Yersinia effector protein (YopO)-mediated phosphorylation of host gelsolin causes calcium-independent activation leading to disruption of actin dynamics

Abstract

Pathogenic Yersinia bacteria cause a range of human diseases. To modulate and evade host immune systems, these yersiniae inject effector proteins into host macrophages. One such protein, the serine/threonine kinase YopO (YpkA in Yersinia pestis), uses monomeric actin as bait to recruit and phosphorylate host actin polymerization-regulating proteins, including the actin-severing protein gelsolin, to disrupt actin filaments and thus impair phagocytosis. However, the YopO phosphorylation sites on gelsolin and the consequences of YopO-mediated phosphorylation on actin remodeling have yet to be established. Here we determined the effects of YopO-mediated phosphorylation on gelsolin and identified its phosphorylation sites by mass spectrometry. YopO phosphorylated gelsolin in the linker region between gelsolin homology domains G3 and G4, which, in the absence of calcium, are compacted but adopt an open conformation in the presence of calcium, enabling actin binding and severing. Using phosphomimetic and phosphodeletion gelsolin mutants, we found that YopO-mediated phosphorylation partially mimics calcium-dependent activation of gelsolin, potentially contributing to a reduction in filamentous actin and altered actin dynamics in phagocytic cells. In summary, this work represents the first report of the functional outcome of serine/threonine phosphorylation in gelsolin regulation and provides critical insight into how YopO disrupts normal gelsolin function to alter host actin dynamics and thus cripple phagocytosis.

Keywords: Yersinia; actin; gelsolin; infectious disease; kinase; mass spectrometry (MS); phagocytosis; phosphorylation.

Figure 1.

YopO phosphorylates gelsolin in the linker region between G3 and G4. A, domain architecture of human cytoplasmic gelsolin with identified phosphorylation sites (indicated in bold). B, in vitro phosphorylation of gelsolin mutants by YopO. YopO WT (2 μm) was incubated with Sf9 actin (2 μm) and gelsolin mutants (12 μm) in the presence of [γ-³²P]ATP in kinase buffer. Proteins were separated by SDS-PAGE and visualized by Coomassie Blue staining (bottom panel). M denotes the molecular weight marker. Destained gels were dried and exposed to X-ray film (top panel). C, sequence alignment of human gelsolin residues 374–392 against gelsolin from different species, highlighting the conservation of Ser-385 (red box) in human, mouse, rat, pig, horse and bovine, but not chicken).

Figure 2.

Molecular dynamics simulations of the effects of phosphorylation by YopO on the gelsolin structure. A, structure of inactive gelsolin showing Ser-381 and Ser-385 and surrounding residues (as sticks) alongside the models of the phosphorylated state. B, change in the radius of gyration (R_g) of WT and phosphorylated gelsolin during MD simulations. In the case of WT, the simulation stabilizes after 50 ns. In contrast, for phosphorylated gelsolin, the R_g continues to increase throughout the course of simulation, indicating an opening of the gelsolin structure, which is a hallmark of its activation. C, evolution of the distance between the center of mass of N- and C-terminal halves of gelsolin during the MD simulation. In the case of WT gelsolin, the distance stabilized after an initial increase of 0.15 nm during the first 50 ns of the simulation. For phosphorylated gelsolin, the distance increased linearly with time, suggesting an opening of its inactive compact arrangement.

Figure 3.

YopO-phosphorylated gelsolin severs F-actin under EGTA conditions. A, sedimentation assay on in vitro-phosphorylated gelsolin under EGTA and calcium conditions. Polymerized F-actin (4 μm) was incubated with various concentrations of phosphorylated gelsolin (gelsolin/WT YopO) or non-phosphorylated gelsolin (gelsolin/KD YopO) in 1.5 mm EGTA or 1 mm CaCl₂ for 1 h at room temperature and subjected to ultracentrifugation. Supernatants (S) and pellets (P) were analyzed by SDS-PAGE. Redistribution of actin from the pellet to the soluble fraction, indicative of severing and/or sequestration, is observed for phosphorylated gelsolin. B, percentages of actin in the supernatant fractions (left) and pellet fractions (right) were plotted as a function of the total concentration of gelsolin in the presence and absence of EGTA conditions. Data represent the average of three experiments.

Figure 4.

Pyrene actin depolymerization assay with gelsolin phosphomimetic mutants in EGTA conditions. Polymerized F-actin (12 μm) was incubated with different gelsolin mutants (6 μm) in 1.5 mm EGTA or 1.0 mm calcium. The loss of pyrenyl fluorescence was measured. In EGTA, the triple mutant PM5 is more active at severing actin filaments, compared with WT and the other mutants. In calcium, phosphomimetic mutants and WT depolymerize actin filaments to similar extents.

Figure 5.

Sedimentation assay on gelsolin WT and the phosphomimetic mutant PM5. A, 4 μm F-actin was incubated with different concentrations of gelsolin WT or PM5 (0–0.3 μm) for 1 h at room temperature and subjected to ultracentrifugation, and the supernatants (S) and the pellets (P) were analyzed by SDS-PAGE. Left, 1.5 mm EGTA. Right, 1 mm calcium chloride. In the EGTA conditions, PM5 severs actin filaments to a greater extent compared with WT. However, at high calcium levels, both WT and PM5 sever filaments to similar extents. B, percentages of actin in the pellet fractions (left) and supernatant fractions (right) were plotted as a function of the total concentration of gelsolin in the presence and absence of EGTA. Data represent the average of three experiments. PM5 severs actin filaments to a greater extent, compared with WT in EGTA conditions. In contrast, at high calcium levels, both WT and PM5 sever filaments to the same extent.

Figure 6.

Pyrene actin depolymerization assay using gelsolin phosphomimetic mutants at different Ca²⁺ concentrations. Polymerized F-actin (12 μm) was incubated with different gelsolin mutants (6 μm) at a molar ratio of 2:1, and depolymerization was observed after 3 h by titrating against EGTA-buffered calcium levels. PM1 and PM5 require lower calcium levels to sever the filaments, in comparison with the other mutants and WT. Data represent the average of six measurements from two independent experiments in which each sample was set up and measured in triplicate. Data were fitted with Hill equation to obtain the K_d values of WT and different gelsolin mutants. p values are expressed relative to WT (***, p < 0.001; ****, p < 0.0001, NS, non-significant) as determined by one-way analysis of variance with Bonferroni post-tests, n = 6.

Figure 7.

Actin nucleation assay with gelsolin phosphomimetic mutants. G-actin was incubated with different gelsolin mutants at a molar ratio of 40:1 for 1 h at 0.1 mm CaCl₂ followed by the addition of 10× KMEI buffer. PM5 nucleated actin filaments less efficiently in comparison with WT. Data represent the average of three experiments.

Figure 8.

Model for the implications of phosphorylation by YopO on native gelsolin regulation.