Fis Is Essential for Yersinia pseudotuberculosis Virulence and Protects against Reactive Oxygen Species Produced by Phagocytic Cells during Infection

Abstract

All three pathogenic Yersinia species share a conserved virulence plasmid that encodes a Type 3 Secretion System (T3SS) and its associated effector proteins. During mammalian infection, these effectors are injected into innate immune cells, where they block many bactericidal functions, including the production of reactive oxygen species (ROS). However, Y. pseudotuberculosis (Yptb) lacking the T3SS retains the ability to colonize host organs, demonstrating that chromosome-encoded factors are sufficient for growth within mammalian tissue sites. Previously we uncovered more than 30 chromosomal factors that contribute to growth of T3SS-deficient Yptb in livers. Here, a deep sequencing-based approach was used to validate and characterize the phenotype of 18 of these genes during infection by both WT and plasmid-deficient Yptb. Additionally, the fitness of these mutants was evaluated in immunocompromised mice to determine whether any genes contributed to defense against phagocytic cell restriction. Mutants containing deletions of the dusB-fis operon, which encodes the nucleoid associated protein Fis, were markedly attenuated in immunocompetent mice, but were restored for growth in mice lacking neutrophils and inflammatory monocytes, two of the major cell types responsible for restricting Yersinia infection. We determined that Fis was dispensable for secretion of T3SS effectors, but was essential for resisting ROS and regulated the transcription of several ROS-responsive genes. Strikingly, this protection was critical for virulence, as growth of ΔdusB-fis was restored in mice unable to produce ROS. These data support a model in which ROS generated by neutrophils and inflammatory monocytes that have not been translocated with T3SS effectors enter bacterial cells during infection, where their bactericidal effects are resisted in a Fis-dependent manner. This is the first report of the requirement for Fis during Yersinia infection and also highlights a novel mechanism by which Yptb defends against ROS in mammalian tissues.

Fig 1. Virulence of mutants in mini-TnSeq assay.

Fitness of knockouts generated in YPIII/pIB1^- (A, C) or IP2666/pIB1⁺ (B, D) at 3 days post-intravenous infection with 10⁴ and 10³, respectively, of mini-TnSeq libraries. Fitness values were obtained by dividing the proportion of sequencing reads for a mutant in the liver (A, B) or spleen (C, D) by its proportion of reads in the inoculum. Each data point for a mutant represents an individual mouse. N = 7–10 mice. Fitness values were log₁₀ transformed and statistical significance was calculated using One Way ANOVA analysis with Dunnett’s multiple comparison post-test. P-values represent comparisons between the fitness scores of individual mutants with the fitness score of the two knockouts of neutral loci (pooled together) in each respective condition. * indicates p≤0.05, ** indicates p≤0.01, *** indicates p≤0.001, and **** indicates p≤0.0001.

Fig 2. Mini-TnSeq accurately predicted outcomes of 1:1 competition experiments.

Mice were inoculated intravenously with 1:1 mixture of 10³ bacteria, comprised of WT-Kan^R and a drug-sensitive mutant. Livers (A) and spleens (B) were dissected at 3 days post-infection, homogenized and plated onto selective and non- selective plates to determine the competitive index (C.I.). Each symbol represents an individual mouse. N = 4–10. C.I. data was log₁₀ transformed and statistical significance was calculated using One Way ANOVA analysis with Dunnett’s multiple comparison post-test comparing the C.I. values of mutant strains with those of the WT strain. * indicates p≤0.05, ** indicates p≤0.01, *** indicates p≤0.001, and **** indicates p≤0.0001.

Fig 3. Four mutants displayed fitness changes following phagocytic cell depletion.

Mice were intraperitoneally injected with RB6-8C5 (A-B) or 1A8 (C-D) 24 hours prior to and post-infection. Mice were inoculated intravenously with 10³ CFU of the IP2666/pIB1⁺ mutant library. Fitness values were obtained by dividing the proportion of sequencing reads for a mutant in the depleted liver (A, C) or spleen (B, D) by its proportion of reads in the inoculum. Each data point for a mutant represents an individual mouse. N = 5–10 mice. Non-depleted fitness values are the same is in Fig 1. Fitness scores values were log₁₀ transformed and an unpaired t-test with the Holm-Sidak correction for multiple comparisons was performed to calculate statistically significant differences between the fitness scores of specific bacterial mutants in depleted versus non-depleted mice. * indicates p≤0.05, ** indicates p≤0.01, *** indicates p≤0.001, and **** indicates p≤0.0001.

Fig 4. ΔdusB-fis is sensitive to neutrophils and inflammatory monocytes in vivo.

C57BL6/J mice were inoculated intravenously with a pool of 10³ bacteria, containing an equal mixture of WT and WT-Kan^R, Δfis-Kan^R and WT, ΔdusB-fis and WT-Kan^R, or ΔdusB-fis::dusB-fis and WT-Kan^R. 24 hours prior to and post-infection, some mice were intraperitoneally injected with either the 1A8 or RB6-8C5 antibody. Other mice were intraperitoneally injected with the MC-21 antibody 1 day prior to infection and each day after until completion of the experiment. Mice were euthanized at the indicated times and livers (A) and spleens (B) were collected, and dilutions of tissue homogenates were plated onto selective and non-selective media to determine the C.I. C.I. data was log₁₀ transformed and statistical significance was calculated using One Way ANOVA analysis with Dunnett’s multiple comparison post-test comparing the C.I. values of mutant strains to WT or C.I. values in non-depleted and depleted mice. * indicates p≤0.05, ** indicates p≤0.01, *** indicates p≤0.001, and **** indicates p≤0.0001.

Fig 5. ΔdusB-fis can colonize, but is unable to sustain growth in systemic tissue sites.

C57BL6/J mice were inoculated intravenously with a pool of 10³ bacteria, containing either an equal mixture of WT yopE::mcherry and ΔdusB-fis-Kan^R (A-B), or with WT yopE::mcherry only or ΔdusB-fis-Kan^R only (C-D). At 4-hour, 24-hour, 48-hour, and 72-hour time-points post-infection, mice were euthanized and spleens and livers were collected. Dilutions of liver (A, C) and spleen (B, D) homogenates were plated and CFU calculated. In co-infection experiments (A-B), the number of bacteria recovered from selective and non-selective plates was used to determine quantity of WT yopE::mcherry and ΔdusB-fis-Kan^R bacteria in each organ sample. CFU/g data was log₁₀ transformed and the values of ΔdusB-fis-Kan^R to WT yopE::mcherry bacteria in each tissue at each time point was compared using an unpaired t-test with the Holm-Sidak correction for multiple comparisons. * indicates p≤0.05, ** indicates p≤0.01, *** indicates p≤0.001, and **** indicates p≤0.0001.

Fig 6. ΔdusB-fis is not defective for Type 3 Secretion or effector translocation.

A) Stationary phase cultures of WT, ΔyscF, ΔdusB-fis, and Δfis strains were diluted 1:40 into L broth with 20mM sodium oxalate and 20 mM MgCl₂. Cultures were grown for 2 hours at 26°C and then shifted to 37°C for 2 hours. After growth, proteins from culture supernatants were precipitated, resolved on a polyacrylamide gel alongside a standard protein ladder, and visualized using commaasie blue. B-C) Stationary phase cultures of WT, ΔyscF, and ΔdusB-fis strains containing YopH-TEM (HTEM) or YopE-TEM (ETEM) beta-lactamase reporters were grown as in A. After growth, the culture supernatant was mixed with nitrocefin at a final concentration of 100 μg/mL. After 10 minutes of incubation, the A₄₉₀ of samples was measured using a BioTek Synergy HT plate reader. Statistical significance was calculated using One Way ANOVA analysis with the Dunnett’s multiple comparison test on log₁₀-transformed values. Each bar represents the mean and standard error of 3 biological replicates. D-E) Stationary phase cultures of WT, ΔyopB, and ΔdusB-fis strains fused containing H-TEM or E-TEM beta-lactamase reporters were grown as in A-C. HEp-2 cells were infected at a multiplicity of infection of 10 for 1 hr and then incubated with CCF4 to determine the percentage of cells containing translocated effectors (% blue). Data was log₁₀ transformed and statistical significance was calculated using One Way ANOVA analysis with Dunnett’s multiple comparison post-test comparing each strain to WT. Each bar represents the mean and standard error of 3–6 biological replicates. F-G) C57BL6/J mice were inoculated intravenously with a pool of 10³ bacteria, containing an equal mixture of ΔdusB-fisΔyscF -Kan^R and ΔyscF. 24 hours prior to and post-infection, mice were intraperitoneally injected with RB6-8C5 antibody. Mice were euthanized at 3 days post-infection and livers (F) and spleens (G) were collected, and dilutions of tissue homogenates were plated onto selective and non-selective media to determine the C.I. C.I. values were log₁₀ transformed and statistical significance was calculated using a Mann-Whitney t-test. * indicates p≤0.05, ** indicates p≤0.01, **** indicates p≤0.0001.

Fig 7. dusB-fis is required for resistance to oxidative stress.

A) Stationary phase cultures of WT and ΔdusB-fis were diluted 1:100 into L broth and L broth adjusted to pH 5.5 and the OD₆₀₀ of cultures was measured at 1-hour-intervals for 12 hours during growth with aeration. Each symbol represents the mean of 2–4 biological replicates. B) Stationary phase cultures were diluted 1:100 into a well of a 96-well plate containing L broth or L broth + 250 μM 2,2’- Bipyridyl and OD₆₀₀ measurements were recorded at 15-minute intervals during growth with aeration. Lines represent the mean of 3 biological replicates. (C-D) Exponential phase cultures were washed and diluted 1:50 into M9 glucose medium or M9 glucose containing 2.5mM DETA NONOate (C) or M9 glucose containing 1.5mM H₂O₂ (D) for 60 minutes. Survival was calculated by determining the number of CFUs recovered following treatment divided by the number of CFUs recovered from untreated cultures. The mean and standard error of 3 biological replicates for DETA NONOate treatment and 6–10 biological replicates for H₂O₂ treatment are shown. Survival values were log₁₀ transformed and statistical significance was calculated using One Way ANOVA analysis with Dunnett’s multiple comparison post-test comparing each strain to WT. E) Δfis fails to up-regulate ROS-responsive genes after exposure to H₂O₂. Exponential phase cultures were washed and diluted 1:50 into M9 glucose medium or M9 glucose containing 20 μM H₂O₂ and were incubated for 10 minutes with aeration. RNA isolated from treated and untreated samples was used to generate cDNA, and qPCR reactions were performed. Relative expression was determined by normalizing to 16S RNA as well as to expression in untreated samples using the ΔΔCT method. Bars represent the mean and standard error of 8 biological replicates. Unpaired Mann-Whitney t-tests were performed to calculate statistical differences between expression of each gene in WT and Δfis. * indicates p≤0.05, ** indicates p≤0.01, **** indicates p≤0.0001, ns indicates not significant.

Fig 8. ΔdusB-fis is sensitive to ROS produced by the NADPH oxidase complex during mouse infection.

C57Bl/6 or C57Bl/6 gp91^phox-/- mice were inoculated intravenously with 1x10³ CFU of a 1:1 mixture of WT and ΔdusB-fis-Kan^R. Livers (A,C,E) and spleens (B,D,F) were collected, weighed, homogenized, and plated for CFUs on selective and non-selective agar at 3 days-post-infection. (A-B) The number of bacteria recovered from selective and non-selective plates was used to determine the C.I. of ΔdusB-fis-Kan^R. Each data point represents an individual mouse. C.I. data was log₁₀ transformed and statistical significance was calculated using the Mann-Whitney t-test. (C-D) The CFU/gram was determined by dividing the total number of recovered CFU on non-selective plates by the weight of the tissue. CFU/g data was log₁₀ transformed and statistical significance was calculated using the Mann-Whitney t-test. (E-F) The number of bacteria recovered from selective and non-selective plates was used to determine the quantity of ΔdusB-fis-Kan^R and WT bacteria in each organ sample. CFU/g data was log₁₀ transformed and statistical significance was calculated using One Way ANOVA analysis with the Dunnett’s multiple comparison test. * indicates p≤0.05, ** indicates p≤0.01, *** indicates 00.001, **** indicates p≤0.0001.