Identification of Francisella tularensis Himar1-Based Transposon Mutants Defective for Replication in Macrophages

Bacterial strains, plasmids, and growth conditions.All bacterial strains and plasmids used in this study are listed in Table 1. Studies involving type A F. tularensis strain Schu S4 were carried out under biosafety level 3 conditions according to federal and institutional select agent regulations. F. tularensis subspecies were routinely grown at 37°C in 2.1% Mueller-Hinton (MH) medium supplemented with glucose, iron, and IsoVitaleX as described previously (41). MH broth contained 1.24 mM CaCl₂ and 1.03 mM MgCl₂, and MH agar contained 0.5% NaCl, 1.6% agar, 1% tryptone (MHT) or protease peptone (MHP), and 2.5% calf or fetal bovine serum (Invitrogen, Carlsbad, CA). For in vivo infections, inocula were prepared in MHP broth. In some experiments, Chamberlain's chemically defined medium (CDM) was used (10). When required, medium was supplemented with kanamycin (10 μg ml⁻¹) or nourseothricin (5 μg ml⁻¹). Selection for nourseothricin resistance was performed at 30°C. Escherichia coli was grown aerobically at 37°C in Luria-Bertani medium supplemented with kanamycin (50 μg ml⁻¹) or nourseothricin (50 μg ml⁻¹) when required. Kanamycin was purchased from United States Biochemical Corporation (Cleveland, OH), and nourseothricin was purchased from WERNER BioAgents (Jena, Germany).

DNA manipulation and transformation.Purification and manipulation of plasmid or genomic DNA, electroporation of F. tularensis, and chemical transformation of E. coli were performed as described previously (41). Custom oligonucleotide primers (Table 1) were synthesized by Operon (Huntsville, AL). DNA maps were constructed using MacPlasmap Pro 3.01 (CGC Scientific, Inc., Ballwin, MO). To determine insertion locations in F. tularensis LVS, genomic DNA from each strain was digested with SpeI and treated with T4 DNA ligase. HimarFT-containing fragments were recovered as plasmid DNA in E. coli DH5αλpir and subsequently used as a template for sequencing (9). Insertion locations for F. tularensis Schu S4 single- and double-crossover strains with mutations in FTT1238c were determined by sequencing directly from genomic DNA. Sequencing reactions and analysis were performed as described previously (41). Southern blotting of SpeI-digested genomic DNA was performed as described previously using either the Random Prime DNA labeling system (Invitrogen) and [α-³²P]dCTP or a digoxigenin High Prime DNA labeling and detection starter kit (Roche Diagnostics Corporation, Indianapolis, IN). Probes constructed for identification of the aminoglycoside 3′-phosphotransferase (aphA-2) (64), the transposase (tnp), or the N terminus of FTL_0706 are described in Table 1.

Genome and bioinformatic analyses.The HimarFT insertion site sequences were compared with the F. tularensis LVS genome (accession no. NC_007880) at http://bbrp.llnl.gov/bbrp/bin/f.tularensis_blast and the F. tularensis Schu S4 genome (NC_006570) available at http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=genome&cmd=Retrieve&dopt=Overview&list_uids=563 . Additional F. tularensis subsp. genomic comparisons were performed using www.francisella.org .

Further analysis and annotation of candidate proteins were performed using the high-resolution tools PhyloBlocks (http://compbio.mcs.anl.gov/ulrich/phyloblock/ ) and Dragonfly (http://compbio.mcs.anl.gov/dragonfly/ ) available via the PUMA2 system (43). Candidate sequences were analyzed using BLAST, InterPro, SignalP, and TMHMM to identify global sequence alignments, local conserved domains, potential signal sequences, and potential transmembrane domains, respectively (1, 5, 47, 48). The results of these analyses were used to assign potential functions to the genes. Further identification of the conserved regions specific for predicted functions was done using PhyloBlocks. In brief, a set of orthologous sequences identified by BLAST were aligned using ClustalW and analyzed by Block Maker for prediction of the conserved motifs (31). Consensus sequences and HMM profiles were also generated for these sets of orthologs to be used for future analysis and classification.

Western blot analysis.Whole-cell lysates (1× 10⁹ cells in 100 μl) were boiled for 5 min in loading buffer, and 5 μl of each sample was loaded and separated on 10% polyacrylamide-sodium dodecyl sulfate gels. Western transfer, blocking, and antibody incubation were performed as described previously (41). Treatment with the mouse monoclonal anti-F. tularensis LPS primary antibody (1:8,000 dilution; Advanced ImmunoChemical Inc.) was followed by treatment with a horseradish peroxidase-conjugated goat anti-mouse immunoglobulin G secondary antibody (1:10,000 dilution; Roche).

F. tularensis LVS HimarFT library construction.The pHimar H3 construct is a modified HimarFT suicide delivery vehicle constructed from temperature-sensitive pFNLTP16 H3 (42). Transposition using pHimar H3 was performed by electroporation of 100 ng of DNA into electrocompetent F. tularensis LVS. After outgrowth at 37°C on a shaker for 4 h, dilutions were plated onto MHT agar containing kanamycin and incubated at 37°C to select for HimarFT. In addition, dilutions were plated onto MHT agar without kanamycin to determine the total number of potential recipients. Colonies grown for 3 or 4 days at 37°C were picked, struck onto MHT agar containing kanamycin, and grown at 37°C to recover individual clones containing HimarFT insertions in the genome. For HimarFT library construction and storage, Km^r colonies were placed into 0.5 ml MH broth containing kanamycin in tissue culture plates with 96 deep wells. The last row in each plate was not inoculated as a control. Sterile lids and Parafilm were used to cover the plates for incubation and freezer storage. After 2 days of growth at 37°C and 200 rpm, 0.5 ml sterile 30% glycerol was added and mixed into each well to obtain a final glycerol concentration of 15%, and the plates were stored at −80°C. Recovery of the library was performed by using a 96-pin replicator and square agar plates (Nalge Nunc International Corp., Rochester, NY) containing MHT agar with kanamycin.

HimarFT library screen utilizing a macrophage infection model.Stage 1 screening with J774A.1 macrophages (ATCC TIB-67) was performed using a 96-well format with a multiplicity of infection (MOI) of ∼50 to 200. Macrophages were seeded at a concentration of 6 × 10⁴ cells per well and incubated overnight at 37°C in 5% CO₂ before infection. Strains containing transposon insertions were grown on MHT agar with kanamycin for 3 days at 37°C and for 0 to 2 days at room temperature and transferred into a 96-well plate containing 100 μl of Dulbecco's modified Eagle's medium (DMEM) (Sigma-Aldrich, St. Louis, MO) supplemented with 10% fetal bovine serum and 4 mM l-glutamate (Invitrogen). The DMEM was aspirated from the macrophages, and 50-μl portions of test bacterial suspensions were placed onto macrophages. These steps were performed one row at a time to minimize drying of the monolayers. A 2-h incubation at 37°C in the presence of 5% CO₂ was performed to allow bacterial attachment and entry. Each row of eight wells was then aspirated to remove the extracellular organisms, washed with 200 μl of Hanks' balanced salt solution, and aspirated again, and the contents were replaced with 200 μl of DMEM containing 5 μg ml⁻¹ of gentamicin (Invitrogen) to kill extracellular bacteria. The plates were then incubated at 37°C in the presence of 5% CO₂ for 2 days. Candidates, which retained the macrophage monolayer similar to uninoculated controls (groups A and B), were subjected to a second-stage 24-well analysis with 3.7 × 10⁵ macrophages per well and defined inocula normalized using optical density to obtain an MOI of 100.

Crystal violet staining of J774A.1 macrophages.The integrity of the macrophage monolayer was assessed by crystal violet staining. DMEM was aspirated from the macrophages, and the monolayers were washed with phosphate-buffered saline and then incubated with a 1% crystal violet solution for 5 min at room temperature. The crystal violet was removed, and the macrophage monolayers were washed with distilled water and allowed to dry before visual assessment of the stain intensity. Crystal violet was purchased from Hardy Diagnostics (Santa Maria, CA).

Quantitation of Francisella cytotoxicity with J774A.1 macrophages by LDH release.The release of cytosolic lactate dehydrogenase (LDH) into the supernatant of infected cells was used as an indicator for lysis (CytoTox96; Promega Corporation, Madison, WI). Bacterial cell suspensions were prepared and normalized to an MOI of 100 based on the optical density at 550 nm (OD₅₅₀) and subsequent plating to determine the actual number of CFU ml⁻¹. The conditions used for J774A.1 macrophage infection were the same as those used for the 24-well stage 2 cytopathic screening. At 48 h postinfection, supernatants were harvested and the maximum lysis was determined (by addition of 100 μl of lysis buffer to an uninfected monolayer with 900 μl DMEM and incubation at 37°C for 45 min). Uninfected medium was used as a background control, and suitable dilutions of the other samples were included. The release of LDH was determined by measuring the OD₄₉₀ using a plate reader and softmax pro software (Molecular Devices Corporation, Sunnyvale, CA).

Growth analysis of F. tularensis LVS mutants in bacteriological medium.Growth in MH or CDM broth was assessed using a 96-well format with a plate reader and softmax pro software. One test colony was emulsified in 100 μl of CDM broth, 10 μl was placed into 190 μl of MH or CDM broth, and the microtiter plate was incubated at 37°C with shaking at 100 to 200 rpm. Turbidity changes in the cultures were monitored at OD₅₅₀ for 24 h (data not shown). Some strains were grown on MH or CDM agar to confirm the observed phenotypes in broth.

In vivo analysis of HimarFT insertion mutants.Six- to 8-week-old female BALB/c mice (Harlan Sprague Dawley, Inc., Indianapolis, IN) were infected by intraperitoneal (i.p.) injection of 0.1 ml (final volume) of phosphate-buffered saline containing the F. tularensis LVS wild type or mutant derivatives. The inocula were prepared from MHP broth cultures in the exponential phase of growth, and the concentrations were adjusted so that they represented approximately 1,000 50% lethal doses (21, 29). The actual number of bacteria delivered was determined by performing plate counts of the inoculum on MHP agar. Infected animals were closely monitored for 21 days, and the number of moribund animals during this time was determined. Animal infection experiments were approved by the Institutional Animal Care and Use Committee and the Institutional Biosafety Committee of the Medical College of Wisconsin.

Complementation of selected HimarFT mutant strains using pFTNAT.The complementation plasmid pFTNAT was constructed by inverse PCR using primers IP1 and IP2 (Table 1) and pFNLTP7 as the template DNA (41). Both the kanamycin and ampicillin resistance genes were removed, and the nourseothricin resistance marker was inserted at a NotI site (26). The pFTNAT plasmid retains the ability to replicate in both E. coli and F. tularensis. The complementation plasmids pFTNAT-FTL_0706, pFTNAT-wbtA_LVS, and pFTNAT-wbtM_LVS were constructed by PCR amplification of genomic DNA with primer pairs H3/H4, WA1/WA2, and WM1/WM2 (Table 1), respectively, resulting in gene fragments with 175-, 472-, and 481-bp upstream sequences (putative promoter regions). All fragments were ligated into pCR-Blunt II-TOPO and subsequently subcloned into pFTNAT using EcoRI, XhoI, and XhoI, respectively.

RNA isolation and RT-PCR analysis of F. tularensis HimarFT strains with insertions in wbtA.RNA was isolated using an RNeasy Protect mini kit (QIAGEN). Isolated RNA was treated with DNase I (Fermentas Inc., Hanover, MD) to remove contaminating genomic DNA for 1 h at 37°C and subsequently purified with the RNeasy Protect mini kit. RNA quality was verified by agarose gel analysis. Reverse transcription (RT)-PCR was performed using the SuperScript III One-Step RT-PCR system with Platinum Taq DNA polymerase (Invitrogen). Control PCRs were performed with Platinum Taq High Fidelity DNA polymerase (Invitrogen). Primers C1, C2, D1, and D2 were used for mRNA detection of either a 366-bp fragment of wbtC or a 410-bp fragment of wbtD downstream of wbtA (Table 1). All reaction mixtures included 50 ng of template DNA or RNA and were incubated at 50°C for 30 min for RT, followed by PCR for 25 cycles.

Statistical analysis.Statistical significance was determined using single-factor analysis of variance.

Transposon mutagenesis of F. tularensis LVS using HimarFT.In previous studies, a temperature-sensitive plasmid and a modified Himar1 transposon (HimarFT) were developed to perform random mutagenesis in F. tularensis (41, 42). While this system resulted in random, single, stable insertions without orientation bias, a drawback was observed when large-scale experiments were performed to saturate the genome. We noted that plasmid sequences were retained at a variable frequency. The failure to lose plasmid sequences may have been due to the length of time required for the plates to achieve the restrictive temperature, the number of generations required to dilute nonreplicating DNA, or the reversion of the point mutation in the replication initiation protein encoded by repA. Sequencing of repA from five isolates retaining the plasmid sequence confirmed the maintenance of the temperature-sensitive mutation. Plasmid rearrangements were not observed (data not shown). We noted that plasmid sequences were lost if strains were passaged at the restrictive temperature two additional times (49).

To avoid the labor-intensive steps of passaging individual isolates, the delivery vehicle was modified by removing the NotI fragment encoding pFNLTP16 to construct pHimar H3 (Fig. 1A). Deletion of the replicon for Francisella converted pFNLTP16 H3 to a suicide delivery system for HimarFT. The efficiency of transposition with pHimar H3 was similar to that obtained for pFNLTP16 H3, with ∼2 × 10⁻⁵ insertion per cell. The stable retention of HimarFT and the loss of plasmid sequences were confirmed by examining 100 mutagenized strains by colony blot analysis with ³²P-labeled probes hybridizing to the aminoglycoside 3′-phosphotransferase (aphA-2) responsible for kanamycin resistance and the transposase (tnp), respectively. Southern blot analysis with genomic DNA isolated from 10 mutants after transposition demonstrated that there were random, single insertions and loss of the plasmid vehicle (Fig. 1B and data not shown). For construction of the insertion library, F. tularensis LVS was mutagenized with pHimar H3 and approximately 7,000 kanamycin-resistant colonies were arrayed into microtiter plates with 96 deep wells and preserved at −80°C.

• Open in new tab
• Download powerpoint

FIG. 1.

(A) Construction of a suicide delivery vehicle for HimarFT transposition. The original HimarFT transposon with the Francisella groEL promoter (pgro) facilitating expression of the aminoglycoside 3′-phosphotransferase (aphA-2) and transposase (tnp) genes on the temperature-sensitive pFNLTP16 vehicle is shown. pHimar H3 was constructed by removal of the NotI fragment containing pFNLTP16. (B) Verification of pHimar H3 insertion in F. tularensis LVS by Southern blot analysis of SpeI-digested genomic DNA. Hybridization with a ³²P-labeled probe specific to aphA-2 was observed for 10 representative strains after pHimar H3 mutagenesis. Wild-type F. tularensis LVS genomic DNA (lane L) was included as a negative control. Three plasmids were included as markers (42): pMiniHimar (4 kb) (lane A), pFNLTP1 (6.9 kb) (lane B), and pFNLTP16 H3 (10 kb) (lane C). A probe specific to the transposase (tnp) gene on the plasmid did not hybridize with genomic DNA but produced a signal for markers A and C (data not shown).

Isolation of mutants with insertions in Francisella genes required for virulence in J774.A1 macrophages. Francisella enters, replicates, and causes apoptosis-induced cytopathogenicity within macrophages (44, 51, 65). To screen for genes required for F. tularensis LVS virulence in J774.A1 macrophages, the integrity of macrophage monolayers infected with individual Km^r clones was assessed by staining with crystal violet. Preliminary experiments indicated that crystal violet staining at 2 days after infection at an MOI of 100 was optimal for the detection of cytopathogenicity (Fig. 2A). The initial screening (stage 1) was performed with an estimated MOI of 50 to 200 using a 96-well format. The retention of the monolayer and dark staining suggested that the clones were unable to enter, replicate, or cause apoptosis during the 2-day incubation. Noncytopathic candidates were retested in duplicate (stage 2) in 24-well plates at an MOI of 100 (Fig. 2B). Mutants were categorized into four groups based on staining intensity: no loss of the monolayer, as in the uninfected control well (group A), a small but detectable loss of monolayer (group B), detectable loss of cells (group C), and severe monolayer destruction, similar to that obtained with the parental strain, F. tularensis LVS (group D). To cull putative auxotrophs or other mutants defective in general metabolism, 140 candidates appearing to possess the most attenuated phenotypes (groups A and B) in macrophages were characterized for growth in complex and minimal bacteriological media (stage 3 screening) (data not shown). Duplicate cultures for each strain were arrayed in 96-well plates in MH or CDM broth, and growth was monitored over time by assessing the increase in optical density. Colony morphology was assessed using cultures transferred from microtiter broth cultures to MHT or CDM agar plates. Twenty-four strains failed to grow in or on CDM and are likely auxotrophic strains. Slower growth or inconsistent growth patterns were observed for another 24 candidates. In summary, the combined screens identified 89 strains with HimarFT insertions in genes potentially important for virulence in macrophages.

• Open in new tab
• Download powerpoint

FIG. 2.

(A) Crystal violet staining of J774A.1 macrophages infected with F. tularensis LVS at various MOIs 2 days postinfection. The retention of the J774A.1 macrophage monolayer after infection is compared to the uninfected control (none). (B) F. tularensis LVS HimarFT mutant stage 2 macrophage screening using an MOI of 100. The uninfected control (well N) retains the monolayer and stains dark, while wild-type F. tularensis LVS (well L) is cytopathic, causing loss of the monolayer. Based on crystal violet retention, strains with HimarFT insertions were classified into four groups (A to D) represented by wells A to D, respectively: group A, similar to uninfected controls; group B, intermediate dye retention; group C, light dye retention; and group D, no dye retention, similar to infection by the parental strain. Phenotypes corresponding to groups A and B were characterized further.

Identification of the site of HimarFT insertion.To recover DNA flanking the HimarFT insertion site, genomic DNA from each of the 89 candidates was digested with SpeI, ligated, and transformed into E. coli DH5αλpir. Of the 89 genomes, 70 (79%) were recovered as independently replicating plasmids containing HimarFT and adjacent genomic DNA, an efficiency similar to that reported elsewhere for transposons in Francisella (69). Recovered plasmids were used as templates for nucleic acid sequencing reactions with primers specific for HimarFT sequences corresponding to aphA-2 and the R6K origin (Table 1 and Fig. 1A). From this analysis, 26 unique genes were identified. Known or predicted functions for these genes included transport processes, metabolism of nucleotides, carbohydrates, amino acids, and coenzymes, capsule and LPS biosynthesis, and posttranslational modification/protein turnover (Table 2). The recovery of several HimarFT insertions in genes previously found to have a role in macrophage growth, including the caseinolytic protease gene clpB, the disulfide bond-related gene dsbB, the aspartate aminotransferase gene aspC2, and the putative capsular genes capBC and wbtA involved in O-antigen biosynthesis, validated the screening process used (28, 55, 58, 63, 67, 69, 70, 76). Several hypothetical proteins were also identified. Insertions appeared to be spread evenly throughout the genome (Fig. 3).

• Open in new tab
• Download powerpoint

FIG. 3.

F. tularensis LVS genomic map of HimarFT insertion mutants recovered after stage 2 and 3 screening (designations above the F. tularensis LVS genome). Hypothetical HimarFT mutants are underlined. Several genes previously reported to be involved in growth inside macrophages were recovered, including clpB, FTL_0439, wbtA, capBC, aspC2, and dsbB (designated by asterisks; above the F. tularensis LVS genome). Regions and genes previously shown to be involved in macrophage growth but not recovered as strongly attenuated after the screening of our HimarFT library are indicated by the gray boxes corresponding to the Francisella pathogenicity islands (FPI#1 and FPI #2) and below the F. tularensis LVS genome, respectively.

In vivo analysis of mutants containing transposon insertions in genes encoding hypothetical proteins.To focus on new gene discovery, strains with insertions in hypothetical genes were selected to test for attenuation in vivo. Groups of five BALB/c mice were given an intraperitoneal (i.p.) injection of mutant derivatives at doses ranging from 1,000 to 5,000 CFU or 3 logs above the reported 50% lethal dose (21, 29). Previous work indicated that BALB/c mice are moribund by day 4 to 8 when they are infected i.p. with 50 CFU or less (39, 53). Mice infected with strains bearing insertions in the hypothetical genes tested survived throughout the 21 days of monitoring (Table 3). We also challenged mice with F. tularensis FTL_0741::HimarFT, with a transposon insertion in a hypothetical locus that exhibited a parental phenotype in vitro. In vivo, LVS FTL_0741::HimarFT exhibited a parental phenotype, while LVS ΔpurMCD failed to induce active infection (53). These results suggest that a HimarFT insertion generates a specific phenotype related to its location in the chromosome.

Characterization of transposon mutants for entry, initial replication, and cytotoxicity.Attenuation in vivo for all mutants tested suggested that the in vitro cellular assay results correlated with pathogenicity. To determine if quantitative defects were associated with different stages of infection, mutants with insertions in hypothetical genes were tested with J774A.1 macrophages to determine their capacities to enter, replicate, and cause the release of LDH, an indicator of cell death (Fig. 4). An intracellular replication-deficient mutant used as a control, F. tularensis LVS ΔpurMCD, exhibited reduced replication and cytotoxicity, consistent with its attenuation in mice (53). Similarly, inactivation of FTL_0741 resulted in no change in replication and cytotoxicity, consistent with the wild-type phenotype in mice. There was no significant difference between LVS and the strains bearing HimarFT insertions in terms of the ability to enter J774A.1 macrophages (Fig. 4A). Statistically significant differences (P < 0.05) from the parental strain were observed when replication at 24 h and cytotoxicity at 48 h for mutants with insertions in FTL_0886, FTL_1414, FTL_0439, FTL_0706, and FTL_0050 were examined (Fig. 4B and C), consistent with the attenuation of these mutants in mice. Thus, defects in macrophage replication generally correlate with the inability to induce cytotoxicity. Interestingly, LVS FTL_0544::HimarFT displayed no significant changes in either cytotoxicity or replication, despite being attenuated in mice, compared to the wild type (Fig. 4B and C).

• Open in new tab
• Download powerpoint

FIG. 4.

(A and B) Entry and replication in J774A.1 macrophages of the F. tularensis LVS wild type and ΔpurMCD, FTL_0741::HimarFT, and attenuated mutants with insertions in loci encoding hypothetical proteins. The experiments were performed using a 24-well format with an MOI of 100 in triplicate. The results are expressed as mean log CFU ml⁻¹; the error bars indicate standard deviations. Mutants found to be significantly impaired in entry at 2 h or in replication at 24 h are indicated by asterisks (P < 0.05). (C) Quantitation of cytotoxicity by measuring LDH release into the supernatant compared to an uninfected monolayer (maximum lysis) at 48 h. LDH release was monitored alongside assessment for entry and replication defects as described above for panels A and B. Experiments were performed in triplicate, The bars indicate mean percentages of cytotoxicity, and the error bars indicate standard deviations. Mutants found to have a significant reduction in the percentage of cytotoxicity at 48 h are indicated by asterisks (P < 0.05).

LPS phenotype of selected HimarFT mutant strains.LPS biosynthesis has been previously shown to be important for intracellular replication (13, 58, 63, 70). Loss of a functional copy of wbtA results in loss of the O-antigen ladder (58, 63). Five F. tularensis LVS wbtA::HimarFT mutants from this study exhibited loss of O antigen, as predicted (Fig. 5A). Two additional mutations in the wbt locus, with HimarFT insertion in either wbtM or wbtC, resulted in complete O-antigen loss and loss of low-molecular-weight O antigen, respectively (Fig. 5A). These phenotypes have not been previously reported and likely correlate to the location in the pathway of LPS biosynthesis. Other mutations in the wbt locus, such as wbtDEF mutations, also result in a lack of O antigen (70). Strains with insertions in predicted capsule-related genes retained the O-antigen ladder, suggesting that mutation of these genes does not affect LPS (Fig. 5B). HimarFT mutations within hypothetical genes were analyzed by Western blotting to assess whether any of these genes represent unrecognized LPS biosynthetic genes. All of the insertions in hypothetical genes resulted in a parental pattern for the O-antigen side chain except for LVS FTL_0706::HimarFT (Fig. 5C). We concluded that the hypothetical protein encoded by FTL_0706 likely functions in the LPS biosynthetic pathway.

• Open in new tab
• Download powerpoint

FIG. 5.

Western blot analysis of F. tularensis LVS or F. novicida U112 cell lysates with a mouse monoclonal anti-F. tularensis LPS antibody. The antibody recognizes LPS from F. tularensis LVS and Schu S4 but not LPS from F. novicida U112. Each HimarFT mutant strain is indicated by the gene in which HimarFT insertion was mapped, as shown in Table 2. Multiple strains with mutations in the same gene are indicated by the gene designation and a candidate number. The previously studied F. tularensis LVS ΔpurMCD mutant is indicated by MCD (53).

Complementation and HimarFT polarity. HimarFT was engineered to contain a strong constitutive groEL promoter to transcribe the aphA-2 kanamycin resistance marker. Transcriptional stop sequences were not added, raising the possibility that the orientation of insertion may affect the transcription of adjacent genes. HimarFT mutants with insertions mapped to the beginning (wbtA) and the end (wbtM) of the wbt region were used to determine if the orientation of HimarFT influenced either the complementation pattern or transcription of downstream genes (Fig. 6A). Gene expression analyses were performed with F. tularensis LVS wbtA1::HimarFT and wbtA2::HimarFT, in which HimarFT had been mapped in the two orientations (Fig. 6A). To detect the presence of mRNA downstream of wbtA, RT-PCR was performed with primers designed to detect message containing wbtC or wbtD (Fig. 6A). Fragments that were the correct size were obtained in PCR and RT-PCR controls with F. tularensis LVS genomic DNA, verifying that there was primer-mediated amplification of the correct region (Fig. 6B). Products were not detectable when template DNA was not included or in PCRs utilizing DNase-treated RNA isolated from wild-type or mutant strains, indicating that the preparations of RNA were not contaminated with genomic DNA (Fig. 6B). RT-PCRs with RNA isolated from LVS, LVS wbtA1::HimarFT, and wbtA2::HimarFT each demonstrated that there was a message containing wbtD on a fragment whose size was similar to that of the genomic DNA control (Fig. 6B). Similarly, a product was detected for all three strains when wbtC was the target sequence (data not shown). The amounts of wbtC and wbtD cDNAs appeared to be reduced in strains containing an upstream HimarFT insertion. These data suggest that HimarFT insertion is partially polar.

• Open in new tab
• Download powerpoint

FIG. 6.

Analysis of F. tularensis LVS HimarFT strains with mutations in the wbt locus by RT-PCR and complementation. (A) Map of the wbt loci (FTL_0592 to FTL_0595) in the F. tularensis LVS genome. The HimarFT insertion sites are indicated for F. tularensis LVS wbtA1::HimarFT, wbtA2, and wbtC mutant strains (indicated by the gene designation and candidate number). FTL_0606 (wbtM) is just downstream of the main wbt operon (FTL_0592 to FTL_0605). The genomic region used for complementation of the wbtA mutants is between the XhoI sites. Bars 1 and 2 correspond to the regions selected for detection of mRNA downstream of wbtA by RT-PCR. HimarFT insertion site maps for F. tularensis LVS wbtA1::HimarFT and wbtA2::HimarFT are shown, with LVS wbtA2::HimarFT containing the groEL promoter (pgro) in the same orientation as the gene. (B) Detection of mRNA containing wbtD downstream of wbtA in LVS wbtA1::HimarFT and wbtA2::HimarFT by RT-PCR. The reaction and template conditions used included RT-PCR with no template (lane 1), LVS genomic DNA (lane 2), and RNA isolated from either LVS wild-type (lane 3), wbtA1::HimarFT (lane 5), or wbtA2::HimarFT (lane 7). Control RT-PCRs without the reverse transcriptase were performed using RNA from LVS wild-type (lane 4), wbtA1::HimarFT (lane 6), and wbtA2::HimarFT (lane 8). (C) Plasmid pFTNAT containing the nourseothricin resistance gene used for complementation of HimarFT mutant strains. (D) Western blot analysis of cell lysates using a mouse monoclonal anti-F. tularensis LPS antibody. The strains analyzed include LVS with pFTNAT (lane 1); LVS wbtA1::HimarFT alone (lane 2), with pFTNAT (lane 3), or with pFTNAT-wbtA_LVS (lane 4); LVS wbtA2::HimarFT alone (lane 5), with pFTNAT (lane 6), or with pFTNAT-wbtA_LVS (lane 7); and LVS wbtM1::HimarFT alone (lane 8), with pFTNAT (lane 9), or with pFTNAT-wbtM_LVS (lane 10).

To complement HimarFT mutants, we constructed a new shuttle vector containing a nourseothricin resistance marker, pFTNAT (Fig. 6C). Nourseothricin is an antibiotic that inhibits a broad spectrum of organisms and is postulated to inhibit protein synthesis by induction of miscoding events (26). It is an attractive marker for Francisella research as it is not used in human or veterinary medicine and has not been shown to display cross-resistance (42). Genomic fragments containing the genes and their predicted promoter regions for complementation were cloned into pFTNAT and transformed into the appropriate HimarFT mutant strains.

Complementation of mutants with insertions in either the wbt region or FTL_0706 was assessed both for O-antigen content and for cytopathogenicity in macrophages. O-antigen expression and cytopathogenicity were restored to F. tularensis LVS wbtM::HimarFT with pFTNAT-wbtM_LVS in trans (Fig. 6D and data not shown). However, complementation of LVS wbtA1::HimarFT was not detected using a similar cloning strategy (pFTNAT-wbtA_LVS) (Fig. 6D). Partial restoration of O-antigen expression occurred when LVS wbtA2::HimarFT was transformed with pFTNAT-wbtA_LVS (Fig. 6D), but this strain still displayed reduced cytopathogenicity compared to LVS (data not shown). The presence of plasmid pFTNAT did not alter the strain phenotypes. The lack of complementation or partial complementation is consistent with results from RT-PCR analyses and suggests that the orientation of HimarFT influences the expression of adjacent genes. These data also support an operon organization for the wbt region.

F. tularensis LVS FTL_0706-1::HimarFT and FTL_0706-2::HimarFT mutants were complemented in trans using pFTNAT-FTL_0706 (Fig. 7A). FTL_0706 does not appear to be part of an operon based on the current annotation of the LVS genome. FTL_0706 provided in trans restored the O-antigen ladder in both FTL_0706::HimarFT mutant strains (Fig. 7B). Cytopathogenicity was restored in LVS FTL_0706-1::HimarFT and FTL_0706-2::HimarFT with pFTNAT-FTL_0706, resulting in loss of the macrophage monolayer, similar to the parental phenotype (Fig. 7C).

• Open in new tab
• Download powerpoint

FIG. 7.

Complementation of F. tularensis LVS FTL_0706::HimarFT mutant strains in trans using pFTNAT-FTL_0706. (A) Two mutants containing HimarFT insertions within FTL_0706 in opposite orientations were assessed for complementation (locations are indicated by the gene designation and candidate number). LVS FTL_0706-1:: HimarFT contains the groEL promoter (pgro) in the same orientation as FTL_0706. F and R indicate the DNA fragment amplified from the genome, and the EcoRI sites indicate the fragment used for complementation. (B) Western blot analysis of cell lysates from LVS and LVS FTL_0706::HimarFT strains using a mouse monoclonal anti-F. tularensis LPS antibody. The strains analyzed include LVS without (lane 1) or with (lane 2) pFTNAT; LVS FTL_0706-1::HimarFT alone (lane 3), with pFTNAT (lane 4), and with pFTNAT-FTL_0706 (lane 5); and FTL_0706-2::HimarFT alone (lane 6), with pFTNAT (lane 7), and with pFTNAT-FTL_0706 (lane 8). (C) Cytopathic analysis with J774A.1 macrophages using crystal violet staining 2 days after infection at an MOI of 100. The retention of the monolayer after infection is compared to the uninfected control (well 1). Wells 2 and 3 contained wild-type LVS without and with pFTNAT. Wells 4 to 6 contained the F. tularensis LVS FTL_0706::HimarFT mutant candidates alone (wells 4), with pFTNAT (wells 5), and with pFTNAT-FTL_0706 (wells 6).

Mutagenesis of F. tularensis Schu S4.FTL_0706 and FTT1238c are highly conserved hypothetical genes having variation in only two nucleotides in F. tularensis LVS and Schu S4, respectively (Fig. 8A). Therefore, we reasoned that introduction of the rescued plasmid carrying FTL_0706::HimarFT could be used to disrupt FTT1238c in Schu S4 (2, 3, 27, 36, 45). Electroporation of the nonreplicative plasmid into Schu S4 resulted in recovery of several Km^r clones, one of which was defective for the synthesis of O antigen like LVS FTL_0706-1::HimarFT. Other clones exhibited a parental O-antigen phenotype and were confirmed to be merodiploids by Southern blotting (Fig. 8B and C). Moreover, nucleotide sequence analysis of the Schu S4 mutants demonstrated that the first of two variable nucleotide differences in the locus was characteristic of LVS.

• Open in new tab
• Download powerpoint

FIG. 8.

Analysis of HimarFT insertions with mutations in FTL_0706 and FTT1238c for F. tularensis LVS and Schu S4, respectively. (A) LVS and Schu S4 wild-type loci are indicated by gray and striped arrows, respectively. HimarFT insertion into FTL_0706 in LVS resulted in the loss of an SpeI site present in the parental chromosomal locus. The sequence for this locus in LVS and Schu S4 is highly conserved, with variation in only two nucleotides, as shown. Introduction of a circularized SpeI fragment containing the HimarFT insertion from LVS FTL_0706-1::HimarFT into Schu S4 promoted homologous recombination resulting in either a single- or double-crossover event as indicated. (B) Western blot analysis of the LVS and Schu S4 strains with a mouse monoclonal anti-F. tularensis LPS antibody demonstrated the presence or absence of the LPS ladder. The strains analyzed were LVS (lane 1), LVS FTL_0706-1::HimarFT (lane 2), Schu S4 (lane 3), Schu S4 FTT1238c::HimarFT merodiploid (lane 4), and Schu S4 FTT1238c::HimarFT (lane 5). (C) Southern blot analysis to assess the presence of both HimarFT and FTL_0706 in the represented strains. SpeI-digested genomic DNA was probed with digoxigenin-labeled fragments of aphA-2 (4-kb signal) or FTL_0706 (with and without HimarFT insertion; 4- and 1-kb signals, respectively). The lane contents are the same as those described above for panel B.