Oxidative stress resistance in the Francisella tularensis live vaccine strain is associated with genetic variability in the ferrous iron uptake gene feoB

Francisella tularensis is a highly virulent bacterial pathogen with an extremely low infectious dose (~10 CFU) and high rates of mortality if left untreated (30-60%). F. tularensis has an extensive history as a bioweapon, and there is no vaccine currently licensed. For these reasons the CDC considers F. tularensis a Tier 1 Select Agent. The unlicensed F. tularensis subsp. holarctica Live Vaccine Strain (LVS) provides moderate protection against virulent strains; however, we have discovered that various “wild type” lab stocks differ in their virulence and ability to confer immunity. Genome sequencing of high virulence (RML, LD50 ~200 CFU) and low virulence (ATCC, LD50 ~9,000 CFU) strains has identified nine differences, of which four are non-synonymous substitutions. One such mutation occurs in the ferrous iron uptake gene feoB in RML. While iron is required for cellular function, ferrous iron (Fe²⁺) can participate in the Fenton reaction with H₂O₂, leading to inactivation of essential iron-sulfur cluster enzymes. Part of the innate immune response involves mitochondria-derived reactive oxygen species in the cytosol. Fully virulent strains of F. tularensis are known to be highly resistant to such host defenses, and have low levels of intracellular iron. Accordingly, the RML strain was highly resistant to exogenous H₂O₂in vitro relative to the ATCC strain. An iron-responsive lacZ reporter had ~2-fold higher induction in the RML strain relative to ATCC during iron limitation. Overexpression of the functional feoB allele, but not the RML allele, leads to significantly increased sensitivity to H₂O₂ and increased killing by primary macrophages stimulated with the cytokine interferon-gamma. Given the connection of iron and H₂O₂ toxicity, I revisited a previously published transposon screen to determine if any of the mutants identified had a role in iron homeostasis and oxidative stress resistance. One such gene was annotated as bacterioferritin (bfr), which in other bacteria forms a hollow, spherical multimer that oxidizes Fe²⁺ to Fe³⁺ and stores the oxidized form in the interior of the sphere. The Δbfr mutant was ~10-fold more sensitive to H₂O₂ and was attenuated nearly 8-fold in murine intranasal infection in terms of LD₅₀ relative to the parental RML strain. Importantly, the Δbfr mutant allowed us to test the hypothesis that H₂O₂ resistance is critical for the RML LVS to stimulate productive immunity. At six weeks post-infection, mice previously infected with either RML or the Δbfr mutant were challenged with an infection of 25 CFU of the fully virulent F. tularensis Schu S4 strain. All mice immunized with RML survived this challenge, while all mice immunized with Δbfr succumbed; only displaying a slight increase in time to death. These results are consistent with the hypothesis that the H₂O₂ resistance of RML LVS mediates increased fitness in a host.