Expression of CD44 in RPM-MC cells is sufficient to enhance protrusion formation. Control RPM-MC cells and RPM-MC cells stably expressing CD44 or a CD44 variant mutated in its ERM-binding site were infected with Listeria, fixed and the number of protrusions per cell quantified. Expression of CD44 enhances protrusion formation compared to control cells and cells expressing the mutated variant of CD44. Error bars indicate one standard deviation from the mean.

Overexpression of the amino- or carboxy-terminus of ezrin does not impair the motility of Listeria. Control HeLa cells and HeLa cells transfected with the GFP-tagged amino- or carboxy-terminal domain of ezrin were infected with Listeria, fixed and stained with fluorescent phalloidin. The actin tails induced by Listeria in HeLa cells expressing the amino- (B) or carboxy-terminus (C) of ezrin are morphologically indistinguishable from those induced by these bacteria in control cells (A). Panels A–C represent phalloidin staining, whereas panels B′–C′ represent GFP fluorescence. Scale bar: 5 μm. (D) Quantification of the length of Listeria actin tails (a parameter that is proportional to bacterial speed) in control HeLa cells and in HeLa cells expressing the amino- or carboxy-terminal domain of ezrin, showing that the expression of both ezrin domains does not impair Listeria motility. Error bars indicate one standard deviation from the mean.

Expression of ezrin T567D in RT4 cells restores the normal morphological features of Listeria protrusion. Control RT4 cells (A) and cells treated overnight with the anti-CD44 monoclonal antibody 1.1asml (B) were infected with Listeria and processed for SEM. In control cells, Listeria induced the formation of normal protrusions (facing arrows in (A)), whereas in cells treated with the 1.1asml antibody, these bacteria formed short and distorted protrusions (arrowhead in (B); compare with Figure 6D). RT4 cells expressing the ezrin T567A (C, D) were treated as for (A, B). The expression of ezrin T567A in RT4 cells (C) caused the formation of distorted Listeria protrusions (arrowhead in (C)). The activation of merlin did not cause further changes in the morphological features of protrusions (arrowhead in (D); compare to (C)). In RT4 cells transfected with the constitutively active ezrin mutant T567D (E, F), Listeria induced the formation of normal protrusions (facing arrows in (E); compare with Figure 6A). The expression of ezrin T567D effectively counteracted the merlin-driven displacement of endogenous ERM proteins, thus resulting in the formation of protrusions having normal morphological features (facing arrows in (F)). White stars indicate the position of bacteria. Scale bar: 1 μm.

ERM proteins are essential for efficient cell-to-cell spread of Listeria. (A–D) Monolayer islet of untransfected LLC-PK1 cells or LLC-PK1 cells expressing GFP-tagged wild-type or mutated CD44 cytoplasmic tail or the ezrin mutant T567A were infected with Listeria. At 6 h after the beginning of the infection, cells were fixed and then stained with fluorescent phalloidin. In control cells (A) as well as in cells expressing the mutated CD44 tail (C), Listeria was able to invade the periphery (arrows in (A, C)) and the centre of a cellular islet (arrowheads in (A, C)). Conversely, in cells expressing wild-type CD44 tail or the ezrin mutant T567A, Listeria did not infect the cells in the centre of a cellular islet (arrowheads in (B, D)) but only the cells localised at the periphery of the islet (arrows in (B, D)). The ability of Listeria to spread from cell to cell correlated with the formation of normal protrusions (a, c), whereas the deficient cell-to-cell spread of these bacteria corresponded to the formation of defective protrusions (b, d). Scale bars: 20 μm (A–D); 1 μm (a–d).

Impairment of Listeria protrusion formation by overexpression of the cytoplasmic tail of CD44. (A–E) SEM analysis of Listeria protrusions in parental RT4 cells (A, C), RT4 cells expressing wild-type CD44 cytoplasmic tail (B, D) and a mutated version of the CD44 cytoplasmic tail (E). In control cells and in cells expressing the mutated CD44 cytoplasmic tail, Listeria induced the formation of long protrusions (arrow in (A)), which are connected to the cell surface through a thin stalk (facing arrowheads in (C, E)). Conversely, in RT4 cells expressing wild-type CD44 cytoplasmic tail, the protrusions are shorter (arrow in (B)) and characterised by a thick and distorted proximal portion (arrowhead in (D)). Stars indicate the position of bacteria. Scale bar: 2 μm (A, B); 0.8 μm (C); 0.3 μm (D); 1 μm (E). (F) Quantification of the number of Listeria protrusions in control RT4 cells and in RT4 cells expressing wild-type or mutated forms of CD44 cytoplasmic tail. Expression of wild-type CD44 cytoplasmic tail impairs protrusion formation, whereas its mutated variant had no effect. Error bars indicate one standard deviation from the mean.

Localisation of ezrin and its GFP-tagged amino- and carboxy-terminal domains during Listeria-induced protrusion formation. HeLa cells were infected with Listeria, fixed and stained with fluorescent phalloidin (A) and a polyclonal antibody against ezrin (B). Listeria-induced protrusions are easily distinguished from intracellular actin tails since they appear slimmer and are connected to the cell periphery by a thin stalk (arrowheads in (A)). Ezrin can be detected at actin comet tails in protrusions (arrowheads in (B); corresponding actin labelling in (A)) but not along the actin tails in the cell body (arrows in (B); corresponding actin labelling in (A)). (C, D) Dynamics of the GFP-tagged amino-terminal (C) and the carboxy-terminal domain (D) of ezrin in HeLa cells infected with Listeria. Panels in (C) represent GFP fluorescence (left) and phase contrast images (right), respectively. The amino-terminal domain of ezrin localised at the membrane surrounding a Listeria-induced protrusion (arrowheads, left panels in (C)) harbouring the bacterium at its tip (arrows, right panels in (C)). The carboxy-terminal domain (D) of ezrin localises at the actin tails associated with motile Listeria (arrows) and can also be detected around nonmotile bacteria within the cell body (arrowheads). Scale bars: 5 μm (A–C); 10 μm (D).

Downregulation of ERM proteins by siRNA impairs the formation of Listeria protrusions. (A) HeLa cells, treated for 30 h with two different siRNA sets specific for ezrin, moesin and radixin, a scrambled siRNA set or left untreated, were infected with Listeria for 6 h. Afterwards, cells were fixed and labelled with fluorescent phalloidin and an antibody for ERM proteins, and the number of protrusions per cells was determined. The treatment of the cells with both siRNA sets caused a marked impairment of protrusion formation compared to untreated cells and cells treated with a scrambled siRNA set. Error bars in A indicate one standard deviation from the mean. (B) Cell lysates were resolved by SDS–PAGE, blotted and probed with an ERM antibody. Both sets of siRNA duplexes caused a strong reduction in the levels of ERM proteins after 36 h, whereas the scrambled siRNA set had no effect. Actin served as the loading control.

Overexpression of ERM protein domains impairs protrusion formation. (A–C) Listeria-infected HeLa cells expressing excess amounts of GFP-tagged amino- or carboxy-terminal domains of ezrin were fixed and processed for SEM. In nontransfected cells, these bacteria induced the formation of slender protrusions that are connected to the cell surface through a thin proximal portion (facing arrowheads in (A)), whereas in cells expressing the amino- (B) or carboxy-terminal (C) domain of ezrin, the protrusions had a distorted proximal portion that was thicker than that in control protrusions (arrowheads in (B, C)). White stars indicate the position of bacteria. Scale bars: 2 μm. (D) Quantification of the number of Listeria protrusions in control HeLa cells and in cells expressing GFP-tagged amino- or carboxy-terminal domain of ezrin. Following infection with Listeria, HeLa cells were fixed and then stained with fluorescent phalloidin to detect actin comet tails. The expression of both ezrin domains decreased the number of protrusions as compared to control cells. Error bars indicate one standard deviation from the mean.

Phosphorylation-dependent activation of ERM proteins is critical for efficient protrusion formation. (A) Control LLC-PK1 cells and cells stably expressing the ezrin mutants T567A or T567D were infected with Listeria and processed for light microscopy. The expression of the ezrin mutant T567A, which mimics a constitutively inactive variant of this protein, reduces the number of protrusions compared to control LLC-PK1 cells and cells expressing the ezrin mutant T567D. (B) Control RT4 cells (WT) and cells stably expressing the ezrin mutants T567A (T/A) or T567D (T/D) were treated overnight with (+HA) or without (−HA) HA and then infected with Listeria. In control cells, the HA-induced activation of merlin reduces the number of protrusions per cell (see also Figure 5A). As expected, in RT4 cells expressing ezrin T567A, the number of protrusions induced by Listeria is reduced and this reduction is not enhanced by the merlin-driven displacement of ERM proteins from the plasma membrane. Conversely, the expression of ezrin T567D is able to counteract the reduction in the number of protrusions caused by merlin activation (compare WT with T567D). (C) Quantification of protrusion length for the experiment described in (B) showing that ezrin T567D counteracts the merlin-induced displacement of endogenous ERM proteins, thus resulting in the formation of protrusions of normal length, as determined by SEM. Error bars (A–C) indicate one standard deviation from the mean.

Displacement and dephosphorylation of ERM proteins correlate with the impairment of protrusion formation. (A) HA-induced activation of merlin impairs protrusion formation. Parental RT4 cells (control) and RT4 cells expressing wild-type merlin (NF2 WT 5/4) or a mutated variant of merlin (NF2 L64P) were infected with Listeria and processed for fluorescent microscopy. In RT4 cells treated with doxycycline and with HA to activate merlin (NF2 WT 5/4 +HA), protrusion formation was decreased as compared to untreated parental cells (control), RT4 cells expressing a mutated variant of merlin (NF2 L64P +HA) or RT4 cells not treated with HA (NF2 WT 5/4 −HA). (B) Specific engagement of CD44 reduces protrusion formation. After overnight treatment with anti-CD44 monoclonal antibodies 1.1asml or 5G8 and with doxycycline, RT4 cells were infected with Listeria and processed for fluorescence microscopy. As compared to control (untreated) cells, RT4 cells treated with the 1.1asml antibody supported less efficiently the formation of protrusions, whereas the treatment with the 5G8 antibody had no effect on this process. Error bars (in A, B and D) indicate one standard deviation from the mean. (C, D) Binding of HA or specific antibodies to CD44 induces NF2 and ERM protein dephosphorylation. Lysates from RT4 cells untreated or treated with HA or CD44 monoclonal antibody 1.1asml were resolved by SDS–PAGE and probed with antibodies against the phosphorylated form of ERM proteins (C) and NF2 (D). Both HA and the monoclonal 1.1asml induced a marked decrease in the level of phosphorylated ERM proteins and NF2 compared to control untreated samples. The third panel from top in (C, D) shows the actin loading control. Bottom panels in (C, D) indicate total NF2 showing that NF2 activation (dephosphorylation) does not cause NF2 degradation.