Salmonella Tol-Pal Reduces Outer Membrane Glycerophospholipid Levels for Envelope Homeostasis and Survival during Bacteremia

Abstract

Salmonellae regulate membrane lipids during infection, but the exact proteins and mechanisms that promote their survival during bacteremia remain largely unknown. Mutations in genes encoding the conserved Salmonella enterica serovar Typhimurium (S Typhimurium) Tol-Pal apparatus caused the outer membrane (OM) sensor lipoprotein, RcsF, to become activated. The capsule activation phenotype for the mutants suggested that Tol-Pal might influence envelope lipid homeostasis. The mechanism involves reducing OM glycerophospholipid (GPL) levels, since the mutant salmonellae similarly accumulated phosphatidylglycerols (PGl) and phosphatidylethanolamines (PE) within the OM in comparison to the wild type. The data support the Escherichia coli model, whereby Tol-Pal directs retrograde GPL translocation across the periplasm. The S Typhimurium mechanism involves contributions from YbgC, a cytoplasmic acyl coenzyme A (acyl-CoA) thioesterase, and CpoB, a periplasmic TolA-binding protein. The functional relationship between Tol-Pal and YbgC and CpoB was previously unresolved. The S Typhimurium Tol-Pal proteins contribute similarly toward promoting OM-GPL homeostasis and Rcs signaling inactivity but differently toward promoting bacterial morphology, rifampin resistance, survival in macrophages, and survival in mice. For example, tolQ, tolR, tolA, and cpoB mutants were significantly more attenuated than ybgC, tolB, and pal mutants in a systemic mouse model of disease. Therefore, key roles exist for TolQ, TolR, TolA, and CpoB during murine bacteremia, which are independent of maintaining GPL homeostasis. The ability of TolQR to channel protons across the inner membrane (IM) is necessary for S Typhimurium TolQRA function, since mutating conserved channel-facing residues rendered TolQ ineffective at rescuing deletion mutant phenotypes. Therefore, Tol-Pal promotes S Typhimurium survival during bacteremia, in part, by reducing OM GPL concentrations, while TolQRA and CpoB enhance systemic virulence by additional mechanisms.

Keywords: CpoB/YbgF; Gram-negative bacteria; Intracellular; Pal; RcsF; Tol-Pal; TolA; TolB; TolQ; TolR; YbgC; anionic glycerophospholipids; bacteria; barrier; cell envelope; constriction; facultative; ion channel; lipid; lysosome; macrophage; membrane curvature; motor; mouse; outer membrane; pathogen; pathogenesis; peptidoglycan hydrolases; periplasm; phosphatidylethanolamines; phosphatidylglycerol; phospholipids; proton-motive force; salmonella; septation; systemic pathogenesis; trafficking; transenvelope complex; translocation; transport; vacuole; virulence.

FIG 1

Tol-Pal spans the Gram-negative cell envelope, and salmonellae produce four major glycerophospholipid (GPL) families within their dual membranes. (A) Enteric bacteria produce the Tol-Pal system of proteins, which are present within the cytoplasm, periplasm, and each membrane of their envelope. The inner membrane (IM) consists exclusively of GPLs, while the outer membrane (OM) is asymmetric and consists mostly of inner-leaflet GPLs and outer-leaflet lipopolysaccharides (LPS). Gram-negative bacteria produce a murein sacculus or peptidoglycan exoskeleton, which is attached to the OM by lipoproteins. (B) Salmonella enterica serovar Typhimurium (S. Typhimurium) generates four major GPL families, which are distinguished by their polar head group structures, indicated by blue and red ovals. Four representative species are depicted for each family. Each species shares an sn-1 esterified C_16:0 and an sn-2 esterified C_18:1 acyl chain and is abundantly present within the S. Typhimurium envelope. The phosphatidylethanolamines (PEs) are diacylated, have a zwitterionic phosphoethanolamine head group, and constitute roughly 60% of the GPL within the IM and 80% of the GPL within the OM (9). Phosphatidylglycerols (PGls) are diacylated, have a singly anionic glycerol-phosphoglycerol head group, and comprise roughly 33% of the GPL within the IM and 17% of the GPL within the OM (9). Cardiolipins (CLs), or diphosphatidylglycerols (diPGls), are tetra-acylated, harbor doubly anionic glycerol-phosphoglycero-phosphoglycerol head group structures, and constitute roughly 7% of the GPL within the IM and 2% of the GPL within the OM (9). Acyl-PGls (aPGls) are triacylated, possess anionic sn-3′ esterified glycerol-phosphoglycerol head groups, and constitute roughly 1% of the GPL within the IM and 2% of the GPL within the OM (10, 65). (C) The enterobacterial Tol-Pal system is transcribed from at least two promoters as two operons, ybgC-tolQRA-tolB-pal-cpoB and tolB-pal-cpoB (28, 66). We previously identified transposon insertion mutations in ybgC, tolQ, and tolB, which caused RcsF-dependent wza-lacZ gene reporter activity to increase for S. Typhimurium (triangles) (47). The Rcs activation phenotype is manifested as a blue colony hue on indicator medium and is depicted in the present study for the site-directed tol-pal mutants (Fig. 4B).

FIG 2

Deleting S. Typhimurium tol-pal genes causes Rcs signaling activity to increase. (A) The chromosomally integrated wza-lacZ gene reporter was used to quantify transcription, which resulted from Rcs signaling activity (47). Briefly, the bacteria were cultured in Luria-Bertani (LB) broth medium to an optical density at 600 nm (OD₆₀₀) of between 0.6 and 0.8, or the mid-exponential growth phase (E phase). The cells were pelleted and lysed so that Miller units could be quantified by a standard β-galactosidase assay. Depicted are the average values ± standard deviation (SD) of results for five independent experiments. An asterisk indicates a statistically significant difference for the mutant relative to the wild type (WT). In all stated instances, two-tailed Student's t tests were used to determine significance (P < 0.05). (B) The ybgC, cpoB, and tolQ genes were cloned and expressed from pBAD24, alongside an empty vector (pBAD) control, in the wild-type and mutant S. Typhimurium genotypes. The bacteria carrying the plasmids were cultured in LB broth plus antibiotic until E phase. Basal pBAD promoter activity drove expression of the gene products. Rcs signaling activity was quantified by a β-galactosidase assay. The experiment was repeated three times, and the average values ± SD are shown. An asterisk indicates a statistically significant difference between the mutant and the wild type (P < 0.05).

FIG 3

tolQ, tolR, and tolA S. Typhimurium mutants accumulate phosphatidylglycerols (PGls) and phosphatidylethanolamines (PEs) within the OM relative to the wild type. The IM and OM fractions for wild-type and mutant S. Typhimurium were isolated by sucrose density gradient ultracentrifugation (63). GPLs were extracted and quantified by reversed-phase liquid-chromatography tandem mass spectrometry (LC-MS/MS) (10). Purified PE and PGl molecules of defined molecular weight, structure, and concentration were used to generate standard curves. Four major S. Typhimurium PE and PGl molecules were then targeted for each membrane. The peak areas for the eight target molecules were compared to the standard curves to calculate the final ng/μl concentration in the membranes (Tables 1 and S3). (A) A graph depicts the IM/OM (ng/ng) ratio calculated for the wild type. The values were obtained using the ng/μl values of the four individual PE and PGl molecules. The average values ± SD were determined for three experiments. A number sign (#) indicates a statistically significant difference in the ng/μl concentration for IM versus the OM fraction of the wild-type S. Typhimurium envelope (P < 0.05) (Tables 1 and S3). The dashed line denotes a 1:1 ratio. (B) A bar graph depicts the IM/OM (ng/ng) ratio for the PGls within the wild-type and tolQ, tolR, and tolA mutant S. Typhimurium cell envelopes. An asterisk indicates a statistically significant difference in the OM PGl level (ng/μl) for the mutant relative to the wild type (P < 0.05) (Tables 1 and S3). (C) A graph depicts the IM/OM (ng/ng) ratio for the PEs within the wild-type and tolQ, tolR, and tolA mutant envelopes. An asterisk indicates a statistically significant difference in the OM PE level (ng/μl) for the mutant relative to the wild type (P < 0.05) (Tables 1 and S3).

FIG 4

Salmonella Tol-Pal proteins differentially promote cell morphology, resistance to rifampin (Rif), and survival in primary mouse macrophages (Mϕs). (A) E-phase bacteria were fluorescently labeled with the lipopholilic membrane dye FM4-64 before being spotted onto agarose pads and visualized by phase-contrast and epifluorescence microscopy at ×100 magnification (47). An arrow indicates an FM4-64-stained vesicle emerging from pal mutant S. Typhimurium. Measurements were also made to determine whether the mutants were altered in their cell size relative to the wild type. For these measurements, we used the FM4-64 images and the fluorescently outlined bacteria (Table S6). Each mutant measured a reduced average cell length relative to the wild type. (B) Susceptibility to Rif was measured by normalizing E-phase bacteria to an OD₆₀₀ of 1.0, or roughly 10⁹ CFU/ml, and spotting aliquots of serially diluted bacteria onto LB agar with or without Rif (5 μg/ml). Four independent experiments were executed, and one representative experiment is shown. The increase in Rcs signaling activity for the mutants is manifested as a blue colony color, which results from increased transcription of the wza-lacZ gene reporter. The increased LacZ levels in the bacteria result in increased cleavage of the indicator substrate, 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside, in the growth medium (X-Gal). (C) Primary C57BL/6 mouse bone marrow-derived macrophages (BMDMϕs) were infected with E-phase salmonellae at a multiplicity of infection (MOI) of roughly 10:1. Each strain was used to infect triplicate wells, and the surviving gentamicin-resistant intracellular CFU were enumerated at 2 and 6 hpi. Five independent experiments were performed. The average values ± SD for each time point are shown. An asterisk indicates a significant difference for the mutant relative to the wild type (P < 0.05). A number sign (#) indicates a significant difference for the tolB and pal mutants relative to the tolQ mutants (P < 0.05).

FIG 5

TolQ ion channel residues are necessary for function. (A) The rescue genotypes harbor a second-site chromosomal copy of the wild type or the mutant tolQ alleles (Fig. S2A and Table S1). Theophylline was added as the small-molecule inducer, which relieves translational repression of the riboswitch upstream of the second-site tolQ allele. The inducer was added at a 0.5 mM final concentration to the LB broth medium where indicated. The levels of Rcs signaling were quantified as Miller units from the gene reporter using E-phase cultures. The average Miller unit values ± SD for five independent experiments are depicted. ns, not significant. (B) Phase-contrast and epifluorescence microscopy was conducted using E-phase cells cultured with or without theophylline. (C) Rifampin susceptibility was determined using S. Typhimurium cultured to the E phase with or without theophylline. An asterisk indicates a bacterial genotype that was cultured with theophylline to the E phase in LB broth media. The inducer was also added to the agar medium in all instances. Four independent experiments were executed, and one representative experiment is shown. (D) Intracellular survival was measured using bacteria cultured to the E phase with or without theophylline. The first three genotypes from left to right on the x axis were cultured to the E phase in LB broth media without inducer. The inducer was added during culture in broth, as well as to the tissue culture medium during the infection period, for the final three genotypes depicted on the x axis. Five independent experiments were performed, and the average values ± SD for each time point are shown. An asterisk indicates a statistically significant difference for the mutant genotype relative to the wild type (P < 0.05).

FIG 6

Tol-Pal proteins differentially promote S. Typhimurium survival during bacteremia. Groups of six female C57BL/6 mice were intraperitoneally (i.p.) injected with roughly 2.5 × 10⁵ wild-type or mutant salmonellae. The mice were euthanized at 48 h. The spleens and livers were dissected and homogenized. Samples were serially diluted, and bacterial CFU/ml were enumerated on LB agar with the appropriate antibiotic. (A) The CFU/ml levels were determined from the livers of the six individual mice infected with each strain. (B) The CFU/ml levels were determined from the spleens of the six individual mice infected with each genotype. An asterisk indicates a statistically significant difference relative to the wild type (P < 0.05). A number sign (#) indicates a significant difference relative to the tolQ mutant genotype (P < 0.05).