Klebsiella pneumoniae DedA family proteins have redundant roles in divalent cation homeostasis and resistance to phagocytosis

Abstract

The DedA superfamily is a highly conserved family of membrane proteins. Deletion of Escherichia coli yqjA and yghB, encoding related DedA family proteins, results in sensitivity to elevated temperature, antibiotics, and alkaline pH. The human pathogen Klebsiella pneumoniae possesses genes encoding DedA family proteins with >90% amino acid identity to E. coli YqjA and YghB. We hypothesized that the deletion of K. pneumoniae yqjA and yghB will impact its physiology and may reduce its virulence. The K. pneumoniae ΔyqjA ΔyghB mutant (strain VT101) displayed a growth defect at 42°C and alkaline pH sensitivity, not unlike its E. coli counterpart. However, VT101 retained mostly wild-type resistance to antibiotics. We found VT101 was sensitive to the chelating agent EDTA, the anionic detergent SDS, and agents capable of alkalizing the bacterial cytoplasm such as bicarbonate or chloroquine. We could restore growth at alkaline pH and at elevated temperature by addition of 0.5-2 mM Ca²⁺ or Mg²⁺ to the culture media. VT101 displayed a slower uptake of calcium, which was dependent upon calcium channel activity. VT201, with similar deletions as VT101 but derived from a virulent K. pneumoniae strain, was highly susceptible to phagocytosis by alveolar macrophages and displayed a defect in the production of capsule. These findings suggest divalent cation homeostasis and virulence are interlinked by common functions of the DedA family.IMPORTANCEKlebsiella pneumoniae is a dangerous human pathogen. The DedA protein family is found in all bacteria and is a membrane transporter often required for virulence and antibiotic resistance. K. pneumoniae possesses homologs of E. coli YqjA and YghB, with 60% amino acid identity and redundant functions, which we have previously shown to be required for tolerance to biocides and alkaline pH. A K. pneumoniae strain lacking yqjA and yghB was found to be sensitive to alkaline pH, elevated temperature, and EDTA/SDS and displayed a defect in calcium uptake. Sensitivity to these conditions was reversed by addition of calcium or magnesium to the growth medium. Introduction of ΔyqjA and ΔyghB mutations into virulent K. pneumoniae resulted in the loss of capsule, increased phagocytosis by macrophages, and a partial loss of virulence. These results show that targeting the Klebsiella DedA family results in impaired divalent cation transport and, in turn, loss of virulence.

Keywords: capsular polysaccharide; divalent cations; membrane transport; phagocytosis.

Fig 1

(A and B) Genomic map of K. pneumoniae yqjA (A) and yghB (B) showing their neighboring genes. Dashed lines indicate gene deletions in VT101 produced using λ red recombination (24). (C) Amino acid sequence alignment showing similarity between K. pneumoniae YqjA and YghB. Amino acid alignment was carried out using Clustal Omega (25). Acidic amino acids (E39, D51) and basic amino acids (R130, R136) in bold fonts, previously shown to be essential in E. coli YqjA and YghB (20, 26) are conserved in K. pneumoniae YqjA and YghB. Solid gray and black lines indicate predicted membrane domains using TMHMM (27). (D) VT101 displays a growth defect on Luria-Bertani plates compared to wild-type K. pneumoniae MKP103. Plates were incubated at 30°C for 24 hours.

Fig 2

VT101 is sensitive to alkaline pH, and Ca²⁺ or Mg²⁺ can suppress alkaline pH sensitivity. (A) VT101 struggles to grow at alkaline pH. VT101 transformed with control vector (vec), pBAD (KpYqjA), or pBAD (KpYghB) was grown overnight, serially diluted, and spotted on LB plates containing 0.1% arabinose and 50 µg/mL apramycin. (B) Ca²⁺ and Mg²⁺ promote the survival of VT101 at alkaline pH. MKP103 and VT101 transformed with control vector (vec) were diluted from overnight culture and spotted on LB plates containing the indicated concentration of CaCl₂ or MgSO₄. The pH of the solid media was adjusted using NaOH. Plates were incubated at 30°C for 24 hours.

Fig 3

VT101 is sensitive to chemical agents causing alkalization of the cytoplasm, and Ca²⁺ or Mg²⁺ can rescue this sensitivity. (A) VT101 struggles to grow in the presence of alkalizing agents. VT101 transformed with control vector (vec), pBAD (KpYqjA), or pBAD (KpYghB) was grown overnight, serially diluted, and spotted on LB plates containing the indicated concentration of sodium bicarbonate or CQ, 0.1% arabinose, and 50 µg/mL apramycin. (B) Ca²⁺ is more efficient in suppressing the effect of CQ compared to Mg²⁺. MKP103 and VT101 transformed with control vector (vec) were diluted from overnight culture and spotted on LB plates containing 4 mM CQ and the indicated concentration of CaCl₂ or MgSO₄. Single control plates containing LB or LB/CQ from the same internally controlled experiment are shown. Plates were incubated at 30°C for 24 hours.

Fig 4

VT101 shows growth sensitivity to elevated temperature, and Ca²⁺ or Mg²⁺ can rescue this sensitivity. (A) VT101 struggles to grow at 42°C. VT101 was transformed with control vector (vec), pBAD (KpYqjA) or pBAD (KpYghB) was grown overnight, serially diluted, and spotted on LB plates containing 0.1% arabinose and 50 µg/mL apramycin. (B) Ca²⁺ or Mg²⁺ can restore the growth of VT101 at 42°C. MKP103 and VT101 transformed with control vector (vec) were diluted from overnight culture and spotted on LB plates with indicated concentrations of CaCl₂ or MgSO₄. Plates were incubated at 30°C and 42°C for 24 hours.

Fig 5

K. pneumoniae VT101 fails to grow in the presence of EDTA or SDS. (A) Sensitivity of VT101 to EDTA. Dilutions of indicated strains from overnight cultures were spotted and grown on LB plates with indicated concentration of EDTA, 0.1% arabinose, and 50 µg/mL apramycin. (B) Sensitivity of VT101 to SDS. Dilutions of indicated strains from overnight cultures were spotted and grown on LB plates with indicated concentration of SDS, 0.1% arabinose, and 50 µg/mL apramycin. Plates were incubated at 30°C for 36 hours.

Fig 6

K. pneumoniae VT101 displays significant hyperpolarization of the membrane potential (ΔΨ). Treatment of MKP103 and VT101 with 25 µM CCCP for 30 min resulted in the loss of ΔΨ was used as the control. Bars represent mean ± SEM of nine biological replicates. The experiment was repeated three times. The Mann-Whitney test was used to perform statistical analysis. Statistical comparisons were considered significant at P < 0.05. ***, P < 0.001.

Fig 7

K. pneumoniae VT101 shows a slower uptake of Ca²⁺. (A) Calcium uptake was measured in MKP103 and VT101 after challenging with 1 mM CaCl₂. (B and C) MKP103 and VT101 were treated with 600 µM lanthanum chloride (LaCl₃), calcium channel blocker, before challenging the cells with 1 mM CaCl₂. (D) Comparing the intracellular calcium between wild type treated with LaCl₃ and VT101. The (apo)aequorin expression was induced for 8 hours using 0.5% arabinose. After reconstitution of aequorin using 5 µM coelenterazine, cells were challenged with a final concentration of 1 mM CaCl₂ to study Ca²⁺ homeostasis. In both wild type and VT101, 1 mM CaCl₂ was introduced at 60 s time point.

Fig 8

K. pneumoniae VT201 possesses less capsule than wild type. To measure the capsular content of VT201 (A) and wild-type ST258 (B), a semiquantitative Percoll gradient assay was used with low-speed centrifugation on a three-layer gradient of 15%, 35%, and 50%. A representative gradient is shown. The experiment was repeated three times.

Fig 9

YqjA and YghB are required to resist phagocytosis by murine alveolar macrophage cells (MH-S). VT201 was phagocytosed in significantly higher numbers compared to wild-type ST258. Overnight cultures of wild type and VT201 were adjusted to OD600 = 1.0 and used to infect the AMs (MH-S) with multiplicity of infection of 1:10. Infected MH-S cells were incubated at 37°C with 5% CO₂. Bacterial cells were harvested at selected time points by lysing MH-S cells, and serial dilution was carried out to estimate the CFU. Bars represent mean ± SEM of nine biological replicates. The experiment was repeated three times. The Mann-Whitney test was used to perform statistical analysis. Statistical comparisons were considered significant at P < 0.05. ***, P < 0.001; ****, P < 0.0001.

Fig 10

YqjA and YghB are required for virulence of K. pneumoniae in Galleria mellonella larvae. (A) Larva injected with sterile PBS pH 7.4 resulted in no death was used as the control. (B) Larvae injected with wild-type ST258 cell suspension made in PBS pH 7.4. (C) Larvae injected with VT201 cell suspension made in PBS pH 7.4. (D) Kaplan-Meier survival curve showing the killing rate of larvae was slower by VT201 compared to wild-type ST258. Each experiment was repeated three times.