N-WASP is required for actin-based motility of E. coli (IcsA). a, N-WASP is not detected in platelet extracts. N-WASP expression was analyzed by SDS-PAGE in brain extracts (lanes 1 and 3) and in platelet extracts (lanes 2 and 4). Lanes 1 and 2, Coomassie blue staining; lanes 3 and 4, immunoblotting with anti-N-WASP. Note the presence of WASP, but no N-WASP, in platelets (lane 4). b, Purified recombinant histidine-tagged human N-WASP protein. SDS-PAGE patterns of insect cell extracts (lanes 1 and 3) and purified recombinant human N-WASP (lanes 2 and 4). Lanes 1 and 2, Coomassie blue staining; lanes 3 and 4, immunoblotting with anti-N-WASP. c, Full-length N-WASP, but not the isolated VCA, binds to IcsA. GST–IcsA_53-508 or GST beads were incubated with N-WASP or VCA in a pull-down assay. Free protein (F) and protein bound to the beads (B) were detected by immunoblotting with C412 anti N-WASP polyclonal antibodies. Lane 1, SDS-PAGE (Coomassie blue staining) of GST–IcsA. d, Binding of N-WASP to E. coli (IcsA) induces bacterial motility in platelet extracts. 1, Actin tails formed by N-WASP–coated E. coli (IcsA) incubated in platelets extracts; 2, absence of actin tails upon incubation of E. coli (IcsA) in platelet extracts; left panels, phase-contrast; right panels: rhodamine-actin fluorescence. Bars, 10 μm. e, Actin polymerization is induced at the surface of Sepharose-coupled GST–IcsA_53-508 beads coated with N-WASP. Actin polymerization occurs at the surface of Sepharose-coupled GST–IcsA_53-508 beads when recombinant N-WASP is added to platelet extracts (1). No polymerization is detected in the absence of N-WASP (2). Left panels: phase-contrast; right panels: rhodamine-actin fluorescence. Bars, 10 μm.

Arp2/3 complex is essential for actin nucleation at the surface of E. coli (IcsA) and Listeria. a, Arp2/3 complex depletion in platelet extracts. Increasing amounts (20, 40, and 80 μg) of glutathione Sepharose-coupled GST or GST-VCA beads were incubated with 25 μl of platelet extracts. Arp2/3 depletion is monitored by immunoblotting using anti-Arp3 polyclonal antibody. 1, Arp3 in untreated platelet extracts; 2–4, Arp2/3-depleted extracts; 5–7, mock-depleted extracts; 8–10, Arp3 associated with GST-VCA beads; and 11–13, Arp3 associated with GST beads. b, VCA affinity-purified Arp2/3 complex from bovine brain. The composition of the Arp2/3 complex is shown. The seven polypeptides are detected by Coomassie blue staining after SDS-PAGE. c, In vitro reconstitution of actin filaments assembly on E. coli (IcsA) and Listeria surface. N-WASP–coated E. coli (IcsA) or Listeria were incubated in a solution of pure rhodamine-labeled F-actin (10 μM) with or without 0.15 μM pure Arp2/3 complex as indicated. Left panels, phase-contrast; right panels, fluorescence. Bar, 5 μm.

Immunolocaliza-tion of N-WASP and Arp2/3 complex in HeLa cells infected with Shigella forming actin tails. Shigella-infected cells were fixed and immunolabeled with NW011 anti-N-WASP polyclonal antibodies (A) or with AR3 anti-Arp3 polyclonal antibodies (C) and with FITC-phalloidin (B and D). A and B represent the sum of all optical sections and C and D represent a single optical section close to focal plaques. Texas red-coupled secondary antibody was used to detect both immunocomplexes. Localization of N-WASP on the bacterium surface is indicated by an arrowhead. Bars, 10 μm.

IcsA mimics Cdc42 in activating N-WASP and stimulating actin polymerization by Arp2/3 complex. Mg-G-actin (2.5 μM, 9% pyrenyl-labeled) was induced to polymerize by addition of 0.1 M KCl and 1 mM MgCl₂ in the presence or absence of Arp2/3 complex (9.5 nM) and other reagents as indicated. a, VCA activates Arp2/3 complex. ×, Control (actin alone). For all other curves, Arp2/3 (9.5 nM) and VCA are at the following concentrations: ○, 0; ▿, 0.06 μM; ▵, 0.12 μM; ⋄, 0.24 μM; □, 0.62 μM; ▾, 1.24 μM; ♦, 2.48 μM; ▪, 3.7 μM. b, N-WASP activates Arp2/3 complex. Same conditions as in a, except N-WASP varied as follows: ○, 0; ⋄, 0.115 μM; □, 0.23 μM; ▵, 0.46 μM; ▴, 0.70 μM; •, 0.93 μM. c, Dependence of the activation of Arp2/3 on VCA and N-WASP concentrations. The maximum rate of polymerization recorded in panels a, b, and f (in the presence of IcsA) is plotted versus the concentrations of VCA (○) or N-WASP (•). d, Cdc42 activates N-WASP–Arp2/3 complex. Same as in b, with 0.2 μM N-WASP, in the absence (▵) or in the presence of GDP-Cdc42 (○) or GTPγS-Cdc42 (•). e, IcsA activates N-WASP–Arp2/3 complex. Actin was polymerized alone (▴); in the presence of 19 nM Arp2/3 (○), 0.25 μM IcsA (⋄) or 0.13 μM N-WASP (small ⋄); in the presence of Arp2/3 and 0.25 μM IcsA (▵); Arp2/3 and 0.13 μM N-WASP (□); Arp2/3, 0.13 μM N-WASP and 30 nM (○), or 0.25 μM (•) IcsA. f, Activation of Arp2/3 complex by N-WASP is more pronounced in the presence of IcsA. Actin was polymerized in the presence of Arp2/3, 0.1 μM IcsA and 0 (○), 0.042 μM (□), 0.087 μM (▵), and 0.55 μM (•) N-WASP.

Isolated VCA acts in a profilin-like fashion in actin assembly. a, Critical concentration plots for assembly of gelsolin-capped filaments (gelsolin:actin = 1:300) in the absence (○) or in the presence of VCA at 1 μM (⋄) or 2 μM (□). Pyrenyl fluorescence was measured at steady state. b, VCA sequesters actin when barbed ends are capped. Gelsolin-capped F-actin (gelsolin:actin = 1:300, 9% pyrenyl-labeled actin) was incubated overnight at 4.5 μM (a), 3.0 μM (b), and 1.5 μM (c), in the presence of VCA at the indicated concentrations. Pyrenyl fluorescence was measured at steady state. c, VCA does not efficiently sequester actin when barbed ends are free. Critical concentration plots were monitored in the absence (○) and in the presence of 5 μM VCA (•). Inset, Increasing amounts of VCA were added to 0.75 μM F-actin. Pyrenyl fluorescence was monitored at steady state. Note that 30 μM VCA appear required to depolymerize 0.6 μM F-actin when barbed ends are capped, whereas 4.6 μM VCA would be expected to be required, assuming a pure sequestering activity and a K_d of 0.8 μM for the VCA–G-actin complex. d, VCA–G-actin does not assemble onto pointed ends, but participates in barbed end assembly with the same kinetic parameters as profilin–actin. Rates of pointed end assembly from gelsolin–actin (○) and of barbed end assembly from spectrin–actin seeds (•) were measured at 1 μM G-actin and the indicated concentrations of VCA.

An NH₂-terminal region in N-WASP binds F-actin in an IcsA-enhanced fashion. a, Critical concentration plots for assembly of gelsolin-capped filaments in the absence (○ and ▵) and in the presence (•) of 3.6 μM N-WASP, or 2 μM GST–N-WASP–Nt fragment (▴). Inset, SDS-PAGE patterns of the pellets (P) and supernatants (S) of samples of 2 μM gelsolin-capped F-actin incubated in the presence of 0 (P₀ and S₀), 1.6 μM (P₁ and S₁), 2.12 μM (P₂ and S₂), and 3.6 μM (P₃ and S₃) N-WASP, and 0 (P′₀ and S′₀), 1 μM (P′₁ and S′₁), 2 μM (P′₂ and S′₂) GST–N-WASP–Nt fragment. P′′ and S′′ represent the pellet and supernatant of 2 μM GST–N-WASP–Nt alone. Note the decrease in actin in the supernatants, and the associated binding of N-WASP and its NH₂-terminal fragment to F-actin. b, N-WASP promotes actin polymerization from a pool of Tβ4–actin complex when barbed ends are free. N-WASP was incubated at the indicated concentrations with 2 μM F-actin (9% pyrenyl-labeled) in the presence of 30 μM Tβ4. Pyrenyl fluorescence was measured at steady state, and converted in F-actin concentrations using a critical concentration plot carried out in parallel using the same F-actin solution. Inset, SDS-PAGE pattern of the pellets (P) and supernatants (S) of sedimented samples (marked with an arrow in the main frame) containing 2 μM F-actin, 30 μM Tβ4, and 0 (P₀ and S₀), 2.09 μM (P₁ and S₁), and 4.65 μM (P₂ and S₂) N-WASP, respectively. c, IcsA strengthens the binding of N-WASP to F-actin. SDS-PAGE patterns of the pellets (P) and supernatants (S) of samples of F-actin (2 μM) incubated in the presence of 2 μM N-WASP and in the absence (P₀ and S₀) or in the presence of 0.15 μM (P₁), 0.3 μM (P₂), and 0.6 μM (P₃) GST–IcsA.

Our findings, combined with results from other laboratories (Machesky and Insall 1998; Miki et al. 1998b; Rohatgi et al. 1999), can be summarized using a thermodynamic scheme expressing that: both Arp2/3 complex and N-WASP exist in solution in two conformations, inactive (X) and active (X*; Fig. 1). The active form of Arp2/3 complex (A*) stimulates actin polymerization. The active form of N-WASP (W*) exposes two functionally active domains involved in actin-based motility, which may be masked in the inactive state by intramolecular interaction (Miki and Takenawa 1998; Miki et al. 1998b). The VCA binds Arp2/3 complex and G-actin, and the NH₂-terminal domain binds F-actin. Our data (Fig. 6 b) indicate that the F-actin binding site is most likely located in the WH1 domain, which also interacts with the verprolin homologue WIP. This formulation accounts for the fact that the Arp2/3 complex by itself has a weak nucleating activity (Mullins et al. 1998; Machesky et al. 1999) and that N-WASP by itself is able to bind F-actin and G-actin weakly (this work). Binding of IcsA (I) to N-WASP shifts the N-WASP equilibrium toward the W* form, exposing the VCA domain and the F-actin binding domain, thus enhancing the affinity of N-WASP for active Arp2/3 complex (Fig. 4 e) and F-actin (Fig. 6 c).

Working model for actin-based movement of Shigella. Left, Binding of N-WASP to IcsA at Shigella surface activates the connector, exposing the Arp2/3 complex, G-actin, and F-actin binding sites. Center, Nucleation: interaction of VCA with G-actin and Arp2/3 complex in a ternary complex in which the G-actin is positioned at the barbed end. Right, Barbed end growth and movement: the VCA domain of N-WASP shuttles G-actin subunits to the growing barbed end. The filament is maintained in close vicinity to the bacterium surface by binding to the NH₂-terminal domain of N-WASP.