(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



The expression of virulence genes increases membrane permeability and sensitivity to envelope stress in Salmonella Typhimurium

['Malgorzata Sobota', 'Biozentrum', 'University Of Basel', 'Basel', 'Pilar Natalia Rodilla Ramirez', 'Alexander Cambré', 'Department Of Microbial', 'Molecular Systems', 'Ku Leuven', 'Leuven']

Date: 2022-04

To further confirm that the PprgH::gfp reporter is coexpressed with HilD-regulated genes, we compared the proteomic profile of WT S.Tm cells sorted according to the bimodal distribution of the green fluorescence signal ( S1 Fig , S1 Table ). The translation of SPI-1 T3SS-1 components and effectors, the SPI-4 T1SS, flagella, and chemotaxis systems was increased in GFP positive cells ( Table 1 ), which was consistent with previously published transcriptomic data describing the HilD regulon [ 8 – 10 ]. This analysis validated the use of the PprgH::gfp fusion as reporter for HilD regulon expression at the single-cell level throughout this study.

Flow cytometry analysis. (A) Proportion of GFP expressing cells (PprgH::gfp) from WT, ΔhilD, and ΔhilE strains not stained by Sytox blue or PI (i.e., alive) in absence of lethal treatment (distilled water control). Data normalized to parallel experiments with the WT strain. When indicated, 4 mM Tris–0.4 mM EDTA was added to the broth as sublethal pretreatment. (B) Frequency of cells stained with either Sytox blue or PI (i.e., dead) after treatment with 100 mM Tris-10 mM EDTA (lethal TE treatment) in WT, ΔhilD, and ΔhilE strains. Data normalized to parallel experiments with the WT strain. When indicated, 4 mM Tris–0.4 mM EDTA was added to the broth. (C) Proportion of GFP expressing cells, unstained by Sytox blue or PI, in distilled water. WT strain carrying the empty vector pBAD24 or the philA plasmid allowing for overexpression of hilA. When indicated, 1 mM arabinose (Ara) and/or 4 mM Tris 0.4 mM EDTA were added to the broth. (D) Frequency of cells stained with either Sytox blue or PI after treatment with 100 mM Tris-10 mM EDTA. When indicated, 1 mM arabinose and/or 4 mM Tris–0.4mM EDTA were supplemented in the medium. (C, D) Data normalized using WT pBAD24 (−Ara or +Ara) as reference, n = 7 replicates. For comparisons against the WT, p-values were calculated using the raw data in paired Wilcoxon tests. For comparisons between mutants or conditions, p-values were calculated using data normalized using corresponding WT or WT pBAD24 as reference in unpaired Mann–Whitney tests. Numbers below the x-axis indicate the number of replicates. Numbers within the graphs are p-values for comparisons discussed in the main text and bars below these numbers indicate the compared groups. Source data are provided as a source data file ( S1 Data ). PI, propidium iodide; TE, Tris-EDTA; WT, wild-type.

(A) Time-lapse microscopy analysis of WT reporter strain (PprgH::gfp, GFP) after HS (51°C, 15 minutes) (white dots) and untreated control (black dots). Violin plots represent the fraction of cells able to form microcolonies among cells expressing the HilD regulon (GFP+), not expressing the HilD regulon (GFP–) and the total population. The p-value was calculated using a paired Wilcoxon test. (B) Representative pictures from time-lapse microscopy experiments. Cells in the upper panel were heat treated. Left picture shows cells at t = 0, and right picture shows cells after 7 hours. Cells in the lower panel are untreated control. Left picture shows cells at t = 0, and right picture shows cells after 3 hours. Scale bar: 2 μm. (C) Cells grown in LB supplemented (+ Ara) or not (no Ara) with 1 mM L-arabinose to induce the overexpression of hilA from philA (derivative of pBAD24) were stained using 30 μm DiOC 2 (3) in the presence of 10 mM Tris-1 mM EDTA (0.1X TE) and analyzed by flow cytometry. Unstained WT control and WT cells treated with 100 μm CCCP and 30 μm DiOC 2 (3) were used to define the population of cells with low membrane potential in each sample ( S4 Fig ). The proportion of cells with low membrane potential was then normalized according to values obtained in parallel experiments using the WT. For comparisons against the WT, p-values were calculated on the raw data using paired Wilcoxon tests. For comparisons between mutants or conditions, p-values were calculated using the normalized dataset and unpaired Mann–Whitney tests. (D) Proportion of cells producing GFP (PprgH::gfp) from WT, ΔhilD, and ΔhilE strains, not stained by Sytox blue or PI (i.e., alive) in distilled water, measured by flow cytometry. (E) Frequency of cells stained with either Sytox blue or PI (i.e., dead) after treatment with 100 mM Tris-10 mM EDTA (TE treated) from WT, ΔhilD, and ΔhilE strains measured by flow cytometry. ( F, G) Reduction of the GFP positive fraction (ΔGFP+) among WT or ΔhilE cells alive after TE treatment compared to distilled water control (F) or 15% sodium cholate compared to PBS control (G). Significance of the deviation of the median from 0 (dashed lines) estimated by Wilcoxon signed rank test (p < 1E-06). (H, I) Normalized frequency of cells stained with either Sytox blue or PI (i.e., dead) after TE treatment from ΔhilD, ΔSPI1, ΔiagB-invG, Δflg, ΔSPI-4, and triple mutant ΔiagB-invG Δflg ΔSPI-4 strains harboring pBAD24 (H) or philA (I) . Data normalized to parallel WT pBAD24 controls. The cultures were supplemented with 1 mM arabinose (+ Ara) to induce hilA expression in the strains carrying philA. For comparisons against the WT, p-values were calculated using the raw data in paired Wilcoxon tests. For comparisons between mutants or conditions, p-values were calculated using the normalized dataset in unpaired Mann–Whitney tests. p-Values for comparisons discussed in the main text are indicated within the panels with bars marking the compared conditions. The S5 Table contains p-values for comparisons of data from panels H and I . Numbers below the x-axis indicate the number of replicates. Source data are provided as a source data file ( S1 Data ). DiOC2(3), 3,3′-diethyloxa-carbocyanine iodide; HS, heat shock; LB, Lysogeny broth; PI, propidium iodide; TE, Tris-EDTA; WT, wild-type.

GFP (from PprgH::gfp) (A) and NPN fluorescence (B–E) were measured from cells treated with 10 μM NPN and were divided by the optical density at 600 nm. Values of each repetition are normalized using a parallel experiment on the WT. When indicated, 0.1% of glucose was added to the broth in order to repress virulence expression (+ glucose, black dots). A control strain constitutively expressing GFP from pM965 carrying PrpsM::gfp (fluorescence distribution shown in S2 Fig ) was used to control for the effect of GFP on the fluorescence readout from NPN uptake. A 2.5-μg/mL Polymyxin B treatment permeabilizing the membrane was used as positive control (C). (D, E) NPN uptake in WT, ΔhilD, ΔSPI-1, ΔiagB-invG, Δflg, ΔSPI-4, and triple mutant ΔiagB-invG Δflg ΔSPI-4 carrying either pBAD24 (D) or philA (E) and treated with 10 μM NPN. Fluorescence values were normalized to the reference WT pBAD24 grown in presence of 1 mM arabinose. Values obtained in the absence of arabinose are shown in S2 Fig . For comparisons against the WT, p-values were calculated using the raw data in paired Wilcoxon tests. For comparisons between mutants or conditions, p-values were calculated using the normalized dataset in unpaired Mann–Whitney tests. p-Values for comparisons discussed in the main text are indicated within the panels with bars marking the compared conditions. S2 Table shows p-values for comparisons between groups from panels D and E . Numbers below the x-axis (n = x) indicate the number of replicates when nonequal between conditions in a given experiment. n = 7 in panels A – C . Source data are provided as a source data file ( S1 Data ). NPN, N-phenyl-1-naphthylamine; WT, wild-type.

In this study, we used the chromosomal PprgH::gfp reporter (in which gfp expression is controlled by a copy of the promoter of the T3SS-1 prg operon) inserted in the locus putPA [ 11 ] as a proxy for the expression of the HilD regulon in S.Tm SL1344 (further referred to as wild type or WT). More specifically, the PprgH::gfp reporter is activated by HilA [ 28 ], itself tightly controlled by HilD [ 29 ], and does not interfere with T3SS-1 expression. The distribution of the gfp expression at late exponential phase in Lysogeny broth (LB) was clearly bimodal with approximately two-third of the population in the OFF state and one-third in the ON state ( S1 Fig ). As previously reported, the ON/OFF ratio results from the production of HilD, whose activity is controlled by the negative regulator HilE at the posttranslational level [ 12 , 17 , 30 ]. Accordingly, the proportion of ON cells was increased to about half of the population in the ΔhilE mutant. Note that the ON/OFF ratio in WT and its ΔhilE derivative varied between experiments, but the latter consistently yielded more ON cells than the WT ( Figs 1A , 2D , and 3A ). Deletion of hilD, on the other hand, locked the cells in the OFF state ( S1 Fig ), validating the PprgH::gfp reporter as a proper proxy for HilD activity.

A reproducible pattern of NPN uptake was observed when comparing various mutants in which hilA was not overexpressed (pBAD24 carrying strains or in absence of arabinose), confirming that the flagella were the most important contributors to membrane permeability ( Fig 1D , S2 Fig ). Unexpectedly, the impact of the T3SS-1 was marginal at WT expression level and production of the SPI-4 T1SS remained relatively neutral in all tested conditions. S2 Table gathers statistical analysis results from this dataset.

Overproducing HilA, a transcriptional regulator of virulence (including T3SS-1 and SPI-4) controlled by HilD [ 33 ], led to a drastic increase in NPN uptake in the WT (p < 1E-06). In these experiments, strains carrying the empty vector pBAD24 ( Fig 1D ) were compared with strains overexpressing hilA from the pBAD24 derivative philA, both growing in the presence of the arabinose inducer ( Fig 1E ).

We then evaluated the relative contribution of T3SS-1, SPI-4 T1SS, and flagella to the increased membrane permeability in S.Tm ( Fig 1D ). The full SPI-1 deletion (including hilD) phenocopied the ΔhilD mutant (p = 0.880), validating our previous observation. However, the deletion of the iagB-invG locus in SPI-1 (i.e., removing operons iag, spt, sic, iac, sip, sic, spa, and inv, but keeping transcriptional regulators hilD, hilC, hilA, and invR intact) or SPI-4 individually had significantly less effect than the ΔhilD mutation ( Fig 1D ). Interestingly, deleting the flgBCDEFGHIJ operon (thereafter shortened flg) or combining deletions of iagB-invG, flg and SPI - 4 phenocopied the ΔhilD mutant (p > 0.05) ( Fig 1D ).

Several functions controlled by HilD are large protein complexes embedded in the envelope of S.Tm (T3SS-1, SPI-4 T1SS, flagella, and chemotactic receptor clusters ( S1 Fig , Table 1 , S1 Table ), potentially affecting envelope integrity. In order to assess permeability of the outer membrane, we used N-phenyl-1-naphthylamine (NPN), a lipophilic dye that is weakly fluorescent in aqueous environments but becomes highly fluorescent in hydrophobic environments such as the inner leaflet of the outer membrane and the innermembrane [ 31 ]. In growth conditions triggering expression of the HilD regulon (i.e., late exponential phase in LB) ( Fig 1A ), the WT and the ΔhilE mutant accumulated significantly more NPN than the ΔhilD mutant ( Fig 1B ). In contrast, adding glucose to the media drastically reduced the expression of the HilD regulon [ 32 ] ( Fig 1A ) and NPN uptake ( Fig 1B ). As a control, polymyxin B, acting as a detergent, increased NPN uptake independently of glucose presence ( Fig 1C ). We tested a possible effect of gfp expression (used to monitor the HilD regulon induction) by constitutively expressing gfp from a plasmid (transcriptional fusion PrpsM::gfp in pM965) ( Fig 1A , S2 Fig ), which demonstrated that the presence of GFP did not affect NPN uptake or the measurement of the NPN fluorescent signal ( Fig 1B ).

The expression of hilD reduces resistance to outer membrane disrupting treatments

In general, permeability to NPN is increased by treatments that disrupt the membrane of gram-negative bacteria like polymyxin B (Fig 1C) [34], indolicidin [35], and aminoglycosides acting as divalent cation binding sites on the outer membrane [36]. Increased NPN uptake in hilD expressing populations of S.Tm (Fig 1B) suggested that the membrane of HilD ON cells could be inherently disrupted, thus making these cells more sensitive to stress targeting the membrane.

Since heat provokes disruption of the outer membrane [37], we first measured survival of S.Tm exposed to a mild heat shock (HS, 51°C, 15 minutes) at the single-cell level with time-lapse fluorescence microscopy. The cells were observed for 16 hours posttreatment. The vast majority of cells expressing the HilD regulon (i.e., GFP positive cells) was unable to resume growth (Fig 2A and 2B HS upper panels, and S3 Fig), while the rest of the population (i.e., GFP negative cells) regrew after a lag period. Untreated cells were able to grow normally regardless of their hilD expression state, with ON cells switching OFF and diluting the GFP (Fig 2B control lower panels, S3 Fig).

To determine if membrane permeability correlated with higher sensitivity to stress in the ON cells, we analyzed HS sensitivity of the triple ΔiagB-invG Δflg ΔSPI-4 mutant, which has already proved less permeable to NPN (Fig 1). This mutant produced an amount of hilD expressing cells comparable to the WT (S3 Fig), with similarly reduced growth rate compared to OFF cells (S3 Fig). However, a higher proportion of these ON cells was able to resume growth after HS than in WT or the ΔhilE mutant (S3 Fig), suggesting that membrane-localized virulence factors increase both membrane permeability and stress sensitivity.

We then measured membrane potential in cells exposed to 10 mM Tris-1 mM EDTA (0.1X Tris-EDTA [TE]), which destabilizes the lipopolysaccharide of gram-negative bacteria [38] and allows entry of the dye 3,3′-diethyloxa-carbocyanine iodide (DiOC 2 (3)) (Fig 2C) [39]. In the presence of 0.1X TE, membrane potential leads to accumulation of DiOC 2 (3) to the point of dye aggregation shifting its fluorescence from green to red. Here, we used flow cytometry because it measures fluorescence in a high number of cells and allows testing multiple conditions in parallel. We followed the gating strategy described in S4 Fig to estimate the proportion of cells with low membrane potential (green only) among the population of stained cells (red and green) in WT and mutant strains. Exposure to the proton ionophore carbonyl cyanide 3-chlorophenylhydrazone (CCCP) was used as control condition in which the membrane potential is abolished and most DiOC 2 (3) stained cells remained green (depolarized membrane). Unstained cells and cells exposed to CCCP and DiOC 2 (3) served as references (S4 Fig). Representative images are provided in S4 Fig.

Again, a pattern consistent with NPN uptake emerged from these experiments. The WT strain and the ΔhilE mutant showed significantly higher proportion of cells with low membrane potential compared to the ΔhilD mutant. Overexpression of hilA in the presence of arabinose increased the proportion of cells with low membrane potential. This was rescued in the ΔiagB-invG Δflg ΔSPI-4 triple mutant (Fig 2C). In the context of endogenous hilA expression, the triple mutant ΔiagB-invG Δflg ΔSPI-4 phenocopied the ΔhilD mutant (p = 0.428).

Membrane potential in absence of stress was measured in control experiments using 3,3′-dipropylthiadicarbocyanine iodide (DiSC 3 (5)) (S5 Fig), a red fluorescent hydrophobic probe accumulates in the polarized membrane of cells [40]. This assay was compatible with the readout for expression of the HilD regulon with the reporter PprgH::gfp. We observed no difference in DiSC 3 (5) staining in hilD ON (GFP+) versus OFF cells (GFP−) (S5 Fig), suggesting that the membrane of hilD expressing cells were not inherently depolarized. Reduction of membrane potential observed with DiOC 2 (3) was therefore due to exposure to 10 mM Tris-1 mM EDTA further destabilizing the outer-membrane, especially when S.Tm expressed the HilD regulon (Fig 2C).

This observation led us to evaluate the susceptibility of the ON cells to a more severe 100 mM TE treatment. We used 2 complementary approaches, flow cytometry and microscopy, to quantify the proportion of dead cells detectable after TE treatment and the fraction of cells expressing the HilD regulon (GFP+ cells) among the survivors. We used 2 dyes to stain the dead cells: propidium iodide (PI) or Sytox blue. Fig 2D shows the proportion of GFP+ cells exposed to distilled water used as solvent for TE. After TE treatment, we observed a clear increase in the proportion of dead cells in WT and the ΔhilE mutant compared to the ΔhilD mutant with both dyes (Fig 2E). Although Sytox blue had the tendency to stain slightly more cells than PI (S6 Fig), we judged the overlap sufficient to pool both staining results in every dataset. The proportion of ON cells decreased in the surviving populations compared to control (Fig 2F). The timing of the experiment (30’ of treatment before cytometry analysis) was too short to allow ON cells to switch OFF and to dilute the GFP by cell division. The constitutive expression of gfp from the promoter PrpsM did not alter the overall pattern of stress sensitivity when comparing WT, ΔhilD, and ΔhilE strains (S6 Fig). For unclear reasons, WT and ΔhilD with PrpsM::gfp were slightly less sensitive than their PprgH::gfp counterparts. This nevertheless suggested that expressing gfp was not increasing stress sensitivity per se. We also ruled out the contribution of nonfluorescent and unstained debris formed during treatment by quantifying the fraction of nonfluorescent events from stressed cells constitutively expressing gfp (S6 Fig). Moreover, a treatment with 15% sodium cholate, a natural detergent present in bile, was even more potent at reducing the proportion of ON cells in WT (−14% versus −6.8% with TE) and the ΔhilE mutant (−14.7% versus −5.2% with TE) (Fig 2G). The effect of lethal TE treatment was further confirmed by live monitoring of cells in a microfluidic device (S1 Movie). In fact, mainly the ON cells from the WT reporter strain exposed to TE died and became stained by PI (red) concomitantly with losing their GFP, while most OFF cells remained apparently intact (quantification in S7 Fig). Imaging of PI and Sytox blue staining upon TE treatment is presented in S6 Fig.

Additional microscopic analysis of WT reporter bacteria after TE treatment confirmed that, although ca. 70% of the cells were able to resume growth (S8 Fig), the fraction of GFP+ cells among the regrowing cells was significantly reduced (17% less GFP+ cells compared to untreated control (S8 Fig)). This indicated that the TE treatment killed a significant amount of bacteria, with a higher probability for the ON cells to die. Flow cytometry and microscopy showed comparable results with 30% of WT cells stained by PI or Sytox blue (Fig 2E) or not regrowing on the agar pad after TE treatment (S8 Fig). Based on this overall death rate and parameters extracted from cytometry experiments (S3 Table), we estimated a 44% death rate for the ON cells, about 2.3 times higher than the death rate of the OFF cells (19%) from WT S.Tm (Formula: Death rate = 100 –((% final × % total survivors)/% initial); median values, S3 Table). These values were comparable in the ΔhilE mutant (41% and 28% respective death rate for ON and OFF cells). However, the death rate of the locked OFF ΔhilD mutant (6.2%) was strikingly lower than for WT or ΔhilE mutant OFF cells. This could be attributable to the method used to discriminate between the ON and OFF cells from WT and ΔhilE mutant strains based on a fluorescent intensity cutoff when using flow cytometry or microscopy. Meaning that, below this detection cutoff, cells in which the HilD regulon is not fully repressed could bias the overall death rate of the so-called “OFF” subpopulations in WT and ΔhilE. Counting colony-forming units (CFUs) posttreatment confirmed the overall death rate being higher in WT and the ΔhilE mutant compared to the ΔhilD mutant (S4 Table). These data also suggested that Sytox blue or PI staining and counting the regrowing cells under the microscope could underestimate the fraction of cells affected by TE treatment and potentially lysing (S4 Table). Alternatively, plating might be an extra stress that kills cells after TE treatment which would otherwise form microcolonies under the microscope or would not be stained by PI or Sytox blue.

As observed for NPN uptake, the deletion of iagB-invG or SPI-4 alone did not change sensitivity to TE. However, the individual deletion of the flg operon significantly increased resistance to TE compared to WT (p = 4.9E-04) (Fig 2H).

Although the cumulative deletions of iagB-invG, flg and SPI-4 increased resistance to TE treatment, they did not fully phenocopy the resistance of the ΔhilD mutant (Fig 2H) (p < 1E-06). This could be due to the higher sensitivity of this assay compared to NPN uptake measurements in batch cultures. It also suggests that other functions controlled by HilD could play a role in increasing stress sensitivity in the ON cells.

The overexpression of hilA (philA + arabinose) drastically enhanced sensitivity to TE (p < 1E-06) (Fig 2I). Under these conditions, the full SPI-1 deletion (p = 2.9E-06) and, to a lesser extent, the deletion of hilD alone (p < 1E-06), of iagB-invG (p = 1.1E-04), and of iagB-invG flg SPI-4 (p < 1E-06) restored some resistance. Individual deletions of flg or SPI-4 had no effect (p > 0.05). Controls in the absence of arabinose reproduce the pattern observed for strains carrying pBAD24 in the presence of arabinose (S9 Fig). S5 Table gathers statistical analysis results from this dataset.

[END]

[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001608

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.