Preferential invasion of mitotic cells by Salmonella reveals that cell surface cholesterol is maximal during metaphase

Abstract

Cell surface-exposed cholesterol is crucial for cell attachment and invasion of many viruses and bacteria, including the bacterium Salmonella, which causes typhoid fever and gastroenteritis. Using flow cytometry and 3D confocal fluorescence microscopy, we found that mitotic cells, although representing only 1-4% of an exponentially growing population, were much more efficiently targeted for invasion by Salmonella. This targeting was not dependent on the spherical shape of mitotic cells, but was instead SipB and cholesterol dependent. Thus, we measured the levels of plasma membrane and cell surface cholesterol throughout the cell cycle using, respectively, brief staining with filipin and a fluorescent ester of polyethylene glycol-cholesterol that cannot flip through the plasma membrane, and found that both were maximal during mitosis. This increase was due not only to the rise in global cell cholesterol levels along the cell cycle but also to a transient loss in cholesterol asymmetry at the plasma membrane during mitosis. We measured that cholesterol, but not phosphatidylserine, changed from a ∼2080 outerinner leaflet repartition during interphase to ∼5050 during metaphase, suggesting this was specific to cholesterol and not due to a broad change of lipid asymmetry during metaphase. This explains the increase in outer surface levels that make dividing cells more susceptible to Salmonella invasion and perhaps to other viruses and bacteria entering cells in a cholesterol-dependent manner. The change in cholesterol partitioning also favoured the recruitment of activated ERM (Ezrin, Radixin, Moesin) proteins at the plasma membrane and thus supported mitotic cell rounding.

Keywords: Bacterial pathogenesis; Cholesterol; Mitosis; Plasma membrane; Salmonella.

Fig. 1.

Salmonella invades mitotic cells preferentially. (A) Representative FACS profiles of RPE1 cells exposed to EGFP-expressing S. Typhimurium SL1344 (MOI 100) for 10 minutes, fixed and stained with propidium iodide (DNA). Gating for EGFP^positive cells identified uninfected and infected cells (left). DNA profiles of total, uninfected and infected cells are shown (middle and right). Arrow shows the enrichment in the infected sample and corresponding depletion in the uninfected population. (B) Interphase and mitotic cells (phospho-Histone H3-negative and -positive, respectively) were gated, and infected and uninfected cells were identified as in A. (C) Representative images of RPE1 cells treated as in A, stained for DNA (blue) and α-tubulin (red). Arrow indicates a mitotic cell. Scale bar: 10 µm. (D) RPE1 cells infected with 12023, SL1344 and LT2, stained and gated as in A and B. (E) Experiments carried out as in D. Ratios of uninfected to infected cells at each stage of the cell cycle. A ratio of 1 (horizontal line) represents no preference. (F) Percentage of interphase cells (left) or mitotic cells (right) infected by one or more than two bacteria [LT2 (light grey), 12023 (grey) or SL1344 (dark grey)] after 10 minutes, scored by immunofluorescence. (G) Ratios of infected cells in mitosis to infected cells in interphase (identified as in B using MPM-2), in adhered or detached (trypsinized) cells. Data are mean ± s.e.m. ns, not significant. *P<0.05; ***P<0.001.

Fig. 2.

Cell-surface cholesterol mediates the targeting of Salmonella to mitotic cells. (A) Scheme depicting the mutants used in the study. (B) RPE1 cells incubated for 10 minutes with SL1344 wild-type or ΔsopE/E2/B, ΔprgH or ΔsipB. EGFP-expressing bacteria (green) stained for DNA (blue), α-tubulin (red) and extracellular bacteria [CSA-1, yellow (green+red double-stained)]. Scale bars: 10 µm. (C) Ratios of bacteria adhered to mitotic cells to bacteria adhered interphase cells, determined by immunofluorescence. (D) Average number of bacteria in interphase cells (grey) or mitotic cells (red) infected as in B scored by immunofluorescence. (E) Percentage of interphase (grey) or mitotic cells (red) infected by SL1344 (WT), or WT and ΔsipB mutant expressing Yersinia invasin protein (‘+ invasin’), scored by flow cytometry. (F) Total cellular cholesterol levels upon cholesterol depletion (‘−cholesterol’) or loading (‘+cholesterol’) in interphase or metaphase cells (‘Mitosis’). (G,H) Percentage of infected cells (G) and ratios of infected cells in mitosis to infected cells in interphase (H) in control, ‘−cholesterol’ or ‘+cholesterol’ samples, determined by flow cytometry. Data are mean ± s.e.m. ns, not significant. *P<0.05; **P<0.01; ***P<0.001.

Fig. 3.

Cell surface cholesterol is maximal during mitosis. (A) Representative FACS profile of live RPE1 cells double-stained for DNA (DRAQ5) and plasma membrane cholesterol (filipin). (B) Quantification of experiments carried out as in A in HeLa or RPE1 cells (N>4 experiments). (C) Representative image of a cell sample used in A,B. Plasma membrane cholesterol (filipin, false-coloured green), DNA (blue). (D) Calibration of fPEG-cholesterol concentrations in living cells. A concentration of 1 µg/ml was chosen for subsequent experiments as it represented the smallest saturating concentration. (E) Representative FACS profile of RPE1 cells from control, cholesterol depleted (‘−cholesterol’) or cholesterol loading (‘+cholesterol’) stained with fPEG-cholesterol. (F) Representative FACS profile of RPE1 cells stained with fPEG-cholesterol, DRAQ5 and phospho-Histone H3. G1, S and G2 cells were identified by DNA gating as in A and mitotic cells were phospho-Histone H3-positive. (G) Scheme of imaging strategy. (H) Single-cell measurements of plasma membrane cholesterol. Representative image of a cell sample used. Plasma membrane cholesterol (filipin, false-coloured green), DNA (blue). The values are integrals of the filipin signals from whole 3D stacks of images. Stages of interphase (G1, S and G2) were determined by gating the integral DNA signals (2n, intermediate and 4n) of each cell, metaphase cells were identified by morphology. (I) Single-cell measurements of cell surface cholesterol. Representative images of a cell sample used, fPEG-cholesterol (green), DNA (blue). The values are integrals of fPEG-cholesterol signals from whole stacks of images. (J) Single-cell measurements of cell surface phosphatidylserine (FITC-Annexin V). Apoptosis was induced by staurosposine (1 µM, 6 hours). (K) DHE fluorescence before (non-permeabilized) and after (permeabilized) TNBS in mitotic cells (rounded) or interphase. The values were normalized to that of ‘interphase before’. (L) Remaining DHE fluorescence after TNBS corresponding to the fraction of inaccessible (cytoplasmic) fraction of cholesterol at the plasma membrane. (M) Fluorescence from various concentrations of DHE loaded (1 hour, 37°C) on giant unilamellar vesicles (GUVs). TNBS was applied to GUVs loaded with the highest concentration. (N) Model proposing the changes in cholesterol distribution at the plasma membrane of mitotic cells. Scale bars: 10 µm. Data are mean ± s.e.m. ns, not significant. *P<0.05; **P<0.01; ***P<0.001.

Fig. 4.

Changes in cholesterol levels support ERM protein recruitment during mitosis. (A) 3D rendering (grid side 13.5 µm) of metaphase cells stained for cholesterol (filipin, false-coloured green), actin (red) and DNA (blue) from control or cholesterol loading (‘+cholesterol’) samples. Bar graph shows the metaphase cells volume in control or ‘+cholesterol’ samples. (B) Immunofluorescence of pERM (green), actin (red) and DNA (blue) in control or ‘+cholesterol’ samples. (C) Recruitment at the plasma membrane of endogenous Rho GTPases and pERM in metaphase cells from control or ‘+cholesterol’ samples. (D) Immunofluorescence of PtdIns(4,5)P2 (green), actin (red) and DNA (blue) in control or ‘+ cholesterol’ samples. Bar graph shows the quantification. Scale bars: 10 µm. Data are mean ± s.e.m. ns, not significant. *P<0.05; **P<0.01.