Analysis of the networks controlling the antimicrobial-peptide-dependent induction of Klebsiella pneumoniae virulence factors

Abstract

Antimicrobial peptides (APs) impose a threat to the survival of pathogens, and it is reasonable to postulate that bacteria have developed strategies to counteract them. Polymyxins are becoming the last resort to treat infections caused by multidrug-resistant Gram-negative bacteria and, similar to APs, they interact with the anionic lipopolysaccharide. Given that polymyxins and APs share the initial target, it is possible that bacterial defense mechanisms against polymyxins will be also effective against host APs. We sought to determine whether exposure to polymyxin will increase Klebsiella pneumoniae resistance to host APs. Indeed, exposure of K. pneumoniae to polymyxin induces cross-resistance not only to polymyxin itself but also to APs present in the airways. Polymyxin treatment upregulates the expression of the capsule polysaccharide operon and the loci required to modify the lipid A with aminoarabinose and palmitate with a concomitant increase in capsule and lipid A species containing such modifications. Moreover, these surface changes contribute to APs resistance and also to polymyxin-induced cross-resistance to APs. Bacterial loads of lipid A mutants in trachea and lungs of intranasally infected mice were lower than those of wild-type strain. PhoPQ, PmrAB, and the Rcs system govern polymyxin-induced transcriptional changes, and there is a cross talk between PhoPQ and the Rcs system. Our findings support the notion that Klebsiella activates a defense program against APs that is controlled by three signaling systems. Therapeutic strategies directed to prevent the activation of this program could be a new approach worth exploring to facilitate the clearance of the pathogen from the airways.

Fig. 1.

Exposure of K. pneumoniae 52145 to PxB increases the resistance to antimicrobial peptides. After PxB treatment, bacteria were washed, and the susceptibility to PxB (A), β-defensin 1 (B), β-defensin 2 (C), HNP-1 (D), or magainin II (E) was tested by the survival assay. Each point represents the mean and standard deviation of eight samples from four independently grown batches of bacteria, and significant survival differences (P < 0.05 [two-tailed t test]) between PxB-pretreated (solid symbols) and untreated (open symbols) bacteria are indicated by asterisks.

Fig. 2.

Treatment of the K. pneumoniae 52145 cps mutant with PxB increases resistance to antimicrobial peptides. (A) Analysis of the expression of the cps operon by Kp52145 carrying the fusion cps::lucFF. The strain was treated with PxB for 1 h (▪) or not treated (□). The data are presented as means ± the standard deviations (n = 6). *, Results are significantly different (P < 0.05; two-tailed t test) from the results for nontreated bacteria. (B to F) After treatment of 52145-Δwca_K2 with PxB for 1 h, the bacteria were washed and exposed to different concentrations of PxB (B), β-defensin 1 (C), β-defensin 2 (D), HNP-1 (E), and magainin II (F). Each point represents the mean and standard deviation of eight samples from four independently grown batches of bacteria, and significant survival differences (P < 0.05 [two-tailed t test]) between bacteria pretreated with PxB (solid symbols) and untreated bacteria (open symbols) are indicated by asterisks.

Fig. 3.

Exposure of K. pneumoniae 52145 to PxB affects the lipid A structure. Negative ion MALDI-TOF mass spectrometry spectra of lipid A isolated from Kp52145 that was treated with 65 ng of PxB/ml for 12 h (B) or not treated (A). The results in the panels are representative of three independent lipid A extractions. (C) Proposed structures corresponding to major peaks and acyl group positions follow previously reported structures for Klebsiella (11) and other Gram-negative bacteria.

Fig. 4.

Lipid A analysis from K. pneumoniae lipid A mutants. Negative-ion MALDI-TOF mass spectrometry spectra of lipid A isolated from the indicated K. pneumoniae strains treated with 65 ng of PxB/ml for 12 h (B, D, and F) or not treated (A, C, and E) are shown. The results in all panels are representative of three independent lipid A extractions.

Fig. 5.

PxB induces the expression of K. pneumoniae 52145 ugd, pmrF, and pagP loci. The expression of the loci implicated in lipid A remodeling was analyzed by measuring the luciferase activity of Kp52145 carrying ugd::lucFF, pmrF::lucFF or pagP::lucFF transcriptional fusions, which were either treated with PxB for 1 h (▪) or not treated (□). The data are presented as means ± the standard deviations (n = 5). *, Results are significantly different (P < 0.05 [two-tailed t test]) from the results for nontreated bacteria.

Fig. 6.

Roles of K. pneumoniae 52145 capsule and lipid A modifications on the resistance to antimicrobial peptides. (A) pmrF mutants were exposed to different concentrations of PxB; and magainin II. (B) pagP mutants were exposed to different concentrations of magainin II. Each point represents the mean and the standard deviation of eight samples from four independently grown batches of bacteria. Symbols: □, Kp52145; •, 52145-Δwca_K2; ▪, 52145-ΔpmrF; ▵, 52145-Δwca_K2-ΔpmrF; ▿, 52145-ΔpagPGB; ▾, 52145-Δwca_K2-ΔpagPGB.

Fig. 7.

Virulence of K. pneumoniae 52145 lipid A mutants. The bacterial counts in mouse organs at 24 h (A) or 96 h (B) postinfection were determined. Mice were infected intranasally with a bacterial mixture containing 4.1 × 10⁴ bacteria of the wild-type strain (Kp52145, •), 4.5 × 10⁴ bacteria of 52145-ΔpmrF (pmrF, ○), 5.1 × 10⁴ bacteria of 52145-ΔpagPGB (pagP, □), and 4.8 × 10⁴ bacteria of 52145-ΔpmrF-ΔpagPGB (pmrF-pagP, ▵), respectively. The results were reported as log CFU per gram of tissue (log CFU/g). *, Results are significantly different (P < 0.05 [two-tailed t test]) from the results for Kp52145.

Fig. 8.

K. pneumoniae PhoPQ and PmrAB two-component systems control PxB-induced transcriptional changes. Analysis of the expression of cps, pmrH, ugd, pagP, mgtA, and pmrD loci by Kp52145 (WT), 52145-ΔrcsB (rcsB), 52145-ΔphoQGB (phoQ), 52145-ΔpmrAB (pmrAB), 52145-ΔpmrD (pmrD), and 52145-ΔphoQGB-ΔpmrAB (phoQ-pmrAB) carrying the transcriptional fusions cps::lucFF (A), pmrF::lucFF (B), ugd::lucFF (C), pagP::lucFF (D), mgtA::lucFF (E), and pmrD::lucFF (F) treated with PxB for 1 h (▪) or not treated (□). The data are presented as means ± the standard deviations (n = 3). *, Results are significantly different (P < 0.05 [two-tailed t test]) from the results for nontreated bacteria. ▵, Results are significantly different (P < 0.05 [two-tailed t test]) from the results for Kp52145 treated in the same manner.

Fig. 9.

There is cross talk between the Rcs and PhoPQ systems in K. pneumoniae 52145. Analysis of the expression of phoP, pagP, mgtA, pmrD, pmrH, ugd, rcsD, and rcsC loci by Kp52145 (WT), 52145-ΔrcsB (rcsB), 52145-ΔphoQGB (phoQ), and 52145-ΔrcsB-ΔphoQGB (rcsB phoQ) carrying the transcriptional fusions phoP::lucFF (A), pagP::lucFF (B), mgtA::lucFF (C), pmrD::lucFF (D), pmrH::lucFF (E), ugd::lucFF (F), rcsD::lucFF (G), and rcsC::lucFF (H) treated with PxB for 1 h (▪) or not treated (□). The data are presented as means ± the standard deviations (n = 3). *, Results are significantly different (P < 0.05 [two-tailed t test]) from the results for nontreated bacteria. ▵, Results are significantly different (P < 0.05 [two-tailed t test]) from the results for Kp52145 treated in the same manner.

Fig. 10.

Role of K. pneumoniae Rcs, PhoPQ, and PmrAB systems in bacterial susceptibility to PxB. Wild-type (Kp52145), 52145-ΔrcsB (rcsB), 52145-ΔphoQGB (phoQ), 52145-ΔpmrAB (pmrAB), 52145-ΔphoQGB-ΔpmrAB (phoQ-pmrAB), and 52145-ΔrcsB-ΔphoQGB (rcsB-phoQ) strains were exposed to different concentrations of PxB. Each point represents the mean and standard deviation of eight samples from four independently grown batches of bacteria.