Evidence for involvement of peptidoglycan in the triggering of an oxidative burst by Listeria monocytogenes in phagocytes

Abstract

We have shown previously that in listeric encephalitis of cattle and rats, nitrotyrosine was produced in microabscesses, implying that both superoxide anion (O(2) (-)) and nitric oxide (NO) are present and react with each other. Evidence of local synthesis of NO by macrophages was provided, but the source of O(2) (-) remained unknown. Here we have examined whether phagocytes exposed to viable and heat-killed Listeria monocytogenes (LMDelta) produce O(2) (-) and, if so, whether this results from direct interaction of phagocytes with the bacterial surface of L. monocytogenes or whether prior opsonization is required. Using lucigenin-enhanced chemiluminescence (LCL) for the measurement of O(2) (-), we show that LMDelta induces an oxidative burst in human neutrophils, monocytes and monocyte-derived macrophages (Mphi). Viability is not required, and opsonization by antibodies and/or complement does not enhance the LCL signal. As Toll-like receptors (TLR) were shown recently to mediate an oxidative burst, TLR agonists representative for pathogen-associated molecular patterns (PAMPs) were tested for their ability to elicit an oxidative burst. These included lipoteichoic acid (LTA), bacterial peptidoglycan (PGN), recombinant flagellin, CpG-containing DNA and double-stranded RNA. Only PGN and flagellin consistently elicited an LCL signal resembling that induced by LMDelta with regard to the kinetics and cell spectrum stimulated. However, flagellin was unlikely to be responsible for the LMDelta-mediated burst, as a flagellin-deficient mutant showed no decrease in LCL. We therefore assume that in LMDelta, core PGN acts as a PAMP and directly induces an oxidative burst in all phagocyte populations. We conclude that in cerebral lesions superoxide anion is generated locally by phagocytes recognizing bacterial PGN.

Fig. 1

LCL of PMN, PBMC and Mφ in response to LMΔ. Addition of superoxide dismutase to LMΔ-stimulated cells (LMΔ + SOD) abrogates the CL signal. Mean temporal traces of two from representative experiments (a) and average signal-to-background ratios of LMΔ-stimulated versus unstimulated cells integrated over the whole experiment (b) are shown. P-values indicate levels of statistical significance of difference.LCL of PMN, PBMC and Mφ by viable and heat-killed LM. Mean temporal traces of two from a representative experiment (a) and mean signal-to-background ratios of viable LM-stimulated versus LMΔ-stimulated cells integrated over the whole experiment (b) are shown for one representative experiment.

Fig. 2

LCL of PMN, PBMC and Mφ by viable and heat-killed LM. Mean temporal traces of 2 from a representative experiment (A) and mean signal-to-background ratios of viable LM-stimulated versus LMΔ-stimulated cells integrated over the whole experiment (B) are shown for one representative experiment.

Fig. 3

The effect of opsonization of bacteria with IgG or fresh-frozen serum on LCL and phagocytosis. (a) Mean temporal traces of two. (b) Signal-to-background ratio of the same LCL experiments. (c) Phagocytosis (average of three experiments; error bars indicate s.d.). Asterisks in (c) denote statistically significant differences when compared with nonopsonized bacteria (*P < 0·05, **P < 0·01).

Fig. 4

LCL in response to LMΔ, PGN, LTA, flagellin and CpG by PMN, PBMC and Mφ. (a) Mean temporal CL traces of a representative experiment are shown. (b) Average signal-to-background ratios (n = 3) for each cell type of LMΔ-stimulated versus unstimulated cells integrated over the whole experiment are shown. Asterisks denote statistically significant differences (*P < 0·05, **P < 0·01).

Fig. 5

Dose-dependent CL in response to the TLR4 ligand LPS. Representative temporal CL traces obtained with Mφ from an individual donor (averages of duplicate measurements) are shown. Symbols of individual curves are given in the figure.