Metabolic interplay between Proteus mirabilis and Enterococcus faecalis facilitates polymicrobial biofilm formation and invasive disease

Abstract

Biofilms play an important role in the development and pathogenesis of catheter-associated urinary tract infection (CAUTI). Proteus mirabilis and Enterococcus faecalis are common CAUTI pathogens that persistently co-colonize the catheterized urinary tract and form biofilms with increased biomass and antibiotic resistance. In this study, we uncover the metabolic interplay that drives biofilm enhancement and examine the contribution to CAUTI severity. Through compositional and proteomic biofilm analyses, we determined that the increase in biofilm biomass stems from an increase in the protein fraction of the polymicrobial biofilm. We further observed an enrichment in proteins associated with ornithine and arginine metabolism in polymicrobial biofilms compared with single-species biofilms. We show that arginine/ornithine antiport by E. faecalis promotes arginine biosynthesis and metabolism in P. mirabilis, ultimately driving the increase in polymicrobial biofilm protein content without affecting viability of either species. We further show that disrupting E. faecalis ornithine antiport alters the metabolic profile of polymicrobial biofilms and prevents enhancement, and this defect was complemented by supplementation with exogenous ornithine. In a murine model of CAUTI, ornithine antiport did not contribute to E. faecalis colonization but was required for the increased incidence of urinary stone formation and bacteremia that occurs during polymicrobial CAUTI with P. mirabilis. Thus, disrupting metabolic interplay between common co-colonizing species may represent a viable strategy for reducing risk of bacteremia.IMPORTANCEChronic infections often involve the formation of antibiotic-resistant biofilm communities that include multiple different microbes, which pose a challenge for effective treatment. In the catheterized urinary tract, potential pathogens persistently co-colonize for long periods of time and the interactions between them can lead to more severe disease outcomes. In this study, we identified the metabolite L-ornithine as a key mediator of disease-enhancing interactions between two common and challenging pathogens, Enterococcus faecalis and Proteus mirabilis. Disrupting ornithine-mediated interactions may therefore represent a strategy to prevent polymicrobial biofilm formation and decrease risk of severe disease.

Keywords: Enterococcus faecalis; Proteus mirabilis; arginine; bacteremia; bacterial metabolism; biofilm; catheter-associated urinary tract infection; ornithine; polymicrobial; urinary tract infection.

Fig 1

P. mirabilis and E. faecalis polymicrobial biofilms exhibit enhanced biofilm biomass characterized by an increase in protein content. (A) Crystal violet staining of biofilms grown for 24 hours in TSB-G. (B) CFUs of biofilms grown for 24 hours in TSB-G. (C–E) Compositional analysis of single or polymicrobial biofilms detailing total protein (C), total carbohydrate (D), and total extracellular DNA (eDNA) (E). For composition analysis, 24 replicate biofilms for each inoculum were suspended in a total volume of 3 mL water. (F) Crystal violet staining of biofilms grown with 50 µg/mL proteinase-K. Data represent the mean ± standard deviation (SD) for at least three independent experiments with at least two replicates each. ns, non-significant, *P < .05, **P < .01;, ***P < .001, and ****P < .0001 by one-way analysis of variance (ANOVA).

Fig 2

Proteins involved in amino acid import, biosynthesis, and metabolism exhibited differential abundance in polymicrobial biofilms compared with single biofilms. Pie charts display all proteins that mapped to KEGG pathways and were either increased or decreased >2-fold in polymicrobial biofilms compared with single-species biofilms. Proteins are grouped by broad functional categories. Note that some proteins mapped to multiple pathways and are therefore represented more than once in the pie charts. The top panels detail P. mirabilis proteins, while the bottom panels show E. faecalis proteins.

Fig 3

L-Ornithine secretion from E. faecalis drives P. mirabilis arginine biosynthesis and a contact-dependent increase in polymicrobial biofilm biomass. (A) Summary of P. mirabilis proteins involved in metabolism that were enriched within polymicrobial biofilms compared with single-species biofilms. Proteins related to ornithine metabolism, arginine biosynthesis, or transport are highlighted in red text. (B) All known genes involved in L-ornithine metabolism and L-arginine biosynthesis in P. mirabilis are displayed. L-Ornithine can either be directly catabolized to putrescine via ornithine decarboxylate (SpeF) or be fed into L-arginine biosynthesis via ornithine carbamoyltransferase (ArgI), which generates L-citrulline. Argininosuccinate synthase (ArgG) uses ATP to generate L-arginino-succinate from L-citrulline and L-aspartate; then, argininosuccinate lyase (ArgH) generates L-arginine and fumarate from L-arginino-succinate. L-Arginine can then be catabolized to putrescine via arginine decarboxylase (SpeA) and agmatinase (SpeB). (C) Crystal violet staining of single-species and polymicrobial biofilms grown for 24 hours in TSB-G. (D) Total protein content as measured by BCA from three pooled biofilms per experiment. (E) Crystal violet staining of biofilms grown for 24 hours in TSB-G with or without 10 mM of L-ornithine. (F) Total protein content as measured by BCA from three pooled biofilms per experiment when established in TSB-G with or without 10 mM of L-ornithine supplementation. (G) Crystal violet staining of biofilms grown for 24 hours in TSB-G with or without 10 mM of L-arginine. (H) Crystal violet staining of biofilms grown for 24 hours in TSB-G with P. mirabilis arginine catabolism mutants, speA::Kan and speB::Kan. (I and J) Crystal violet staining of biofilms grown for 24 hours in TSB-G with or without 10 mM of agmatine (I) or putrescine (J). Data represent the mean ± SD for three to five independent experiments with at least two replicates each. ns, non-significant, one-way ANOVA multiple comparisons, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Fig 4

Metabolite profiles of single and polymicrobial biofilms established in TSB-G or pooled human urine. Metabolomics analysis of supernatants collected from single and polymicrobial biofilms at 2, 6, and 24 hours post-inoculation in either TSB-G or dilute human urine. Average metabolite concentrations from two independent experiments are represented as log₂ fold change compared with an uninoculated media-alone control (either TSB-G or urine) collected at the 2-hour time point.

Fig 5

Arginine/ornithine antiport by E. faecalis and L-arginine biosynthesis by P. mirabilis contribute to enhanced biofilm biomass on silicone catheters under flow conditions. (A) Image of the glass bladder model setup with five bladders running in tandem. (B and C) CFUs of P. mirabilis (B) or E. faecalis (C) wild-type and mutant bacteria were enumerated from effluent collected through the catheter port for each bladder at 0, 3, 6, 9, 12, and 24 hours post-inoculation. (D and E) Crystal violet staining of bacterial biomass on either 10-mm catheter segments (D) or the catheter eyelet (E) 24 hours post-inoculation. (F) CFUs enumerated from 10-mm catheter segments at 24 hours post-inoculation. Data represent mean ± SD for at least three independent glass bladder experiments with three replicate catheter segments each. ns, non-significant, one-way ANOVA multiple comparisons, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Fig 6

Metabolic interplay between E. faecalis and P. mirabilis contributes to secondary bacteremia. Bacterial CFUs in urine, bladder, kidneys, and spleen collected 96 hours post-infection in a CAUTI murine model. For single-species infections, animals were infected with 10⁵ CFUs of either wild-type P. mirabilis or P. mirabilis argI::Kan (A) or wild-type E. faecalis or E. faecalis arcD::Tn (B). For polymicrobial infections, mice were coinfected with a 50:50 mixture of the wild-type strains or their respective mutants (C). Each data point represents total CFUs per indicated organ. Data from three independent studies were analyzed via non-parametric Kruskal- Wallis one-way ANOVA. (D–G) The CFUs of each specific bacterial strain are displayed for single-species infection and each coinfection, with statistical significance determined by non-parametric Kruskal-Wallis one-way ANOVA comparison to CFUs from single-species infection, *P < 0.05 and **P < 0.01.

Fig 7

Working model of how metabolic interplay between E. faecalis and P. mirabilis facilitates biofilm enhancement. E. faecalis imports arginine and secretes ornithine via the ArcD antiporter. P. mirabilis imports the now-abundant ornithine and generates arginine through the sequential action of ArgI, ArgG, and ArgH. The resulting arginine has three possible fates: (i) catabolism to the polyamine putrescine, consequently resulting in production of urea; (ii) incorporation into proteins, potentially including those involved in enhanced biomass production; and (iii) export out of the cell, which would provide additional arginine for import by E. faecalis. Notably, E. faecalis has additional arginine import mechanisms that are not linked to ornithine antiport, such as the Art ABC transporter. This metabolic interplay ultimately increases the total protein content of the polymicrobial biofilm and is associated with increased abundance of potential adhesins in both species.