Autoinducer 2 (AI-2) Production by Nontypeable Haemophilus influenzae 86-028NP Promotes Expression of a Predicted Glycosyltransferase That Is a Determinant of Biofilm Maturation, Prevention of Dispersal, and Persistence In Vivo

Abstract

Nontypeable Haemophilus influenzae (NTHi) is an extremely common human pathobiont that persists on the airway mucosal surface within biofilm communities, and our previous work has shown that NTHi biofilm maturation is coordinated by the production and uptake of autoinducer 2 (AI-2) quorum signals. To directly test roles for AI-2 in maturation and maintenance of NTHi biofilms, we generated an NTHi 86-028NP mutant in which luxS transcription was under the control of the xylA promoter (NTHi 86-028NP luxS xylA::luxS), rendering AI-2 production inducible by xylose. Comparison of biofilms under inducing and noninducing conditions revealed a biofilm defect in the absence of xylose, whereas biofilm maturation increased following xylose induction. The removal of xylose resulted in the interruption of luxS expression and biofilm dispersal. Measurement of luxS transcript levels by real-time reverse transcription-PCR (RT-PCR) showed that luxS expression peaked as biofilms matured and waned before dispersal. Transcript profiling revealed significant changes following the induction of luxS, including increased transcript levels for a predicted family 8 glycosyltransferase (NTHI1750; designated gstA); this result was confirmed by real-time RT-PCR. An isogenic NTHi 86-028NP gstA mutant had a biofilm defect, including decreased levels of sialylated matrix and significantly altered biofilm structure. In experimental chinchilla infections, we observed a significant decrease in the number of bacteria in the biofilm population (but not in effusions) for NTHi 86-028NP gstA compared to the parental strain. Therefore, we conclude that AI-2 promotes NTHi biofilm maturation and the maintenance of biofilm integrity, due at least in part to the expression of a probable glycosyltransferase that is potentially involved in the synthesis of the biofilm matrix.

Keywords: Haemophilus influenzae; biofilm; otitis media; quorum signal.

FIG 1

Generation of NTHi 86-028NP luxS xylA::luxS, with xylose-inducible luxS expression. (A) Schematic showing the generation of NTHi 86-028NP luxS xylA::luxS (WES204) depicting luxS under the control of the xylA promoter and flanking regions used for homologous recombination into the xyl region of the NTHi 86-028NP luxS chromosome. (B) H. influenzae strain WES204 was cultured in sBHI medium supplemented with increasing concentrations of d-xylose. Supernatant samples were taken at three times during growth for V. harveyi bioluminescence to compare AI-2 production under inducing conditions to that of 86-028NP. Values represent means, and error bars represent standard errors of the means (n = 6).

FIG 2

Induction of luxS expression increases phosphorylcholine expression. H. influenzae strains NTHi 86-028NP, NTHi 86-028NP luxS, and WES204 were cultured overnight on sBHI agar with or without 5 mM xylose. (A) Bacterial colonies were lifted onto nitrocellulose membranes and incubated with anti-PCho monoclonal antibody for determination of the percentage of colonies positive for PCho expression. (B) Whole-bacterium ELISA using anti-PCho monoclonal antibody for detection of surface-accessible PCho. Values represent means (n = 4), and error bars depict standard deviations. ***, P  < 0.0001.

FIG 3

Induction of luxS expression promotes biofilm development in a stationary system. Stationary biofilms were established by H. influenzae strains NTHi 86-028NP, NTHi 86-028NP luxS, and WES204 in sBHI medium with or without 5 mM xylose, as indicated, and stained with a BacLight live/dead reagent (Molecular Probes) for confocal microscopy. (A to D) Vertical z-series images were compiled to generate representative volume views of 12-h biofilms formed by NTHi 86-028NP (A), WES204 induced by xylose for 12 h (B), NTHi 86-028NP luxS (C), and uninduced WES204 (D). (E to J) z-series images of biofilms formed for 12 h (E and F), 24 h (G and H), or 48 h (I and J) were exported into Matlab for COMSTAT analysis of biofilm average thickness (E, G, and I) and total biomass (F, H, and J). Graphs were derived from 10 individual fields of view per well in 4 independent biofilms per wells, and error bars represent means and standard deviations.

FIG 4

Modulation of luxS expression alters biofilm development and dispersal. Biofilms were established by gfp-expressing WES204 under continuous-medium-flow conditions in sBHI medium or sBHI medium supplemented with 5 mM xylose, as indicated, and imaged by confocal microscopy at 24, 27, and 48 h postinoculation for COMSTAT analysis of total biofilm biomass (A) and average thickness (B). At 24 h postinoculation, the medium source was switched, when indicated, from sBHI medium to sBHI medium supplemented with xylose or vice versa and maintained for the remainder of the experiment. Data were derived from 3 independent continuous-flow biofilms per variable, and error bars represent means and standard deviations.

FIG 5

Transcription of luxS decreases as NTHi 86-028NP biofilms approach dispersal. NTHi 86-028NP stationary biofilms were established in sBHI medium for 6, 12, 24, 48, and 72 h for isolation of RNA and measurement of luxS transcripts by real-time RT-PCR. Data were normalized and are expressed as ratios of luxS to gyrA. Each measure represents data for 5 independent samples, and error bars represent means and standard deviations.

FIG 6

Continuous luxS expression prevents NTHi 86-028NP biofilm dispersal. (A) Biofilms were established by H. influenzae gfp-expressing strains NTHi 86-028NP and WES204 for 96 h under continuous medium flow, with confocal microscopy every 24 h for COMSTAT analysis of average biofilm thickness. NTHi 86-028NP biofilms were cultured in sBHI medium, and WES204 was cultured in sBHI medium supplemented with 5 mM xylose. Values were derived from 3 independent replicates per variable, and error bars represent means and standard deviations. **, P  < 0.01; ***, P  < 0.001. (B) Biofilms were fixed and processed for SEM (see Materials and Methods) at 96 h postinoculation.

FIG 7

Impact of the putative glycosyltransferase gene gstA on NTHi 86-028NP biofilm structure. Static biofilm cultures of NTHi 86-028NP and NTHi 86-028NP gstA were incubated in sBHI medium overnight in 4-well microscope chamber slides (Nunc), stained with BacLight live/dead reagent (Molecular Probes) essentially as described in the legend of Fig. 6, and visualized by confocal laser scanning microscopy. (A and B) Three-dimensional biofilm images reconstructed from vertical z-series of NTHi 86-028NP (A) and NTHi 86-028NP gstA (B). Biofilm structural parameters were quantified by using COMSTAT (56), essentially as we described previously (9, 12, 15, 16, 57). (C and D) Comparisons of thicknesses and biomasses of NTHi 86-028NP and NTHi 86-028NP gstA biofilms. Graphs were derived from 10 individual fields of view per well in 4 independent biofilms per well, and error bars represent means and standard deviations.

FIG 8

Impact of gstA on NTHi 86-028NP biofilm matrix. Static biofilm cultures of NTHi 86-028NP and NTHi 86-028NP gstA were incubated overnight (∼12 h) in microscopy chamber slides, essentially as described in the legend of Fig. 7, and stained with FilmTracer SYTO ruby red matrix stain reagent (Thermo Fisher) according to the manufacturer’s instructions. Biofilms were examined by confocal laser scanning microscopy as described in the legend of Fig. 7. (A and B) Comparison of three-dimensional biofilm images of NTHi 86-028NP (A) and NTHi 86-028NP gstA (B) showing a smaller amount of matrix in the NTHi 86-028NP gstA biofilms. (C and D) Comparison of biofilm structural parameters. Values were derived from 10 independent fields of view in three independently generated biofilm cultures. Error bars represent the standard deviations for each group.

FIG 9

gstA has a significant impact on NTHi 86-028NP tissue-associated (biofilm) bacterial loads in experimental chinchilla otitis media infections. Chinchillas (6 per group) were experimentally infected via the transbullar route with ∼10³ CFU/ear of NTHi 86-028NP or NTHi 86-028NP gstA (see Materials and Methods). Groups of animals were euthanized at 1 week, 2 weeks, and 3 weeks postinfection, and bacterial loads in effusion and bullar tissue homogenate samples were determined by plate counts. *, statistically significant differences between groups (P < 0.05).