The antimicrobial polymer PHMB enters cells and selectively condenses bacterial chromosomes

Abstract

To combat infection and antimicrobial resistance, it is helpful to elucidate drug mechanism(s) of action. Here we examined how the widely used antimicrobial polyhexamethylene biguanide (PHMB) kills bacteria selectively over host cells. Contrary to the accepted model of microbial membrane disruption by PHMB, we observed cell entry into a range of bacterial species, and treated bacteria displayed cell division arrest and chromosome condensation, suggesting DNA binding as an alternative antimicrobial mechanism. A DNA-level mechanism was confirmed by observations that PHMB formed nanoparticles when mixed with isolated bacterial chromosomal DNA and its effects on growth were suppressed by pairwise combination with the DNA binding ligand Hoechst 33258. PHMB also entered mammalian cells, but was trapped within endosomes and excluded from nuclei. Therefore, PHMB displays differential access to bacterial and mammalian cellular DNA and selectively binds and condenses bacterial chromosomes. Because acquired resistance to PHMB has not been reported, selective chromosome condensation provides an unanticipated paradigm for antimicrobial action that may not succumb to resistance.

Figure 1. PHMB effects on cell membrane permeability and entry into bacteria.

(a) Structure of polyhexamethylene biguanide (PHMB; CAS# 27083-27-8); alternative chemical names: polyhexanide; example trade names: Vantocil™, Cosmocil™, Baquacil™, Prontosan^®. See Supplementary Table 1 for further details. PHMB is composed of repeating basic biguanidine units connected by hexamethylene hydrocarbon chains, providing a cationic and amphipathic structure with a high capacity for hydrogen bonding, electrostatic and hydrophobic interactions. PHMB preparations typically comprise polymers of mixed length with amine, guanidine and cyanoguanidine end groups (eg. average n = 13.8, 3,035 g/mol44) (b) Effects of PHMB, heat, polymyxin B (positive control) and triclosan (negative control) on cell permeability to SYTOX^®Green. The MIC values for the antibacterials tested are indicated with colour-coded vertical arrows, at top. (c) Fluorescence microscopy of PHMB-FITC entry into diverse bacteria. PHMB-FITC (2 μg/mL) was added to bacterial cultures and the cells were counter stained with DAPI. (d) Confocal image showing localisation of PHMB-FITC (green) in B. megaterium; bacteria were counterstained with the membrane localising probe wheat germ agglutinin (WGA) conjugated to Alexa Fluor-555 (red) and visualized as live (top) and fixed (bottom) cells; Bar = 5 μm. (e) Fluorescence intensity profile plot analysis of cellular localisation of PHMB-FITC and WGA fluorescence (the white line indicated the cross section used for analysis). The green line indicates the levels of FITC and position (mainly within the cell). The red line indicates the levels of WGA and position (mainly within the membrane).

Figure 2. PHMB-mediated cell elongation and chromosome condensation in bacteria.

(a) E. coli was treated with PHMB for 90 minutes and examined by bright field microscopy. (b) Mean cell length as a function of PHMB concentration. MIC (arrow) is indicated. (c) Pattern of chromosome distribution in cells following PHMB-FITC treatment. Cultures of E. coli, strain K-12 were treated with PHMB-FITC, counter stained with DAPI and examined using fluorescence microscopy. Chromosomes appear as condensed DAPI-stained foci; more apparent in the enlarged image. (d) Pattern of chromosome distribution in filamentous/multinucleated E. coli following PHMB exposure. RNA silencing of ftsZ expression was used to arrest cell division, and cells were then untreated or treated with PHMB, stained with DAPI and examined by fluorescence microscopy. (e) Pattern of chromosome distribution in B. megaterium cells that were untreated or treated with PHMB, stained with DAPI and WGA-red and examined using fluorescence microscopy.

Figure 3. Effects of PHMB on bacterial SOS responses.

(a) Chromosome condensation in E. coli strains SS996 (sulB103; FtsZ mutant insensitive to SulA), JW2669 (recA⁻; knock-out of recA) and AB2474 (lexA1, mutation that prevents SOS response induction) following treatment with PHMB for 2 hours. Cells were DAPI stained to reveal DNA. (b) SOS response reporter expression, quantified by fluorimetry. The SOS reporter E. coli strain SS996 carrying a chromosomal sulAp-gfp fusion was untreated or treated with PHMB, mitomycin C, a known SOS inducer, or triclosan, which does not induce an SOS response. The MIC values against SS996 were PHMB, 0.75 μg/mL; triclosan, 2 μg/mL; mitomycin C, 0.06 μg/mL, and these values were used to calculate %MIC.

Figure 4. PHMB binding to bacterial chromosomal DNA in vitro.

(a) PHMB or PHMB-FITC was mixed with isolated chromosomal DNA from E. coli strain K-12, and samples were analysed by EMSA. Patterns of retarded in-gel mobility indicate DNA binding by PHMB. (b) PHMB-mediated exclusion of SYTOX®Green binding to isolated E. coli chromosomal DNA, where reduced fluorescence indicates DNA binding by PHMB. (c) Circular dichroism spectroscopy of mixtures of PHMB and isolated E. coli chromosomal DNA. (d) Plot of ellipticity as a function of PHMB:DNA ratios.

Figure 5. Model for the antibacterial mechanism of action of PHMB, and suppression of growth inhibition.

(a) Proposed antibacterial mechanism of action of PHMB. (b) Pairwise growth inhibition interactions between PHMB and Hoechst 33258 and negative control non-DNA-binding ligands (triclosan and trimethoprim) in diverse bacterial species. (c) Relationship between bacterial genome AT-content and antibacterial interactions with PHMB. Plot of growth inhibition interactions and DNA AT-content in diverse species. Interaction values are fractional inhibitory concentration indicies (FICI) between PHMB and Hoechst 33258 or negative control non-DNA-binding ligands (trimethoprim and triclosan). (d) Bacillus megaterium growth inhibition by PHMB and suppression by combinations with the DNA ligand Hoechst 33258 (blue lines). See Supplementary Table 4 for species list and inhibition values.

Figure 6. PHMB entry into mammalian cells.

(a) Primary fibroblasts were treated with PHMB-FITC (3.5 μg/mL), counter stained with Hoechst 33258 and observed by fluorescence microscopy. (b) Flow cytometry analysis of a panel of mammalian cells treated with PHMB-FITC. Inset: a representative example of a flow cytometry histogram of HeLa cell populations that were untreated (purple population) or treated with PHMB-FITC (0.4 μg/mL) (green population). (c) HeLa cells were treated with PHMB-FITC (3.5 μg/mL) and chloroquine (0–20 μM) for 2 hours. The effects of chloroquine on fluorescence were measured by flow cytometry, using geometric mean fluorescence intensity (arbitrary units (A.U.), log scale).