Chemical communication of antibiotic resistance by a highly resistant subpopulation of bacterial cells

Abstract

The overall antibiotic resistance of a bacterial population results from the combination of a wide range of susceptibilities displayed by subsets of bacterial cells. Bacterial heteroresistance to antibiotics has been documented for several opportunistic Gram-negative bacteria, but the mechanism of heteroresistance is unclear. We use Burkholderia cenocepacia as a model opportunistic bacterium to investigate the implications of heterogeneity in the response to the antimicrobial peptide polymyxin B (PmB) and also other bactericidal antibiotics. Here, we report that B. cenocepacia is heteroresistant to PmB. Population analysis profiling also identified B. cenocepacia subpopulations arising from a seemingly homogenous culture that are resistant to higher levels of polymyxin B than the rest of the cells in the culture, and can protect the more sensitive cells from killing, as well as sensitive bacteria from other species, such as Pseudomonas aeruginosa and Escherichia coli. Communication of resistance depended on upregulation of putrescine synthesis and YceI, a widely conserved low-molecular weight secreted protein. Deletion of genes for the synthesis of putrescine and YceI abrogate protection, while pharmacologic inhibition of putrescine synthesis reduced resistance to polymyxin B. Polyamines and YceI were also required for heteroresistance of B. cenocepacia to various bactericidal antibiotics. We propose that putrescine and YceI resemble "danger" infochemicals whose increased production by a bacterial subpopulation, becoming more resistant to bactericidal antibiotics, communicates higher level of resistance to more sensitive members of the population of the same or different species.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Figure 1. Heterogeneous response of B. cenocepacia to PmB.

Population analysis profiling (PAP) of B. cenocepacia strains K56-2, K56-2ΔrpoE, K56-2ΔrpoE/500, CF706-J, K56-2ΔhldA, and K56-2ΔarnBC, and P. aeruginosa PAO1 by agar dilution at 24 h except for CF706-J by broth dilution at 18 h. n = 6. Panels (B) and (C), E-test showing discrete colonies at otherwise clear zones of inhibition, indicating heterogeneous response to PmB.

Figure 2. Protective effects of B. cenocepacia ΔrpoE/500 on P. aeruginosa PAO1, exposed to PmB, in co-culture.

The dotted line represents the limit of detection (50 cfu/ml). Three independent experiments each done in duplicate.

Figure 3. Proteins released into the supernatant of B. cenocepacia K56-2 and Δ rpoE/500 treated with 500 µg/ml of PmB compared to those released from untreated K56-2 and naïve ΔrpoE.

Proteins were run on 14% SDS-PAGE and detected by silver staining and those that were differentially expressed were further identified by LC-MS.

Figure 4. Contribution of the polyamine putrescine in the response to PmB.

(A) Polyamines biosynthetic pathway. (B) Exogenous putrescine increases the resistance of the parental K56-2 to PmB and restores the resistance to PmB to the parental level in ΔBCAL2641; n = 6 (C) The deletion of BCAL2641 leads to reduced resistance to PmB relative to the parental K56-2. (D) Single-copy complementation of ΔBCAL2641. (E) The polyamine synthesis inhibitor dicyclohexylamine (blue) reduces the resistance of B. cenocepacia K56-2 to PmB, with little to no effect of 3-(methylthio)propylamine (red), shown in a turbidimetric PAP at 24 h; n = 5. (F) TLC analysis of polyamines released in the supernatants of 20 h old M9 cultures compared to standards. (G) Relative amounts of putrescine released from the wild-type and mutants, n = 4.

Figure 5. Role of putrescine in the protective effects of B. cenocepacia against PmB.

(A) Involvement of putrescine in the protective effects of B. cenocepacia on P. aeruginosa PAO1 at 24 h. The dotted line represents the limit of detection (50 cfu/ml). Three independent experiments each done in duplicate. (B) Putrescine protects the bacterial surface from binding to PmB; 50 mM of putrescine reduced binding of PmB-Oregon green 514 conjugate (25 µg/ml) when both added together, whereas it could replace already bound PmB; n = 6.

Figure 6. Contribution of YceI in the response to PmB and its role in protection against PmB.

(A) The deletion of BCAL3310 and BCAL3311 (ΔyceI) leads to reduced resistance to PmB relative to the parental K56-2. (B) PAP by agar dilution showing complementation of the reduced resistance in ΔyceI mutant by yceI under the control of the rhamnose promoter on pSCrhaB2 to the parental level at 0.4% rhamnose. (C) Involvement of YceI in the protective effects of B. cenocepacia on P. aeruginosa PAO1 at 24 h. Three independent experiments each done in duplicate. (D) Purified YceI homologues, BCAL3310 and BCAL3311. (E) Binding of BCAL3310 and BCAL3311 to PmB-Oregon green 514 conjugate. BSA was used as a control for binding. n = 6.

Figure 7

Figure 8. Homogenous response of B. cenocepacia K56-2 to bacteriostatic antibiotics. n = 6.