Fig 1.
K. pneumoniae ethanolamine utilization (eut) loci provide growth advantages when EA is provided as a carbon or nitrogen source.
(A) Graphic representation of ethanolamine (EA) breakdown via the EutBC ammonia lyase into respective carbon and nitrogen sources. Enzymes are highlighted in green (B) Genetic organization of the eut long- and short-locus of K. pneumoniae isolate KPPR1S with the deleted genes highlighted. (C) In vitro aerobic (i & iii) and anaerobic (ii & iv) growth kinetics as depicted by CFU/mL over time of the wild-type (WT) K. pneumoniae or single locus mutants (ΔeutCs & ΔeutCL) grown with EA as the sole nitrogen (i & ii) or the sole carbon (iii & iv) source at 37°C. Strains were grown in M9 minimal media (MM) supplemented with 0.4% Gly and 2.5 mM EA (i & ii) or 20 mM EA and 10 mM NH4Cl (iii & iv). All strains grown with EA as a nutrient source included 200nM B12. Mean ± SEM for ≥ 3 independent experiments is shown. N = nitrogen; C = carbon.
Fig 2.
EA present in the gut is metabolized by gut commensals, and it provides K. pneumoniae with a growth advantage.
(A) EA metabolizing capacity of K. pneumoniae WT, the isogenic mutants ΔeutCL/CS, ΔeutCS, ΔeutCL, and gut commensals using an agar plate colorimetric assay detecting acetaldehyde, a byproduct of EA metabolism. Strains were grown for 20 hours at 37°C on M9 MM (0.4% Gly, 10 mM NH4Cl), M9 MM with casamino acids (CAA), and De Man–Rogosa–Sharpe (MRS) agar, all containing 10 mM EA, and acetaldehyde quantified as described in Material and Methods section. E. coli Nissle (Nissle), E. coli healthy scientist (HS), murine E coli isolate (GH2), Acinetobacter sp. (GH4), Acinetobacter junii (GH5), Lactococcus lactis (L. lactis), Lactobacillus fermentum (L. fermentum), Lactobacillus animalis (L. animalis), and Lactobacillus gasseri (L. gasseri). Mean ± SEM for ≥ 3 independent experiments shown. (B) In vitro competition experiments between the WT and the ΔeutCL/CS K. pneumoniae. Both strains were inoculated 1:1 in cecal filtrate (CF), and grown at 37°C under aerobic or anaerobic conditions. CFU were enumerated, and CI values determined at 4, 8 and 24 hours post-inoculation. Shown is mean ± SEM from ≥ 3 independent experiments. Statistical significance of CIs were calculated using Wilcoxon signed-rank test with a theoretical median of 0. (C) LC MS/MS data showing the concentration of EA in the colonic extracts of mice either mock-infected (uninfected) or infected with K. pneumoniae. Boxes and whiskers with data points from individual mice shown with mean and minimum to maximum values. Each symbol represents a single mouse. (D) qRT-PCR showing eut loci expression in the murine GI tract. Shown is fold-change in transcription of eat, eutBS, eutCS (short eut locus) and eutBL, eutCL (long eut locus) comparing RNA isolated from cecal contents from either K. pneumoniae-inoculated mice or grown in M9 MM. Mice were inoculated by oral feeding of 106 CFU, and cecal contents harvested 15 days post-inoculation. KPPR1S 16S (rrsA)-specific primers were used as housekeeping gene for 2-ΔΔCT analysis. The values were further normalized to gyrA expression. For each biological replicate (n ≥ 3 for in vitro and in vivo samples), qRT-PCR was conducted in triplicate. Shown is mean ± SEM. *, P ≤ 0.05; ns, not significant.
Fig 3.
EA metabolism provides K. pneumoniae an advantage in the GI tract.
(A) Fecal shedding of K. pneumoniae inoculated mice. Mice were orally inoculated with 106 CFU of the WT K. pneumoniae, or the isogenic double locus mutant (ΔeutCL/CS) (n ≥ 10 for each group). On the indicated days, K. pneumoniae was enumerated from collected feces. Each symbol represents a single mouse on a given day. Bars, indicate medial bacterial shedding, and dashed line indicates limit of detection. A Mann-Whitney U test was performed between the WT and the ΔeutCL/CS at each time point. (B-F) Fecal shedding from in vivo competition experiments. Mice were orally inoculated with a 1:1 mixture of the WT and the ΔeutCL/CS strain (B), the WT and the eut short locus mutant ΔeutCS strain (C), the WT and the eut long locus mutant ΔeutCL strain (D), the WT and the chromosomally complemented strain (eutCL+) (E), the long locus mutant (ΔeutCL::kan) and the short locus mutant (ΔeutCS) (F), and fecal shedding was enumerated on the indicated days (n ≥ 8). The competitive index (CI) for each day was determined as described in Materials and Methods section. Each symbol represents a single mouse on a given day. Values above the dashed line indicate the mutant outcompeted the WT, whereas values below the dashed line indicate WT outcompeted the mutant. Bars indicate median value. Statistical significances of CIs were calculated using Wilcoxon signed-rank test with a theoretical median of 0. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001.
Fig 4.
K. pneumoniae EutR regulates the long but not the short eut locus (A) In vitro qRT-PCR measuring the transcripts of genes from the eut long (eutBL & eutCL) and short (eat, eutBS, eutCS) locus present in the WT and the eutR- isogenic mutant. Strains were either grown with EA provided as a nitrogen source (test) or ammonia (control). KPPR1S gyrA-specific primers were used as housekeeping gene for 2-ΔΔCT analysis. Mann-Whitney U test was performed comparing eut loci transcript abundance in the WT and eutR- strain background. (B-C) GFP kinetic assay of the WT and the eutR- mutant strain carrying plasmid with eutL promoter transcriptional gfp fusion (eutL΄-gfp+). (D-E) GFP kinetic assay of the WT and the eutR- mutant strain carrying plasmid with eutS promoter transcriptional gfp fusion (eutS΄-gfp+). (B&D) Strains were grown in M9 MM + 0.4% Gly and 10 mM NH4Cl or M9 MM + 20 mM EA and 10 mM NH4Cl (EA as C). (C&E) Strains were grown in M9 MM + 0.4% Gly and 10 mM NH4Cl or M9 MM + 0.4% Gly and 2.5 mM EA. All strains grown with EA as a nutrient source included 200nM B12. ****, P ≤ 0.0001; ns, not significant.
Fig 5.
K. pneumoniae eut short locus expression is regulated in an NtrC-dependent manner.
(A) (i) Shown are the putative NtrC and RpoN binding sites present in the promoter region upstream of the eut short locus. The conserved NtrC and RpoN binding elements are shown, with the conserved -12(GC) and the -24 (GG) RpoN binding sites capitalized. NtrC binding site selected for site-directed mutagenesis (SDM) highlighted within box. (ii) Proposed binding of NtrC and RpoN when EA is present as a nitrogen source. (B-C) GFP kinetic assay of the WT and the ntrC- mutant strain carrying plasmid with gfp transcriptional fusion to either (B) the eutS promoter (euts΄-gfp+) or (C) eutL promoter (eutL΄-gfp+). Strains were grown in M9 MM + 0.4% Gly and 2.5 mM EA (experimental) or M9 MM + 0.4% Gly & 10 mM NH4Cl (control). (D) qRT-PCR measuring transcript abundance of the eut short (eat) and long (eutBL, eutCL) locus in the WT and the ntrC- isogenic mutant background. Strains were grown in M9 MM + 0.4% Gly and 2.5 mM EA (experimental) or M9 MM + 0.4% Gly & 10 mM NH4Cl (control). KPPR1S gyrA-specific primers were used as housekeeping gene for 2-ΔΔCT analysis. Mann-Whitney U test was performed comparing eut loci transcript abundance in the WT and ntrC- strain background. (E) GFP kinetic assay of WT strain carrying plasmid with gfp fused to either the WT eutS promoter region (eutS΄-gfp+) or the eutS promoter with the putative NtrC binding element mutated through SDM. Strains were grown in M9 MM + 0.4% Gly and 2.5 mM EA (experimental) or M9 MM + 0.4% Gly & 10 mM NH4Cl (control). All strains grown with EA as a nutrient source included 200nM B12. Mean ± SEM for ≥ 3 independent experiments is shown. *, P ≤ 0.05; ns, not significant.
Fig 6.
The short eut locus is conserved among the K. pneumoniae species complex but presence of the long locus varies by taxa.
Heatmap gene arrows showing the distribution of the eut loci across the entire K. pneumoniae Species Complex (KpSC), ordered by phylogeny. The darker blue indicates higher proportion of loci presence within a given taxon. Genes are shown as arrows, while putative regulator binding sites are shown as raised directional arrows. Red crosses indicate complete absence of a genetic feature in a taxon: Ka—K. africana. Kp–K. pneumoniae. Kqq–K. quasipneumoniae subsp. quasipneumoniae. Kqs–K. quasipneumoniae subsp. similipneumoniae. Kqv–K. quasivariicola. Kvt–K. variicola subsp. tropica. Kvv–K. variicola subsp. Variicola.
Fig 7.
The eut long locus is present within a metabolism-focused region of the chromosome.
Gene comparison map showing typical arrangements between the ypfJ and yybH flanking genes. Two out of 5 K. variicola subsp. tropica isolates lacked the operon, along with 235 out of 291 K. quasipneumoniae subsp. similipneumoniae genomes.19Ka—K. africana. Kp–K. pneumoniae. Kqq–K. quasipneumoniae subsp. quasipneumoniae. Kqs–K. quasipneumoniae subsp. similipneumoniae. Kqv–K. quasivariicola. Kvt–K. variicola subsp. tropica. Kvv–K. variicola subsp. variicola.
Table 1.
Strains used in the studya.