Combinational deletion of three membrane protein-encoding genes highly attenuates yersinia pestis while retaining immunogenicity in a mouse model of pneumonic plague

Abstract

Previously, we showed that deletion of genes encoding Braun lipoprotein (Lpp) and MsbB attenuated Yersinia pestis CO92 in mouse and rat models of bubonic and pneumonic plague. While Lpp activates Toll-like receptor 2, the MsbB acyltransferase modifies lipopolysaccharide. Here, we deleted the ail gene (encoding the attachment-invasion locus) from wild-type (WT) strain CO92 or its lpp single and Δlpp ΔmsbB double mutants. While the Δail single mutant was minimally attenuated compared to the WT bacterium in a mouse model of pneumonic plague, the Δlpp Δail double mutant and the Δlpp ΔmsbB Δail triple mutant were increasingly attenuated, with the latter being unable to kill mice at a 50% lethal dose (LD50) equivalent to 6,800 LD50s of WT CO92. The mutant-infected animals developed balanced TH1- and TH2-based immune responses based on antibody isotyping. The triple mutant was cleared from mouse organs rapidly, with concurrent decreases in the production of various cytokines and histopathological lesions. When surviving animals infected with increasing doses of the triple mutant were subsequently challenged on day 24 with the bioluminescent WT CO92 strain (20 to 28 LD50s), 40 to 70% of the mice survived, with efficient clearing of the invading pathogen, as visualized in real time by in vivo imaging. The rapid clearance of the triple mutant, compared to that of WT CO92, from animals was related to the decreased adherence and invasion of human-derived HeLa and A549 alveolar epithelial cells and to its inability to survive intracellularly in these cells as well as in MH-S murine alveolar and primary human macrophages. An early burst of cytokine production in macrophages elicited by the triple mutant compared to WT CO92 and the mutant's sensitivity to the bactericidal effect of human serum would further augment bacterial clearance. Together, deletion of the ail gene from the Δlpp ΔmsbB double mutant severely attenuated Y. pestis CO92 to evoke pneumonic plague in a mouse model while retaining the required immunogenicity needed for subsequent protection against infection.

FIG 1

Ail and Lpp production and transmission electron microscopy analysis. (A) Y. pestis cultures grown overnight (at 37°C) were collected, and the production of Ail and Lpp in the whole-cell lysates was analyzed by immunoblotting using antibodies to Ail and Lpp. Anti-DnaK antibodies were used as a loading control for Western blots. (B) WT CO92 and its Δlpp ΔmsbB Δail triple mutant were grown to the exponential growth phase at 28°C and subjected to transmission electron microscopy analysis. o, bacterial outer membrane; i, bacterial inner membrane; p, periplasmic space. Bar = 0.5 μm.

FIG 2

Functionality of the T3SS and production/enzymatic activity of the Pla protease. Cultures of various Y. pestis strains grown in HIB overnight were diluted 1:20 in modified M9 medium or fresh HIB, and growth was continued at 28°C for 3 h, followed by an additional 2 h of incubation at 37°C. (A and B) The production of YopH, LcrV, and YopE in M9 medium (A) or YopH in HIB chelated for calcium (B) was measured by Western blotting using specific antibodies. Anti-DnaK antibodies were employed to examine bacterial pellets to ensure that bacterial supernatants were obtained from similar numbers of bacteria across the tested strains. For the translocation studies, cultures of various Y. pestis strains grown overnight in HIB were diluted and sensitized to DMEM by growth at 28°C for 30 min, followed by an additional 1 h of incubation at 37°C. HeLa cells were then infected with the above-mentioned cultures at an MOI of 30. (C) After 4 h of infection, the cytosolic fraction of the host cells was separated from the pellet and probed with anti-YopE antibodies. Antiactin and anti-DnaK antibodies were also used on the supernatant and pellet fractions, respectively, to monitor equal loading of the samples during Western blot analyses. (D) Pla production in various Y. pestis CO92 stains was examined with specific antibodies to Pla, and anti-DnaK antibodies were used as a loading control. For measurement of Pla activity, the tested Y. pestis CO92 strains were mixed with the Pla substrate [DABCYL-Arg-Arg-Ile-Asn-Arg-Glu(EDANS)-NH₂], and the kinetics of substrate cleavage was measured. The Δpla single mutant of CO92 was employed as a negative control during the assay. (E) The kinetics of each reaction is plotted as arithmetic means ± standard deviations. Statistical analysis of Pla activity data was performed by one-way ANOVA with a Bonferroni post hoc test. Statistically significant P values between the groups are indicated by a vertical line.

FIG 3

Production of F1 antigen. Selected Y. pestis cultures grown overnight were diluted 1:20 in fresh HIB, and growth was continued at 28°C for 3 h, followed by an additional 2 h of incubation at 37°C. F1 production was either examined by using immunochromatographic reaction dipsticks (A) or probed by immunofluorescence staining with anti-F1 antibodies followed by flow cytometric analysis (B) and microscopy (C). The Δcaf1 mutant of CO92 was employed as a negative control. Magnification, ×400 (C). FITC, fluorescein isothiocyanate.

FIG 4

Survival analysis and antibody responses of mice infected with WT Y. pestis strain CO92 and its mutant strains in a pneumonic plague model. Female Swiss Webster mice (10 per group) were challenged with 1.3 × 10⁴ CFU of WT Y. pestis strain CO92 or its various mutants by the i.n. route. (A) Survival of mice was plotted and analyzed by Kaplan-Meier survival estimates. Statistically significant P values for comparisons of various mutant- and WT CO92-infected mice are indicated under each curve. (B) Mice were bled at 14 days p.i., and the total IgG responses to F1-V antigen were determined by an ELISA. The arithmetic means ± standard deviations are plotted. ** indicates statistical significance (P < 0.001) compared to preimmune serum.

FIG 5

Virulence potential of and subsequent protection conferred by the Δlpp ΔmsbB Δail triple mutant of Y. pestis CO92 in a pneumonic plague mouse model. (A) Female Swiss Webster mice (n = 9 or 10/group) were challenged with the indicated doses of the Δlpp ΔmsbB Δail triple mutant or 1.6 × 10⁴ CFU of WT Y. pestis CO92 by the i.n. route. The surviving animals and age-matched naive mice were then rechallenged on day 24 p.i. with 1 × 10⁴ CFU of the WT CO92 luc2 strain. Statistically significant P values are for comparisons to WT CO92-infected mice during the initial challenge or to naive control animals during WT CO92 luc2 rechallenge. (B) Total IgG responses to the F1-V antigen were examined in sera at day 14 after initial infection. ** indicates statistical significance (P < 0.001) compared to preimmune serum or between different doses of the triple mutant used for initial infection. (C) The animals after rechallenge were imaged at 72 h for bioluminescence. The bioluminescence scale is shown at the right and ranges from most intense (red) to least intense (violet). The animal on the left of each imaging panel represents an uninfected control.

FIG 6

Survival analysis and subsequent protection conferred by high doses of the Δlpp ΔmsbB Δail triple mutant of Y. pestis CO92 in a pneumonic plague mouse model. (A) Female Swiss Webster mice (5 to 10 per group) were infected with various doses of the Δlpp ΔmsbB Δail triple mutant or 1.3 × 10⁴ CFU of WT Y. pestis CO92 by the i.n. route. Surviving mice with age-matched naive animals were then rechallenged on day 24 p.i. with 1.4 × 10⁴ CFU of the WT CO92 luc2 strain. Statistically significant P values are for comparisons to the WT CO92-infected mice in the initial challenge or to naive control mice during the WT CO92 luc2 rechallenge. (B) Total IgG responses to the F1-V antigen or whole bacteria were examined in sera at day 14 after initial infection. The titers of antibody isotypes to F1-V antigen were further delineated by using isotype-specific secondary antibodies. *** indicates statistical significance (P < 0.0001) compared to preimmune serum. (C) Animals were imaged on day 3 and/or day 7 after rechallenge for bioluminescence. The bioluminescence scale is shown on the right and ranges from most intense (red) to least intense (violet).

FIG 7

Dissemination of WT CO92 and its Δlpp ΔmsbB Δail triple mutant in a mouse model of pneumonic plague. Female Swiss Webster mice were challenged with 2.5 × 10⁶ CFU of WT Y. pestis CO92 or its Δlpp ΔmsbB Δail triple mutant by the i.n. route. Organs and blood were harvested from mice (n = 5) on days 2, 3, and 6 p.i. The bacterial loads in different organs and blood from each individual mouse were plotted, and the arithmetic means are indicated by the horizontal bars. ** indicates statistical significance (P < 0.001) compared to WT CO92 on each day (day 6 was compared to day 3 for WT CO92).

FIG 8

Histopathology of mouse tissues following pneumonic infection with WT CO92 or its Δlpp ΔmsbB Δail triple mutant. Female Swiss Webster mice were challenged with 2.5 × 10⁶ CFU of WT Y. pestis CO92 or the Δlpp ΔmsbB Δail triple mutant by the i.n. route. On days 2, 3, and 6 p.i., a portion of the lungs, liver, and spleen (n = 3 to 5) was stained with H&E and evaluated by using light microscopy in a blind fashion. Only data for day 3 are shown. The presence of bacteria, neutrophilic infiltration, hemorrhage/necrosis, and rarefied red pulp is indicated by asterisks, arrows, arrowheads, and plus signs, respectively.

FIG 9

Cytokine/chemokine analysis of sera (B) and lung homogenates (A) of mice in a pneumonic plague mouse model. Mice were challenged with 2.5 × 10⁶ CFU of WT Y. pestis CO92 or its Δlpp ΔmsbB Δail triple mutant by the i.n. route. At 2 and 3 days p.i., 5 mice from each group (at each time point) were euthanized. The lungs were harvested and homogenized, and blood was collected via cardiac puncture. The production of various cytokines/chemokines was measured by using a multiplex assay. Only the cytokines/chemokines showing statistically significant differences in mutant- compared to WT CO92-infected mice are plotted as arithmetic means ± standard deviations. * and ** indicate statistical significance (P < 0.01 and P < 0.001, respectively) compared to WT CO92 on each day.

FIG 10

Serum resistance and Ail production by various Y. pestis CO92 strains. Various Y. pestis strains (∼5 × 10⁶ CFU) grown overnight were mixed with either unheated or heated sera from human, NHP, and mouse. (A) After incubation for 2 h at 37°C, the number of surviving bacteria (CFU) in each sample was determined. * indicates statistical significance (P < 0.01) compared to WT CO92 for each type of serum. ND, not detectable. (B) The levels of Ail protein and DnaK in these strains were analyzed by immunoblotting using specific antibodies to Ail and DnaK.

FIG 11

Adherence and invasion of WT Y. pestis CO92 and its mutant strains. HeLa cells (A) and A549 cells (B) were infected with various Y. pestis CO92 strains at an MOI of 100 at 37°C for 2 h. The percentages of adherent (I) and invading (II) bacteria compared to the total number of bacteria used to infect epithelial cells were calculated. The arithmetic means ± standard deviations are plotted. * and ** indicate statistical significance (P < 0.05 and P < 0.001, respectively) compared to both WT CO92 and the Δlpp ΔmsbB mutant.

FIG 12

Intracellular survival of various Y. pestis CO92 mutant strains in epithelial cells and macrophages. Murine MH-S macrophages (A), human monocyte-derived macrophages (HMDM) (B), HeLa epithelial cells (C and D), and human A549 alveolar epithelial cells (E) were infected with various Y. pestis CO92 strains at MOIs of 10, 1, 100, and 100, respectively. After 45 to 60 min of incubation at 37°C and following an hour of gentamicin treatment, the cells were harvested at 2 and 4 h post-gentamicin treatment for macrophages and at 12 h for epithelial cells. The number of bacteria surviving intracellularly was assessed, and percent survival was calculated. *, **, and *** indicate statistical significance (P < 0.05, P < 0.005, and P < 0.001, respectively) compared to WT CO92 at each time point or between two tested strains, as indicated by the horizontal bars.

FIG 13

Inflammatory cytokine production by macrophages infected with various Y. pestis CO92 strains. Murine alveolar macrophages (A) and human monocyte-derived macrophages (B) were infected with various Y. pestis strains. Supernatants from infected macrophages were collected at 0, 2, or 4 h after gentamicin treatment. The levels of various cytokines in the supernatants were measured by using a multiplex assay. Only the cytokines/chemokines showing statistically significant differences compared to WT CO92-infected macrophages are plotted as arithmetic means ± standard deviations. *, **, and *** indicate statistical significance (P < 0.05, P < 0.01, and P < 0.0001, respectively) compared to WT CO92 on each day.