Clustering of Nck by a 12-residue Tir phosphopeptide is sufficient to trigger localized actin assembly

Abstract

Enteropathogenic Escherichia coli (EPEC) translocates effector proteins into mammalian cells to promote reorganization of the cytoskeleton into filamentous actin pedestals. One effector, Tir, is a transmembrane receptor for the bacterial surface adhesin intimin, and intimin binding by the extracellular domain of Tir is required for actin assembly. The cytoplasmic NH2 terminus of Tir interacts with focal adhesion proteins, and its tyrosine-phosphorylated COOH terminus binds Nck, a host adaptor protein critical for pedestal formation. To define the minimal requirements for EPEC-mediated actin assembly, Tir derivatives were expressed in mammalian cells in the absence of all other EPEC components. Replacement of the NH2 terminus of Tir with a viral membrane-targeting sequence promoted efficient surface expression of a COOH-terminal Tir fragment. Artificial clustering of this fusion protein revealed that the COOH terminus of Tir, by itself, is sufficient to initiate a complete signaling cascade leading to pedestal formation. Consistent with this finding, clustering of Nck by a 12-residue Tir phosphopeptide triggered actin tail formation in Xenopus egg extracts.

Figure 1.

Intimin-expressing bacteria attach efficiently to mammalian cells expressing membrane-targeted Tir. (A) Full-length Tir (TirFL) and TirMC, a Tir derivative lacking the NH₂-terminal cytoplasmic domain, are depicted. TirMC was directed to the plasma membrane by an NH₂-terminal targeting sequence derived from a viral HN-protein. Arrows indicate the positions of HA-epitope tags and Y474. (B) HEp-2 cells transfected with plasmids encoding TirFL or TirMC were stained with an mAb to the HA epitope to identify Tir-expressing cells, with an mAb to visualize phosphotyrosine (Y-PO₄), and with phalloidin to visualize F-actin. The inset shows several TirFL-expressing cells. (C) Mock-transfected HeLa cells, or cells expressing TirFL, TirMC, or TirMC(Y474F) (identified by anti-HA fluorescence) were challenged with a laboratory strain of E. coli carrying a plasmid-encoding intimin or a pUC19 vector control. Cell binding index, defined as the percentage of cells with at least five bacteria bound, was measured microscopically, and data represent the mean of two experiments.

Figure 2.

Tir expressed in mammalian cells allows a tir-deficient EPEC strain to form actin pedestals in the absence of the NH₂-terminal cytoplasmic domain. (A) HEp-2 cells expressing TirFL were infected with either EPECΔtir (top panels) or EPECΔtirΔeae (bottom panels) and stained as described in the Fig. 1 legend. The arrow indicates positions of several bound bacteria. (B) HEp-2 cells expressing either TirMC (top panels) or a Y474F mutant of TirMC (bottom panels) were infected with EPECΔtir and stained as described in the Fig. 1 legend. The arrow indicates positions of several bound bacteria. (C) HeLa cells transfected with plasmids encoding TirFL or TirMC were infected with wild-type EPEC, EPECΔtir + pTir, or EPECΔtir. Cells with at least five bound bacteria were examined, and the percentage of Tir-expressing cells (identified by anti-HA fluorescence) or Tir-nonexpressing cells generating actin pedestals was quantitated. Data represent the means of triplicate samples of 200 cells each. Similar results were observed in independent experiments.

Figure 3.

Clustering of the COOH terminus of EPEC Tir beneath the plasma membrane is sufficient to trigger actin pedestal formation. (A) HEp-2 cells expressing TirMC were challenged with a laboratory strain of E. coli harboring either an intimin-expressing plasmid (top panels) or a vector control (bottom panels). Tir and F-actin were visualized with an anti-HA antibody and phalloidin, respectively. The arrow indicates positions of bound bacteria. (B) HeLa cells expressing TirMC were treated with antibodies to either TirM, the extracellular domain of Tir (top panels), to the HA epitope (middle panels), or left untreated (bottom panels). Then, they were challenged with S. aureus particles that carry protein A on their surface (top panels) or anti–mouse IgG-coated S. aureus (middle and bottom panels). Tir and F-actin were visualized as described in A. S. aureus particles are visible when staining for TirMC because they bind the fluorescently labeled secondary antibody. The inset shows a magnification of several pedestals.

Figure 4.

Actin pedestal formation resulting from clustering of TirMC is quantitatively, kinetically, and morphologically similar to pedestal formation after EPEC infections. (A) TirMC-expressing HeLa cells were challenged with particles of the indicated type, and the percentage of bound particles generating actin pedestals was quantitated. Data represent the means ± SD of three experiments in which 50 cells were examined for each type of particle in each experiment. A range of 330–635 particles was counted per 50 cells. (B) HeLa cells were either transfected with a plasmid encoding TirMC or were primed with an EPEC strain that delivers TirFL and other Esps to host cells but does not express intimin. Cells were subsequently challenged with E. coli expressing intimin, fixed at various time points, and quantitated for pedestal formation as described in the Fig. 2 legend. Each data point represents the mean of two independent experiments in which 200 cells were examined. (C) The transfected or EPEC-primed HeLa cells described in B were examined at 10 or 60 min after challenge. Tir and F-actin were visualized with an anti-HA antibody and phalloidin, respectively. Arrows indicate examples of Tir localizations at the tips of pedestals.

Figure 5.

Clustering results in tyrosine phosphorylation of TirMC and recruitment of Nck, N-WASP, Arp2/3 complex, and α-actinin. (A) HeLa cells were either mock transfected and infected with wild-type EPEC (top panels) or were transfected with plasmids encoding TirMC or TirMC(Y474F) and challenged with E. coli expressing intimin (middle and bottom panels). Cells were stained with antibodies to detect phosphotyrosine (Y-PO₄), Nck, N-WASP, Arp3, or α-actinin (all shown in green). Bacterial DNA was visualized by DAPI staining (blue). (B) HeLa cells were mock transfected and infected with EPECΔtir + pTir (left lane), or were transfected with a plasmid encoding TirMC and treated with particles coated with anti-TirM antibodies (right lane). HA-tagged Tir was immunoprecipitated from cell lysates and Western blotted for HA-Tir and phosphotyrosine.

Figure 6.

Clustering of TirMC does not effectively trigger actin pedestal formation in cells that lack Nck. (A) TirMC was expressed in MEFs that express Nck1 (Nck ^+/− cells; top panels) and MEFs that do not express either Nck1 or Nck2 (Nck ^−/− cells; bottom panels). Cells were treated with anti-TirM antibodies and S. aureus particles and visualized as described in the Fig. 3 legend. (B) Nck-proficient MEFs (top panels) and Nck-deficient MEFs (bottom panels) were either mock transfected and infected with EPECΔtir + pTir (left panels) or were transfected with a plasmid encoding TirMC and challenged with E. coli expressing intimin (right panels). Cells were stained with antibodies to detect phosphotyrosine (green) and with DAPI to visualize bacterial DNA (blue).

Figure 7.

A clustered Nck-binding peptide from EPEC Tir is sufficient to trigger actin assembly in vitro. (A) Latex beads were coated with 12aa(+PO₄), a peptide that binds human Nck; 12aa, a nonphosphorylated peptide that lacks Nck-binding activity; or 7aa(+PO₄), a peptide that binds Nck poorly (Campellone et al., 2002). EPEC Tir residues present within the depicted peptides are underlined. Peptide-coated beads were added to Xenopus egg extracts supplemented with rhodamine-actin and examined microscopically. (B) Xenopus egg extracts were supplemented with the peptides described in A. Biotinylated peptides and associated proteins were collected from extracts using streptavidin-labeled particles and subjected to anti-Nck immunoblot analysis (all lanes except Crude Extract). Crude Extract lane is equivalent to 1/20 of the amount of extract represented in the bead-associated lanes.

Figure 8.

Nck is critical for actin assembly initiated by Tir peptide 12aa(+PO₄). Xenopus extracts were left untreated or were subjected to two rounds of immunodepletion with beads coated with either an anti-Nck antibody or with a rabbit IgG control antibody. Nck depletion was confirmed by immunoblot (not depicted). Recombinant GST-Nck was added to the Nck-depleted extract in the bottom panels. Extracts were supplemented with rhodamine actin, and beads coated with peptide 12aa(+PO₄) were added.