Differential mast cell outcomes are sensitive to FcεRI-Syk binding kinetics

Abstract

Cross-linking of immunoglobulin E-bound FcεRI triggers multiple cellular responses, including degranulation and cytokine production. Signaling is dependent on recruitment of Syk via docking of its dual SH2 domains to phosphorylated tyrosines within the FcεRI immunoreceptor tyrosine-based activation motifs. Using single-molecule imaging in live cells, we directly visualized and quantified the binding of individual mNeonGreen-tagged Syk molecules as they associated with the plasma membrane after FcεRI activation. We found that Syk colocalizes transiently to FcεRI and that Syk-FcεRI binding dynamics are independent of receptor aggregate size. Substitution of glutamic acid for tyrosine between the Syk SH2 domains (Syk-Y130E) led to an increased Syk-FcεRI off-rate, loss of site-specific Syk autophosphorylation, and impaired downstream signaling. Genome edited cells expressing only Syk-Y130E were deficient in antigen-stimulated calcium release, degranulation, and production of some cytokines (TNF-a, IL-3) but not others (MCP-1, IL-4). We propose that kinetic discrimination along the FcεRI signaling pathway occurs at the level of Syk-FcεRI interactions, with key outcomes dependent upon sufficiently long-lived Syk binding events.

FIGURE 1:

Syk Recruitment to FcεRI. Syk-KO RBL cells expressing Syk^mNG (green) were primed with AF647-IgE (magenta) and imaged after cross-linking with 0.1 µg/ml DNP-BSA. (A) Sample images from a confocal time series showing the redistribution of AF647-IgE-FcεRI and Syk^mNG upon FcεRI stimulation (see also Supplemental Video 1). Resting cross-section shows homogeneous distribution of AF647-IgE-FcεRI at the plasma membrane and Syk^mNG in the cytosol. Upon cross-linking (5 min), FcεRI aggregation and Syk^mNG colocalization is readily seen at the adherent cell surface. Scale bar: 10.3 µm. White boxes in the “Overlay” panels are enlarged in the “Zoom” panels. Scale bar: 2 µm. (B) Treatment with 1 µM dasatinib results in a loss of Syk^mNG recruitment (bottom) to FcεRI aggregates (top). Images of the adherent cell membrane acquired in TIRF. Scale bar: 5 µm. (C) Both FcεRI and Syk^mNG aggregates (top) are disrupted upon addition of 100 mM monovalent DNP-lysine (middle). Scale bar: 5 µm. Selected images from a time series (bottom) show the dispersion of an individual Syk aggregate within seconds of DNP-lysine addition. Individual Syk^mNG molecules can be seen diffusing away from the original diffraction-limited aggregate (see also Supplemental Video 2). Images acquired in TIRF. Scale bar: 1 µm. (D) Plot of positive correlation between Syk^mNG and AF647-IgE intensity within each AF647-IgE aggregate. (E) Selected images from a confocal time series before and after photobleaching (at t = 0 s) of an individual Syk^mNG aggregate. Scale bar: 1 µm. Bottom curve quantifies the rapid recovery of mNG fluorescence intensity within the bleached region (white circles).

FIGURE 2:

Quantifying FcεRI-Syk interactions using single-particle tracking of Syk^mNG. Single Syk^mNG molecules were tracked using TIRF microscopy as they bound and dissociated from the membrane. (A) Example of Syk^mNG trajectories detected at the plasma membrane (left). Projection of trajectories in the time dimension (middle) shows the trajectory lengths. Enlargement of a single trajectory that lasts ∼10 s (right). (B) Cumulative probability distribution of Syk^mNG trajectory lengths before (black) and 4–5 min after addition of 1 µg/ml DNP-BSA (green). Solid gray lines represent the fit to the data. (C) Bar graph depicts fast and slow off-rates found when fitting the distributions in B. See Materials and Methods for details. (D) Fraction of the slow off-rate component (α_s) increases from ∼5% to 23% after addition of 1 µg/ml DNP-BSA. Treatment with 1 µM dasatinib (Das) reduces α_s in both resting and activated cells. Error bars in C and D are a 68% credible interval as described in Materials and Methods. See also Supplemental Table S1.

FIGURE 3:

FcεRI-Syk off-rate is independent of antigen dose or aggregate size. (A) AF647-IgE–labeled RBL cells cross-linked for 5 min with indicated DNP-BSA concentration and imaged (left) at the adherent surface in TIRF (scale bar: 5 µm) and using (right) dSTORM (scale bar: 2 µm). (B) Clustering of localizations in dSTORM images from A using a hierarchical clustering algorithm (Matlab, MathWorks). Cluster sizes are shown as gray dots, and the distribution is summarized by the mean (red cross), the median (red line), and the 25th and 75th percentiles (blue box). Kolmogorov-Smirnov tests show significant (p < 0.01) differences between resting (0 ng) and other DNP-BSA doses. (C) Mobility of Syk^mNG represented as the cumulative distribution of squared displacements (r², ΔT = 0.3 s or 3 frames) for each DNP-BSA dose. See Supplemental Table S1 for values. (D) Cumulative distribution of Syk^mNG track lengths for each DNP-BSA dose. (E) The fraction of the slow off-rate component (α_s) increases with DNP-BSA dose. Data collected between 1 and 5 min after the addition of DNP-BSA. Error bars are a 68% credible interval as described in Materials and Methods. See also Supplemental Table S1. (F) Comparison of Syk^mNG trajectory localization with AF647-IgE aggregates in cells stimulated with 0.1 µg/ml DNP-BSA for up to 5 min. Left, overlay of Syk^mNG trajectories (green lines) with AF647-IgE aggregates (magenta). Scale bar: 1 µm. See also Supplemental Video 4. Middle, comparison of Syk^mNG track-length distributions inside and outside the AF647-IgE mask. A Kolmogorov-Smirnov test shows the distributions are significantly different (p < 0.01). Right, no correlation is observed between Syk^mNG track length and the intensity of the corresponding AF647-IgE aggregate.

FIGURE 4:

Syk^mNG-Y130E exhibits a faster FcεRI off-rate. (A) TIRF images of AF647-IgE (magenta) and Syk^mNG (green) membrane localization in Syk-KO cells reconstituted with Syk^mNG-WT (left), Syk^mNG-Y130E (middle), or Syk^mNG-Y130F (right) after 4–5 min stimulation with 0.1 µg/ml DNP-BSA. Scale bar: 5 µm. (B) Quantification of FcεRI recruitment capacity for Syk. Individual aggregates of AF647-IgE were masked, and the total intensity within the mask for both the AF647-IgE and Syk^mNG channels is plotted per aggregate. Recruitment is similar for Syk^mNG-WT and each mutant. (C) Cumulative probability distributions of trajectory lengths for Syk^mNG-WT, Syk^mNG-Y130E, and Syk^mNG-Y130F both before (dashed lines) and after (solid lines) addition of 1 µg/ml DNP-BSA. (D) Bar graph depicts the slow off-rate (k_s) found when fitting the distributions in C. Fraction of slow off-rate component (α_s) increases with DNP-BSA dose for Syk^mNG-WT and each mutant (right). Error bars are a 68% credible interval (see Materials and Methods).

FIGURE 5:

Key mast cell outcomes are impaired in Syk^mNG-Y130E cells. (A) Degranulation measured by relative β-hexosaminidase released after 30 min of incubation with 0.1 µg/ml DNP-BSA in Syk-KO cells reconstituted with Syk^mNG-WT, Syk^mNG-Y130E, or Syk^mNG-Y130F. (B) Comparison of cytokine concentration in cell media of Syk-KO cells reconstituted with Syk^mNG-WT, Syk^mNG-Y130E, or Syk^mNG-Y130F before and after 3-h stimulation with 0.1 µg/ml DNP-BSA. Results were repeated at least three times for all four cytokines. Bar plots represent mean and SD of technical replicates in one representative sample preparation. (C) The formation of actin plaques at the basolateral surface (left) and ruffling on the apical surface (right) in response to 0.1 µg/ml DNP-BSA treatment in both Syk^mNG-WT and Syk^mNG-Y130E reconstituted Syk-KO cells. Filamentous actin labeled with phalloidin-AF647 (red). Scale bar: 2 µm. (D, E) Heat maps of relative changes in intracellular calcium concentration upon either addition of a (D) low dose (0.001 µg/ml) or (E) high dose (1 µg/ml) of DNP-BSA. Each row represents the ratio of Fura-2 emission using 350-nm/380-nm excitation for a single cell over time. Ratio color bar scale range, one- to twofold increase: blue–red. (F) Relative increase in Fura-2 ratio per cell after addition. (G) Time between antigen addition and response for each cell. (F,G) Stimulated with 1 µg/ml DNP-BSA and calculated as described in Materials and Methods.

FIGURE 6:

Phosphorylation kinetics of Syk and downstream signaling partners. (A) Western blot detection of the Syk phosphorylation profile in Syk-KO cells reconstituted with Syk^mNG-WT, Syk^mNG-Y130E, or Syk^mNG-Y130F in response to stimulation with 0.1 µg/ml DNP-BSA for 5 min. Bar plots (right) represent mean and SD of relative Syk phosphorylation level from at least three experiments. Phosphorylation of tyrosines associated with Syk autocatalytic activity (pY519/520 and pY342) is significantly reduced (t test: *, p < 0.05; **, p < 0.01) in Syk^mNG-Y130E. No significant difference (n.s) in phosphorylation is seen at sites susceptible to Lyn catalytic activity (pY317, pY346). (B) Western blot detection of the phosphorylation time course of downstream signaling molecules Vav1, LAT, and ERK. Cells were stimulated with 0.1 µg/ml DNP-BSA for 0, 2, or 5 min. Bar plots (right) represent mean and SD of the relative phosphorylation levels from three experiments (t test: *, p < 0.05).