Heterogeneity of Salmonella enterica lipopolysaccharide counteracts macrophage and antimicrobial peptide defenses

Abstract

Salmonella enterica is comprised of over 2,500 serovars, in which non-typhoidal serovars (NTS), Enteritidis (SE), and Typhimurium (STM) are the most clinically associated with human infections. Although NTS have similar genetic elements to cause disease, phenotypic variation including differences in lipopolysaccharide (LPS) composition may control immune evasion. Here, we demonstrate that macrophage host defenses and LL-37 antimicrobial efficacy against SE and STM are substantially altered by LPS heterogeneity. We found that SE evades macrophage killing by inhibiting phagocytosis while STM survives better intracellularly post-phagocytosis. SE-infected macrophages failed to activate the inflammasomes and subsequently produced less interleukin-1β (IL-1β), IL-18, and interferon λ. Inactivation of LPS biosynthesis genes altered LPS composition, and the SE LPS-altered mutants could no longer inhibit phagocytosis, inflammasome activation, and type II interferon signaling. In addition, SE and STM showed differential susceptibility to the antimicrobials LL-37 and colistin, and alteration of LPS structure substantially increased susceptibility to these molecules. Collectively, our findings highlight that modification of LPS composition by Salmonella increases resistance to host defenses and antibiotics.

Keywords: LL-37; Salmonella; colistin; inflammasomes; interferons; lipopolysaccharide; macrophages; phagocytosis.

Fig 1

Human macrophages differentially ingest, kill, and respond to Salmonella enterica serovars. (A) Intracellular bacterial counts were enumerated at indicated time points by plating serial dilutions of macrophage lysates on Luria-Bertani (LB) agar plates. (B) Percent survival of Salmonella enterica was calculated by the following formula (CFU_24hpi/CFU_2hpi) × 100, which represents the percent CFU at 24 hpi relative to initial phagocytosis (2 hpi). (C, D) Macrophages were infected with STM using an MOI of 10 and SE using an MOI of 100. Intracellular bacteria counts and percent survival were determined as in Fig. 1A and B, respectively. (E–H) IL-1β, IL-18, TNF-α, or IL-6 cytokine levels present in culture media of untreated (Mock), STM-infected, or SE-infected macrophages, respectively, using the indicated MOI at 6 hpi. Cytokines were analyzed by flow cytometry using a bead-based multiplex assay. (I) Macrophage death was quantified using the cell-impermeant SYTOX Green nucleic acid dye. Human macrophages were infected with Salmonella for 2 h or left untreated (Mock). Cells were stained with 500 nM of SYTOX Green in RPMI media with no phenol red. Fluorescence intensity was measured by Cytation 5 plate reader at 6 hpi. Data represent fold change relative to untreated cells (Mock). Graphs are represented as means of n ≥ 4 independent experiments ± standard deviation (SD). P-values were calculated using Mann-Whitney U-test for panels A–D or one-way analysis of variance with Holm-Sidak’s post-test for multiple comparisons for E–I. P-value: *< 0.05, **< 0.01, *** <0.001, and **** <0.0001. CFU, colony forming units.

Fig 2

SE induces robust NF-κB signaling but does not activate the inflammasome. (A) Representative fluorescence confocal microscopy images of human macrophages left untreated (Mock) or infected with STM or SE (red). Cells were labeled with anti-p65 (green) and stained with DAPI (blue) to label DNA. (B) Number of bacteria per macrophage was enumerated by manually counting the number of red fluorescent bacteria per infected macrophage. Data are presented as a violin plot of the number of bacteria per macrophage of at least 110 infected cells from three independent experiments. (C) Nuclear p65 fluorescence intensity was quantified by ImageJ. DAPI staining was used to identify the area of nuclei. (D, E) Quantitative real-time PCR of IL1β (D) and TNFα (E) transcript levels of macrophages when left untreated (Mock) or infected with STM or SE for 6 h. (F) Representative fluorescence microscopy images of human macrophages mock infected or infected with STM or SE for 4 h. Cells were labeled with anti-Human ASC antibody (green), stained with DAPI (blue) to label DNA, and imaged using OLYMPUS FV3000 confocal laser scanning confocal microscopy. (G) Percentage of cells that formed ASC specks were quantified from microscopy images. Cells with at least one ASC puncta were considered positive. Graphs are represented as means ± SD of n ≥ 3 independent experiments. P-value was calculated by Mann-Whiney U-test for the graph in panel B or by one-way analysis of variance with Holm-Sidak’s post-test for multiple comparisons for the other graphs. P-value: *< 0.05, **< 0.01, *** <0.001, and **** <0.0001.

Fig 3

Diverse magnitudes of pro-inflammatory genes are transcriptionally induced by SE and STM. (A) Volcano plots of differentially expressed genes in SE-infected macrophages relative to Mock, STM-infected macrophages relative to Mock, or SE-infected macrophages relative to STM. Red dots represent statistically significant gene changes. (B) Number of significantly changed genes from SE- or STM-infected macrophages over untreated (Mock). Upregulated (red) or downregulated (blue) genes were identified using fold change (log₁₀ ratio of normalized counts). Significant genes were determined based on P-value <0.05. (C) Quantification of IL23A, CSF2, IL36RN, and IFNγ transcripts using NanoString technology. Expression of all genes is presented as normalized counts. (D) Quantification of CSF2 and IFN-γ levels produced in the culture media of macrophages mock infected, infected with STM, or infected with SE for 6 h. Cytokine levels were analyzed by flow cytometry using a bead-based multiplex assay. Graphs are represented as means of n ≥ 3 independent experiments ± SD. P-value was calculated using one-way analysis of variance with Holm-Sidak’s post-test for multiple comparisons. P-value: *< 0.05, **< 0.01, *** <0.001, and **** <0.0001.

Fig 4

SE LPS biosynthesis genes are required for formation of the rugose colony morphology and phagocytosis inhibition. (A) Phagocytosis of multiple strains of STM and SE was quantified by enumerating CFU of lysed macrophages at 2 hpi. (B) Representative images of colony morphology formed by wild-type STM, wild-type SE, and SE isogenic LPS biosynthesis gene mutants. Images were acquired by the brightfield using an OLYMPUS MVX10 MacroZoom fluorescence microscope. (C) LPS was extracted from the indicated bacterial strain, separated by SDS-PAGE, and silver stained. The representative image was scanned using a BioRad ChemiDoc XRS Imager. (D) Proposed structure of SE LPS and genes involved in biosynthesis based on previous publications (22, 23). (E) Phagocytosis of SE transposon mutants by macrophages were determined by enumerating CFU of intracellular bacteria at 2 hpi. (F) Normalized bacterial inoculum of indicated strains was quantified by enumeration of CFU. (G) Phagocytosed bacteria were quantified by fluorescent live cell imaging measurements using a pH-sensitive pHrodo probe through IncuCyte technology. Bacteria were pre-labeled with pHrodo Green STP Ester and used to infect human macrophages. Bacteria became highly fluorescent when phagocytosed into the low pH phagolysosome. Graphs are represented as means of n ≥ 4 independent experiments ± SD. P-value was calculated using one-way analysis of variance (ANOVA) with Holm-Sidak’s post-test for multiple comparisons. P-value: *< 0.05, **< 0.01, *** <0.001, and **** <0.0001.

Fig 5

LPS biosynthesis gene mutations augment inflammasome activation and interferon signaling. (A) Representative fluorescence microscopy images of human macrophages mock infected or infected with wild-type SE or various isogenic LPS biosynthesis gene mutants for 6 h. Cells were labeled with anti-Human ASC antibody (green), stained with DAPI (blue) to label DNA, and imaged using an OLYMPUS FV3000 confocal microscope. Inflammasome assembly was monitored by ASC speck formation. (B, C) IL-1β and IFN-γ in culture medium were measured at 6 hpi by flow cytometry using a bead-based multiplex assay. (D) Volcano plots of differentially expressed genes in infected macrophages by SE LPS biosynthesis rfaI or rfbU mutants relative to the wild-type parental strain. (E, F) Quantification of IFN-β and IFN-α produced in the culture media at 6 hpi was performed as in panels B and C. Graphs indicate the mean of n ≥ 5 independent experiments ± SD. P-value was calculated using one-way analysis of variance with Holm-Sidak’s post-test for multiple comparisons. P-value: *< 0.05 and **< 0.01.

Fig 6

Inactivation of LPS biosynthesis genes increases Salmonella enterica serovars susceptibility to antimicrobial peptides. (A) Percent survival of SE and STM at different concentrations of colistin. Bacteria were treated with different concentrations of colistin in PBS, incubated at 37°C for 6 h, and CFUs were enumerated by serial dilution and plating on LB agar plates. Percent survival was calculated based on CFU at the indicated concentration relative to untreated bacteria. (B) Percent survival of SE and STM at different concentrations of LL-37. Percent survival was determined as in panel A. (C) Bacterial turbidity of wild-type SE and its isogenic LPS biosynthesis gene mutants when grown in LB broth in the presence of colistin (10 µg/mL). Bacterial growth was determined based on OD₆₀₀ at 24 h. (D) Percent survival in colistin of SE and its isogenic LPS biosynthesis gene mutants were calculated as in panel A. (E) Bacterial growth of wild-type STM and its isogenic rfaI deletion strain (STMΔrfaI) in the presence or absence of colistin (10 µg/mL) were determined based on OD₆₀₀ at 24 h. (F) Percent survival of wild-type STM and its isogenic rfaI deletion strain (STMΔrfaI) in the presence or absence of LL-37 (20 µg/mL). Percent survival was calculated as in panel A, which represents percent CFU obtained after 6 h post LL-37 treatment relative to untreated bacteria. Graphs indicate the mean of n ≥ 4 independent experiments ± SD. P-values in panels A, B, and F were calculated using one-way analysis of variance (ANOVA) with Holm-Sidak’s post-test for multiple comparisons. P-values in C, D, and E panels were calculated using two-way ANOVA and followed up by Dunnett’s multiple comparisons test. P-value: *< 0.05, **< 0.01, *** <0.001, and **** <0.0001.

Fig 7

Model for host::pathogen interaction when LPS structure is modified. On the left, STM modifies LPS structure to be phagocytosed by human macrophages to avoid extracellular killing by the antimicrobial peptide (LL-37), survives intracellularly, and triggers a high level of cytokines (IL-1β, IL-8, and IFN-γ). In the middle, SE modifies LPS structure to block phagocytosis, inhibit LL-37 extracellular killing, and trigger a low level of IL-1β, IL-8, and IFN-γ. In the right, mutations in the LPS biosynthetic genes render the bacteria highly susceptible to phagocytosis, LL-37 extracellular killing, and elevation of inflammatory responses.