A Francisella tularensis Live Vaccine Strain (LVS) Mutant with a Deletion in capB, Encoding a Putative Capsular Biosynthesis Protein, Is Significantly More Attenuated than LVS yet Induces Potent Protective Immunity in Mice against F. tularensis Challenge

Cell lines, bacteria, and mice.Human macrophage-like THP-1 cells (ATCC TIB-202) were cultured in RPMI 1640 medium containing penicillin (100 μg/ml) and streptomycin (100 U/ml) and supplemented with 10% fetal bovine serum (FBS). F. tularensis LVS and SchuS4 were obtained from the Centers for Disease Control and Prevention (Atlanta, GA). To prepare stocks, we passaged the bacteria once on phorbol-12-myristate-13-acetate (PMA)-differentiated monolayers of THP-1 cells, cultured them on chocolate II agar plates (BD BBL, Sparks, MD) for 3 days, scraped the colonies into sterile saline, resuspended the bacteria in the presence of 20% glycerol, and stored them frozen at −80°C. Before each use in animals, one vial of LVS or SchuS4 was removed from the freezer, immediately thawed in a 37°C water bath, diluted in sterile saline, and kept on ice until use.

Six- to eight-week-old specific-pathogen-free female BALB/c mice were purchased from Charles River Laboratory (Wilmington, MA) and used according to protocols approved by the animal research committees of the University of California—Los Angeles (UCLA) and Colorado State University (CSU).

Construction of an LVS mutant deficient in a putative capsule synthesis gene.An antibiotic-marker-free LVS mutant with a deletion in a putative capsule biosynthesis gene, capB (FTL_1416/FTT0805), LVS ΔcapB, was constructed by allelic exchange. Briefly, approximately 1-kb upstream and downstream DNA fragments flanking the capB coding sequence were amplified from the genomic DNA of F. tularensis LVS. The two PCR fragments were ligated via an EcoRI site to form an in-frame capB deletion cassette. This cassette was cloned into suicide plasmid pMP590 (27), bearing a kanamycin resistance gene and a sacB gene, and subsequently delivered into LVS by electroporation. LVS transformants were subjected to selection on chocolate agar made from GC medium base (BD Difco, Sparks, MD) supplemented with 1% hemoglobin in the presence of 10 μg/ml kanamycin and subsequently on chocolate agar supplemented with 5% sucrose. The resulting LVS ΔcapB mutant does not carry any antibiotic resistance gene. To confirm the deletion of capB from the LVS ΔcapB chromosome, we amplified the mutated capB region by PCR using Phusion high-fidelity DNA polymerase (Finnzyme, Inc., Woburn, MA) and primer pair 5′-CGTAAAAAGGCGTTGGTGAT-3′ and 5′-AGTCCTATCGAAGCCATATTAC-3′. We determined the nucleotide sequences of the amplified mutated capB region and its upstream 218-bp and downstream 370-bp regions and found them to be identical to the corresponding sequences of the LVS genome (www.francisella.org ) but with an in-frame deletion of capB as expected.

A lipopolysaccharide (LPS)-deficient LVS mutant (LVS ΔwbtDEF) was constructed and confirmed by genomic nucleotide sequencing of the mutated locus, using a strategy similar to those described above.

Southern blot analysis of bacterial genomic DNA.LVS and LVS ΔcapB were grown on chocolate agar at 37°C for 2 days. The genomic DNA was isolated by using a genomic DNA isolation kit according to the manufacturer's protocol (Qiagen, Valencia, CA). Bacterial genomic DNAs were subjected to EcoRV, HindIII, or AvaII digestion and separated on a 0.7% agarose gel. The gel was subjected to denaturation in alkaline fixation/transfer buffer (0.4 M NaOH, 1 M NaCl). The denatured DNA in the treated gel was transferred onto a charged nylon membrane (GE Healthcare, Pittsburgh, PA) by capillary transfer and hybridized with biotin-labeled probes at 55°C using the North2South chemiluminescent hybridization and detection kit (Pierce Biotechnology, Rockford, IL). Probes were generated by a North2South biotin random prime labeling kit (Pierce Biotechnology) according to the manufacturer's protocol, using a 1-kb Plus DNA ladder (Invitrogen) and an LVS genomic DNA fragment as templates. The DNA fragment for the probe was amplified from LVS genomic DNA by using primer pair 5′-AGGCGTCGACTGGTGATATCTATTATTATGATG-3′ and 5′-AGAGCGGCCGCTAATACCTAAGATATCATAA-3′ (restriction enzyme sites are underlined). Chemiluminescence signals were detected with a Chemi Doc XRS imaging system (Bio-Rad, Hercules, CA).

Quantitative RT-PCR.Colonies of F. tularensis grown on chocolate agar for 2 days were scraped into Chamberlain defined medium (5), the bacterial suspension was adjusted to an initial optical density at 540 nm (OD₅₄₀) of 0.05, and the bacteria were subcultured aerobically to the mid-logarithmic growth phase (OD₅₄₀ of 0.6). Total RNA from approximately 6 × 10⁸ bacteria of each strain was extracted by using a RiboPure-Bacteria RNA isolation kit (Applied Biosystems/Ambion, Austin, TX) and subsequently treated with DNase according to the manufacturer's guidelines. The primers used for the quantitative PCR (qPCR) assays were designed by using Primer3 software (34). Quantitative reverse transcription (RT)-PCR assays were performed in a two-step reaction. First, cDNA was synthesized by using a high-capacity cDNA reverse transcription kit (Applied Biosystems, Carlsbad, CA) with an iCycler instrument (Bio-Rad) under the following conditions: 25°C for 10 min, 37°C for 120 min, and 85°C for 5 s. For each 20-μl reaction mixture, 2 μl 10× RT buffer, 0.8 μl 100 mM deoxynucleoside triphosphate (dNTP) mix, 2.0 μl random primers, 1.0 μl MultiScribe reverse transcriptase, 1.0 μl of RNase inhibitor, 3.2 μl of nuclease-free water, and 10 μl of RNA template were added. After cDNA synthesis, qPCR assays were performed by using an ABI 7900 sequence detection system (Applied Biosystems). For each 25-μl qPCR mixture, 12.5 μl 2× Power SYBR green PCR master mix (Applied Biosystems), 300 nM each primer pair, and 2 μl of cDNA were added. Primer pairs used were 5′-CGGGTCATACCCAGCTTTAG-3′ and 5′-GCTCCTGGAGTCATCCCATA-3′ for capA, 5′-TCGAGTGTATGGCTTTGCAG-3′ and 5′-AGCCTCGGTCTTTACAAGCA-3′ for capB, 5′-ATGGGGGCTCCTGATAGAGT-3′ and 5′-CCAATTACCTGAACTTGGTTTGA-3′ for capC, 5′-CGTAAAAAGGCGTTGGTGAT-3′ and 5′-ATAAAGCCACGACGATTTGG-3′ for FTL_1417, and 5′-CTGTCGTCAGCTCGTGTTGT-3′ and 5′-CGTAAGGGCCATGATGACTT-3′ for FTL_R0003 (16S rRNA). The amplification of cDNA with each pair of primers was performed under the following conditions: 1 cycle at 95°C for 10 min and 40 cycles at 95°C for 15 s and 60°C for 1 min. The 16S rRNA gene was used as an internal control.

Data collection was performed by using ABI Sequence Detection 1.3 software (Applied Biosystems). The expression levels of the target genes (capA, capB, capC, and FTL_1417) were normalized to levels of 16S rRNA and compared with LVS (baseline) by using the 2^−ΔΔC_T method (Applied Biosystems User Bulletin no. 2 [pN4303859]) for the relative quantitation of gene expression. Real-time data are presented as the fold change compared with LVS levels at the mid-logarithmic growth phase. Error bars represent the standard deviations of the ΔΔC_T values. Statistical significance was determined by the Student t test.

Complement sensitivity assay.LVS ΔcapB and the parental LVS were grown on chocolate agar overnight, suspended in RPMI medium, incubated with either 10% fresh normal human serum (type AB) or heat-inactivated human serum at 37°C for 10 min, serially diluted, and plated onto chocolate agar plates.

Competition assay for growth in human macrophage-like THP-1 cells.To monitor the intramacrophage growth of an individual bacterial strain in a competitive assay, we introduced a pair of plasmids, pMP607 and pMP633, which harbor the aphA-1 gene and the hyg gene, respectively (27), into the parental LVS and the capB deletion strain. This resulted in an LVS strain carrying a kanamycin resistance gene (LVS-kan) or a hygromycin resistance gene (LVS-hyg) and a comparable pair of LVS ΔcapB strains, each carrying one of the same antibiotic resistance genes (LVS ΔcapB-kan and LVS ΔcapB-hyg). Bacteria were grown on chocolate agar for 2 days, scraped into RPMI medium, and opsonized with 10% human serum (type AB). Human monocytic THP-1 cells were differentiated by treatment with 100 nM PMA for 3 days at a density of 3 × 10⁵ cells/ml. Bacteria were added to the differentiated THP-1 cells at a multiplicity of infection (MOI) of 10:1 (bacteria to cells), and the cultures were centrifuged at 1,000 × g at 4°C for 30 min and subsequently incubated at 37°C for 60 min to allow bacterial uptake. Human macrophage-like THP-1 cells coinfected with LVS ΔcapB-hyg and LVS-kan or coinfected with the same two strains carrying the opposite resistance markers (LVS ΔcapB-kan and LVS-hyg) were washed three times to remove extracellular bacteria and cultivated in Dulbecco's modified Eagle's medium (DMEM) containing 10% heat-inactivated FBS and 0.1 μg/ml gentamicin, a culture condition that we have determined suppresses extracellular bacterial growth but has a minimal impact on the growth of the intracellular bacteria. At various times postinfection, monolayers were scraped into homogenization buffer containing 250 mM sucrose and 20 mM HEPES (pH 7.2) and lysed with glass beads by vortexing 10 times in 2-s pulses. The lysates were serially diluted and plated onto chocolate agar containing 10 μg/ml kanamycin or 250 μg/ml hygromycin as appropriate. CFU were enumerated, and the ratio between the numbers of LVS ΔcapB CFU and LVS CFU was calculated at each time point.

Analysis of bacterial LPS and protein expression by Coomassie blue staining and Western blotting.Bacterial lysates were separated by 15% SDS-PAGE and transblotted onto nitrocellulose membranes. The membrane was incubated with a monoclonal antibody specific to F. tularensis LPS (FB11; Abcam, Inc., MA) at a dilution of 1:5,000 and subsequently with horseradish peroxidase-conjugated goat anti-mouse immunoglobulin G (Bio-Rad, Hercules, CA) at a dilution of 1:25,000. Signals on the blot were developed by using the Supersignal West Pico chemiluminescent substrate (Pierce, Rockford, IL). Proteins in the bacterial lysates were visualized by Coomassie blue staining as a control for equal loading.

Complementation of truncated capB.The DNA fragment encoding wild-type capB was PCR amplified from LVS genomic DNA by using Phusion high-fidelity DNA polymerase (Finnzyme) and primer pair 5′-CTCATATGTTGGATTTTTGGTTAATTGT-3′ and 5′-ACAGGATCCCTATAAATTATGCTTCTTTT-3′ (the restriction enzyme sites are underlined). The PCR product was cloned into pZErO (Invitrogen, Carlsbad, CA) and verified by nucleotide sequencing. The confirmed capB coding sequence was subcloned into the NdeI and BamHI sites downstream of the groE promoter in the shuttle vector pFNLTP6/gro-gfp (28), in which the gfp coding sequence was replaced by the capB coding sequence. The resultant plasmid, pFNLTP/gro-capB, was electroporated into LVS ΔcapB. Transformants were selected on chocolate agar supplemented with 7.5 μg/ml kanamycin and verified by colony PCR using a primer pair specific to the vector: 5′-GTATTAGTTCGTCGTGC-3′ and 5′-GAGCCATCTTTATTGCG-3′. One of the verified clones, LVS ΔcapB/CapB, was amplified on chocolate agar supplemented with 7.5 μg/ml kanamycin and used for further characterization studies.

Vaccination and challenge of mice.Vaccination and challenge of mice were conducted as described previously (23). Briefly, mice were vaccinated by the i.d. or intranasal (i.n.) route. For i.d. vaccination, mice were shaved, decontaminated with 70% ethanol at the base of the tail, and injected by using 27-gauge needles and 1/2-ml tuberculin syringes (Becton Dickinson, NJ) with 50 μl sterile phosphate-buffered saline (PBS) (negative control) or with various doses of LVS (positive control) or LVS ΔcapB diluted in 50 μl PBS. For i.n. vaccination, mice were anesthetized by intraperitoneal (i.p.) injection with ketamine (80 mg/kg of body weight) and xylazine (10 mg/kg), and vaccines serially diluted in sterile PBS were administered into the nostrils in a total volume of 20 μl. At various times after vaccination, mice were either euthanized for immunology studies or challenged by the i.n. route at UCLA with 4,000 CFU LVS (>5× the 50% lethal dose [LD₅₀]), administered in the same way for i.n. vaccination, or by the aerosol route at CSU with 10× the LD₅₀ of type A F. tularensis strain SchuS4. The aerosol challenge was conducted in a chamber of 5 cubic feet with the mice conscious and active, using a Glas-Col inhalation exposure system (Glas-Col LLC, Terre Haute, IN). The dose of 10× the LD₅₀ for strain SchuS4 was obtained by aerosolizing 5 ml of a suspension containing 3.2 × 10⁷ to 3.4 × 10⁷ CFU/ml of bacteria over a period of 15 min. The actual number of bacteria in the nebulizer was confirmed by culturing the bacterial suspension in duplicate on chocolate agar. Challenged mice were weighed and monitored for illness and death for 3 weeks. Mice that met predetermined humane endpoints for euthanasia were euthanized and counted as a death. The mean survival time was calculated by dividing the sum of the surviving days of all mice by the total number of mice examined, with animals surviving until the end of the experiment given a survival time of 21 days, when the experiment was terminated.

To quantitate bacterial burden in host tissues, mice were euthanized at 5 days postinfection. Liver, spleen, lung, and, in the case of i.d. immunization, a 1-cm² area of local skin at the injection site were removed aseptically and immediately weighed, immersed in 2 ml sterile PBS, and kept on ice until homogenization. The homogenization was performed by using a Pro200 homogenizer (Pro Scientific, Inc., Oxford, CT). Serial 10-fold dilutions of the homogenates were plated onto cysteine heart agar supplemented with 1% (wt/vol) hemoglobin and sulfamethoxazole (40 μg/ml), trimethoprim (8 μg/ml), and erythromycin (50 μg/ml) (24) and incubated in a 5% CO₂ atmosphere at 37°C for 3 days before colonies were enumerated (23). The supplementation of sulfamethoxazole, trimethoprim, and erythromycin in the agar was done to prevent contamination from the mouse tissues. We have determined that the doses of these three antibiotics used in our experiments have a minimal inhibitory effect on the growth of LVS.

Assay for lymphocyte proliferation.Groups of four BALB/c mice were sham immunized or immunized i.d. with 1 × 10⁵ CFU LVS or 1 × 10⁶ CFU LVS ΔcapB or immunized i.n. with 112 CFU LVS or 1 × 10⁵ CFU LVS ΔcapB. At 4 weeks postimmunization, mice were anesthetized by the i.p. injection of ketamine and xylazine, bled, and then euthanized. Sera were collected for serum antibody measurements as described below. Spleens were removed, and a single-cell suspension of splenocytes was prepared. Lymphocyte proliferation in response to heat-inactivated LVS antigen was assayed as described previously, with modifications (23). Splenocytes from each animal were cultured in triplicate and stimulated with heat-inactivated LVS for 48 h. Two hours before harvesting, the stimulated splenocytes were pulsed with [methyl-³H]thymidine and harvested with a Combi cell harvester (Molecular Devices/Skatron, Sunnyvale, CA). The amount of incorporated [³H]thymidine was determined by counting with an LS 6500 scintillation system (Beckman Coulter, Fullerton, CA) and the values were expressed as counts per minute (cpm).

Heat-inactivated LVS antigen was prepared by growing LVS on chocolate agar for 2 days, scraping the bacteria into PBS, and incubation at 80°C for 1 h. The loss of viability was confirmed by plating the antigen onto chocolate agar and incubation at 37°C for 2 days. No colonies were detected on the plates, indicating that no viable bacteria survived heat inactivation.

Serum antibody detection by enzyme-linked immunosorbent assay (ELISA).Sera collected from sham-immunized mice or mice immunized with LVS or LVS ΔcapB were analyzed for levels of IgM, IgA, IgG, IgG1, and IgG2a antibodies specific for LVS as described previously (23). The heat-inactivated LVS antigen was diluted in carbonate-bicarbonate buffer (50 mM NaHCO₃, 50 mM Na₂CO₃) to an optical density of 0.025 at 540 nm (equivalent to 5 × 10⁶ CFU/0.1 ml), and 0.1 ml was used to coat 96-well high-binding-capacity plates (Corning, NY) for 2 h at 37°C. Excess antigen was removed by washing three times with PBS. Sera at a starting dilution of 1:32 were diluted further through a 2-fold series with PBS containing 1% bovine serum albumin. The diluted sera were incubated with heat-inactivated LVS coated onto 96-well plates at an ambient temperature for 3 h. The plates were subsequently incubated for 90 min at an ambient temperature with alkaline phosphatase-conjugated goat anti-mouse IgA (Sigma, St. Louis, MO), IgM, IgG1, or IgG2a (Invitrogen) at a dilution of 1:1,000. The plates were washed three times with PBS containing 0.05% Tween 20 after each incubation. One hundred microliters of p-nitrophenyl phosphate substrate in diethanolamine buffer (phosphatase substrate kit; Bio-Rad, Hercules, CA) was added to each well. The yellow color that developed was read at 414 nm for absorbance by using a Multiscan microplate reader (TiterTek, Huntsville, AL). The endpoint antibody titer was calculated as the reciprocal of the highest serum dilution that was a minimum of 0.05 optical density units above that of the sham-immunized control serum plus 2 standard errors (SE).

Statistical analysis.One-way analysis of variance (ANOVA) with Tukey's posttest or two-way ANOVA with Bonferroni's posttest was performed by using GraphPad (San Diego, CA) Prism software, version 5.01, to determine significance in comparisons of mean CPMs for lymphocyte proliferative responses, mean antibody titers, and mean log₁₀ CFU organ counts among mice in vaccinated and control groups. A log-rank analysis (Mantel-Cox test) using GraphPad Prism, version 5.01, was used to determine the significance of survival curves among mice in immunized and in sham-immunized control groups.

Construction and in vitro characterization of LVS ΔcapB, an LVS mutant deficient in a putative capsule synthesis gene.We used allelic exchange to construct a putative capsule-deficient mutant (LVS ΔcapB) in which the capB gene (FTL_1416/FTT0805) was partially deleted without interrupting its neighboring genes. The resultant LVS ΔcapB mutant is free of an antibiotic resistance marker, and the truncated capB contains 30 nucleotides encoding 6 amino acids from the N terminus and 4 amino acids from the C terminus of CapB (Fig. 1 A), as confirmed by nucleotide sequencing of the mutated genomic region. To confirm further the genomic structure of the capB locus, we performed Southern blotting on the genomic DNA of LVS ΔcapB and the parental LVS isolated from bacteria grown on chocolate agar for 2 days. The genomic DNA was treated with EcoRV, HindIII, or AvaII; separated by electrophoresis on a 0.7% agarose gel; blotted; and probed with a biotin-labeled 3.4-kb genomic DNA fragment spanning capB, most of the upstream region of FTL_1417, the entire downstream region of capC, and most of the further downstream region of capA. As shown at the top of Fig. 1B for LVS, the probe is predicted to hybridize with two DNA fragments of 1.6 and 1.7 kb in EcoRV digests; six fragments of 5.8, 0.8, 0.5, 1.0, 0.3, and 1.5 kb in HindIII digests; and three fragments of 1.6, 1.2, and 3.1 kb in AvaII digests. As shown at the bottom of Fig. 1B for LVS ΔcapB, due to the loss of capB, the same probe is predicted to hybridize with one fragment of 2.2 kb in EcoRV digests; four fragments of 5.8, 1.1, 0.3, and 1.5 kb in HindIII digests; and two fragments of 1.7 and 3.1 kb in AvaII digests. As shown in Fig. 1C, the probe detected each of the predicted DNA fragments in EcoRV-, HindIII-, or AvaII-treated genomic DNA of LVS and LVS ΔcapB. These results further confirmed that capB was deleted from the LVS ΔcapB genome and that there were no additional DNA rearrangements in the chromosomal region where capB resides.

• Open in new tab
• Download powerpoint

FIG. 1.

Construction and characterization of LVS ΔcapB. (A) Construction of LVS ΔcapB. LVS ΔcapB was constructed by allelic exchange between LVS (top) and a plasmid carrying an exchange cassette containing truncated capB (FTL_1416/FTT0805) and its flanking nucleotide sequences, including the upstream region of FTL_1417 (FTL_1417′) and the downstream region of capC and capA (capCA′) (middle). The resultant LVS ΔcapB strain (bottom) retains 6 amino acids from the N terminus and 4 amino acids from the C terminus of CapB and contains no antibiotic resistance marker (unmarked). (B) Schematic representations of the structure of the capB locus and its upstream and downstream regions in the genomes of LVS (top) and LVS ΔcapB (bottom), the position of a probe spanning the EcoRV restriction site located in FTL_1417 and the EcoRV site located in capA, and the predicted EcoRV, HindIII, and AvaII restriction fragments (in kilobases) detected by the probe by Southern blot analysis. A, restriction enzyme AvaII; E, EcoRV; H, HindIII. (C) Southern blot analysis of LVS ΔcapB (Δ) and the parental LVS genomic DNA. The genomic DNAs of LVS (lanes 1, 3, and 5, indicated below the panels) and LVS ΔcapB (lanes 2, 4, and 6) were isolated from bacteria cultured on chocolate agar for 2 days. The DNA was digested with EcoRV, HindIII, or AvaII, as indicated at the top of each panel; electrophoresed with a DNA ladder (L); blotted; and hybridized with the biotin-labeled probe as shown in B and the biotin-labeled DNA ladder. The left of each panel shows a 1-kb Plus DNA ladder, and to the right of each panel are the fragments resulting from restriction enzyme digestion of the bacterial DNA detected by the probe. (D and E) Quantitative RT-PCR analysis of capB and its neighboring genes. Total RNA was isolated from LVS, LVS ΔcapB, and the capB-complemented strain LVS ΔcapB/CapB; reverse transcribed in the presence (+) or absence (−) of reverse transcriptase; and amplified by using primer pairs specific to capB, its upstream region of FTL_1417, and downstream regions of capC and capA. 16S rRNA (16S) was used as an internal control. Transcript levels are presented as the mean fold changes versus LVS ± SE. ***, P < 0.001 by two-way ANOVA. qRT-PCR products were analyzed on a 2% agarose gel. The predicted sizes of the qRT-PCR product for capA, capB, capC, FTL_1417, and 16S rRNA are 231, 231, 212, 209, and 156 bp, respectively. The sizes of the DNA in the DNA ladder are shown at the left of each panel in E. F. tularensis strains from which the RNA was isolated are shown at the right of each panel in E.

To determine whether the deletion of capB affects the transcription of its upstream and downstream genes, we isolated total RNA of LVS ΔcapB, the parental LVS, and capB-complemented LVS ΔcapB (LVS ΔcapB/CapB) from mid-logarithmic-phase cultures in Chamberlain defined medium and treated them with DNase to eliminate possible genomic DNA contamination. We then analyzed the transcript levels of FTL_1417, capB, capC, capA, and the 16S rRNA gene (serving as an internal control) in all three strains by using a two-step qPCR assay. We compared the fold changes of these transcripts between LVS ΔcapB and LVS and between LVS ΔcapB/CapB and LVS. As shown in Fig. 1D, we found that (i) the transcript of capB was not detectable in LVS ΔcapB; (ii) the transcript level of capB in LVS ΔcapB/CapB was significantly (∼16-fold) higher than that in LVS, indicating that multiple copies of pFNLTP/gro-capB likely reside in LVS ΔcapB/CapB; and (iii) there were no significant changes in the transcript levels of capA and capC between LVS ΔcapB and LVS or between LVS ΔcapB/CapB and LVS. In addition, we verified that there was no genomic DNA contamination in each of the RNA samples, since in the absence of reverse transcription, no PCR product was generated in each of the samples (Fig. 1E). These results demonstrate that the deletion of capB minimally affected the transcript levels of its upstream and downstream genes FTL_1417, capC, and capA.

To determine whether LVS ΔcapB produces LPS, we analyzed bacterial lysates of LVS and LVS ΔcapB by Coomassie blue staining and Western blotting. We observed similar profiles of protein expression (Fig. 2 A, right) and similar ladder-like patterns of LPS immunoreactivity (Fig. 2A, left) for LVS and LVS ΔcapB. To assess serum sensitivity, we incubated the mutant and the parental LVS in fresh human serum. We observed that LVS ΔcapB, like the parental LVS, is not sensitive to serum-mediated killing (Fig. 2B). In contrast, an LPS-deficient LVS mutant (LVS ΔwbtDEF) was readily killed by complement in human serum (Fig. 2B).

• Open in new tab
• Download powerpoint

FIG. 2.

In vitro characterization of LVS ΔcapB. (A) Analysis of LVS and LVS ΔcapB LPS and proteins. (Left) LPS was analyzed by immunoblotting using a monoclonal antibody to LPS (note the characteristic LPS ladder pattern). (Right) Proteins were evaluated by Coomassie blue staining to provide a loading control. (B) Complement sensitivity assay. Each bacterial strain was incubated with either fresh human AB serum (ABS) or heat-inactivated AB serum (HI-ABS) for 10 min and then assayed for CFU on chocolate agar. This experiment was performed three times, with similar results. (C) Competition for growth in human macrophage-like THP-1 cells. THP-1 cells were coinfected with LVS ΔcapB carrying a hygromycin resistance gene (ΔcapB-h) and with LVS carrying a kanamycin resistance gene (LVS-k) (solid circles, left vertical axis) or with the same two strains carrying the opposite resistance markers (solid squares, right vertical axis). At 0, 7, and 22 h postinfection, the cell monolayer was lysed, serial dilutions of the lysate were plated onto chocolate agar supplemented with either kanamycin or hygromycin, CFU were enumerated, and the ratio between the numbers of LVS ΔcapB and LVS CFU was calculated. The ratios of LVS ΔcapB to LVS CFU on the left and right vertical axes differ because of the variability among the four strains in the numbers of CFU recovered from THP-1 monolayers at the start of the experiment. For the competition data shown on the left axis, the initial number of LVS ΔcapB organisms was greater than the initial number of LVS organisms (initial ratio of ∼10:1), and for the competition data shown on the right axis, the initial number of LVS ΔcapB organisms was fewer than the initial number of LVS organisms (initial ratio of ∼0.4:1). In both cases, LVS ΔcapB was outcompeted by LVS, and therefore, the ratio of LVS ΔcapB CFU to LVS CFU declined with time in culture. This experiment was performed twice, with similar results.

To evaluate the attenuation of LVS ΔcapB in vitro, we compared the intramacrophage growth of LVS ΔcapB with that of LVS in human macrophage-like THP-1 cells in a growth competition assay. We found that the LVS ΔcapB mutant carrying either a hygromycin or kanamycin resistance marker grew more slowly than the parental LVS carrying the opposite antibiotic resistance marker (kanamycin or hygromycin, respectively) (Fig. 2C). This result indicated that the LVS ΔcapB mutant is more attenuated than the parental LVS in human macrophage-like THP-1 cells.

LVS ΔcapB is more attenuated than parental LVS in mice.To compare the virulence of LVS ΔcapB with that of the parental LVS in vivo (Fig. 3 A1 and A2), we immunized BALB/c mice (four mice per group) with LVS ΔcapB at doses ranging from 1 × 10⁶ to 1 × 10⁸ CFU i.d. or from 1 × 10³ to 1 × 10⁷ CFU i.n. Mice immunized with LVS at doses ranging from 1 × 10⁵ to 1 × 10⁷ CFU i.d. or 1.5 × 10² to 4.1 × 10³ CFU i.n. and nonimmunized mice served as controls. The mice were weighed daily from day 2 postchallenge and observed for signs of illness, death, and, for mice immunized i.d., the extent of a local lesion for 3 weeks. As shown in Fig. 3A1, for mice administered LVS i.d., 100% of mice survived after immunization at the doses tested in this experiment. However, mice immunized i.d. with 1 × 10⁷ CFU LVS developed severe local lesions. In an experiment performed separately, we immunized BALB/c mice (nine mice per group) i.d. with 1 × 10⁴ to 1 × 10⁸ CFU LVS/mouse and monitored for 2 weeks. One hundred percent of mice survived after i.d. immunization with 1 × 10⁴ to 1 × 10⁷ CFU LVS, consistent with the results shown in Fig. 3A1. However, all nine mice immunized with 1 ×10⁸ CFU of LVS died on days 4 to 6 postimmunization, with a mean survival time of 4.25 days. In contrast, for mice administered LVS ΔcapB i.d., 100% of the mice survived doses of 10⁶ to 10⁸ CFU (versus 0% survival at a dose of 10⁸ CFU for LVS). Although mice immunized with 10⁸ CFU LVS ΔcapB i.d. had early weight loss, these mice had no other signs of illness and showed only mild local inflammation at the injection site (Fig. 3A1).

• Open in new tab
• Download powerpoint

FIG. 3.

LVS ΔcapB is more attenuated than parental LVS and complemented LVS ΔcapB/CapB in mice. Groups of four BALB/c mice were nonimmunized, immunized with PBS (sham), or immunized i.n. or i.d. with LVS, LVS ΔcapB, or LVS ΔcapB/CapB at the indicated doses (CFU) and monitored for survival, weight change, local lesions (after i.d. immunization), and other signs of illness for 2 or 3 weeks. (A1 and A2) Mice were weighed daily from days 2 to 9 postimmunization and monitored for survival for 3 weeks. (B1 to B3) Mice were weighed daily from day 0 (the day of immunization) to day 7 postimmunization and monitored for survival for 2 weeks. Values are means ± SE. RF, ruffled fur; HB, hunched back; SE, swollen eyes.

For mice administered LVS i.n., deaths occurred and at relatively low doses such that the LD₅₀ i.n. was ∼700 CFU (versus >10⁷ CFU for LVS ΔcapB). All mice immunized i.n. with LVS at each dose tested showed signs of illness (ruffled fur and hunched back), and they had significant weight loss between days 4 and 9 after immunization (Fig. 3A2). In contrast, for mice immunized i.n. with LVS ΔcapB, 100% of mice survived after immunization at all tested doses, and only transient weight loss occurred between day 2 and day 6 postimmunization, at doses of 1 × 10⁶ CFU and 1 × 10⁷ CFU/mouse (Fig. 3A2).

To evaluate further the safety of LVS ΔcapB and to confirm that the attenuated phenotype of LVS ΔcapB was due to the loss of only capB, we constructed a capB-complemented strain, LVS ΔcapB/CapB, and compared its virulence with those of LVS ΔcapB and the parental LVS via the i.n. route, the route by which LVS is most virulent. BALB/c mice in groups of four were immunized with LVS (6.7 × 10¹ to 6.7 × 10⁴ CFU), LVS ΔcapB (10⁴ to 10⁷ CFU), or LVS ΔcapB/CapB (10⁴ to 10⁷ CFU) (Fig. 3B1 to B3). Mice were monitored and weighed for 2 weeks daily starting from the day of immunization (day 0). As shown in Fig. 3B1 to B3, mice immunized with LVS at each of the tested doses had significant weight loss and an LD₅₀ of approximately 700 CFU (Fig. 3B1), and mice immunized with LVS ΔcapB had 100% survival at each of the tested doses, with transient weight loss in mice immunized with 10⁶ or 10⁷ CFU (Fig. 3B2). In contrast to LVS ΔcapB-immunized mice, mice immunized with LVS ΔcapB/CapB at doses ranging from 10⁵ to 10⁷ CFU showed significant weight loss and death. Thus, the attenuated phenotype of LVS ΔcapB was partially restored by providing CapB in trans. Taken together with the results from genomic nucleotide sequencing, Southern blotting, the quantitative RT-PCR assay, and the in vivo virulence test, the attenuated phenotype of LVS ΔcapB was due to the loss of capB.

In subsequent experiments evaluating the immunogenicity and efficacy of the i.d. route of immunization, we generally used doses of 10⁶ CFU of LVS ΔcapB and 10⁵ CFU of LVS. At these doses, LVS ΔcapB was nontoxic in mice and caused no apparent inflammation at the site of injection. In contrast, LVS retained significant toxicity, e.g., causing ruffled fur. In subsequent experiments evaluating the i.n. route of immunization, we generally used doses of 10⁵ CFU LVS ΔcapB and approximately 200 CFU LVS. Again, at these doses, LVS ΔcapB was completely nontoxic, whereas LVS retained significant toxicity, resulting in ruffled fur and a hunched back in immunized animals; moreover, in two experiments, even this low dose of LVS proved to be lethal to 25% of immunized animals.

These results demonstrated that LVS ΔcapB is at least 10,000-fold-more attenuated than LVS after i.n. immunization and substantially more attenuated after i.d. immunization.

LVS ΔcapB is cleared faster than LVS after intranasal infection.To assess the dissemination and clearance of LVS ΔcapB, we immunized groups of four BALB/c mice i.n. with 112 CFU LVS or 1 × 10⁵ CFU LVS ΔcapB or i.d. with 1 × 10⁵ CFU LVS or 1 × 10⁶ CFU LVS ΔcapB. At various times postimmunization, mice were euthanized, and spleen, liver, lung, and skin at the injection site (in the case of i.d. immunization) were assayed for bacterial burden. As shown in Fig. 4 A, following i.n. immunization, LVS replication peaked at day 4 in the lung and at day 7 in the spleen and liver; the LVS was cleared by day 22 postimmunization in all three organs examined. In contrast, although the LVS ΔcapB mutant was administered at a much higher dose i.n. (10⁵ CFU versus 112 CFU for LVS), LVS ΔcapB replicated at levels significantly lower than those of LVS at the peak of infection, and it was cleared faster than LVS in all three organs. Following i.d. immunization (Fig. 4B), LVS ΔcapB replicated less efficiently in the local injection area than did the parental LVS. However, LVS ΔcapB spread systemically to the liver and spleen, where it was cleared at a rate similar to or greater than that of the parental LVS. The replication of both LVS ΔcapB and its parental LVS in the lung was limited (Fig. 4B).

• Open in new tab
• Download powerpoint

FIG. 4.

Dissemination and clearance of LVS ΔcapB in mice. Groups of four BALB/c mice were immunized i.n. with 112 CFU LVS or 1 × 10⁵ CFU LVS ΔcapB or i.d. with 1 × 10⁵ CFU LVS or 1 × 10⁶ CFU LVS ΔcapB. At the indicated times postimmunization, mice were euthanized, and numbers of CFU in the spleen, liver, lung, and skin at the site of injection (after i.d. immunization) were assayed. LVS ΔcapB is cleared faster than LVS in spleen, liver, and lung after i.n. immunization and at the site of injection in the skin after i.d. immunization. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (by ANOVA).

These results demonstrated that LVS ΔcapB is cleared much faster than LVS in mice after i.n. inoculation and at least as fast as LVS after i.d. inoculation.

Immunization with LVS ΔcapB induces strong cell-mediated and humoral immune responses.To examine whether immunization with LVS ΔcapB induces cell-mediated and humoral immune responses, we immunized groups of four mice with LVS ΔcapB at doses of 1 × 10⁵ CFU by the i.n. route or 1 × 10⁶ CFU by the i.d. route, doses that were nontoxic (Fig. 3) and that provided excellent protection against i.n. LVS challenge (see below). At 4 weeks postimmunization, mice were anesthetized, bled, and euthanized (Fig. 5 A), and cell-mediated and humoral immune responses were assayed as described in Materials and Methods. With respect to cell-mediated immunity, immunization with LVS ΔcapB i.n. or i.d. induced LVS-specific splenic lymphocyte proliferative responses that were significantly higher than those of sham-immunized mice and comparable to those of LVS-immunized mice (Fig. 6). With respect to humoral immune responses (Fig. 7), following i.n. immunization, LVS ΔcapB induced significantly higher levels of IgG, IgG2a, and IgM antibodies than LVS by 7 days postimmunization (Fig. 7A, C, and E). At 14 days after immunization, mice immunized i.n. with LVS had a transiently significantly elevated IgA antibody level (Fig. 7D). After 14 days postimmunization, there were no significant differences in IgG, IgG1, IgG2a, IgA, and IgM antibody levels between mice immunized with LVS and those immunized with LVS ΔcapB via the same route (i.e., either the i.n. or i.d. route) (Fig. 7). Interestingly, by 4 days after immunization, both LVS ΔcapB and LVS induced significantly elevated titers for IgG, IgG2a, and IgM by the i.d. route and IgM by the i.n. route compared with titers on day 1 postimmunization.

• Open in new tab
• Download powerpoint

FIG. 5.

Immunization and challenge protocols. Groups of BALB/c mice were immunized i.n. or i.d. with LVS ΔcapB or LVS. Mice immunized with PBS (sham) or LVS served as controls. At the indicated times postimmunization, mice were either euthanized for immunology studies (A), challenged with approximately 4,000 CFU LVS (>5× the LD₅₀) by the i.n. route (B and C), or challenged with 10× the LD₅₀ of type A F. tularensis strain SchuS4 by the aerosol route (D). For assessments of organ bacterial burden, four or eight mice per group challenged with LVS i.n. were euthanized at 5 days postchallenge, and the organs were homogenized and assayed for CFU of LVS (B). For assessments of survival, eight mice per group challenged with LVS i.n. or SchuS4 by aerosol were monitored for signs of illness and death for 3 weeks after challenge (C and D). LPA, lymphocyte proliferation assay.

• Open in new tab
• Download powerpoint

FIG. 6.

Immunization with LVS ΔcapB induces cell-mediated immune responses comparable to those induced by LVS. Groups of four BALB/c mice were immunized i.n. with 112 CFU LVS or 1 × 10⁵ CFU LVS ΔcapB or i.d. with 1 × 10⁵ CFU LVS or 1 × 10⁶ CFU LVS ΔcapB. At 4 weeks postimmunization, mice were euthanized, and single-cell suspensions of splenocytes were prepared. A total of 1 × 10⁶ splenocytes were incubated with 2 × 10⁵ or 2 × 10⁶ CFU heat-inactivated (HI) LVS as indicated, and splenic lymphocyte proliferation was assayed. The lines above the bars indicate a statistical comparison between the bar beneath the left end of the line and the bar beneath the right end of the line. Only comparisons where differences are statistically significant are shown. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (by two-way ANOVA).

• Open in new tab
• Download powerpoint

FIG. 7.

Immunization with LVS ΔcapB induces potent antibody responses. Groups of four BALB/c mice were immunized i.n. with 112 CFU LVS or 1 × 10⁵ CFU LVS ΔcapB or i.d. with 1 × 10⁵ CFU LVS or 1 × 10⁶ CFU LVS ΔcapB. At the indicated times postimmunization, mice were euthanized, and serum was collected and analyzed for IgG, IgA, IgM, IgG1, and IgG2 antibodies specific to heat-inactivated LVS. The antibody level was calculated as the log₁₀ of the reciprocal of the endpoint dilution of the test serum. Data represent means ± SE. The lines above the bars indicate a statistical comparison between the bar beneath the left end of the line and the bar beneath the right end of the line. Only comparisons where differences were statistically significant between LVS and LVS ΔcapB by the same route are shown. *, P < 0.05; ***, P < 0.001 (by two-way ANOVA).

Thus, LVS ΔcapB, while highly attenuated, induces cellular and humoral immune responses comparable to those of LVS after i.n. or i.d. immunization.

Immunization with LVS ΔcapB protects mice from lethal intranasal challenge with LVS.To examine the capacity of LVS ΔcapB to induce protective immunity against lethal challenge with F. tularensis, we immunized mice i.n. or i.d with LVS ΔcapB at various doses (Fig. 5B and C and 8). Sham-immunized mice and mice immunized with LVS (150 CFU i.n. and 1 × 10⁵ or 1 × 10⁶ CFU i.d.) served as controls. At 4 weeks postimmunization, mice were challenged i.n. with 4,000 CFU LVS (>5× the LD₅₀). Five days after challenge, approximately the peak time of bacterial growth in the host with an LVS i.n. challenge dose of 4,000 CFU, mice were euthanized, and the liver, spleen, and lung were harvested and assayed for bacterial burden (Fig. 5B). As shown in Fig. 8, in mice immunized with LVS ΔcapB i.n. or i.d., the bacterial burden in the spleen, liver, and lung was significantly (3 to 4 logs) lower than that in sham-immunized animals. In the spleen of mice immunized i.n. with various doses of LVS ΔcapB or 150 CFU of the parental LVS, there were no significant differences in bacterial burdens between mice immunized i.n. with LVS and mice immunized with various doses of LVS ΔcapB, except for the dose of 10³ CFU (P < 0.01) (Fig. 8A). In the spleens of mice immunized i.d., there were no significant differences in F. tularensis LVS burden between mice immunized with 10⁶ to 10⁸ CFU of LVS ΔcapB and mice immunized with 10⁵ to 10⁶ CFU of LVS (Fig. 8B). In the livers of mice immunized i.n. or i.d., there were no significant differences in bacterial burden between mice immunized with LVS ΔcapB and mice immunized with LVS (Fig. 8C and D). In the lungs of mice immunized i.n., there were no significant differences in the bacterial burden between mice immunized with LVS and mice immunized with various doses of LVS ΔcapB, except for the dose of 10³ CFU (P < 0.01) (Fig. 8E). In the lungs of mice immunized i.d. with LVS ΔcapB and the parental LVS, although the bacterial burden in mice immunized with 10⁶ or 10⁷ CFU LVS ΔcapB was significantly higher (P < 0.01 and P < 0.05, respectively) than that in mice immunized with 10⁵ CFU LVS, the bacterial burden in these LVS ΔcapB-immunized mice was comparable to that of mice immunized with 10⁶ CFU LVS (Fig. 8F).

• Open in new tab
• Download powerpoint

FIG. 8.

Immunization with LVS ΔcapB induces protective immunity against F. tularensis LVS i.n. challenge in a dose-dependent manner. BALB/c mice (four mice per group) were immunized i.n. (left) or i.d. (right) with LVS ΔcapB at the doses indicated on the horizontal axes. Mice immunized with PBS (sham) (eight per group) or LVS (four per group) served as controls. Four weeks later, the mice were challenged i.n. with 4,000 CFU LVS. At 5 days postchallenge, the spleen (A and B), liver (C and D), and lung (E and F) were removed and assayed for bacterial burden. Symbols represent each animal in a group. CFU data were compared by one-way ANOVA (Prism 5.01 software) with Bonferroni's posttest. Dashed lines indicate the limit of detection. Differences in CFU between the sham group and each of the groups vaccinated with either LVS or LVS ΔcapB were statistically significant (P < 0.001). The lines above the bars indicate a statistical comparison between the bar beneath the left end of the line and the bar beneath the right end of the line. Only comparisons between the LVS-immunized group and the LVS ΔcapB-immunized group where differences were statistically significant are shown.

Thus, the protection induced by immunization with LVS ΔcapB was dose dependent. Among all the doses of LVS ΔcapB tested, we found that doses of 1 × 10⁵ CFU i.n. and 1 × 10⁶ CFU i.d. were both well tolerated and induced protection in mice against F. tularensis challenge comparable to that of LVS. Therefore, in subsequent challenge experiments and the experiments described above in which immune responses were evaluated, we used these doses to immunize mice.

The above-described results indicated that in mice immunized with LVS ΔcapB and then challenged with F. tularensis LVS, both local replication in the lung and systemic dissemination of F. tularensis are strongly inhibited.

To determine further if immunization with LVS ΔcapB can provide protection against lethal LVS challenge in mice, we immunized groups of 12 mice once i.n. with approximately 200 CFU LVS (the maximum dose at which the majority of mice consistently survive) or 1 × 10⁵ CFU LVS ΔcapB or once i.d. with 1 × 10⁵ CFU LVS or 1 × 10⁶ CFU LVS ΔcapB (Fig. 9). Notably, 25% (3/12) of mice immunized i.n. with just 200 CFU LVS died within 10 days of immunization, further underscoring the toxicity of LVS. In contrast, mice immunized with LVS ΔcapB i.n. or i.d. did not show any signs of illness. At 4 weeks postimmunization, surviving mice were challenged i.n. with 4,000 CFU LVS (>5× the LD₅₀). Four of 12 mice in each group were euthanized at 5 days postchallenge, and lung, liver, and spleen were removed for assays of bacterial burden (Fig. 5B and 9A). The remaining eight mice in each group (except the LVS i.n. group, where only five mice remained) were monitored for signs of illness, weight loss, and death for 3 weeks (Fig. 5C and 9B). Mice immunized i.n. or i.d. with LVS ΔcapB and challenged 4 weeks later had lung bacterial burdens >4 logs lower after i.n. immunization and 2 logs lower after i.d. immunization than those of sham-immunized mice; the number of CFU in the spleen was below the limit of detection after i.n. immunization and was ∼4 logs lower than that in sham-immunized mice, similar to mice immunized with LVS, and F. tularensis levels in the liver were below the limit of detection after both i.d. and i.n. immunization, similar to mice immunized with LVS (Fig. 9A). Importantly, the bacterial burden in mice immunized with LVS ΔcapB was 3 logs lower than that in mice immunized with the LPS deletion mutant LVS ΔwbtDEF (data not shown). Although mice immunized with LVS ΔcapB had a somewhat higher bacterial burden in the lung than mice immunized with LVS, these mice showed no signs of illness after challenge. In the survival part of the study, 100% of mice immunized i.n. or i.d. with LVS ΔcapB and challenged with LVS i.n. at 4 weeks postimmunization survived, the same as what was found for mice immunized i.n. or i.d. with LVS (Fig. 9C).

• Open in new tab
• Download powerpoint

FIG. 9.

Immunization with LVS ΔcapB induces protective immunity against i.n. F. tularensis LVS challenge comparable to that of the parental LVS. Groups of 12 mice were sham immunized, immunized i.d. with 1 × 10⁵ CFU LVS or 1 × 10⁶ CFU LVS ΔcapB, or immunized i.n. with 200 CFU LVS or 1 × 10⁵ CFU LVS ΔcapB. Four weeks (A and C) or 8 weeks (B and D) later, mice were challenged i.n. with 4,000 CFU LVS. (A and B) At 5 days postchallenge, four mice per group were euthanized, and the spleen, liver, and lung were assayed for bacterial burden. CFU values are shown as means ± SE. The dashed line indicates the limit of detection. The mean CFU in each organ among different groups was analyzed by grouped ANOVA with Bonferroni's posttest (Prism 5.01 software). The lines above the bars indicate a statistical comparison between the bar beneath the left end of the line and the bar beneath the right end of the line. Only comparisons where differences were statistically significant are shown. **, P < 0.01; ***, P < 0.001. (C and D) The remaining mice in each group were monitored for survival and signs of illness for 3 weeks. The differences in survival between the mice in the vaccinated groups and the mice in the sham-vaccinated group were evaluated by using a log-rank (Mantel-Cox) test (Prism 5.01). ***, P < 0.0001 versus Sham. ^a, 3 of 12 mice (25%) died after i.n. immunization with LVS; of the remaining 9 mice, 4 were studied for the data shown in A, and 5 were studied for the data shown in C.

To evaluate the protection provided by immunization with LVS ΔcapB after a longer immunization-challenge interval, we immunized groups of 12 mice i.n. or i.d. with LVS or LVS ΔcapB as described above and challenged them i.n. with 4,000 CFU LVS at 8 weeks postimmunization (Fig. 9B and D). Again, 25% (3/12) of mice immunized i.n. with LVS died within 10 days of immunization, consistent with the results described above. At 5 days postchallenge, mice immunized with LVS ΔcapB i.n. or i.d. had significantly lower bacterial burdens than sham-immunized mice in all three organs examined (Fig. 9B). In mice immunized i.n., LVS-immunized animals had significantly fewer CFU in the spleen and lung than did LVS ΔcapB-immunized animals. In the liver, there was no significant difference in CFU between animals immunized with LVS and those immunized with LVS ΔcapB. In mice immunized i.d., there was no significant difference in CFU between LVS ΔcapB- and LVS-immunized animals in any of the three organs assayed (Fig. 9B). In the survival part of the study (Fig. 9D), all eight sham-immunized animals died, whereas all animals immunized with LVS ΔcapB or LVS i.n. or i.d. survived (eight mice per group except for the group immunized i.n. with LVS, where only five animals survived immunization).

Thus, LVS ΔcapB, while much less virulent than LVS, induces potent protective immunity against lethal LVS i.n. challenge, comparable to that induced by LVS.

Immunization with LVS ΔcapB protects mice from aerosol challenge with virulent F. tularensis strain SchuS4.To further define the protective immunity induced by immunization with LVS ΔcapB, we evaluated the capacity of the vaccine to protect against highly virulent F. tularensis strain SchuS4. We immunized groups of eight mice once i.n. with 1 × 10⁵ CFU LVS ΔcapB or i.d. with 1 × 10⁶ CFU LVS ΔcapB, challenged the mice 6 weeks later by aerosol with 10× the LD₅₀ of F. tularensis SchuS4, and monitored the animals for weight change, signs of illness, and death for 3 weeks (Fig. 5D and 10). Mice sham immunized with PBS or immunized with LVS (200 CFU i.n. or 1 × 10⁵ CFU i.d.) served as controls. Mice immunized with LVS ΔcapB developed strong protective immunity to F. tularensis SchuS4 challenge (Fig. 10). In mice immunized i.d. (Fig. 10A and C), protection was incomplete. All LVS ΔcapB- and sham-immunized mice succumbed to challenge; however, LVS ΔcapB-immunized mice survived twice as long (10 days) as sham-immunized mice (5 days; P < 0.0001). LVS-immunized mice were better protected than LVS ΔcapB-immunized mice (62.5% versus 0% survival); however, all these mice showed substantial weight loss after challenge (Fig. 10A).

• Open in new tab
• Download powerpoint

FIG. 10.

Immunization with LVS ΔcapB induces protective immunity against F. tularensis SchuS4 aerosol challenge. Groups of eight mice were sham immunized, immunized i.d. (A and C) with 1 × 10⁵ CFU LVS or 1 × 10⁶ CFU LVS ΔcapB, or immunized i.n. (B and D) with 200 CFU LVS or 1 × 10⁵ CFU LVS ΔcapB. Six weeks later, all mice were challenged by the aerosol route with 10× the LD₅₀ of F. tularensis strain SchuS4. Mice were monitored for weight change (A and B) and survival (C and D) for 3 weeks. Weights at each time point are shown as means ± SE. Mean survival time was calculated by dividing the sum of the survival times of all mice in a group by the total number of mice challenged, with animals surviving until the end of the experiment given a survival time of 21 days, when the experiment was terminated. The differences in survival between the mice in the vaccinated groups and mice in the sham-vaccinated group were evaluated by using a log-rank (Mantel-Cox) test (Prism 5.01 software).

In mice immunized with LVS ΔcapB i.n. (Fig. 10D), 100% of the mice survived challenge, versus 0% for sham-immunized animals (P < 0.0001). Similarly, 100% of LVS-immunized animals survived. Notably, the LVS ΔcapB- and LVS-immunized animals showed no signs of illness after challenge, e.g., no weight loss (Fig. 10B).

Thus, LVS ΔcapB induces potent protective immunity against F. tularensis SchuS4 aerosol challenge, especially after i.n. immunization. Protection after i.n. immunization is stronger than that after i.d. immunization and comparable to that induced by LVS.

Killed vaccines.Studies have shown that killed whole-cell vaccines do not afford high-level protection against type A F. tularensis in animal models (4, 10, 15, 29, 37, 38).

Subunit vaccines.Subunit vaccines comprising an F. tularensis protein in an adjuvant formulation or F. tularensis LPS purified from LVS have not demonstrated high-level efficacy against non-type A or virulent type A F. tularensis in animal models (8, 13, 14, 40). Also, capsule material prepared from virulent strain SchuS4 was unable to induce protective immunity in guinea pigs and mice against SchuS4 challenge (17). Intraperitoneal immunization with an F. tularensis LVS outer membrane protein preparation protected 50% of mice challenged i.n. with type A SchuS4 (22).

Live attenuated type A vaccines.Live attenuated single-deletion F. tularensis type A strains carry a potential risk of reversion to virulence, and multiple deletions in type A strains have a tendency to reduce the immunogenicity of the vaccine. An F. tularensis SchuS4 mutant deficient in IglC was significantly attenuated but was unable to protect against type A F. tularensis challenge despite its ability to persist and disseminate to the liver and spleen of the host (9, 46). An FTT1103 deletion mutant of F. tularensis SchuS4 was found to be highly attenuated, but it provided protection against SchuS4 challenge only when administered i.n. at a dose of 10⁸ CFU (9, 32). A purified auxotroph of F. tularensis SchuS4 was found to give poor protection against i.n. challenge with the homologous parental organism and no better protection than an analogous purine auxotroph of the less virulent type B LVS strain (31). An F. tularensis SchuS4 mutant with a deletion in clpB gave partial protection by the i.d. and oral routes (9, 46). An F. tularensis SchuS4 mutant deficient in FTT0918 was highly protective after i.d. administration (41); however, an additional deletion in capB gave reduced immunoprotection (9, 46).

Live attenuated non-type A vaccines.Previously constructed live attenuated non-type A mutants have not afforded potent immunoprotection against virulent F. tularensis challenge. O-antigen-deficient mutants in F. tularensis subsp. tularensis, holarctica, and novicida were overattenuated in mice and therefore unable to induce protective immunity against type A strains (26, 33, 39, 44). Similarly, our initial effort with O-antigen-deficient LVS strains (LVS ΔwbtDEF and LVS Δwzy) failed to provide potent protection against respiratory challenge with F. tularensis. An F. tularensis subsp. novicida iglB mutant is significantly attenuated in mice; oral but not i.n. administration provided limited protection against F. tularensis SchuS4 challenge to mice (7). An F. tularensis LVS deficient in iron superoxide dismutase was found to be protective against F. tularensis SchuS4 challenge following i.n. immunization with 5 × 10³ CFU of this mutant (3). However, i.n. inoculation of 1 × 10⁴ CFU of this mutant caused 17% and 40% deaths in C57BL/6 and BALB/c mice, respectively, indicating an underattenuation of this strain (2).

In contrast to these vaccines, the LVS ΔcapB vaccine is neither overattenuated nor underattenuated. At the same time, because LVS ΔcapB contains multiple defined noncontiguous deletions, including most of the FTT0918, pilA, and capB genes, there is extremely little likelihood of reversion to virulence to the wild-type F. tularensis subsp. holarctica strain, a less virulent subspecies than F. tularensis subsp. tularensis to begin with.

Note that LVS ΔcapB shares some common genetic features with the SchuS4 ΔFTT0918 ΔcapB strain (9). FTT0918 has been shown to be essential for the virulence of strain SchuS4 (46). In LVS, only the C terminus of FTT0918 remains, which forms a fusion protein with the N terminus of FTT0919. The reintroduction of FTT0918 in cis partially restores virulence to LVS, and the reintroduction of both FTT0918 and pilA fully restores the virulence of LVS (35) to that of wild-type F. tularensis subsp. holarctica. Given that three major deletions (most of the FTT0918, pilA, and capB genes) were necessary to reduce the virulence of the parent of LVS to a nontoxic level and that LVS contains additional gene deletions, it is perhaps not surprising that a SchuS4 mutant with only a double deletion in FTT0918 and capB remains partially virulent in BALB/c mice and fully virulent in C3H/HeN mice (9).

In summary, LVS ΔcapB, while >10,000-fold-less virulent than LVS by the respiratory route, induces potent protective immunity against lethal F. tularensis respiratory challenge comparable to that of LVS after i.n. administration. By the i.d. route of administration, LVS ΔcapB offers significant protection although less than that of LVS against F. tularensis SchuS4 aerosol challenge. Balancing safety and efficacy, LVS ΔcapB, an antibiotic-marker-free vaccine that is immunoprotective in mice at doses that are completely nontoxic, already has the potential to offer a reasonable alternative to the LVS vaccine.

In future studies, we shall evaluate the efficacy of recombinant versions of LVS ΔcapB overexpressing F. tularensis immunoprotective proteins. The concept of a vaccine vector overexpressing native immunogenic proteins, as opposed to expressing foreign antigens, is a new paradigm in vaccinology pioneered in this laboratory; the value of the concept was demonstrated in the development of novel tuberculosis vaccines (20, 21). Preliminary studies indicate that the expression of native F. tularensis proteins in LVS ΔcapB will significantly enhance its efficacy.