Portal de Periódicos da CAPES

Listeria monocytogenes exploits ERM protein functions to efficiently spread from cell to cell

2005; Springer Nature; Volume: 24; Issue: 6 Linguagem: Inglês

10.1038/sj.emboj.7600595

1460-2075

Sascha Pust, Helen Morrison, Jürgen Wehland, Antonio Sechi, Peter Herrlich,

3D Printing in Biomedical Research

Article24 February 2005free access Listeria monocytogenes exploits ERM protein functions to efficiently spread from cell to cell Sascha Pust Sascha Pust Department of Cell Biology, Gesellschaft für Biotechnologische Forschung (GBF), Braunschweig, Germany Present address: Albert-Ludwigs Universität Freiburg, Institut für Experimentelle und Klinische Pharmakologie und Toxikologie I, Albertstraße 25, 79104 Freiburg, Germany Search for more papers by this author Helen Morrison Helen Morrison Forschungszentrum Karlsruhe, Institute of Toxicology and Genetics, Karlsruhe, Germany Institute of Molecular Biotechnology, Jena, Germany Search for more papers by this author Jürgen Wehland Jürgen Wehland Department of Cell Biology, Gesellschaft für Biotechnologische Forschung (GBF), Braunschweig, Germany Search for more papers by this author Antonio S Sechi Corresponding Author Antonio S Sechi Department of Cell Biology, Gesellschaft für Biotechnologische Forschung (GBF), Braunschweig, Germany Present address: Institut für Biomedizinische Technologien—Zellbiologie, Universitätsklinikum Aachen, Rheinisch-Westfälische Technische Hochschule (RWTH), Pauwelsstrasse 30, 52057 Aachen, Germany Search for more papers by this author Peter Herrlich Peter Herrlich Forschungszentrum Karlsruhe, Institute of Toxicology and Genetics, Karlsruhe, Germany Institute of Molecular Biotechnology, Jena, Germany Search for more papers by this author Sascha Pust Sascha Pust Department of Cell Biology, Gesellschaft für Biotechnologische Forschung (GBF), Braunschweig, Germany Present address: Albert-Ludwigs Universität Freiburg, Institut für Experimentelle und Klinische Pharmakologie und Toxikologie I, Albertstraße 25, 79104 Freiburg, Germany Search for more papers by this author Helen Morrison Helen Morrison Forschungszentrum Karlsruhe, Institute of Toxicology and Genetics, Karlsruhe, Germany Institute of Molecular Biotechnology, Jena, Germany Search for more papers by this author Jürgen Wehland Jürgen Wehland Department of Cell Biology, Gesellschaft für Biotechnologische Forschung (GBF), Braunschweig, Germany Search for more papers by this author Antonio S Sechi Corresponding Author Antonio S Sechi Department of Cell Biology, Gesellschaft für Biotechnologische Forschung (GBF), Braunschweig, Germany Present address: Institut für Biomedizinische Technologien—Zellbiologie, Universitätsklinikum Aachen, Rheinisch-Westfälische Technische Hochschule (RWTH), Pauwelsstrasse 30, 52057 Aachen, Germany Search for more papers by this author Peter Herrlich Peter Herrlich Forschungszentrum Karlsruhe, Institute of Toxicology and Genetics, Karlsruhe, Germany Institute of Molecular Biotechnology, Jena, Germany Search for more papers by this author Author Information Sascha Pust1,4,‡, Helen Morrison2,3,‡, Jürgen Wehland1, Antonio S Sechi 1,5 and Peter Herrlich2,3 1Department of Cell Biology, Gesellschaft für Biotechnologische Forschung (GBF), Braunschweig, Germany 2Forschungszentrum Karlsruhe, Institute of Toxicology and Genetics, Karlsruhe, Germany 3Institute of Molecular Biotechnology, Jena, Germany 4Present address: Albert-Ludwigs Universität Freiburg, Institut für Experimentelle und Klinische Pharmakologie und Toxikologie I, Albertstraße 25, 79104 Freiburg, Germany 5Present address: Institut für Biomedizinische Technologien—Zellbiologie, Universitätsklinikum Aachen, Rheinisch-Westfälische Technische Hochschule (RWTH), Pauwelsstrasse 30, 52057 Aachen, Germany ‡These two authors contributed equally to this work *Corresponding author. Department of Cell Biology, Gesellschaft für Biotechnologische Forschung (GBF), Mascheroder Weg 1, 38124 Braunschweig, Germany. Tel.: +49 531 6181241; Fax: +49 531 6181444; E-mail: [email protected] The EMBO Journal (2005)24:1287-1300https://doi.org/10.1038/sj.emboj.7600595 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Cell-to-cell spread is a fundamental step in the infection cycle of Listeria monocytogenes that strictly depends on the formation of bacteria-induced protrusions. Since Listeria actin tails in the protrusions are tightly associated with the plasma membrane, we hypothesised that membrane–cytoskeleton linkers would be required for initiating and sustaining their formation and the subsequent cell-to-cell spread. We have found that ezrin, a member of the ezrin, radixin and moesin (ERM) family that functions as a key membrane–cytoskeleton linker, accumulates at Listeria protrusions. The ability of Listeria to induce protrusions and effectively spread between adjacent cells depends on the interaction of ERM proteins with both a membrane component such as CD44 and actin filaments. Interfering with either of these interactions or with ERM proteins phosphorylation not only reduces the number of protrusions but also alters their morphology, resulting in the formation of short and collapsed protrusions. As a consequence, Listeria cell-to-cell spread is severely impaired. Thus, ERM proteins are exploited by Listeria to escape the host immune response and to succeed in the development of the infection. Introduction Listeria monocytogenes is a Gram-positive, facultative intracellular bacterium that causes food-borne infections, which can lead to abortion and diseases as severe as meningitis, septicaemias and gastroenteritis. At the cellular level, the infection cycle of Listeria is characterised by four major steps: adhesion to, and invasion of, host cells, escape from phagocytic vacuoles, actin-based intracellular movement and cell-to-cell spread. Except for the escape from phagocytic vacuoles, the other steps are characterised by an interplay between bacteria and the host cell actin cytoskeleton (Tilney and Portnoy, 1989; see also Vazquez-Boland et al, 2001; Portnoy et al, 2002). Several studies have demonstrated that Listeria induces its own uptake into nonphagocytic cells and accomplishes to move within them by subverting the function of key cytoskeletal components such as Ena/VASP proteins and the Arp2/3 complex (see Frischknecht and Way, 2001; Vazquez-Boland et al, 2001; Cossart et al, 2003). These studies not only culminated in the reconstitution of Listeria motility in cell-free systems (Loisel et al, 1999), but have also been instrumental for understanding actin cytoskeleton dynamics in general. In contrast, the mechanisms underlying Listeria cell-to-cell spread have not yet been intensively studied. This process begins when motile bacteria approach the inner face of the plasma membrane, thus triggering the formation of finger-like structures called protrusions, which harbour the bacteria at their tips. Subsequently, these protrusions, due to the force generated by actin polymerisation, can penetrate into adjacent cells (Tilney and Portnoy, 1989; Robbins et al, 1999), leading to the direct cell-to-cell transfer of the bacteria that allows them to escape both the humoral and the cytotoxic T-cell host immune responses (see O'Riordan and Portnoy, 2002). Since Listeria actin tails are closely juxtaposed to the plasma membrane during protrusion formation and cell-to-cell spread, it is reasonable to conceive that actin cytoskeleton–membrane interactions contribute to the onset and progression of these processes. Ezrin, radixin and moesin (ERM) proteins are a family of widely distributed membrane-associated proteins responsible for linking the plasma membrane to the underlying actin cytoskeleton (see Bretscher et al, 2002). The N-terminus of ERM proteins can interact with membrane components such as CD44, CD43, intercellular adhesion proteins (ICAMs), PtdIns (4,5) P2, and with the phosphoprotein EBP50 (Tsukita et al, 1994; Niggli et al 1995; Reczek et al, 1997; Serrador et al, 1997; Legg and Isacke, 1998; Yonemura et al, 1998). The C-terminal domain harbours a stretch of ∼30 amino acids that can bind to F-actin (Turunen et al, 1994; Pestonjamasp et al, 1995). ERM proteins exist in two functionally different states. In their inactive state, the C-terminal domain is thought to be associated with the N-terminal domain causing ERM proteins to acquire a 'closed' conformation (Gary and Bretscher, 1995; Henry et al, 1995; Pearson et al, 2000). Upon activation by, for example, the phosphorylation of a critical threonine residue in the C-terminal domain of ERM proteins, the interactions between the N- and C-terminal domains are disrupted, thus exposing critical binding sites for the membrane–cytoskeleton interactions (see Bretscher et al, 2002). Merlin, a tumour suppressor protein closely related to ERM proteins, exhibits a similar N-terminus but lacks actin-binding sites in its C-terminus. Apart from these distinct structural properties, merlin requires, in contrast to ERM proteins, dephosphorylation to be active. Moreover, ERM proteins and merlin have antagonistic functional properties: ERM proteins promote cell growth, whereas merlin suppresses growth and inhibits signal transduction (Eldridge, 1981; Trofatter et al, 1993; Morrison et al, 2001 and our unpublished results; see also Bretscher et al, 2002). ERM proteins have also been implicated in crucial steps of the life cycle of invasive bacteria that are characterised by the rearrangement of the actin cytoskeleton. Ezrin can be detected at the entry sites of the enteroinvasive bacterium Shigella flexneri in epithelial cells and the overexpression of its N-terminus impaired the ability of these bacteria to invade these cells (Skoudy et al, 1999). Furthermore, the hyaluronan (HA)-mediated interaction of Streptococcus pyogenes with host cells triggers ERM protein-dependent cytoskeletal changes and disruption of intercellular junctions, processes that precede their invasion of soft tissues (Cywes and Wessels, 2001). These changes are likely based on ERM protein dephosphorylation. Interestingly, ezrin also localises to Listeria protrusions but not to Listeria actin tails within the cell body (Sechi et al, 1997). In light of the key role of ERM proteins as membrane–cytoskeleton linkers, we hypothesised that they may favour the formation of Listeria protrusions and the subsequent cell-to-cell spread by crosslinking the actin tails to the surrounding plasma membrane. Here, we demonstrate that the interaction of active (phosphorylated) ERM proteins with both membrane components and actin tails is essential for efficient protrusion formation and cell-to-cell spread. Results ERM proteins link actin tails to the plasma membrane in Listeria protrusions Listeria protrusion formation and cell-to-cell spread can be ideally analysed in mature epithelial monolayers, which are expected to closely mimic the onset and development of these processes in vivo (see Temm-Grove et al, 1994, and references therein). However, protrusions inside mature epithelial monolayers are difficult to observe since the cells are tall and tightly packed against each other. We chose, therefore, to study protrusion formation in single or low-confluent cells, where they are easily detectable by light and electron microscopic techniques. Although the mechanisms underlying the formation of both types of protrusions may be different, their highly similar behaviour (see Robbins et al, 1999) makes the conclusions drawn here most likely appropriate for both types of protrusions. The observation that ezrin accumulates at Listeria-induced protrusions but not at the actin tails within the cell body (Figure 1A and B; see also Sechi et al, 1997) suggests that active ERM proteins may associate with Listeria protrusions by simultaneously binding to the actin comet tails and to the membrane surrounding them. To test this possibility, we generated GFP-tagged amino- and carboxy-terminal domains of ezrin that can be defined as independent entities based on biochemical and structural data (see Pearson et al, 2000; Bretscher et al, 2002). To reduce the potential interference of GFP with the binding activity of both ezrin domains, the GFP moiety was cloned at the COOH and the NH2 ends of the amino- and carboxy-terminal domains of ezrin, respectively. In addition, to minimise a potential inhibitory effect of these constructs on protrusion formation, cells expressing low levels of both fusion proteins were analysed. According to previous observations (Algrain et al, 1993; Amieva et al, 1999), in uninfected HeLa cells, GFP-tagged amino- and carboxy-terminal domains of ezrin localised to the membrane and actin filaments, respectively (not shown), indicating that GFP did not alter their binding properties. In Listeria-infected HeLa cells, the GFP-tagged amino-terminal domain of ezrin localised to the membrane surrounding the protrusion but was not detectable along actin comet tails within the cytoplasm (Figure 1C), whereas the GFP-tagged carboxy-terminal domain of ezrin localised to the actin tails of motile Listeria (Figure 1D). Figure 1.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). Download figure Download PowerPoint Interaction of ERM proteins with the plasma membrane and actin filaments is essential for efficient protrusion formation Since ERM protein domains expressed separately act as dominant-negative elements on ERM protein functions (Henry et al, 1995; De Joussineau et al, 2003), we reasoned that the expression of sufficiently high levels of ezrin domains would impair protrusion formation by interfering with endogenous ERM protein functions. Since the low magnification and resolution provided by the light microscope did not allow an unequivocal assessment of protrusion morphological features, we analysed these structures by scanning electron microscopy (SEM). Control Listeria protrusions had a slender shape and were connected to the cell surface by a thin stalk (facing arrowheads in Figure 2A). Conversely, the protrusions formed in HeLa cells expressing GFP-tagged NH2 or COOH ezrin had a less slender shape and were usually connected to the cell surface by a thick and distorted end (arrowhead in Figure 2B and C; see Table I). Figure 2.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. Download figure Download PowerPoint Table 1. Quantification of the number of normal and atypical protrusions related to the SEM analysisa Cells examined Pseudopodia examined Normal pseudopodia Atypical pseudopodia HeLa cells 10 17 15 (88%) 2 (12%) HeLa cells/NH2 ezrinb 16 24 2 (8%) 22 (92%) HeLa cells/COOH ezrinb 15 23 3 (13%) 20 (87%) RT4 cells 35 52 48 (92%) 4 (8%) RT4 cells/CD44 tail wtc 25 33 4 (12%) 29 (88%) RT4 cells/CD44 tail mutc 19 38 32 (85%) 6 (15%) RT4 cells 15 20 17 (85%) 3 (15%) RT4 cells/+1.1asmld 15 26 4 (15%) 22 (85%) RT4 cells/ezrin T567Ae 20 47 21 (45%) 26 (55%) RT4 cells/ezrin T567A/+1.1asmld 26 56 28 (50%) 28 (50%) RT4 cells/ezrin T567De 19 30 25 (83%) 5 (17%) RT4 cells/ezrin T567D/+1.1asmld 23 43 35 (81%) 8 (19%) a Two to three experiments were conducted to quantify the number of protrusions. b Transient expression of GFP-tagged ezrin domains. c Stable expression of CD44 cytoplasmic tails. d Overnight treatment with 5 μg/ml 1.1asml anti-CD44 antibody. e Stable expression of ezrin T567A or T567D point mutants. We determined whether these morphological changes were accompanied by a reduction in the number of protrusions per cell. As the number of protrusions directly correlates with the number of motile bacteria, only cells that contained an equivalent number of motile bacteria (corresponding to the number of actin tails per cell) were analysed. Compared to control cells, the expression of either GFP-tagged NH2 or COOH ezrin caused a significant reduction in the number of protrusions per cell (Figure 2D; see Table II). Notably, the expression of ERM protein domains did not alter the morphology of Listeria actin tails within the cytoplasm (Figure 3A–C′) nor their length, which is proportional to the bacterial speed (see Theriot et al, 1992) (Figure 3D), indicating that intracellular Listeria motility per se was not affected in these assays. The specificity of our approach was further supported by the observation that protrusion formation was greatly impaired by the siRNA-induced downregulation of all endogenous ERM proteins in HeLa cells (Figure 4). Figure 3.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. Download figure Download PowerPoint Figure 4.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. Download figure Download PowerPoint Table 2. Statistical analysis (Student's t-test ) of the data presented in this worka Data set #1 Data set #2 P-value Significant difference Figure 2D HeLa control (n=50) HeLa NH2 ezrin (n=50) <0.0001 Yes HeLa control (n=50) HeLa COOH ezrin (n=50) <0.0001 Yes Figure 3D HeLa control (n=50) HeLa NH2 ezrin (n=50) 0.0443 Yes HeLa control (n=50) HeLa COOH ezrin (n=50) 0.006 Yes Figure 5A RT4 control (n=60) RT4 NF2 +HA (n=60) <0.0001 Yes RT4 control (n=60) RT4 NF2 −HA (n=102) 0.7306 No RT4 control (n=60) RT4 L64P +HA (n=60) 0.0013 Yes Figure 5B RT4 control (n=64) RT4 NH2 NF2 (n=61) <0.0001 Yes RT4 control (n=64) RT4 COOH NF2 (n=61) 0.5954 No Figure 5D RT4 control (n=120) RT4 1.1asml (n=126) <0.0001 Yes RT4 control (n=120) RT4 5G8 (n=120) 0.1915 No Figure 6F RT4 control (n=103) RT4 CD44 tail wt (n=108) <0.0001 Yes RT4 control (n=103) RT4 CD44 tail mut (n=103) 0.1422 No Figure 7 RPMC (n=90) RPMC CD44wt (n=90) <0.0001 Yes RPMC (n=90) RPMC CD44mut (n=90) 0.2753 No Figure 8A LLC-PK1 (n=100) LLC-PK1 ezrin T/A (n=100) <0.0001 Yes LLC-PK1 (n=100) LLC-PK1 ezrin T/D (n=100) 0.0017 Yes Figure 8B RT4 control −HA (n=100) RT4 control +HA (n=100) <0.0001 Yes RT4 ezrin T/A −HA (n=100) RT4 ezrin T/A +HA (n=100) 0.0242 Yes RT4 ezrin T/D −HA (n=100) RT4 ezrin T/D +HA (n=100) 0.1694 No RT4 control −HA (n=100) RT4 ezrin T/A −HA (n=100) <0.0001 Yes RT4 control −HA (n=100) RT4 ezrin T/D −HA (n=100) 0.0054 Yes RT4 control +HA (n=100) RT4 ezrin T/A +HA (n=100) 0.3343 No RT4 control +HA (n=100) RT4 ezrin T/D +HA (n=100) <0.0001 Yes Figure 8C RT4 control −asml (n=22) RT4 control +asml (n=28) <0.0001 Yes RT4 ezrin T/A −asml (n=57) RT4 ezrin T/A +asml (n=74) 0.2491 No RT4 ezrin T/D −asml (n=32) RT4 ezrin T/D +asml (n=39) 0.8670 No RT4 control −asml (n=22) RT4 ezrin T/A −asml (n=57) <0.0001 Yes RT4 control −asml (n=22) RT4 ezrin T/D −asml (n=32) 0.9216 No RT4 control +asml (n=28) RT4 ezrin T/A +asml (n=74) 0.6632 No RT4 control +asml (n=28) RT4 ezrin T/D +asml (n=39) <0.0001 Yes aThe statistical analysis was carried out by comparing data set #1 with data set #2. Displacement and dephosphorylation of ERM proteins correlate with the impairment of protrusion formation In another approach towards antagonising ERM protein localisation, we took advantage of the observation that the activation of merlin can displace ERM proteins from membrane components such as CD44 (Morrison et al, 2001). Since the COOH-terminus of merlin does not harbour F-actin-binding sequences, the merlin-induced displacement of ERM proteins would impair protrusion formation by interfering with ERM proteins–membrane, ERM proteins–actin interactions or both. In our system, the overexpression of merlin is under the tetracycline response promoter, and its activation can be induced by HA (Morrison et al, 2001). Active merlin decreased protrusion formation compared to control cells, cells in which merlin was inactive or cells that expressed an inactive merlin mutant (L64P) (Figure 5A and Supplementary data). Consistent with these data, the overexpression of the N-terminal domain of merlin (which, due to its structural similarity to ERM proteins, competes with them for binding to the membrane), but not the C-terminal domain of merlin (which is unable to bind to F-actin), reduced protrusion formation (see Supplementary data). Figure 5.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. Download figure Download PowerPoint Since HA activates merlin (characterised by its dephosphorylation) predominantly by signalling through CD44 (Morrison et al, 2001), we determined whether the engagement of CD44 with specific antibodies could affect protrusion formation. The CD44 monoclonal antibody 1.1asml, which activates merlin as efficiently as HA (Morrison et al, 2001) did, indeed, reduce protrusion formation. By contrast, the CD44 monoclonal antibody 5G8, which does not activate merlin (Morrison et al, 2001), had no effect on protrusion formation (Figure 5B). The engagement of CD44 by HA or 1.1asml clearly reduced the phosphorylation of both merlin and ERM proteins without affecting their total abundance (Figure 5C and D), suggesting that the impairment of protrusion formation depends not only on the displacement of ERM proteins but also on their dephosphorylation (see below). Overexpression of the ERM-binding site of CD44 impairs Listeria protrusion formation and morphology Since ERM proteins need docking sites at the plasma membrane to function properly as membrane–cytoskeleton linkers, the displacement of ERM proteins from the membrane by overexpressing the ERM-binding site of the cytoplasmic tail of CD44 (Legg and Isacke, 1998; Morrison et al, 2001) would impair protrusion formation. In control RT4 cells, Listeria protrusions were morphologically similar to those induced in HeLa cells (Figure 6A; facing arrowheads in Figure 6C). In contrast, protrusions formed in RT4 cells expressing CD44 cytoplasmic tail were reduced in number (Figure 6F), much shorter (4.71±1.85 μm (n=27) versus 9.19±2.83 μm (n=47) in control cells) and linked to the cell surface by a thick and collapsed end (Figure 6B; arrowhead in Figure 6D). The expression of comparable levels of the CD44 cytoplasmic tail mutated in its ERM-binding site (see Legg and Isacke, 1998) did not affect both protrusion number and morphology (Figure 6E and F). Figure 6.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. Download figure Download PowerPoint CD44 is sufficient to enhance protrusion formation but cannot rescue normal protrusion morphology in RPM-MC cells Since CD44 localises to Listeria protrusions (see Supplementary data), and sequestration of ERM proteins away from the membrane proteins such as CD44 could impair protrusion formation, we sought to determine whether CD44 plays