The Salmonella kinase SteC targets the MAP kinase MEK to regulate the host actin cytoskeleton

Abstract

After host cell entry, Salmonella replicate in membrane-bound compartments, which accumulate a dense meshwork of F-actin through the kinase activity of the Salmonella SPI-2 type III secretion effector SteC. We find that SteC promotes actin cytoskeleton reorganization by activating a signaling pathway involving the MAP kinases MEK and ERK, myosin light chain kinase (MLCK) and Myosin IIB. Specifically, SteC phosphorylates MEK directly on serine 200 (S200), a previously unstudied phosphorylation site. S200 phosphorylation is predicted to displace a negative regulatory helix causing autophosphorylation on the known MEK activatory residues, S218 and S222. In support of this, substitution of S200 with alanine prevented phosphorylation on S218 and S222, and phosphomimetic mutations of S200 stimulated phosphorylation of these residues. Both steC-null and kinase-deficient mutant strains displayed enhanced replication in infected cells, suggesting that SteC manipulates the actin cytoskeleton to restrain bacterial growth, thereby regulating virulence.

Graphical abstract

Figure 1

Myosin II Is Recruited to Sites of SteC-Dependent F-Actin Reorganization in Its Phosphorylated, Activated Form (A and B) Swiss 3T3 fibroblasts expressing SteC-myc (blue) were labeled for Myosin IIB (green in A) or phospho-myosin light chain (PMLC, green in B) and stained for F-actin (red). Scale bars represent 40 μm (A) or 20 μm (B). (C–E) Swiss 3T3 fibroblasts were infected with different strains of S. Typhimurium (blue) as indicated and were labeled for Myosin IIB (green in C) or PMLC (green in E) and F-actin (red). Scale bars represent 8 μm. The percentage of microcolonies associated with Myosin IIB was quantified in (D). Results are expressed as means ± SEM of four independent experiments. ^∗∗p < 0.01. See also Figure S1.

Figure 2

Myosin IIB and MLCK Are Involved in SteC-Dependent F-Actin Remodeling RNAi-mediated knockdown of Myosin IIA in WT or Myosin IIB KO cells (A–C) or of myosin light chain kinase (MLCK) (D–F). (A) WT or Myosin IIB KO mouse embryonic fibroblasts (MEFs) were transfected with scrambled (Scr) or Myosin IIA (Myo IIA)-specific pools of siRNA oligos, as indicated. Myosin IIA and Myosin IIB levels were assessed by WB. The same blots were probed for tubulin as a loading control. (D) Swiss 3T3 fibroblasts were transfected with a Scr oligo or three individual siRNA oligos against MLCK (denoted MLCK1, MLCK2, and MLCK3). Whole-cell lysates were analyzed by WB against MLCK or tubulin. (B and E) siRNA-treated cells were infected with GFP-expressing WT S. Typhimurium (green) and stained for F-actin (red). Scale bars represent 8 μM. (C and F) Percentage of Salmonella microcolonies associated with F-actin in the conditions indicated. Results are expressed as means ± SEM of a least three independent experiments. ^∗∗∗p < 0.001. See also Figure S2.

Figure 3

MEK and ERK Contribute to F-Actin Meshwork Formation (A–C) WT MEFs were depleted of ERK1 and ERK2 with a mixture of siRNA oligos. Cells were infected with GFP-expressing S. Typhimurium (green in C), stained for F-actin (red) and scored for F-actin association with microcolonies (B). (D–F) Similarly, WT or MEK-1 KO MEFs were transfected with a scrambled (Scr) oligo or three individual MEK2-targeting siRNAs (denoted MEK2-1, MEK2-2, and MEK2-3) as indicated (D). Cells were infected with Salmonella (green in F) and stained for F-actin (red), and F-actin association with microcolonies was quantified (E). (G and H) MEFs depleted of ERK1/2 or MEK1/2 were infected for 8 hr with WT S. Typhimurium (blue in H) and were labeled for Myosin IIB (green) and stained for F-actin (red). Myosin IIB recruitment to Salmonella microcolonies was quantified (G). Results are expressed as means ± SEM of three independent experiments (B, E, and G). Scale bars represent 8 μm (C, F, and H). ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001. See also Figure S3.

Figure 4

Phosphorylation of MEK1 by SteC (A) Swiss 3T3 fibroblasts were infected with the indicated strains of S. Typhimurium (S. T) for 8 hr. Cells were then lysed and MEK1/2 or ERK1/2 phosphorylation was assessed by WB with antibodies specific for MEK1/2 phosphorylated on S218/S222 or ERK phosphorylated at the MEK (T202/T204) consensus site. Total MEK or ERK levels were assessed as loading controls. (B) WT or MEK1 KO MEFs were transfected with scrambled (Scr) or MEK2-targeting siRNA oligos. Cells were then infected with WT S. Typhimurium for 8 hr and ERK phosphorylation was measured as described above. (C) MEK1 was incubated with SteC or SteCK256H in kinase assay buffer with [γ^-32P] ATP at 30°C for 30 min. Proteins were subjected to SDS-PAGE followed by Coomassie blue staining and autoradiography. (D) Phosphorylation site identification of MEK phosphopeptide DVKPSNILVNSRGEIK. Tandem mass spectrum obtained for precursor ion at m/z 617.280³⁺ was used in conjunction with MS-Product (ProteinProspector v 5.7.2) to annotate both C-terminal (y) ions and N-terminal (b) ions present and confirm the site of phosphorylation (denoted by S). The y1 ion indicates that the peptide has a Lys (K) residue at its C terminus and the b2/b3 ions correlate with the N terminus of the peptide, Asp-Val-Lys (D-V-K). The loss of phosphoric acid (MH-H3PO4³+) from the triply charged precursor (MH³⁺) indicates that the peptide is phosphorylated and the site of phosphorylation is confirmed by the ions b6/b7 and y7-H3PO4 and y8-H3PO4. m/z, mass to charge ratio in atomic mass units (amu); cps, counts per second. See also Figure S4.

Figure 5

MEK1 Phosphorylation on S200 Induces Helix A Displacement and Autophosphorylation on S218/222 (A–C) Molecular dynamics (MD) simulations were carried out on native MEK1 (cyan), MEK1 phosphorylated on S218/222 (magenta), or MEK1 phosphorylated on S200 (salmon). All snapshots were taken at 35.6 ns. (A) Superposition of three snapshots showing helix A displacement when S200 is phosphorylated. Helix A was excluded for the calculation of the superposition. (B) Separation between the alpha carbon atoms of R49 and R201 over the course of the simulation. (C) Cartoon representation of the region around the T loop. S218, S222 and ATP are shown as sticks, Mg²⁺ as a white sphere. In the middle panel, two stabilizing Na⁺ ions are shown as yellow spheres. (D) MEK1 or MEK1K97R were incubated with SteC or SteCK256H in kinase assay buffer with [γ^-32P] ATP at 30°C for 30 min. Proteins were subjected to SDS-PAGE followed by Coomassie blue staining and autoradiography or WB with a phospho-specific antibody against MEK1 phosphorylated on S218 and S222. (E) MEK1 variants were immunoprecipitated from MEK1 KO cells expressing MEK1-Flag (MEK1-WT) or MEK1 S200A-Flag. Cells containing no transgene (KO) were used as a control. IP mixtures were incubated with recombinant SteC-His in kinase assay buffer at 30°C for 30 min. MEK1 phosphorylation on S218/222 was assessed by WB. The membrane was reprobed with anti-Flag and anti-His antibodies as loading controls. (F) REF52 cells expressing HA-MEK1, HA-MEK1 S200D, or HA-MEK1 S200E were serum starved for 16 hr. The indicated concentrations of fetal calf serum (FCS) were added to the growth medium for 10 min. MEK proteins were immunoprecipitated with an anti-HA antibody and assessed for their phosphorylation on S218/222 by WB. Membranes were reprobed with an HA antibody, as a loading control. See also to Figure S5.

Figure 6

SteC Controls Salmonella Replication (A and B) HeLa cells (A) or bone marrow (BM)-derived macrophages (B) were infected with the indicated S. Typhimurium strains. At the time points indicated, cells were lysed, and released bacteria were grown on rich medium for enumeration of colony-forming units (cfu). Values of fold increase in bacterial numbers were calculated from the ratio of the numbers of cfu/ml at each time point, and the number of intracellular bacteria at 2 hr after entry. Results are expressed as mean fold increase ± SEM of at least three independent experiments. ^∗p < 0.05. (C) Model of SteC-dependent F-actin reorganization: SteC activates MEK through phosphorylation of S200, causing displacement of helix A and autophosphorylation on S218/222. MEK1 phosphorylation leads to the activation of a signal transduction pathway involving ERK, MLCK, and Myosin IIB to induce cytoskeletal rearrangements in the vicinity of Salmonella microcolonies. SteC also induces a Myosin-independent signaling pathway to remodel the actin cytoskeleton.