling P.F. 1993. Neisseria meningitidis produces ironregulated proteins related to the RTX family of exoproteins. J. Bacteriol. 175:811818. 73. Lin W., Full ner K.J., Clayton R., Sexton J.A., Rogers M.B., Calia K.E., Calderwood, S.B., Fraser C., and Mekalanos J.J. 1999. Identification of a Vibrio cholerae RTX toxin gene cluster that is tightly linked to the cholera toxin prophage. Proc. Natl. Acad. Sci. U.S.A 96:10711076. 74. Welch R.A., Forestier C., Lobo A., Pellett S., Thomas W., Jr. and Rowe G. 1992. The synthesis and function of the Escherichia coli hemolysin and related RTX exotoxins. FEMS Microbiol. Immunol. 5:2936. 75. Lally E.T., Hill, R.B., Kieba I.R., and Korostoff J. 1999. The interaction between RTX toxins and target cells. Trends Microbiol. 7:356361. 76. Forestier C. and Welch, R.A. 1990. Nonreciprocal complementation of the hlyC and lktC genes of the Escherichia coli he molysin and Pasteurella haemolytica leukotoxin determinants. Infect. Immun. 58:828 832.
168 77. Stanley P., Packman L.C., Koronakis V., and Hughes C. 1994. Fatty acylation of two internal lysine residues required for the toxic activity of Escherichia col i hemolysin. Science 266:19921996. 78. Westrop G.D., Hormozi E.K., Da Costa N.A., Parton R., and Coote, J.G. 1996. Bordetella pertussis adenylate cyclase toxin: proCyaA and CyaC proteins synthesised separately in Escherichia coli produce active toxi n in vitro. Gene 180:9199. 79. Trent M.S., Worsham L.M., and Ernst Fonberg, M.L. 1998. The biochemistry of hemolysin toxin activation: characterization of HlyC, an internal protein acyltransferase. Biochemistry 37:46444652. 80. Boardman B.K. and S atchell K.J. 2004. Vibrio cholerae strains with mutations in an atypical type I secretion system accumulate RTX toxin intracellularly. J. Bacteriol. 186:8137 8143. 81. Lee, B.C., Lee J.H., Kim M.W., Kim B.S. Oh M.H., Kim K.S., Kim T.S., and Choi S.H. 2008. Vibrio vulnificus rtxE is important for virulence, and its expression is induced by exposure to host cells. Infect. Immun. 76:15091517. 82. Fullner K.J. and Mekalanos J.J. 2000. In vivo covalent cross linking of cellular actin by the Vibri o cholerae RTX toxin. EMBO J. 19:53155323. 83. Fullner K.J., Lencer W.I., and Mekalanos J.J. 2001. Vibrio choleraeinduced cellular responses of polarized T84 intestinal epithelial cells are dependent on production of cholera toxin and the RTX toxin. Infect. Immun. 69:63106317. 84. Sheahan, K.L., Cordero, C.L., and Satchell K.J. 2004. Identification of a domain within the multifunctional Vibrio cholerae RTX toxin that covalently cross links actin. Proc. Natl. Acad. Sci. U.S.A 101:97989803. 85. Sheahan, K.L. and Satchell K.J. 2007. Inactivation of small Rho GTPases by the multifunctional RTX toxin from Vibrio cholerae. Cell Microbiol. 9:1324 1335. 86. Sheahan, K.L., Cordero, C.L., and Satchell K.J. 2007. Autoprocessing of the Vibrio cholerae RTX toxin by the cysteine protease domain. EMBO J. 26:25522561. 87. Kim Y.R., Lee S.E., Kook H., Yeom J.A. Na H.S., Kim S.Y., Chung S.S., Choy H.E., and Rhee J.H. 2008. Vibrio vulnificus RTX toxin kills host cells only after contact of the ba cteria with host cells. Cell Microbiol. 10:848862. 88. Lee, J.H., Kim M.W., Kim B.S., Kim S.M., Lee B.C., Kim T.S., and Choi S.H. 2007. Identification and characterization of the Vibrio vulnificus rtxA essential for cytotoxicity in vitro and virul ence in mice. J. Microbiol. 45:146 152.
169 89. Liu M., Alice A.F., Naka H., and Crosa, J.H. 2007. The HlyU protein is a positive regulator of rtxA1 a gene responsible for cytotoxicity and virulence in the human pathogen Vibrio vulnificus Infect. Immun. 75:32823289. 90. Lee, B.C., Choi, S.H., and Kim T.S. 2008. Vibrio vulnificus RTX toxin plays an important role in the apoptotic death of human intestinal epithelial cells exposed to Vibrio vulnificus Microbes. Infect. 10:1504 1513. 91. Kashimoto, T ., Ueno S., Hanajima M., Hayashi H., Akeda Y., Miyoshi S., Hongo, T., Honda T., and Susa N. 2003. Vibrio vulnificus induces macrophage apoptosis in vitro and in vivo. Infect. Immun. 71:533535. 92. Kashimoto, T., Ueno S., Hayashi H., Hanajima, M ., Yoshioka K., Yoshida, K., Mutoh, K., and Susa, N. 2005. Depletion of lymphocytes, but not neutrophils, via apoptosis in a murine model of Vibrio vulnificus infection. J. Med. Microbiol. 54:1522. 93. Kothary M.H. and Kreger, A.S. 1985. Production and partial characterization of an elastolytic protease of Vibrio vulnificus Infect. Immun. 50:534540. 94. Dowling A.J., Daborn P.J., Waterfield N.R., Wang, P., Streuli C.H., and ffrenchConstant R.H. 2004. The insecticidal toxin Makes caterpillars floppy (Mcf) promotes apoptosis in mammalian cells. Cell Microbiol. 6:345353. 95. Dowling A.J., Waterfield N.R., Hares M.C., Le Goff G., Streuli C.H., and ffrenchConstant R.H. 2007. The Mcf1 toxin induces apoptosis via the mitochondrial pathway and apoptosis is attenuated by mutation of the BH3like domain. Cell Microbiol. 9:24702484. 96. Narayanan, S.K., Nagaraja, T.G., Chengappa, M.M., and Stewart, G.C. 2002. Leukotoxins of gram negative bacteria. Vet. Microbiol. 84:337356. 97. Horton, R.M ., Hunt H.D., Ho S.N., Pullen J.K., and Pease, L.R. 1989. Engineering hybrid genes without the use of restriction enzymes: gene splicing by overlap extension. Gene 77:6168. 98. Chou, L.F., Peng H.L., Yang Y.C., Kuo M.C., and Chang H.Y. 2009. Localization and characterization of VVA0331, a 489kDa RTX like protein, in Vibrio vulnificus YJ016. Arch. Microbiol. 191:441450. 99. Westrop G., Hormozi K., da Costa N., Parton R., and Coote, J. 1997. Structure function studies of the adenylate cyclas e toxin of Bordetella pertussis and the leukotoxin of Pasteurella haemolytica by heterologous C protein activation and construction of hybrid proteins. J. Bacteriol. 179:871879. 100. Pukatzki S., Ma A.T., Sturtevant, D., Krastins B., Sarracino D., N elson W.C., Heidelberg J.F., and Mekalanos J.J. 2006. Identification of a conserved
170 bacterial protein secretion system in Vibrio cholerae using the Dictyostelium host model system. Proc. Natl. Acad.Sci. U.S. A 103:15281533. 101. Mougous J.D., Cuff M.E., Raunser S., Shen A., Zhou M., Gifford C.A., Goodman A.L., Joachimiak G., Ordonez C.L., Lory S. et al 2006. A virulence locus of Pseudomonas aeruginosa encodes a protein secretion apparatus. Science 312:15261530. 102. Suarez G., Sierra J.C., Sha J., Wang S., Erova T.E., Fadl A.A., Foltz S.M., Horneman, A.J., and Chopra, A.K. 2008. Molecular characterization of a functional type VI secretion system from a clinical isolate of Aeromonas hydrophila. Microb. Pathog. 44:344 361. 103. Lud u, J.S., de Bruin O.M., Duplantis B.N., Schmerk C.L., Chou A.Y., Elkins K.L., and Nano, F.E. 2008. The Francisella pathogenicity island protein PdpD is required for full virulence and associates with homologues of the type VI secretion system. J. Bact eriol. 190:45844595. 104. Wang, X., Wang Q., Xiao J., Liu Q., Wu H., Xu L., and Zhang Y. 2009. Edwardsiella tarda T6SS component evpP is regulated by esrB and iron, and plays essential roles in the invasion of fish. Fish. Shellfish. Immunol. 27:469 477. 105. Schell M.A., Ulrich R.L., Ribot W.J., Brueggemann, E.E., Hines H.B., Chen D., Lipscomb L., Kim H.S., Mrazek J., Nierman W.C. et al 2007. Type VI secretion is a major virulence determinant in Burkholderia mallei Mol. Microbiol. 64:1 4661485. 106. Wu H.Y., Chung P.C., Shih H.W., Wen S.R., and Lai E.M. 2008. Secretome analysis uncovers an Hcpfamily protein secreted via a type VI secretion system in Agrobacterium tumefaciens J. Bacteriol. 190:28412850. 107. Fauvart M. and M ichiels, J. 2008. Rhizobial secreted proteins as determinants of host specificity in the rhizobium legume symbiosis. FEMS Microbiol. Lett. 285:19. 108. Boyer F., Fichant G., Berthod J., Vandenbrouck Y., and Attree I. 2009. Dissecting the bacterial type VI secretion system by a genome wide in silico analysis: what can be learned from available microbial genomic resources? BMC. Genomics 10:104. 109. Williams, S.G., Varcoe L.T., Attridge S.R., and Manning P.A. 1996. Vibrio cholerae Hcp, a secreted protein coregulated with HlyA. Infect. Immun. 64:283 289. 110. Zheng J. and Leung, K.Y. 2007. Dissection of a type VI secretion system in Edwardsiella tarda. Mol. Microbiol. 66:11921206.
171 111. Ballister E.R., Lai A.H., Zuckermann R.N., Cheng Y., and Mougous J.D. 2008. In vitro self assembly of tailorable nanotubes from a simple protein building block. Proc. Natl. Acad. Sci. U.S.A 105:37333738. 112. Pukatzki S., Ma A.T., Revel A.T., Sturtevant D., and Mekalanos J.J. 2007. Type VI secretio n system translocates a phage tail spikelike protein into target cells where it cross links actin. Proc. Natl. Acad. Sci. U.S.A 104:1550815513. 113. Ma A.T., McAuley S., Pukatzki S., and Mekalanos J.J. 2009. Translocation of a Vibrio cholerae type VI secretion effector requires bacterial endocytosis by host cells. Cell Host. Microbe 5:234243. 114. Bonemann, G., Pietrosiuk A., Diemand A., Zentgraf H., and Mogk A. 2009. Remodelling of VipA/VipB tubules by ClpV mediated threading is crucial for type VI protein secretion. EMBO J. 28:315 325. 115. Pukatzki S., McAuley S.B., and Miyata S.T. 2009. The type VI secretion system: translocation of effectors and effector domains. Curr. Opin. Microbiol. 12:1117. 116. Robinson J.B., Telepnev M.V. Zudina, I.V., Bouyer D., Montenieri J.A., Bearden, S.W., Gage K.L., Agar S.L., Foltz S.M., Chauhan S. et al 2009. Evaluation of a Yersinia pestis mutant impaired in a thermoregulated type VI like secretion system in flea, macrophage and murine mod els. Microb. Pathog. 47:243251 117. Ishikawa T., Rompikuntal P.K., Lindmark B., Milton D.L., and Wai S.N. 2009. Quorum sensing regulation of the two hcp alleles in Vibrio cholerae O1 strains. PLoS. One. 4:e6734. 118. Khajanchi B.K., Sha J., Ko zlova E.V., Erova T.E., Suarez G., Sierra J.C., Popov V.L., Horneman, A.J., and Chopra A.K. 2009. N Acyl homoserine lactones involved in quorum sensing control type VI secretion system, biofilm formation, protease production, and in vivo virulence fr om a clinical isolate of Aeromonas hydrophila Microbiology 155: 3518 3531. 119. Mattinen, L. Nissinen R., Riipi, T., Kalkkinen, N., and Pirhonen, M. 2007. Host extract induced changes in the secretome of the plant pathogenic bacterium Pectobacterium at rosepticum Proteomics. 7:35273537. 120. Mougous, J.D., Gifford C.A., Ramsdell T.L., and Mekalanos J.J. 2007. Threonine phosphorylation post translationally regulates protein secretion in Pseudomonas aeruginosa. Nat.Cell Biol. 9:797 803. 121. Harri sYoung L., Tamplin, M.L., Mason J.W., Aldrich H.L., and Jackson, J.K. 1995. Viability of Vibrio vulnificus in association with hemocytes of the American oyster ( Crassostrea virginica). Appl.Environ.Microbiol. 61:5257.
172 122. Arias Barrau E., Olivera E.R., Luengo J.M., Fernandez C., Galan B., Garcia J.L., Diaz E., and Minambres B. 2004. The homogentisate pathway: a central catabolic pathway involved in the degradation of Lphenylalanine, Ltyrosine, and 3 hydroxyphenylacetate in Pseudomonas puti da. J. Bacteriol. 186:50625077. 123. Keith K.E., Killip L., He P., Moran G.R., and Valvano, M.A. 2007. Burkholderia cenocepacia C5424 produces a pigment with antioxidant properties using a homogentisate intermediate. J. Bacteriol. 189:90579065. 124. Wintermeyer, E., Flugel M., Ott, M., Steinert M., Rdest, U., Mann K.H., and Hacker J. 1994. Sequence determination and mutational analysis of the lly locus of Legionella pneumophila. Infect. Immun. 62:11091117. 125. Valeru, S.P., Rompikuntal P. K., Ishikawa T., Vaitkevicius K., Sjoling A., Dolganov N., Zhu J., Schoolnik G., and Wai S.N. 2009. Role of melanin pigment in expression of Vibrio cholerae virulence factors. Infect. Immun. 77:935942. 126. Istivan T.S. and Coloe, P.J. 2006. Phospholipase A in Gram negative bacteria and its role in pathogenesis. Microbiology 152:1263 1274. 127. Schmiel D.H., Wagar E., Karamanou, L., Weeks D., and Miller V.L. 1998. Phospholipase A of Yersinia enterocolitica contributes to pathogenesis in a m ouse model. Infect. Immun. 66:39413951. 128. Shinoda, S., Matsuoka H., Tsuchie T., Miyoshi S., Yamamoto S., Taniguchi H., and Mizuguchi Y. 1991. Purification and characterization of a lecithindependent haemolysin from Escherichia coli transformed by a Vibrio parahaemolyticus gene. J. Gen. Microbiol. 137:27052711. 129. Brown R.N. and Gulig P.A. 2008. Regulation of fatty acid metabolism by FadR is essential for Vibrio vulnificus to cause infection of mice. J. Bacteriol. 190:76337644.
173 BIOGR APHICAL SKETCH Jennifer Joseph was born and raised in Jacksonville, FL. After graduating from Stanton College Preparatory School, she attended Florida State University in Tallahassee, FL. She graduated cum laude with a B.S. in b iological sciences and mov ed to Gainesville, F L to work at the United States Department of Agriculture for a short while before pursuing a graduate degree through the Interdisciplinary Program in Biomedical Sciences at the University of Florida. During her graduate career, Jennif er has been funded by the Alumni Fellowship and an NIH funded Training Grant in Biodefense and Emerging Infectious Disease