Disruption of yfjD Gene in Francisella tularensis By Japheth Omonira B.S. University of Notre Dame, 2016 Thesis Submitted in partial fulfillment of the requirements for the degree of Master of Science in the Degree of Master of Science in the Center of Biomedical Engineering at Brown University PROVIDENCE, RHODE ISLAND MAY 2018 i This thesis by Japheth Omonira is accepted in its present form by the Center for Biomedical Engineering as satisfying the thesis requirements for the degree of Master of Science Date:___________________ Signature:____________________________________ Dr. Gerard Nau, Advisor Date:___________________ Signature:____________________________________ Dr. Kareen Coulombe, Reader Date:___________________ Signature:____________________________________ Dr. Peter Belenky, Reader Approved by the Graduate Council Date:___________________ Signature:____________________________________ Dr. Andrew G. Campbell, Dean of the Graduate School ii Acknowledgements I sincerely thank Dr. Gerard Nau for giving me the opportunity to conduct my research in his laboratory at Rhode Island. I am deeply grateful for his support and mentorship throughout my time at Brown University. I would also like to thank Ana Hernandez for all of her guidance, hard work, and instruction over the years. Additionally, I would like to thank the members of my thesis committee, Dr. Kareen Coulombe and Dr. Peter Belenky, for serving as my readers. I would like to thank Dr. Jacquelyn Schell for her guidance throughout my time at Brown University. Finally, I am deeply grateful to my mother and family for their endless love and encouragement. iii Table of Contents Acknowledgements………………………………………………………………………………...………iii List of Figures...…………………………………………………………………………………………….v List of Tables ……………………………………………………………………………………………...vi Abstract……………………………………………………………………………………………………..1 Introduction…………………………………………………………………………………………………2 Materials and Methods…………………………………………………………………………………….10 Designing the disruption plasmid…………………………………………………………………………………….10 Isolating the pJH1 vector……………………………………………………………………………………………...10 Construction and confirming PCR products for use for Gibson Assembly……………………………………..11 Ligating the FTL_0951 fragment into pJH1, issues and troubleshooting………………………………………12 Ligation of FTL_0951 fragment cut from pGEM into pJH1 and electroporation into E. coli. ……....…….16 Transformation of LVS with disruption construct via triparental mating………………………………………17 PCR tests to confirm construct integration………………………………………………………………………….17 Colony quantification experiment testing integrated construct removal………………………………………..18 Immunoblot of protein release………………………………………………………………………………………..18 Isolating pFNLTP8 and pFNLTP8/pGRO vectors…………………………………………………………………19 Constructing a genetic complement plasmid……………………………………………………………………….19 Complement plasmid transformation into LVS…………………………………………………………………….20 Infection Assay………………………………………………………………………………………………………….20 Results……………………………………………………………………………………………………..22 Construction and Confirmation of the disruption plasmid………………………………………………………..22 Disruption of FTL_0951 in LVS………………………………………………………………………………………23 Disruption mutant d0951 is stable……………………………………………………………………………………25 Constructing a genetic complement…………………………………………………………………………………26 FTL_0951 disruption mutant initiates cytokine response in mouse macrophage cells ..……………………..27 Disruption mutant has reduced protein release compared to wildtype………………………………………….29 Troubleshooting the complementing constructs…………………………………………………………………….30 Discussion…………………………………………………………………………………………………32 References…………………………………………………………………………………………………37 iv List of Figures Figure 1: TNF-α production by human macrophages co-cultured with different bacterial strains……………5 Figure 2: LVS ΔFTL_0883 stimulates TNF-α …………………......…………………………………………………………………….6 Figure 3: Diagram of predicted functional regions of genes ybeX and yfjD………………………………………………8 Figure 4: Diagram showing the integration of the disruption construct into the yfjD gene………………………9 Figure 5: Construction of suicide vector to disrupt FTL_0951………………………………………….. 22 Figure 6: Confirming plasmid integration via PCR testing……………………………………………….24 Figure 7: Illustration of colony quantification experimental procedure…………………………………..25 Figure 8: Calculated number of CFUs/mL …………………………………………………………….....26 Figure 9: Release of TNF-α from infected macrophages. ……………………………………………….28 Figure 10: Infection assay ...……………………………………………………………………………... 29 Figure 11: Immunoblots of protein secreted by LVS and mutants………………………………………..30 v List of Tables Table 1: Disruption Plasmid Primer Sequences…...……………………………………………………... 10 Table 2: d0951 Complement Primer Sequences ………………………………………………………… 19 vi Abstract In the fight against rising antibacterial resistance, we are studying protein of unknown function in Francisella tularensis to find potential new drug targets and vaccines. Francisella tularensis is the causative agent of tularemia, a zoonotic disease that affects animals and humans. F. tularensis has the ability to infect and replicate in host cells, including macrophages. Infected macrophages produce less cytokines and chemokines, allowing F. tularensis to avoid the immune response. Previous work has shown that a mutant for the gene ybeX stimulates more cytokine production and attenuates intracellular growth and virulence. The ybeX gene is one of four genes in the live vaccine strain (LVS) of F. tularensis that encode for proteins containing cystathionine β-synthase (CBS) pair domains and a corC protein domain, another being FTL_0951 (yfjD). In this study, the function of YfjD was investigated using a genetic approach. A suicide vector containing ~500 base pairs (bp) of DNA homologous to the yfjD gene was introduced into LVS to create an insertional mutant. Disruption of the yfjD gene resulted in an LVS strain that elicited a stronger immune response of macrophages to infection than wildtype, paralleling the ΔybeX strain. The disruption mutant also reduced protein released from the bacteria, a phenotype opposite to that of the ΔybeX mutant. Genetic complementation, in which a functional copy of the yfjD gene was introduced to return the phenotypes to wild type, was not achieved. Sequencing and immunoblotting show the complementing construct to not be created and functioning as designed. The phenotypic differences, therefore, have not been definitively proven to be caused by the disruption. The yfjD gene is at the end of an operon, however, making it less likely the disruption caused effects downstream of the insertion. Due to this, it is asserted that the disruption mutant has distinct phenotypes related to the ΔybeX strain, but these findings need to be confirmed with successful genetic complementation. 1 Introduction The rise of antibacterial resistance is causing increasing public health issues [1]. Antibiotics are classes of drugs that target and kill bacteria. Their use in treating infections has drastically reduced the threat of diseases caused bacterial pathogens. However, continued use, overuse, and misuse have led to the rise of antibacterial resistant strains of bacteria. This increased resistance has increased the burden on overall health and economic concerns [1]. Illnesses that were once made trivial by antibiotics are threatening to become dangerous again. To fight this, new drug targets and classes of antibiotic drugs and vaccines are needed. New antibacterial targets are found by studying pathogenic bacteria. By studying the unique properties of bacteria, new targets can be found and the functions of bacteria better understood. One such property is evading the immune system. Some bacteria, such as Francisella tularensis, have the ability to infect a host without eliciting an immune response. By studying how F. tularensis does this, the pathways that interact and affect the host’s immune response can be better understood. Understanding this may uncover a potential target that removes the bacteria’s ability to avoid detection, and allows the body to recognize and eliminate the infecting bacteria. The goal of this study was to investigate F. tularensis’ ability to evade an immune response and identify a new target for antibiotic drugs and vaccines. Francisella tularensis was first isolated from ground squirrels in 1911. The squirrels had been found dying of a plague-like illness in Tulare County in California. Originally called Bacterium tularense, it was renamed after Dr. Edward Francis who pioneered its research [2]. It is the causative agent of the disease tularemia, which was discovered by Dr. Francis to manifest itself as different human illnesses such as rabbit fever, tick fever, lemming fever, and deer fly fever [3]. Humans can become infected through tick and deer fly bites, skin contact with an infected animal, ingestion of contaminated water, inhalation of a contaminated aerosol, and through laboratory exposure [4]. The symptoms of tularemia 2 depend on the strain, dose, and route of infection. The most common presentation is ulceroglandular tularemia, which is a result of infection through the skin. In this case, an ulcer develops at the site of infection, and within a few days the patient develops fever, chills, malaise, headaches and a sore throat [2]. Antibiotic therapy is the recommended treatment, with gentamicin, streptomycin, doxycycline, and ciproflaxin being the most common antibiotics used [4]. F. tularensis has been classified as a Tier 1 select agent, a subset of biological agents that the Department of Health and Human Services (HHS) and Agriculture (USDA) have determined to potentially pose a severe threat to public health and safety [5], and a Category A bioterrorism agent by the Centers for Disease Control and Prevention (CDC) [6]. Tier 1/ Category A agents have the highest risk of intentional abuse with major potential for mass casualties and grave effect to the public. F. tularensis falls under this category due to its low infectious dose, ability to infect via aerosol, and its history of being developed as a bioweapon [7]. Although the live vaccine strain (LVS) is unable to induce complete protection and its mechanism of attenuation is incompletely understood, it is used to study tularemia due to its virulence in murine models and its minimal virulence in humans [8]. LVS is the strain used in this study. Francisella is an intracellular pathogen with a complex lifecycle. After infection, it resides transiently in phagosomal compartments before entering the cytosol. It is able to infect and multiply within macrophages in vivo, along with a number of other cell types in vitro and in vivo. The uptake of F. tularensis is done via pseudopod loops involving several macrophage receptors including the complement receptor CR3, the mannose receptor, and the scavenger receptor A [9]. The bacterium stops the maturation of the phagosome at a late endosomal-like stage, and in 15-30 minutes, the phagosome is transiently acidified. This acidification allows the bacterium to escape into the cytosol of the macrophage and replicate. After another 15-30 minutes, the phagosomal membrane is degraded and the bacteria escape into the cytoplasm, where the bacteria multiply to higher levels [2]. Cell-to-cell dissemination is thought to occur after the bacteria are released into the cytosol. Another mechanism suggests a different 3 method of dissemination that allows bacteria to be transferred directly from infected cells to uninfected cells [9]. This is done by exploiting natural host cell processes that transfer cytosolic material through contact-dependent mechanisms. Macrophages are known to use one of these contact-dependent mechanisms, trogocytosis, to briefly fuse membranes with a nearby cell [9]. During the fusion, some of the proteins in the membranes are transferred. When the cells split apart, the proteins transferred are kept by the other cell. Using live cell imaging, Steele et al. show F. tularensis in infected macrophage-like J774A.1 (J774) cells transfer to uninfected cells after cell to cell contact [10]. Both donor and recipient cells remained viable after transfer. In addition to this, by monitoring protein transfer, as a way to measure trogocytosis rates, and bacterial transfer at the same time, Steele et al. found infection increased trogocytosis rates [10]. After interacting with bacteria or components from them like lipopolysaccharide (LPS), macrophages normally produce cytokines and chemokines that trigger inflammation [11]. This is typically a pro-inflammatory response by the innate immune system. The innate immune system refers to the immunological systems comprised of general responses to pathogens. When stimulated, a pro- inflammatory response triggers inflammation that increases blood flow and draws immune cells to a site of infection [11]. However, when infected with Francisella tularensis, macrophages manifest an altered response [8]. Figure 1 also shows this altered response by measuring the production of tumor necrosis factor α (TNF-α), a cytokine involved in the immune response and systemic inflammation. When not infected, human macrophages produce low levels of TNF- α. When co-cultured with E. coli, however, a large amount of TNF- α is produced, signifying that the macrophage initiated an inflammatory response. In contrast, Francisella philomiragia and Francisella novicida, two bacteria in the same genus as F. tularensis, elicits lower levels of TNF- α. LVS, which is derived from a virulent subspecies of F. tularensis, stimulates very little TNF- α production. From this, the observation is made that LVS stimulates a smaller pro-inflammatory response compared to other bacteria. 4 Figure 5: TNF-α production by human macrophages co-cultured with different bacterial strains. Paul Carson, Unpublished One reason F. tularensis avoids eliciting a pro-inflammatory response is its unique LPS. LPS is a carbohydrate found on the outer surface of gram-negative bacteria. It elicits a release of the cytokines like interleukin 1 (IL1) and TNF-α. When compared to the LPS of Escherichia coli, the LPS of F. tularensis elicits a 1000-fold weaker response [8]. This is due to unusual modifications F. tularensis makes to the structure of LPS, specifically to the O antigen, that causes immunological differences. The LPS of F. tularensis does not bind to host molecules such as LPS-binding protein, TLR4, or TLR2, which are crucial in triggering the pro-inflammatory response [2]. E. coli LPS is also responsible for inhibiting NO production [12]. In some animals, NO production is used as a form of innate immunity used by macrophages to limit infections by intracellular bacterial pathogens [12]. NO production makes macrophages cytostatic or cytotoxic, preventing viruses, bacteria, fungi, protozoa, helminths, and tumor cells from growing [13]. Therefore, F. tularensis effectively hides from the innate immune system. F. tularensis’ ability to evade the host immune response may also be due to an ability of the bacteria to adapt to the intracellular environment. Studies have shown that F. tularensis that had replicated in macrophages elicit of an immune response in following infections when compared to F. 5 tularensis that had been grown in bacterial growth media. Spermine was found to be an important polyamine whose presence alters F. tularensis and affects the infected macrophage’s ability to produce pro-inflammatory cytokines [14]. To determine the importance of spermine to F. tularensis’ virulence, Russo et al. created a deletion mutant unable to respond to extracellular spermine [14]. They found the gene FTL_0883 (ybeX) to be necessary for F. tularensis to respond to spermine [14]. After mutating ybeX, strains elicited higher levels of cytokines from macrophages than wildtype. For example, TNF-α production by mouse and human macrophages infected with LVS and their ΔybeX mutant showed marked differences (Figure 2 from [14]). This indicated that the YbeX plays a role in avoiding detection by the immune response, and that inhibiting its function could lead to attenuation and better host immunological responses to bacterial infection. Figure 6: LVS ΔFTL_0883 stimulates TNF-α. Bacteria cultivated overnight in TSB-C were used to infect either human macrophages or murine BMDM at an MOI of 10. After 24 h, the supernatants were removed, and an ELISA was performed to quantify the levels of TNF-α. The data are presented as the means ± the SD of triplicate wells within one experiment representative of at least three independent experiments. Figure is from reference 14. The function of YbeX, also known as corC, is still largely unknown. YbeX contains a corC domain thought to be involved in magnesium and cobalt efflux of the CorA transporter system [15]. LVS, however, does not have a functional corA gene [14]. The experiments that established corC as a transporter have some issues. The experiments tested corC by introducing mutations to an operon. The 6 phenotypes seen in the mutant could be due to effects on other genes in the operon and not corC itself. Also, the experiments were never complemented. Complementation reintroduces a functional copy of the affected gene to restore expression of that gene. Restoring gene expression should return phenotypic changes back to wild type. Without genetic complementation, it is uncertain if the phenotypes seen were due to the mutation in that gene or due to other, unexpected effects. Previous studies in the Nau lab into the function of YbeX has shown that it is involved in a number of cell functions. YbeX is necessary for F. tularensis to respond to spermine. In the presence of spermine, LVS elicits lower TNF-α production; this phenotype is not seen in the ΔybeX mutant [14]. YbeX is important for intracellular growth. The ΔybeX mutant grew at least 10-fold less than wild type in human and mouse macrophages. This growth difference is not attributable to varying invasion rates, as the number of bacteria in a host cell differed by less than 2-fold between LVS and the mutant [14]. The ΔybeX mutant also had attenuated virulence. Mice infected with wild-type LVS lost a significant amount of body weight while mice infected with the mutant lost less than 5% of their body weight over the same time period [14]. Francisella tularensis has a relatively small genome size, 1.8 Mb compared to E coli’s 4.6 Mb [16]. Perhaps it is for this reason that LVS has only four genes that encode CBS domains, one being FTL_0883 and another FTL_0951 (yfjD). The work in this thesis focuses on yfjD. The gene yfjD is very similar to ybex; both contain a CorC region and CBS pair domains. Figure 3 diagrams the functional regions encoded by each gene. 7 Figure 7: Diagram of predicted functional regions of genes ybeX and yfjD. ATR – Amino terminal region. CBS – cystathionine beta-synthase. Duf-21 – domain of unknown function 21, which is predicted to encode a transmembrane domain. Due to these similarities, it was hypothesized that the proteins YfjD and YbeX may function in the same biological process. To test this, a genetic approach to create an LVS disruption mutant, d0951, generating a mutant with no functioning YfjD protein. A complement, in which the gene functionality in the mutant is restored, would also be created. This would verify that any phenotypic changes seen were due to the genotypic change in the yfjD gene alone and not due to any off site effects. The mutant would then be tested and compared to the wild-type and complement to try to elucidate the role yfjD has in Francisella tularensis. We utilized homologous recombination to generate a disruption mutant. A plasmid vector was constructed containing the 500 bp region with homology to the native LVS yfjD gene. This region of homology targets the plasmid to a specific site in the genome. The plasmid’s insertion is intended to interfere with transcription of the gene, thereby disrupting yfjD expression. The plasmid chosen for this was pJH1. pJH1 encodes for hygromycin resistance and contains an E. coli origin of replication, pBBR1 oriV. This allows it to replicate in E. coli but not LVS. Because of this, only LVS bacteria that have successfully integrated the disruption construct into the genome will grow in the presence of hygromycin. 8 It also allows for the construction to be done in E. coli. Figure 4 shows an overview diagraming the process. Figure 8: Diagram showing the integration of the disruption construct into the yfjD gene. The dark blue indicates the genes before and after yfjD. The green section is the yfjD gene. The striped green in the region of homology. The gray rectangle is the ori T site of pJH1; the yellow is the oriV site of pJH1. 9 Materials and Methods Designing the disruption plasmid A suicide vector to introduce a disruption into the FTL_0951 (yfjD) gene was designed first. The primers for a yfjD disruption insert were determined using the Snapgene and ApE programs. The primer sequences are shown in table 1. The coding sequence for the yfjD gene was taken from NCBI. The section of yfjD to be inserted into the plasmid pJH1 spans from bp 68 to bp 724, making its length 657 bp. This DNA segment was designed to be inserted at the BamH1 site in the suicide vector pJH1. This results in a yfjD disruption construct with a 657 bp region homologous to the gene of interest inside pJH1. Table 1: Disruption Plasmid Primer Sequences yfjD Forward Primer CCTGCAGGTCGACTCTAGAGCTGAGACAGCGATGAT yfjD Reverse Primer CTCGGTACCCGGGGATCCGCCATTCTCACAGAG pJH1 Forward Primer GATCCCCGGGTACCGAG pJH1 Reverse Primer CTCTAGAGTCGACCTGCAG Isolating the pJH1 vector The pJH1 vector was obtained from a previous Nau Lab stock, EC-XL /pJH1, using a miniprep procedure. The procedure recommends growing a 2 mL overnight culture in LB and selective antibiotic. 1.5 mL of the culture is then poured into an Eppendorf tube and spun for 30 seconds. The supernatant is aspirated, and the pellet resuspended via vortexing in 100 µL of solution I (50 mM glucose, 10 mM EDTA, and 25 mM Tris pH 8.0). 200 µL of freshly made solution II (0.2 M NaOH, 1% SDS in H2O) is then added and mixed gently, then held on ice for 5 minutes. Then, 150 µL of ice cold solution III (3M potassium acetate, 2M acetic acid) is added and the mixture again mixed gently and held on ice for 5 minutes. This is then spun for 5 minutes and the supernatant transferred to a tube containing 1 mL 100% ethanol. This is mixed and spun again for 5 minutes. The supernatant is then removed, and the pellet 10 washed with 1 mL of 70% ethanol. The supernatant is again removed, and the pellet allowed to air dry before being suspended in 50 µL microbiology water and 1 µL of 10 mg/µL boiled RNase A. DNA was quantified via spectrophotometry. Constructing and confirming PCR products for use for Gibson Assembly Initially, the disruption construct was to be made using Gibson Assembly [17]. For this, the four primers in table 1 were designed to amplify the vector and the yfjD fragment via PCR reactions to be inserted into the vector. Two inverse reactions were done to amplify pJH1. One reaction used the Accuprime protocol. This was a mixture of 2.5 μL of 10X AccuPrime PCR Buffer I, 0.5 μL of Primer Mix each, 1 μL of the template DNA (1:50 pJH1) , 0.5 μL of AccuPrime Taq DNA polymerase (added last) and nuclease free water to bring the solution to 25 μL (20 μL in this case). The other pJH1 inverse PCR reaction used the Phusion protocol detail by New England Biolabs. This solution had 4 μL of 5X Phusion HF, 0.4 μL of 10mM dNTPs, 1 μL of primer each, 1 μL of the template DNA (same as before) and the nuclease free water to bring the solution to 20 μL (12.4 μL in this case). One yfjD PCR reaction was done using the AccuPrime protocol. It was done in the same manner as the pJH1 AccuPrime protocol, substituting 10X AccuPrime PCR Buffer II for buffer I and with the designed primers and template DNA (LVS genomic DNA). The two pJH1 PCR reactions were evaluated together by agarose gel electrophoresis. Denaturing was done at 98°C for 30 seconds, annealing at 54°C for 30 seconds and extension at 72°C for 3.5 minutes for a total of 35 cycles. The yfjD PCR reaction was run in a separate thermocycler. Denaturing was done at 98°C for 30 seconds, annealing at 54°C for 30 seconds and extension at 72°C for 30 seconds for a total of 30 cycles. Gel electrophoresis and imaging was used to assess the success of the reactions. We used 1% agarose in Tris-borate-EDTA buffer, with 1X Sybr Safe from Invitrogen for visualization. After 30 11 minutes to an hour of electrophoresis, the gels were imaged using a Chemidoc digital imager. No bands were seen for the three PCR mixtures. I performed troubleshooting using different PCR template concentrations, as well as a temperature gradient PCR reaction to test a number of different annealing temperatures at once. Eight PCR reactions were run. The annealing temperatures used were 50, 50.7, 51.9, 53.8, 56.1, 58, 59.2, and 60ºC. This did not yield amplicons of the appropriate size. Ligating the FTL_0951 fragment into pJH1, issues and troubleshooting Because amplicons were not generated for Gibson assembly, we undertook a conventional cloning strategy using restriction digests and ligation. More pJH1 plasmid was obtained using the same protocol outlined before. The 567 bp PCR product from yfjD was purified using the QIAquick PCR purification protocol. 125 μL of buffer PB was added to the PCR mix. This sample was applied to the column (in a collection tube) and was spun for 1 minute. The flow through was discarded after. To wash, 750 μL of buffer PE was applied to the column and the column spun for 1 minute. The flow through was discarded again. The column was spun again for 1 minute. The column was moved to a different tube for collection. 50 μL of buffer EB was added to the column and the column spun for 1 minute. The flow through was saved as the purified product. A PstI, Bam HI double digest was done on the purified yfjD amplicon and the pJH1 plasmid separately. The digest was done by mixing 1.5 μL of 10X buffer, 1.5 μL bovine serum albumin (BSA), 2 μL water, 8 μL DNA (pJH1 or purified yfjD amplicon), 1 μL PstI, and 1μL Bam HI double-digest. The mixtures were incubated at 37ºC for 2 hours. 2 µL of loading buffer was added to each digest and was loaded into a gel, along with a marker, and gel electrophoresis was done. Imaging showed bands of the proper size, which were then cut out and the DNA recovered using a Wizard Gel Cleanup System (Promega). The gel slices were weighed and 10 μL of membrane binding solution per 10 mg of gel was added to each sample. They were then both vortexed and put into a 65ºC water bath, vortexing every few 12 minutes. The samples were then transferred to appropriately labeled columns in labeled collecting tubes. They were incubated at room temperature for 1 minute then spun at 16000g for 1 minute. 700 μL of membrane wash solution was then added to each solution and the columns in the collection tubes spun again for 1 minute. The flow-through was treated as a biohazard waste and removed accordingly. Another wash was done using 500 μL of wash solution, spinning for 1 minute. This flow-through was not treated as hazardous and was discarded accordingly. The column was then centrifuged again for 1 minute. 25 μL of nuclease free water was added to the columns, now in new, labeled centrifuge tubes to collect the flow- through. It was allowed to sit for 1 minute and then spun for 1 minute in the new tubes, to collect the DNA. The DNA fragments were then enzymatically ligated together. To set up the ligation, the concentration of each sample was measured using a spectrophotometer (Nanovue Plus, GE Healthcare). The pJH1 sample had concentrations of 3.4 and 3.5 ng/μL. The yfjD amplicon had concentrations of 2.4, 3.5, and 2.2 ng/μL. The average of the values was taken as the working concentrations. For the ligation, 12 μL of pJH1 (42 ng), 3.4 μL of amplicon (9.18 ng), 2.6 μL of nuclease free water, and 2 μL of T4 DNA ligase buffer were mixed together. 1 μL of T4 ligase (NEB) was then added and incubated in a thermocycler using a program that alternates between 10ºC and 30ºC, 30 seconds at each temperature, for 99 cycles. Visualization of the ligation reaction was performed on an agarose gel. The image showed a faint band for the pJH1 sample but none for the yfjD amplicon or the ligation mix. Troubleshooting this process allowed for the production of a yfjD amplicon that was purified. This was then used in two pJH1-0951 ligations. The first contained 2 μL of 0951 (25.6 ng), 0.5 μL pJH1 (1.85 ng) with 2 μL ligase buffer and water bringing the volume up to 20 μL. The second contained 0.32 μL (4 ng) 0951, 5 μL (18.5 ng) pJH1 along with 2 μL buffer and water up to 20 μL. 1 μL of ligase was added to the solutions and they were run through the thermocycler in a program that alternated between 30 seconds at 10ºC and 30 seconds at 30ºC for 99 cycles. 13 Transformations of Escherichia coli cells were done using 2 μL of each ligation, adding the ligation to 10 μL of chemically competent DH5α cells. This was put on ice quickly after and incubated for 20 minutes. They were then moved to a 42ºC water bath for 50 seconds then put back on ice for 2 minutes. They were then transferred into 1 mL of LB each and put in the shaker at 37.7ºC for one hour. 500 μL of each were transferred to new centrifugation tubes. They were then spun for 1 minute at 16000 g. 350 μL were removed and the pellet was resuspended in the 150 μL left. This was then plated on hygromycin coated plates using plating beads. Troubleshooting efforts included the use of more culture for the miniprep and a new yfjD forward primer. This primer was used in a gradient PCR ranging from an annealing temperature of 50°C to 60°C. Four positive controls were created in the same proportions using the old forward yfjD primer. 2 µL of these PCRs were run through an agar gel electrophoresis and the gel imaged. Several of the PCRs gave strong bands in line with the positive control. The brightest band was cut out, purified, and Gibson Assembly was used to insert the yfjD fragment to previously extracted pJH1 vector. Transformations were done using this mix, but did not yield colonies. A new attempt was adopted using the pMB1 vector created by an Xho I digest of pJH1, to eliminate the unneeded parts of the 7 kB pJH1 plasmid. The digest was done using 1.5 μL of buffer 4, 1.5 μL of BSA, 1 μL of water, 10 μL of plasmid, and 1 μL of enzyme. It was incubated at 37ºC for 2 hours. The plasmid was run on an agar gel and imaged to confirm the plasmid had been digested as predicted. It was then cut from the gel and purified into 25 µL of nuclease-free water. The plasmid was then re-ligated by mixing 12 μL water, 5 μL of plasmid, 2 μL of buffer, and 1 μL of T4 ligase, and placing the mixture in the thermal cycler to cycle between 30ºC and 10ºC for 30 seconds each for 2 hours. The pMB1 plasmid was then digested with Bam HI and PSTI, and this digested plasmid used in transformations. Three transformations of the pMB1 plasmid were done, one using TOP10 cells (Invitrogen), one using chemically-competent DH5αL cells, and another using electro-competent EC100 cells. Heat shock was done for the Top 10 using 10 μL of cells and DH5αL using 50 μL cells (ice for 20 minutes, 42ºC water 14 bath for 50 seconds, then on ice for 2 minutes). For the EC100, cell transformations were done by electroporation using 50 μL cells and 2 μL of ligation reaction. All three cell types gave colonies. Liquid cultures were made from colonies: 2 from the Top 10, two from the DH5α, and one from the EC100 cells. DNA was extracted from bacteria grown from isolated colonies using the miniprep protocol previously explained (see Isolating pJH1 vector), again using water with RNase to resuspend the DNA pellet. A digest with XhoI was then done on all 5 and a control pJH1. This used 10 μL of plasmid, 1.5 μL of BSA, 1.5 μL of buffer 4, 1 μL of water, and 1 μL of XhoI enzyme. They were incubated at 37ºC for 2 hours before being run on a gel. This time, they all showed bands of the correct size. Because the region of pJH1 that anneals to the pJH1 primers should still be in the pMB1 plasmid, the cut pMB1 plasmid was used in a PCR using the same primers used in the pJH1 PCR. A subsequent gel of the PCR products showed a band around 1500-2000 bp, not the expected range. This was similar to the results from a previous pJH1 PCR, suggesting the same problems experience there. It was decided to alter the cloning strategy and instead insert the yfjD fragment into pGEM using TA-cloning, then subclone into pJH1. The yfjD fragment was first cloned into the pGEM T-vector using a kit. They were transformed into 10 µL of Top10 cells via chemical shock, and plated onto LB agar plates with100 µg/ml ampicillin and coated with ChromoMax (Fisher) for blue-white screening. Plates were cultured overnight in a 37°C incubator. White colonies were then grown overnight in 5 mL liquid cultures, and the plasmid extracted. A double digest of the pGEM-0951 plasmid was done using 1 µL PstI and Bam HI each with 10 μL of plasmid used. The whole 15 µL digest was then run on a gel, were two clear bands were seen, one the appropriate size of the pGEM vector and another the appropriate size for the 0951 fragment. The fragment band was cut out and purified from the gel. 15 Ligation of FTL_0951 fragment cut from pGEM into pJH1 and electroporation into E. coli The yfjD amplicon and pJH1 plasmid concentrations were determined using a nanodrop machine. For the amplicon, the readings were 16.6 and 15.4, giving an average of 16 ng/μL. For pJH1, the readings were 83.5 and 79.0, giving an average of 81.3 ng/μL. A gel was run of these three to back up these readings. The relative band intensities match the readings fairly well. Ligations were then set up at different ratios, a 7:1 ratio using 100 ng (1.23 μL) pJH1 and 50 ng (3.125 μL) 0951, another was 3:1 using 100 ng pJH1 and 20 ng (1.25 μL) 0951, and another was 1:1 using 100 ng pJH1 and 7.1 ng (0.44 μL) 0951. The 3:1 and 1:1 ratio ligations both gave colonies, and grown in overnight cultures in 5 mL LB with 5 μL of hygromycin. In preparation of a triparental mating, one helper culture was made in LB with kanamycin and 1 LVS culture was made in TSB-C. The subsequent triparental mating, however, did not yield any colonies. The cloning process was restarted to address the lack of colonies yielded from the mating. This approach used a larger plasmid preparation starting with 25 ml of overnight culture. 10 µL of the midiprep was digested, run on a gel, and the bands cut out. The bands were purified, along with a new batch of pJH1plasmid obtained via a 5 mL miniprep. The relative band intensity when viewed in a gel suggested significantly higher concentrations. More ligations were set-up, this time at a 1:1 and a 2:1 insert to pJH1 volume ratio. These were then used in electroporations, each using 2 μL ligation mix with 50 μL EC100 cells with 1 mL of LB was used as recovery media. The resulting colonies were picked for 5 mL liquid cultures and then miniprepped using the GeneJet column miniprep (Thermofisher) using 2 mL of culture. After the plasmid was isolated, a Bam HI/ PstI double digest was done. An agarose gel of the digests was done to separate the digests, and the image showed 3 of the 5 digests showed bands. The three clones showed inserts of varying sizes, all larger than expected. Sequencing found E. coli DNA had been introduced to the end of the ~500 base pair yfjD insert. Since the goal is to introduce “random” DNA 16 in the middle of the 0951 gene to disrupt it, the extra DNA was not expected to alter this outcome, and the sequenced plasmid was used moving forward. As long as the region of homology is still in the fragment, the “random” DNA will be introduced and the gene disrupted. Transformation of LVS with disruption construct via triparental mating Triparental mating was done using the pJH1-0951f created. The OD of the E. coli cultures, the pJH1-0951f and the helper culture, were measured and adjusted to 0.5 in 1 mL. The 1 mL adjusted E. coli and 1.5 mL of LVS were pelleted by centrifugation for 3 minutes and the supernatant aspirated. The E. coli cells were then washed with 1 mL of fresh LB and resuspended in 25 μL LB each. The two E. coli cell suspensions were then used to resuspend the LVS pellet. The resuspension was then pipetted onto a LB plate and incubated at 37°C for 18-24 hrs. The triparental mating mix was plated chocolate agar plates containing hygromycin and polymixin. This was done by scooping the entire mix into 800 μL of TSB-C media and dispersing using a vortexer. 200 μL was then transferred onto each of the four plates and spread using plating beads. The plates were incubated at 37°C with 5% CO2 for 3 days. The resulting colonies were the patched onto hygromycin and polymixin chocolate agar plates and allowed to grow again for 3 days in an incubator at 37ºC with 5% CO2. Liquid cultures derived from these patches were grown in 5 mL of TSB-C with hygromycin. PCR tests to confirm construct integration The liquid cultures derived from the triparental mating were miniprepped to extract the genome of the bacteria to test integration. This was done via a set of PCRs. These PCRs used a combination of primers that would recognize the plasmid pJH1 DNA and the 0951 gene DNA. By doing a PCR that amplifies a section of the plasmid, the incorporation of the plasmid in the LVS genome was determined. 17 Colony quantification experiment testing integrated construct removal Another experiment was set up to test the frequency that LVS removed the integrated plasmid from its genome and returned to wildtype. A single colony is used to grow two broth cultures, one in selective media (hygromycin) and one in non-selective media. After overnight culture, serial dilutions were made and then plated onto two plates, one selective and one non-selective. The bacteria were then counted on the plates, using the dilution to estimate the number of colony forming units (CFUs). Colony forming units per milliliter (CFUs/mL) were determined by taking the product of the number of individual colonies (N), a correction factor of 100, and ten to the power of the dilution factor (df): 𝐶𝐹𝑈 = 100 ∗ 𝑁 ∗ 10^(𝑑𝑓) Immunoblot of protein release An immunoblot tested the relative amount of protein released by various bacterial strains into an overnight culture. The bacterial strains were grown in 5 mL TSB-C overnight and spun and the supernatant taken. 30 μL of the supernatant was mixed with 10 μL of 4X buffer with lithium dodecyl sulfate (LDS). 20 μL of this was loaded into the polyacrylamide gel and run at 150 V for an hour. The separated proteins were then transferred to a polyvinylidene difluorine (PVDF) membrane using an electrical current of 0.15 amps for one hour. The membrane was blocked for an hour with 5% milk in Phosphate-Buffered Saline Tween-20 (PBST) for immunoblots testing total protein release. Immunoblots using anti-streptavidin antibodies were blocked for an hour with 3% BSA. The blocking solution was then removed and the membranes were incubated with either polyclonal anti-Francisella antibodies to test total protein release, or anti-streptavidin antibody to test for the StrepTagII tagged protein. The membranes were incubated overnight at 4ºC. They were washed three times with PBST for 15 minutes each wash. After the washes, the membranes were incubated for 1.5 hours with the probing antibody. Immunoblots testing for total protein release used anti-rabbit antibodies in 5% milk in PBST. Immunoblots testing for the StrepTagII used anti-mouse antibody in PBST. The membranes were washed 3more times with PBST 18 for 15 minutes each wash. After the last wash, the fluorescence on the membrane was imaged using a BioRad Chemidoc imager. Isolating pFNLTP8 and pFNLTP8/pGRO vectors The pFNLTP8 and pFNLTP8/pGRO starting vectors were grown from Nau lab stocks, DH5α cells with pFNLTP8 and DH5α cells with pFNLTP8/pGRO-0883 respectively, and miniprepped using the GeneJet protocol, eluted into 50 μL of elution buffer each. They were then digested. Two digests were done: pFNLTP8 with Bam HI and NotI, and pFNLTP8-pGRO with Bam HI. The digests were run in a gel, cut out of the gel, and gel purified, eluted into 25 µL of water each. Constructing a genetic complement plasmid The procedure used to construct a genetic complement was very similar procedure to the construction of the disruption plasmid. Four primers were designed and used to create four different amplicons capped with restriction sites. Their sequences are seen in table 2. One pair amplified the coding sequence of FTL_0951, the other pair included the 96 bps upstream the gene to capture the DNA sequence that could contain a native promoter. Each pair had a version designed to add a sequence for the tag Strep Tag II (IBA LifeSciences) to the end of the amplicon. Table 2: d0951 complement primer sequences. d0951 complement F GCGGCCGCACTATTTCATCTTTATATTTATTTCCATAATATATTTTTC pGRO.d0951 complement F CGGATCCATGAGTACTTATACTGTAGTTATTATCATATTTATAC d0951 complement R GGATCCTTACTAAGTTTCTGTAGTAATAGTTAACTTTATAG d0951 StrepTagII GGATCCTTATTTTTCAAACTGCGGATGGCTCCACTAAGTTTCTGTAGTAATAGTTAAC complement R 19 PCR reactions were used to create the amplicons, which were then TA-cloned into pGEM. Restriction digesting was used to retrieve the amplicons from pGEM, and the product ligated into pFNLTP8 or pFNLTP8-pGRO. Electroporations were done to move the ligation products into E. coli. However, only the complements with the native promoter successfully incorporated the yfjD gene in the vector; the pGRO complements have yet to be completed. The successful ligation for the native untagged complement used 2 μL of buffer, 2 μL of vector, 5 μL of insert, 1 μL of T4 ligase, and 10 μL of water. The pFNLTP8 plasmid digested with NotI and Bam HI was used as a vector. The product of the ligation reaction was then introduced into EC100s via electroporation. 2 μL of ligation was mixed with ~100 μL of EC100 cells. The time constants were 4.8 ms for the negative control and 4.7 ms for the native complement. The recovery was done in SOC and then plated onto LB + kanamycin plates. The cells were then grown, the plasmids extracted via the GeneJet miniprep, and a confirmation digest and gel done. Complement plasmid transformation into LVS Once both the complementing constructs incorporating the upstream sequence were constructed, they were used to transform LVS and the disruption mutant, d0951. LVS and d0951 were electroporated. Before pulsing, the bacteria were washed 3 times with sucrose. For both LVS and d0951, there were 4 electroporations: one control with no DNA, one with an irrelevant plasmid TC3D [18], one with the untagged complementing construct incorporating the upstream sequence, and one with the tagged construct incorporating the upstream sequence. These were all plated onto chocolate agar plates containing kanamycin; the LVS no DNA control was also plated on a chocolate agar plate with no antibiotics as a no selection control. Infection assay Bacterial strains were used in in vitro infection experiments. These experiments initially used RAW 264.7 mouse macrophage-like cells that had been altered by their manufacturer to express GFP 20 when the NF-κB pathway is activated [19]. These cells were recovered from flasks by scraping, and then centrifuged, resuspended, and counted. A dilution was made to give 2.1x106 cells, enough for 5x104 cells per well for 21 wells in a 96 well plate. The dilution was put on ice while the bacteria were prepared. An OD reading was done of the bacteria to estimate the bacterial density from overnight cultures where OD600 = 1 is approximately 109 bacteria/ml. Multiplicity of infection (MOI) calculations were done to determine the amount needed to achieve an MOI of 10 based on the number of RAW cells plated in the wells.. The bacteria were prepared according to these calculations. 100 μL of the bacterial suspension were added to the wells, followed by the macrophages. In one experiment, macrophages were not resuspended at this point. This caused uneven distribution of cells in suspension, leading to unequal amounts of cells added to each well. The experiment was repeated to correct this. 100 µL of media was added after the macrophages cells. A media negative control and LPS positive control were also included. The plate was incubated overnight, and the cells imaged the next day. The complement plasmids were sent to sequence to confirm the plasmid in the mutant is the same as designed. 21 Results Construction and Confirmation of the disruption plasmid Construction of the pJH1.d0951 disruption construct was first attempted to be constructed using the Gibson Assembly method [17]. For this, four primers were designed to amplify the pJH1 vector and the FTL_0951 fragment to be inserted into the vector. The AccuPrime and phusion protocols were used to amplify pJH1; only AccuPrime was used for the insert. No bands were seen for the PCR reactions, despite repeated attempts. A set of new PCR reactions with different annealing temperatures to amplify the FTL_0951 insert were successful. New PCR reactions were run for the pJH1 vector as well; the AccuPrime protocol gave multiple bands of varying size while the phusion protocol yielded no DNA. To troubleshoot, a gradient PCR reaction was run using the AccuPrime protocol. When all were run on a gel, none of the PCRs yielded bands of the correct size; some lanes had one band of incorrect size while others had multiple bands. This could be due to the primers possibly annealing to improper spots. In an attempt to remedy this, a new pMB1 vector was created via an Xho I digest of pJH1. PCR reactions using this vector gave similar results, suggesting the same problem. Figure 9: Construction of suicide vector to disrupt FTL_0951. Left: Gel images of (A) 0951 fragment PCR product, (B) 0951 fragment digested from the pGEM vector, (C) pJH1 vector digested with Bam HI and PstI, and (D) 0951 fragment digested from pJH1. All gels contain a DNA ladder in the 22 leftmost lane. Right: Diagrams of (A) the primers designed to amplify the 0951 fragment, (B) 0951 fragment in pGEM, (C) the pJH1 vector, and (D) the pJH1.0951 vector. Since we were unsuccessful generating a PCR product using pJH1 as a template, we used a more traditional approach of cloning restriction digest fragments. Our initial attempts to clone PCR amplicons digested with restriction enzymes proved unsuccessful. Therefore, the PCR product amplified from FTL_0951 was first ligated into pGEM by TA cloning. This method proved to be successful. As shown in figure 5A, the 567 bp region was amplified. The amplicon was then TA cloned into pGEM and then recovered by restriction digest, shown in 5B. The restriction fragment was purified from the gel and ligated with digested pJH1 plasmid. The resulting ligation was moved into E. coli via electroporation. The plasmid was then extracted from the subsequent colonies and confirmed to be the pJH1.0951 construct via a Bam HI PstI double digest, seen in 5D. Disruption of FTL_0951 in LVS After the pJH1.d0951 plasmid had been constructed, it was transformed into LVS via triparental mating. This process involves co-incubation of LVS with E. coli containing the pJH1.0951 construct and E. coli with a “helper” plasmid that codes for conjugated pili that allow transfer of DNA. The helper conjugative plasmid transfers itself into the E. coli with the disruption plasmid, and then transfers both plasmids into LVS. Selective antibiotics are used to allow for the growth of transformed LVS only. A series of PCR tests were done to confirm the plasmid integrated into the genome as designed (Fig. 6). Three PCR reactions were done on both LVS and the disruption mutant: one amplifying a region from the Ori T site in the pJH1 plasmid to end of the yfjD fragment, one amplifying across the entire yfjD gene, and one amplifying the internal 500 bp yfjD fragment region (Fig 6A). The first PCR reaction, the junction reaction spanning from the Ori T site to the edge of the fragment (Fig. 6A, purple arrows), was expected to yield, for the disruption mutant, a band around ~2000 when accounting for the plasmid DNA that had inserted itself into the plasmid during cloning. The LVS control PCR was expected to yield nothing because the Ori T site used does not exist in its normal genome (Fig. 6B). The second PCR, which used primers flanking the yfjD gene on either end (Fig. 6A, red arrows), would yield a band the 23 size of the gene for the wildtype but not for the disruption, because the region would be too large to amplify due to the addition of the disruption construct (Fig. 6B). The final PCR, amplifying the fragment (Fig. 6A, blue arrows), was expected to show ~500 bp for both strains (Fig. 6B). Figure 10: Confirming plasmid integration via PCR testing. A. The expected results from PCR testing. + indicates presence of a DNA band, - indicates its absence. B. Gel images of testing PCR reactions. C. Diagram of the expected amplified regions in pJH1.0951. Arrows indicate primers. The same forward primer was used in the flanking and junction reactions. The results of the genomic PCR testing indicated the resulting strain had integrated the disruption construct into the correct locus of the chromosome (Fig 6C). Because the Ori T primer cannot anneal to anything near yfjD in the wildtype, the band for the mutant in the Ori T spanning PCR shows that the pJH1 plasmid is in the mutant genome (Fig. 6C, junction). The flanking PCR shows that the plasmid was integrated in the mutant genome in the middle of the yfjD gene, since the PCR of the wildtype yielded a band yet the mutant did not (Fig. 6C, flanking). This is most likely due to the integrated DNA causing the region of interest to be too large to be amplified under the conditions given. The internal PCR served as a positive control that the genomic DNA could function as a PCR template (Fig. 6C, internal). These 24 results show that the disruption construct integrated into the expected locus, generating a strain with an insertion mutation in FTL_0951. This strain will be referred to as d0951. Disruption mutant d0951 is stable An experiment was set-up to investigate the rate, if any, at which the disruption mutant removed the inserted DNA and reverted back to wildtype. Chromosomal integration of a plasmid has the potential to recombine, looping out the plasmid and reconstituting the wild-type locus. Figure 7 shows a schematic for the experimental procedure. In this experiment, a colony is used to grow two broth cultures, one in selective media, in this case hygromycin is present, and one in non-selective media. The selective media should have a lower frequency of wild-type bacteria because wildtype cannot grow well in the selective media, while the non-selective media should have a higher frequency of wildtype. After growing the cultures, serial dilutions were made and then plated onto two plates, one selective and one not. The plates growing d0951 from the selective media culture should not vary too much in number; due to the selective media allowing only the mutant to grow. In contrast, the plates from the unselective media could vary between the two, if the mutant was reverting back to wildtype. If the bacteria are reverting back to wildtype, these reverted bacteria would be able to proliferate in the unselective media, and not the selective media, causing a discrepancy in the number of CFUs. Figure 11: Illustration of colony quantification experimental procedure. 25 This colony quantification experiment was done for three disruption clones. The results can be seen in Figure 8. Though there was variability in the number of CFUs counted between clones, each clone had no significant difference between the number of CFUs counted on the final plates. This is true whether the starting culture contained hygromycin or not. The consistency in CFUs indicates that the disruption mutant is stable. Therefore, the bacteria retain the insertion mutation and do not revert back to wildtype. The variability between clones may be due to different amounts of starting bacteria used to grow the initial liquid culture since these are displayed on a linear scale. Figure 12: Calculated number of CFUs/mL after growth without selection. Three individual colonies of d0951 were grown overnight without hygromycin. Serial dilutions of these cultures were made and plated on to agar plates with and without hygromycin. CFU/mL were calculated as described in Methods. Constructing a genetic complement Four PCR primers for four different constructs were designed to complement strain with the FTL_0951 disruption. One pair of complementing constructs attempted to capture a native promoter by including 96 bp directly upstream of the yfjD gene. The other pair of constructs incorporated the promoter from the groEL locus in LVS (pGRO). One of each design included a carboxy-terminal tag using Strep Tag II the other without a tag. The overall strategy for complementation involved amplifying 26 the yfjD gene, with and without the upstream region, cloning the PCR product into a pGEM-T vector, and then subcloning into a pFNLTP8 vector. The pGRO promoter was used when there was no upstream sequence The resulting constructs were used to transform the yfjD disruption mutant (d0951). This cloning process was not completed for the construct designed to incorporate the pGRO promoter. In the trouble shooting process, it was postulated that the pGRO-pFNLTP8 vector was re- ligating with itself, since the vector was being cut open in a double-digest with Bam HI. The solution to this problem was treating the vector with shrimp alkaline phosphate (SAP) immediately after the Bam HI digest. This, however, failed to result in a successful ligation when used with a fresh batch of insert that had been cut from a successful pGEM ligation. A closer look at these complement plasmids in a gel revealed that there was leftover FTL_0883 (ybeX) DNA from the starting plasmid, which was meant to be removed in the Bam HI digest. To remedy this, a Bam HI digest was run for 24 hours before SAP treatment. Though this seemed to eliminate the ybeX sequence, the pGRO complements have yet to be made successfully. Complementing constructs incorporating the 96 bp upstream of FTL_0951 were successfully assembled based on restriction digest confirmation of the resulting plasmids. FTL_0951 Disruption mutant initiates cytokine response in mouse macrophage cells We first tested the effects of FTL_0951 disruption on the interaction of LVS with macrophages. TNF-α production was tested as a measure of macrophage stimulation and inflammation. For these, experiments, non-transformed macrophages were differentiated from monocytes derived from peripheral blood of healthy human blood donors. These primary macrophages were infected with LVS, ΔybeX, and d0951, with a media-only negative control. The cells were incubated overnight, the supernatants collected and analyzed via ELISA. The results, shown in Figure 9, show relatively low TNF-α production when macrophages are infected with LVS (Fig. 9). In comparison, the deletion and disruption mutant show higher production levels. 27 Figure 9: Release of TNF-α from infected macrophages. Primary human macrophages were cultured with different bacterial strains, where MOI = 10. Delta YbeX refers to the LVS strain with a deletion in the FTL_0883/ybeX gene. FTL_0951-4 is one of the d0951 strains. Supernatants were harvested after 24 hours and tested for TNF by ELISA. RAW 264.7 mouse macrophage cells with the NF-κB reporter were infected with LVS, d0951, two complements, and a d0951 with the irrelevant TC3D construct in the pFNLTP8 backbone. Uninfected wells in which the RAW cells were given media or LPS served as negative and positive controls, respectively. Media was used as a negative control, to show the baseline level of GFP production. The media images show very little GFP produced, indicating that the NF-κB pathway was not activated (Fig. 10, media). LPS was used as a positive control to show the levels of GFP production when the pathway is stimulated. The cells incubated with LPS were visibly brighter than the media control, indicating high GFP production levels and activation of the NF-κB pathway (Fig. 10, LPS). When the cells incubated with LVS were imaged, slightly more GFP was seen when compared to the media control. The level of GFP produced, however, was less than the LPS control (Fig. 10, LVS). In contrast, the d0951 disruption mutant stimulated GFP levels similar to the LPS control. These results indicate that, while LVS stimulates low NF-κB pathway activation, the disruption mutant activates the pathway. Both the tagged and the untagged complement strains showed GFP levels closer to the disruption mutant than the wild-type LVS strain (Fig. 10, Tagged Complement, Untagged Complement). This indicates the mutants with the 28 complementing construct are still stimulating the NF-κB pathway to the same levels as the mutant, suggesting that genetic complementation was not achieved. Figure 10: Infection assay. Raw cells were infected with different strains to a multiplicity of 10. The cells were GFP imaged bright field imaged at the same time. Disruption mutant has reduced protein release compared to wildtype Thus far, d0951 shared similar pro-inflammatory phenotypes with ΔybeX (Figs. 9 and 10). These results were consistent with the hypothesis that the similar domain architecture predicted involvement in similar processes in the bacteria (Fig. 3). An additional phenotype was tested to explore this further. Prior work in the Nau lab identified that more cellular material is released into the liquid phase of overnight ΔybeX broth cultures compared to wildtype (Hernandez, unpublished). Therefore, an immunoblot was done to compare the amount of protein released by LVS, d0951, and ΔybeX and its complements. Each sample was normalized to bacterial number based on the optical density of the liquid culture from which the protein was taken. A follow-up western blot was done measuring the released 29 protein of LVS, d0951, and the two native complements, one tagged and one not tagged. As seen previously, disruption of FTL_0951 resulted in less immuno-reactive material being released into the supernatant (Fig. 11, right, LVS versus d0951). However, plasmids with full-length FTL_0951 failed to return phenotypes back to the levels seen with LVS (Fig. 11, right). Figure 11: Immunoblots of protein released by LVS and mutants. All samples were normalized to optical density as a means of approximating bacterial number. Left: Immunoblot of wild-type LVS, ΔybeX mutant, the ΔybeX with a complementing vector, ΔybeX with an empty vector, and two 0951 disruption mutants. Right: Immunoblot of LVS, the disruption mutant, and two versions of the disruption complement. Troubleshooting the complementing constructs To troubleshoot the lack of complementation, an immunoblot was also done to test the amount of YfjD produced by the complement. In designing the complementing construct, a Strep Tag II tag was added to the end of yfjD. Probing for this tag allows the production of YfjD to be measured. This immunoblot used whole cell lysate. The resulting blot produced no visible strep tagged protein for the complement. This indicates that the complement was not producing YfjD, giving a possible explanation for the lack of complementation seen. It is also possible the tag is clipped off the protein at some point in its production. 30 The complementing construct was also sent to sequence. Comparing the sequencing results to the designed plasmid showed multiple in-frame deletions in the constructed plasmid. These gaps could alter the expression of yfjD in the complementing plasmid. An incorrect construct might produce a dysfunctional or nonfunctional form of YfjD, or might not produce protein at all. 31 Discussion The goal of this study was to investigate the impact YfjD production has on Francisella tularensis. This was done by introducing a genetic mutation into the yfjD, disrupting gene expression. By exploring the effects of the mutation, the role yfjD plays in the bacterium can be probed. The function of yfjD is interesting because of the similarities between the predicted structure of yfjD and ybeX, a protein previously found to be important to virulence in Francisella tularensis. By investigating the role of similar proteins, the role of YbeX can be investigated by proxy. A pJH1.d0951 suicide vector was created and integrated into the LVS genome. Being a suicide vector, the pJH1.d0951 construct is unable to replicate inside Francisella tularensis, which means only bacteria that have successfully integrated the plasmid into its genome will be resistant to hygromycin. To test if the disruption construct stayed stably integrated, colony counts were performed in bacteria growing with or without selective pressure. By altering the growth of bacterial cultures with and without antibiotics, the relative growth can be compared to determine if antibiotic resistance is lost over passages. Since resistance is imparted by the integrated DNA, loss of resistance can be attributed to loss of the integrated DNA or its functionality. The results of the experiment showed no significant difference between strains grown with or without hygromycin. This suggests that the mutant does not remove the integrated DNA from its genome under the conditions tested. Since pJH1 imparts hygromycin resistance, consistent growth in media containing hygromycin also ensures that only the mutant LVS is present. A potential limitation to this study is that stability was tested only in broth culture. F. tularensis has a wide range of habitats in cells and animals. It is possible that selection for removal of the pJH1.d0951 suicide vector could occur under some conditions that have yet to be defined. Nevertheless, it was important to demonstrate stability in standard lab conditions first. Previous work had shown ΔybeX strains release more material than wild-type LVS during culture in broth, suggesting there is instability of the cell envelope. This immunoreactive material is likely 32 comprised of proteins and other cellular constituents. Since d0951 behaved similar to ΔybeX in assays of macrophage stimulation, the amount of released protein between wild-type LVS and mutants were compared in this study. Here we found divergent phenotypes between ΔybeX and d0951. While ΔybeX released more material than wildtype into culture supernatants, the FTL_0951 disruption mutants, released qualitatively less than wildtype. This has some interesting implications on the possible functions and interactions of YbeX and YfjD. Because the mutants seem to have opposite effects on protein release, it is possible to propose a model where YbeX and YfjD counterbalance protein release and/or membrane stability. In this model, the YbeX protein inhibits protein release, possibly by promoting membrane stability. In the absence of YbeX, YfjD acts to promote membrane instability and protein release. In contrast, loss of YfjD has the opposite effect. In the FTL_0951d strains, the absence of YfjD results in more stability and less release because of unopposed YbeX activity. When creating genetic mutants, it is important to complement the mutations. A complement is a mutant whose defective gene expression has been restored by the reestablishment of another functional copy. Ideally, this new copy is introduce into the bacterium in a plasmid with the natural promoter, to match the natural environment of the original gene, including gene copy number and level of expression incited by the promoter. Another strategy is to introduce a plasmid with a copy of the gene being promoted by a known and understood promoter. Any approach can lead to variable or partial complementation because of incorrect levels of gene expression. As a result, phenotypes may or may not be restored to wild-type. The results obtained thus far do not make a case for complementation. One must examine these results keeping in mind the question of possible offsite effects. It must be determined whether it is more likely the phenotype differences are caused by the desired gene disruption, or if the process of making the insertion mutant affected other parts of the bacteria. 33 There are several reasons why complementation could have failed. It is possible that in creating the disruption mutant, other genes may have been affected. Genetic complementation of the yfjD gene would not restore these offsite mutations or correct effects on genes near the yfjD locus. Another explanation has to do with the nature of the mutation. Since the gene was disrupted and not deleted, the remaining sequence in the disruption mutant might produce a truncated version of the YfjD. This could have unexpected effects on the bacteria, such as being a dominant negative. Alternatively, the complementing construct may not be expressing appropriate levels of yfjD RNA. Abnormal expression levels or inappropriate regulation of gene expression could result in improper protein causing poor complementation. The answer to this question may lie in the genome of F. tularensis. When looking at the genome, yfjD gene and the genes that come before it read in one direction. The genes that come after read in the opposite. This means that, when reading in the direction that would properly codes for yfjD, there is no gene downstream that would be transcribed. This is why it is possible to disrupt yfjD in the manner it was; since downstream effects are not expected. As described above, the location of the integration of the disruption construct was determined to be in the yfjD gene as designed. Because of this, it is unlikely that the phenotype differences between the disruption mutant and the wildtype are attributed to downstream effects. To evaluate why complementation was not successful, two tests were performed. An immunoblot for the Strep Tag II added to the end of the tagged complement showed no tagged YfjD. While this suggests that no protein was being produced, it is possible that the tag was cleaved off the protein by the bacterium. Subsequent analysis of the construct’s DNA sequence revealed a stop codon, a three base pair DNA sequence that signals the end of translation, to be present in the complement construct just before the tag sequence. Because of this, it is possible that yfjD was transcribed, but translation ended before the tag sequence could be added. This would result in YfjD production without incorporating the Strep Tag II. 34 The constructed complement plasmids were also sequenced. Sequencing of the complement construct showed four in-frame deletions in the construct sequence. These gaps could alter protein expression, stability, or create a dysfunctional or nonfunctioning form of the protein to be made. These results show the lack of complementation was likely due to an incorrect complementing plasmid. While it is possible that there are offsite effects or a dominant negative, but that cannot be assessed until the proper complementing construct is assembled. Despite the lack of genetic complementation, certain conclusions can still be made. A disruption mutant was successfully created as designed, as confirmed by the PCR tests. This disruption mutant stimulated a larger pro-inflammatory response compared to wild-type LVS, both higher levels of TNF-α and more NF-κB activation. The disruption mutant also causes less protein to be released compared to wild-type LVS and the ΔybeX mutant. These results indicate that, while both the ybeX deletion and yfjD disruption mutant stimulate an immune response, they seem to have opposite effects on protein release and membrane stability. Follow-up studies should work on genetic complementation. As stated previously, complementation plays an important role in in genetic studies by showing that the phenotypes observed can be reversed by restoring proper gene expression. In addition, a yfjD deletion mutant can also be created, and the experiments retested with that strain. The likelihood of a dominant negative is lessened with deletion mutants. In a deletion strain there is little leftover DNA sequence to produce a bioactive protein. Experiments should also be done to further compare the impact and function of YfjD and YbeX in Francisella tularensis. Previous studies have shown more ΔybeX phenotypes, such as attenuated intracellular growth and virulence. Complementary studies can be done with FTL_0951d to further compare the impacts of YfjD and YbeX. Francisella tularensis is a Gram-negative bacterium, meaning it has an inner and an outer membrane. LPS is found on the outer leaflet of most Gram-negative bacteria. As previously stated, LPS 35 is immunostimulatory, triggering an inflammatory response and NO production in macrophages. It has also been reported to be involved in host cell phagocytosis, a process the F. tularensis uses to infect macrophages. LVS O-antigen mutants were phagocytosed by tight pseudopod layers compared to wild type LVS, which were phagocytosed by more spacious pseudopod loops [19]. Francisella tularensis also has secretion systems that contribute to bacterial virulence. F. tularensis has two TolC orthologs localized at the outer membrane that play roles in drug resistance, virulence, and suppressing pro-inflammatory cytokines [19]. Type VI secretion systems in F. tularensis are presumed to be anchored in the outer membrane and interacts with an inner membrane protein complex. They allow the transfer of effector proteins into host of competing bacterial cells. These systems are initiated by environmental stimuli such as macrophage uptake, and play roles in virulence and intracellular replication. Gram-negative bacteria are also known to spontaneously release proteins via outer membrane vesicles. Bacterial outer membrane proteins can also play roles in virulence, intracellular growth, phagosomal escape, uptake, and secretion [19]. In addition to outer membrane proteins, periplasmic and inner membrane proteins also perform a variety of functions that promote cell survivability and viability; energy production, intake, secretion, signal transduction, and virulence [19]. Membrane instability may involve the malfunction of the secretion systems and affect the function of the membrane proteins, leading to attenuated growth, lower virulence, or both. Therefore, disrupting membrane stability and function could be a valuable approach to altering the virulence of Gram negative pathogens like F. tularensis. Further study is needed to refine our understanding of CBS domain-containing proteins like YbeX and YfjD. Ultimately, these may be new therapeutic targets. 36 References [1] World Health Organization, "Antimicrobial resistance: global report on surveillance," World Health Organization, Geneva, 2014. [2] P. C. F. Oyston, "Francisella tularensis: unravelling the secrets of an intracellular pathogen," Journal of Medical Microbiology, pp. 921-930, 2008. [3] K. L. Elkins, S. C. Cowley and C. M. Boiso, "Innate and adaptive immune responses to an intracellular bacterium, Francisella tularensis live vaccine strain," Microbes and Infections, pp. 135-142, 2003. [4] Centers for Disease Control and Prevention, "Tularemia," 27 September 2016. [Online]. Available: https://www.cdc.gov/tularemia/. [Accessed 2 April 2018]. [5] Centers for Disease Control and Prevention, "Select Agents & Toxins FAQ's," 2017. [Online]. Available: https://www.selectagents.gov/faq-general.html. [Accessed 2 April 2018]. [6] Centers for Disease Control and Prevention, "Bioterrorism Agents/Diseases," 17 August 2017. [Online]. Available: https://emergency.cdc.gov/agent/agentlist-category.asp. [Accessed 2 April 2018]. [7] Centers for Disease Control and Prevention, "Tularemia - United States, 2001-2010," Morbidity and Mortality Weekly Report, vol. 62, no. 47, pp. 963-966, 2013. [8] P. E. Carlson, J. A. Carroll, D. M. O'Dee and G. J. Nau, "Modulation of virulence factors in Francisella tularensis determines human macrophage responses," Microbial Pathogenesis, pp. 204-214, 2007. [9] J. Pizarro-Cerda, A. Charbit, J. Enninga, F. Lafont and P. Cossart, "Manipulation of host membranes by the bacterial pathogens Listeria, Francisella, Shigella and Yersinia," Seminars in Cell & Developmental Biology, pp. 155-167, 2016. [10] S. Steele, L. Radlinski, S. Taft-Benz, J. Bruton and T. H. Kawula, "Trogocytosis-associated cell to cell spread of intracellular bacterial pathogens," eLife Sciences, pp. 1-17, 2016. [11] L. Ferrero-Miiani, O. H. Nielsen, P. S. Andersen and S. E. Girardin, "Chronic inflammation: importance of NOD2 and NALP3 in interleukin-1B generation," Clinical & Experimental Immunology, vol. 147, no. 2, pp. 227-235, 2007. [12] J. Ellis, P. C. F. Oyston, M. Green and R. W. Titball, "Tularemia," Clinical Microbiology Reviews, pp. 631-646, 2002. [13] J. MacMicking, Q.-w. Xie and C. Nathan, "Nitric Oxide and Macrophage Function," Annual Review of 37 Immunology, pp. 323-350, 1997. [14] B. C. Russo, J. Horzempa, D. M. O'Dee, D. M. Schitt, M. J. Brown, P. E. Carlson, Jr., R. J. Xavier and G. J. Nau, "A Francisella tularensis Locus Required for Spermine Responsiveness Is Necessary for Virulence," Infection and Immunity, pp. 3665-3676, 2011. [15] N. Zhang, X. Ren, D. Zhu, D. Li and D. Wang, "Crystallization and preliminary crystallographic studies of CorC, a magnesium-ion transporter," Structural Biology and Crystallization Communications, pp. 681-683, 2010. [16] L. C. Kingry and J. M. Peterson, "Comparative review of Francisella tularensis and Francisella novicida," Frontiers in Cellular and Infection Microbiology, pp. 1-12, 2014. [17] L. Young, D. G. Gibson, R.-Y. Chuang, J. C. Venter, C. A. Hutchison III and H. O. Smith, "Enzymatic Assembly of DNA molecules up to Several Hundred Kilobases," Nature Methods, vol. 6, no. 5, pp. 343-345, 2009. [18] New England BioLabs, Gibson Assembly Master Mix Instruction Manual, Ipswich, MA: New England Biolabs. [19] H. M. Rowe and J. F. Huntley, "From the Outside-In: The Francisella tularensis Envelope and Virulence," Frontiers in Cellular and Infection Microbiology, vol. 5, 2015. [20] M. S. Trent, L. M. S. Worsham and M. L. Ernst-Fonberg, "HlyC, the Internal Protein Acyltransferase That Activates Hemolysin Toxin: Role of Conserved Histidine, Serine, and Cysteine Residues in Enzymatic Activity As Probed by Chemical Modification and Site-Directed Mutagenesis," Biochemistry, pp. 3433-3439, 1999. 38