The Antimicrobial Peptide Human Beta-Defensin 2 Inhibits Biofilm Production of Pseudomonas aeruginosa Without Compromising Metabolic Activity

Abstract

Biofilm production is a key virulence factor that facilitates bacterial colonization on host surfaces and is regulated by complex pathways, including quorum sensing, that also control pigment production, among others. To limit colonization, epithelial cells, as part of the first line of defense, utilize a variety of antimicrobial peptides (AMPs) including defensins. Pore formation is the best investigated mechanism for the bactericidal activity of AMPs. Considering the induction of human beta-defensin 2 (HBD2) secretion to the epithelial surface in response to bacteria and the importance of biofilm in microbial infection, we hypothesized that HBD2 has biofilm inhibitory activity. We assessed the viability and biofilm formation of a pyorubin-producing Pseudomonas aeruginosa strain in the presence and absence of HBD2 in comparison to the highly bactericidal HBD3. At nanomolar concentrations, HBD2 - independent of its chiral state - significantly reduced biofilm formation but not metabolic activity, unlike HBD3, which reduced biofilm and metabolic activity to the same degree. A similar discrepancy between biofilm inhibition and maintenance of metabolic activity was also observed in HBD2 treated Acinetobacter baumannii, another Gram-negative bacterium. There was no evidence for HBD2 interference with the regulation of biofilm production. The expression of biofilm-related genes and the extracellular accumulation of pyorubin pigment, another quorum sensing controlled product, did not differ significantly between HBD2 treated and control bacteria, and in silico modeling did not support direct binding of HBD2 to quorum sensing molecules. However, alterations in the outer membrane protein profile accompanied by surface topology changes, documented by atomic force microscopy, was observed after HBD2 treatment. This suggests that HBD2 induces structural changes that interfere with the transport of biofilm precursors into the extracellular space. Taken together, these data support a novel mechanism of biofilm inhibition by nanomolar concentrations of HBD2 that is independent of biofilm regulatory pathways.

Keywords: Pseudomonas aeruginosa; airways; antimicrobial peptides; biofilm; cystic fibrosis; epithelial cells; innate immunity; mucosa.

FIGURE 1

Metabolic activity of P. aeruginosa in the presence and absence of HBD2 and HBD3 over 18 h. Bacteria were incubated in 10% Mueller-Hinton/140 mM NaCl supplemented with 0.01% resazurin and fluorescence emitted by resorufin reflecting the production of reducing metabolites was measured every 3 h (530 nm_ex, 616 nm_em). Shown are the means ± SD of three independent experiments conducted in duplicates. RFU: relative fluorescence units. p < 0.001 for HBD2 (A) at 1, 2, and 4 μM and for HBD3 (B) at 0.25, 0.5, and 1 μM compared to the solvent control in univariate ANOVA with Bonferrroni post hoc analysis. All other concentrations were not significantly different from the solvent controls.

FIGURE 2

Comparative effects of HBD2 and HBD3 on P. aeruginosa biofilm and metabolic activity. Shown are biofilm formation and accumulated resorufin fluorescence after 18 h of incubation with HBD2 (A) and HBD3 (B) at the concentrations given. Data are expressed relative to the control and represent means ± SD of three independent experiments conducted in triplicates. ***p = 0.0004 in Two-way ANOVA. N.S: not significant (p = 0.7721).

FIGURE 3

ATP quantification in P. aeruginosa after 18 h incubation in the presence or absence of HBD2 and HBD3 at the concentrations given. ATP concentrations are in nM and were calculated based on a standard curve. Shown are means ± SD of three independent experiments conducted in duplicates. p = 0.01 in One way ANOVA with Bonferroni post hoc analysis for 2 μM HBD2 compared to 2 μM HBD3.

FIGURE 4

Effects of HBD2 on A. baumannii biofilm formation and metabolic activity. Shown are crystal violet absorbance and accumulated resorufin fluorescence expressed as % of the control after 18 h of incubation with HBD2 at the concentrations given. Data represent means ± SD of three independent experiments conducted in triplicates. **p = 0.004 for biofilm reduction versus reduction of metabolic activity in two tailed Paired Samples Test. In Oneway ANOVA with Bonferroni post hoc analysis, p = 0.001 for resazurin reduction at 4 μM HBD2, and p < 0.001 for biofilm reduction at 1, 2, and 4 μM HBD2, compared to the solvent control. All other data points were not significantly different from the control.

FIGURE 5

Comparative effects of D-HBD2 and linear HBD2 on P. aeruginosa biofilm and metabolic activity. Shown are biofilm formation and accumulated resorufin fluorescence expressed as % of the control after 18 h of incubation with all D-HBD2 (A) and linear HBD2 (B) at the concentrations given. Data represent means ± SD of three independent experiments conducted in triplicates. In paired T test comparing biofilm reduction and reduction of metabolic activity, ***p < 0.001 for D-HBD2 (A) and not significant (N.S.) for linear HBD2 (B). In Oneway ANOVA with Bonferroni post hoc analysis, biofilm formation (p = 0.033) but not metabolic activity (p = 0.473) is significantly reduced by D-HBD2. For linear HBD2, none of the data is significantly different from the solvent control.

FIGURE 6

In silico docking and binding energies (ΔG) of various QS molecules calculated for LasR and HBD2. AutoDock Vina was used to predict binding sites and potential hits for HBD2 and quorum sensing molecules in comparison to LasR. (A) Test N-(3-oxohexanoyl) homoserine lactone (3-oxo-C12-HSL, green) lies inside the LasR binding pocket in the same region as co-crystallized 3-oxo-C12-HSL (blue) with LasR (RSCB 3IX3). (B) HBD2 does not contain a binding pocket for test 3-oxo-C12-HSL (green). Free energy of binding (ΔG) for various hits were determined for phosphorylcolamine (NEtP), 3-oxo-C12-HSL, N-butyryl homoserine lactone (C4-HSL), and 2-heptyl-3-hydroxy-4-quinolone (PQS) as ligands with either LasR (C) or HBD2 (D) as rigid receptors. Dashed lines indicate the –6 kcal/mol threshold for actively bound molecules.

FIGURE 7

Relative gene expression of flgF and pslA in the presence and absence of 0.25 μM HBD2 as determined by qPCR. Gene expression of flgF (A) and pslA (B) in P. aeruginosa was calculated relative to the reference gene gapA after incubation in the presence or absence of HBD2 for up to 12 h. Shown are means ± SEM, n = 6. In multivariate ANOVA with Bonferroni post hoc analysis (*p < 0.05 and **p < 0.01), gene expression of flgF and pslA changed over time (Control: p < 0.01 for flgF 0.5 h versus 6 h and 12 h, and p < 0.05 for pslA 2 h versus 6 h and 12 h; HBD2: p < 0.05 for flgF 0.5 h versus 12 h) but there was no significant difference between the control and HBD2 treated bacteria.

FIGURE 8

Outer membrane protein profile of P. aeruginosa after 18 h incubation in the presence and absence of HBD2. (A) Four μL of concentrated outer membrane preparations from HBD2 treated (0.125–1 μM) or solvent control exposed bacteria (0) were resolved by SDS Tris-Tricine PAGE and visualized by silver stain. (Med) indicates medium only processed like bacteria-containing samples. The band migrating between 10 and 15 kDa in all samples is consistent with the expected molecular weight of lysozyme (14 kDa) that was added to the extraction buffer. (B) Approximate molecular weight and intensities of bands were quantified with Image Lab software and band intensities detected in both replicates were normalized to the intensity of the presumptive lysozyme band. Each data point represents the average of replicates. Each line represents the protein profile for the indicated HBD2 concentration (in μM).

FIGURE 9

Atomic force microscopy of P. aeruginosa after 18 h incubation in the presence and absence of 0.25 μM HBD2. Bacteria were incubated on glass slides and fixed with 2.5% glutaraldehyde prior to imaging. Images taken with the atomic force microscope were first order flattened before extracting measurement for bacterial roughness. (A) Representative images. CTRL: solvent control exposed bacteria. (B) Box and whisker chart (with inner points and outliers) of roughness measurements from multiple images of solvent exposed control bacteria (CTRL, n = 85) and 0.25 μM HBD2 treated bacteria (n = 69). ***p < 0.001 in independent samples T test.