Early emergence of Yersinia pestis as a severe respiratory pathogen

Abstract

Yersinia pestis causes the fatal respiratory disease pneumonic plague. Y. pestis recently evolved from the gastrointestinal pathogen Y. pseudotuberculosis; however, it is not known at what point Y. pestis gained the ability to induce a fulminant pneumonia. Here we show that the acquisition of a single gene encoding the protease Pla was sufficient for the most ancestral, deeply rooted strains of Y. pestis to cause pneumonic plague, indicating that Y. pestis was primed to infect the lungs at a very early stage in its evolution. As Y. pestis further evolved, modern strains acquired a single amino-acid modification within Pla that optimizes protease activity. While this modification is unnecessary to cause pneumonic plague, the substitution is instead needed to efficiently induce the invasive infection associated with bubonic plague. These findings indicate that Y. pestis was capable of causing pneumonic plague before it evolved to optimally cause invasive infections in mammals.

Figure 1. pPCP1 is required by ancestral Y. pestis to cause primary pneumonic plague.

(a) Genomic maximum parsimony tree and divergence based on 16 Y. pestis genomes. The division between modern, pandemic strains and ancestral isolates is indicated. Tree was adapted from Morelli et al. (b) Bacterial burden within the lungs and spleens of mice (n=10) infected i.n. with the indicated Y. pestis strains. Each point represents the numbers of bacteria recovered from a single mouse at 48 h post inoculation. The limit of detection is indicated by the dashed line and symbols in the dotted line indicate c.f.u. below the limit of detection. Symbols below the limit of detection represent mice that did not have detectable numbers of bacteria. A solid line indicates the median of c.f.u. recovered. The presence or absence of pPCP1 in each strain is indicated below. (c) Immunoblot analysis of whole-cell lysates of the indicated Y. pestis strains with antibodies to Pla and RpoA (as a loading control). The lower band represents the autoprocessed form of Pla (see Supplementary Fig. 5a). Full blots are shown in Supplementary Fig. 6. Panel is representative of three independent replicates. Data are combined from two independent experiments and error bars represent the s.e.m. (*P≤0.05, **P≤0.01, ***P≤0.001, NS, not significant by Mann–Whitney U-test).

Figure 2. Pestoides F resembles Δpla Y. pestis during intranasal infections.

(a) Survival of mice (n=20) infected i.n. with Y. pestis CO92, CO92 Δpla or Pestoides F. (b) Bacterial burden within the lungs and spleens of mice (n=10) infected i.n. with the indicated Y. pestis strains, as described in Fig. 1. By 72 h, mice begin to succumb to the CO92 infection (indicated by “X” on the x axis). (c) Pathology of mouse lung sections stained with H&E at 48 h post inoculation with PBS (mock), Pestoides F, CO92 Δpla or CO92 Y. pestis. Representative images of inflammatory lesions are shown (arrows; n=3). Scale bar, 200 μm. (d) Enumeration of total immune cells present in BAL fluid 48 h post inoculation with PBS (mock), Pestoides F, CO92 Δpla or CO92 Y. pestis (n=10). (e) Abundance of the indicated inflammatory cytokines present in BAL fluid at 48 h post inoculation. Data are combined from two independent experiments and error bars represent the s.e.m. (*P≤0.05, **P≤0.01, NS, not significant by Log-Rank test (survival), Mann–Whitney U-test (c.f.u.)). H&E, haematoxylin/eosin.

Figure 3. Pestoides F is competent to carry pPCP1 and produce active Pla.

(a) Relative copy number of pPCP1 (represented by the pPCP1-encoded genes pla and pst) in the pPCP1-reintroduced Pestoides F and CO92 strains, compared with CO92 (set at 1). pspA was used as a control for a chromosomal gene. Relative copy number for each gene was measured by quantitative PCR from gDNA isolated from cultures grown overnight at 37 °C and normalized to gyrB. Data are combined from three independent biological replicates repeated twice; error bars represent the s.e.m. (b) Immunoblot analysis of whole-cell lysates of indicated Y. pestis strains grown at 37 °C with antibodies against Pla and RpoA (as a loading control). Panel is representative of three independent replicates. The presence or absence of pPCP1 in each strain is indicated. Full blots are shown in Supplementary Fig. 6. (c) The Plg-activating ability of the indicated Y. pestis CO92 or Pestoides F strains cultured at 37 °C is shown. Data are representative of three independent experiments performed in triplicate; error bars represent the s.e.m.

Figure 4. The acquisition of Pla by Pestoides F is sufficient to cause primary pneumonic plague.

(a) Survival of mice (n=10) infected i.n. with the indicated strains of Y. pestis. (b) Bacterial burden within the lungs and spleens of mice (n=10) infected i.n. with the indicated Pestoides F or CO92 strains, as described in Fig. 1. The presence or absence of pPCP1 is indicated. (c) Pathology of mouse lung sections stained with H&E at 48 h post inoculation with PBS (mock) or the indicated Y. pestis strains. Representative images of inflammatory lesions (arrows; n=3) are shown. Scale bar, 200 μm. (d) Enumeration of total immune cells present in BAL fluid 48 h post inoculation with PBS (mock) or the indicated Y. pestis strains (n=10). Data are combined from two independent experiments. Error bars represent the s.e.m. (*P≤0.05, **P≤0.01, ***P≤0.001, NS, not significant; Log-Rank test (survival), Mann–Whitney U-test (c.f.u.)). H&E, haematoxylin/eosin.

Figure 5. The ancestral variant of Pla is sufficient to cause pneumonic plague but not optimal systemic infection.

(a) Bacterial burden within the lungs and spleens of mice (n=10) infected i.n. with Y. pestis CO92 or Pestoides F carrying pPCP1 with either the T259 or I259 variant of Pla, as described in Fig. 1. (b) Bacterial burden within the inguinal lymph nodes and spleens of s.c. infected mice (n=10) with wild-type Pestoides F, or CO92 or Pestoides F carrying pPCP1 with either the T259 or I259 variant of Pla. The variant of Pla in each strain is indicated. Data are combined from two independent experiments and error bars represent the s.e.m. (*P≤0.05, NS, not significant by Mann–Whitney U-test).