Fig 1.
Diversity of loop 3 (L3) insertions among a collection of 16,086 public KP genomes.
A. Proportion of genomes encoding an ompK36 gene that is truncated or possessing a L3 insertion (if intact). Bars are coloured by the proportion of genomes carrying one or more carbapenemase genes. Only genomes with a single copy of ompK36 that could be unambiguously characterised were included (n = 15,193). B. Proportion of L3 insertions identified that comprise amino acid insertions of Glycine-Aspartate (GD), Threonine-Aspartate (TD), Aspartate (D), Serine-Aspartate (SD) or others (N/GG/TYD/YGS). C. Distribution of genomes possessing each L3 insertion type across major high-risk sequence types (ST). ST258 and ST512 are grouped together as they form a single clone.
Fig 2.
Multiple acquisitions of GD and other L3 insertions among the high-risk ST258/512 clone.
A. Phylogenetic tree of 3629 isolates with public genome data belonging to sequence types (ST) 258 and 512. The tree was rooted using an ST895 isolate (accession SRR5385992) that was subsequently removed. Isolate tips are coloured by the type of OmpK36 L3 insertion (if applicable). Metadata columns from left to right show whether the ompK35 and ompK36 genes are intact (i.e., not truncated), the type of OmpK36 L3 insertion (if applicable), the carbapenemase gene type (if applicable) and the country of origin. Carbapenemases and countries are shown only for those with ≥15 isolates. Carbapenemase gene variants that have imperfect matches to known variants are grouped together with the most closely-related known variant. A similar interactive visualisation with more detailed metadata is available using Microreact at https://microreact.org/project/exB9brEAsQcpg7vKXMJtoF-k-pneumoniae-st258512 B. A zoomed-in visualisation of the clade highlighted in (A). Scale bars show the number of SNPs.
Fig 3.
The D and TD insertions are associated with large clonal expansions in ST16 and ST231.
Phylogenetic trees of 446 and 302 isolates with public genome data belonging to ST16 (A) and ST231 (B), respectively. The trees were rooted using an ST17 outgroup (accession ERR1228220) and an ST101 outgroup (accession ERR1216956), respectively, each of which were subsequently removed. Isolate tips are coloured by the type of OmpK36 L3 insertion (if applicable). Metadata columns from left to right show whether the ompK35 and ompK36 genes are intact (i.e., not truncated), the type of OmpK36 L3 insertion (if applicable), the carbapenemase gene type (if applicable) and the country of origin. Carbapenemases and countries are shown only for those with ≥10 isolates. Carbapenemase gene variants that have imperfect matches to known variants are grouped together with the most closely-related known variant. Scale bars show the number of SNPs. Similar interactive visualisations with more detailed metadata are available using Microreact at https://microreact.org/project/m8qd8j1YmfMapiPJ7prAEh-k-pneumoniae-st16 (ST16) and https://microreact.org/project/pZRm6DsxvZVYPQ2Ea33Buw-k-pneumoniae-st231 (ST231).
Fig 4.
L3 insertions reduce OmpK36 pore diameter and restrict the diffusion of meropenem.
A-D. Cartoon illustrations of OmpK36 (grey) in which L3 is coloured for all the variants; the view is from the extracellular space and perpendicular to the membrane (top panels). Insertions in L3 reduce the relative pore radius as calculated by the program HOLE [58]; surface representation to show the impact of the mutations in the pore size (bottom panels). E-F. Rate of diffusion (ΔOD400/t(s) of meropenem (E) and glucose (F) as determined by liposome swelling assays for different OmpK36 variants. G-I: In vitro competition against KP36WT and KP36WT+D (G), KP36WT+GD (H) and KP36WT+TD (I). No competitive advantage is observed when glucose is added to the M9 minimal media. Lactose and cas-aminoacids supplementation results in outcompetition of L3 insertion expressing strains by KP36WT. J. Median meropenem MIC values for strains encoding different ompK36 variants (n = 3 replicates). E-F: All comparisons not shown were non-significant. Error bars ±SEM. * p<0.332; ** p<0.0021, *** p<0.0002, **** p<0.00001, statistical significance determined by ordinary one-way ANOVA with Tukey’s multiple comparison test. G-I: Error bars ±SD. *, * p<0.332; ** p<0.0021, *** p<0.0002, **** p<0.00001, statistical significance determined by ordinary one-way ANOVA with Dunnett’s multiple comparison to glucose supplementation with a single pooled variance.
Table 1.
Isogenic strains used in this study.
The parental strain ICC8001 [16] was genetically modified in a seamless and markerless recombineering approach to generate the ompK36 variants.
Fig 5.
OmpK36 L3 insertions maintain virulence but confer a competitive disadvantage in a preclinical murine pneumonia model.
A-H. Pneumonia was induced by the intratracheal administration of 250 CFU of isogenic KP strains expressing D (KP36WT+D), GD (KP36WT+GD) and TD (KP36WT+TD) OmpK36 L3 insertions. A strain lacking any L3 insertion (KP36WT) and PBS (uninfected) were used as controls. A schematic of the infection is outlined in panel A. At 48 hours post infection significant weight loss was induced by infection with all strains, irrespective of OmpK36 variant (B) and no significant differences were observed between strains. No significant differences were observed in the lung (C) and blood (D) bacterial burdens between infection with any strain. Serum TNF was only significantly increased following infection with KP36WT compared to uninfected (PBS) controls (E). Serum IL-6 (F), CXCL-1 (G) and lung neutrophils (H) significantly increased following infection with all strains compared to uninfected (PBS) controls, with no significant differences between strains observed. I-L. Competition assays were employed to stringently assess the fitness of OmpK36 L3 insertion mutations. A schematic of the infection is outlined in panel I. 250 CFU of KP36WT was competed against 250 CFU of KP36WT+D (J), KP36WT+GD (K) or KP36WT+TD (L) and bacterial burdens were assessed at 36 hours post infection in the lungs. Each graph shows the % CFU recovered in the lungs in individual mice followed by a summary bar with the mean competition result across infections. KP36WT significantly outcompeted all the L3 insertions tested. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p<0.0001. All experiments were conducted in biological duplicate with 4–5 mice per group. B-H Significance was determined by ordinary one-way ANOVA followed by Tukey’s multiple comparison post-test, except in E where Kruskal-Wallis test was employed as data was not normally distributed. J-L Mean competition was assessed by Mann-Whitney T-test. The diagrams in A and I were created with BioRender.com.
Table 2.
Minimum inhibitory concentrations of different antibiotics for isogenic KP strains.