Heme cross-feeding can augment Staphylococcus aureus and Enterococcus faecalis dual species biofilms

Abstract

The contribution of biofilms to virulence and as a barrier to treatment is well-established for Staphylococcus aureus and Enterococcus faecalis, both nosocomial pathogens frequently isolated from biofilm-associated infections. Despite frequent co-isolation, their interactions in biofilms have not been well-characterized. We report that in combination, these two species can give rise to augmented biofilms biomass that is dependent on the activation of E. faecalis aerobic respiration. In E. faecalis, respiration requires both exogenous heme to activate the cydAB-encoded heme-dependent cytochrome bd, and the availability of O₂. We determined that the ABC transporter encoded by cydDC contributes to heme import. In dual species biofilms, S. aureus provides the heme to activate E. faecalis respiration. S. aureus mutants deficient in heme biosynthesis were unable to augment biofilms whereas heme alone is sufficient to augment E. faecalis mono-species biofilms. Our results demonstrate that S. aureus-derived heme, likely in the form of released hemoproteins, promotes E. faecalis biofilm formation, and that E. faecalis gelatinase activity facilitates heme extraction from hemoproteins. This interspecies interaction and metabolic cross-feeding may explain the frequent co-occurrence of these microbes in biofilm-associated infections.

Fig. 1. Single- and dual-species in vitro biofilms.

A E. faecalis (Ef) and S. aureus (Sa) biofilms were grown in 96-well plates alone or in combination (1:1) over five days before biofilm was quantified by crystal violet (CV). Day 0 refers to 1 h post-inoculation. Data shows mean and SD (N ≥ 3) with ***p < 0.001 and ****p < 0.0001 when compared to E. faecalis-only biofilms of the same time point using 1-way ANOVA with Tukey’s post-hoc test. B Day 5 biofilm and (C) planktonic cells of Ef or Sa or both were collected and CFU/well determined using selective agar. Data shows mean and SD (N ≥ 6) with ***p < 0.001 and **** p < 0.0001 using 1-way ANOVA and Tukey’s post-hoc test.

Fig. 2. Augmentation of biofilms with multiple E. faecalis and S. aureus laboratory strains and clinical isolates.

A Six S. aureus laboratory strains (USA300LAC–ISP479) and ten patient isolates (C01–C50) were grown alone or with E. faecalis (OG1RF) for five days before biofilms were quantified by crystal violet (CV). Results show biofilm levels normalized to OG1RF-only control, with mean and SD displayed (N ≥ 3). Points colored green are significantly different to respective single species biofilms by p < 0.05 using 1-way ANOVA with Bonferroni’s post-hoc test (details in Supplementary Table 1). B Four laboratory strains of E. faecalis (OG1RF-V583), together with 28 E. faecalis patient isolates (VRE122-TTSHW-EF43) were grown for five days alone or with S. aureus (USA300LAC) before biofilm was quantified by CV. Results show biofilm levels normalized to OG1RF-only controls, with mean and SD (N ≥ 3). Points colored green are significantly different to respective single species biofilms by p < 0.05 using 1-way ANOVA with Bonferroni’s post-hoc test (details in Supplementary Table 2).

Fig. 3. Validation of E. faecalis transposon library screen hits.

A The parental strain (OG1RF) and nine transposon mutants were grown alone (circles), or together with S. aureus (Sa, USA300LAC, triangles) for five days before biofilm levels were determined by crystal violet (CV). Data shows mean and SD relative to OG1RF-only control (N ≥ 3). **p < 0.01, ***p < 0.001 and ****p < 0.0001 when compared to respective OG1RF-only or OG1RF + Sa control biofilms using 1-way ANOVA with Dunnett’s post-hoc test. B Oxidative respiration of E. faecalis requires MenA, CydA, CydB, heme and O₂.

Fig. 4. E. faecalis biofilms augmented under oxic and anoxic conditions.

E. faecalis (Ef) biofilms were grown for five days under A anoxic and (B) oxic conditions, alone or in the presence of S. aureus (Sa), hemin or hemoglobin (Hb) before biofilm was quantified by crystal violet (CV). Data shows mean and SD of CV absorbance. ****p < 0.0001 when compared to Ef-alone control using Dunnett’s post-hoc test (N ≥ 4).

Fig. 5. The E. faecalis cydABCD operon is required for biofilm augmentation by hemin and S. aureus, and for heme uptake.

A Transposon mutants for cydA, cydB, cydC, cydD and menA, deletion mutants of cydB and cydD, their chromosomally complemented strains, along with parental wild-type OG1RF, were grown alone, in the presence of hemin (25 μg/ml), or with S. aureus (Sa, USA300LAC) for five days before biofilm was quantified with crystal violet (CV). Insert shows cydABDC operon and arrows indicate the direction of transcription. B Transposon and deletion mutants for cydA, cydB, cydC and cydD, along with complementary strains and parental wild-type OG1RF, were grown overnight in the presence of hemin (5 μg/ml) before cells were pelleted, lysed and analyzed by LC–MS for intracellular heme. Data shows mean values and SD. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001 when compared to respective OG1RF control (first of each group) using Tukey’s post-hoc tests (N ≥ 3).

Fig. 6. Effect of heme-deficient S. aureus mutant on dual-species biofilm augmentation.

Parental strain S. aureus USA300LAC, the hemB transposon mutant and the atlA transposon mutant were grown alone or in combination with E. faecalis (Ef) for five days before biofilm was quantified with crystal violet (CV). Data shows mean and SD (N = 6). *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001 when compared to Ef alone, or as otherwise indicated by brackets, using 1-way ANOVA with Tukey’s post-hoc tests.

Fig. 7. Effects of gelatinase E (gelE) deletion on heme-source utilization.

Parental strain E. faecalis (OG1RF) and respective gelE deletion mutant (∆gelE), were grown alone or in combination with hemin (25 μg/ml), hemoglobin (Hb, 10 μg/ml) or S. aureus (Sa, USA300LAC) for five days before biofilm was quantified with crystal violet (CV), Data shows mean and SD (N ≥ 4). ****p < 0.0001 when compared to respective Ef controls using 1-way ANOVA with Bonferroni’s post-hoc tests.