Stoichiometry of Nck-dependent actin polymerization in living cells

Abstract

Regulation of actin dynamics through the Nck/N-WASp (neural Wiskott-Aldrich syndrome protein)/Arp2/3 pathway is essential for organogenesis, cell invasiveness, and pathogen infection. Although many of the proteins involved in this pathway are known, the detailed mechanism by which it functions remains undetermined. To examine the signaling mechanism, we used a two-pronged strategy involving computational modeling and quantitative experimentation. We developed predictions for Nck-dependent actin polymerization using the Virtual Cell software system. In addition, we used antibody-induced aggregation of membrane-targeted Nck SH3 domains to test these predictions and to determine how the number of molecules in Nck aggregates and the density of aggregates affected localized actin polymerization in living cells. Our results indicate that the density of Nck molecules in aggregates is a critical determinant of actin polymerization. Furthermore, results from both computational simulations and experimentation support a model in which the Nck/N-WASp/Arp2/3 stoichiometry is 4:2:1. These results provide new insight into activities involving localized actin polymerization, including tumor cell invasion, microbial pathogenesis, and T cell activation.

Figure 1.

Aggregation of CD16/7-mCherry–Nck SH3 domains induces actin comet tails. (A) Representative confocal image of an NIH-3T3 cell cotransfected with GFP-actin and CD16/7-mCherry–Nck SH3 domains. Antibody-induced aggregation of Nck SH3 domains results in robust actin comet tails. A higher magnification of clusters is shown in the inset. Bar, 10 µm. (B) Time-lapse confocal images of two representative (n = 145) aggregate/actin comet tails (white and yellow arrows) demonstrate aggregate motility. Images were taken at 30-s intervals over the course of 4.5 min. Bars, 1 µm. (C) Quantitative analysis of individual actin comets. A plot of Nck SH3 molecules in aggregate versus aggregate velocity (n = 145) shows no apparent correlation.

Figure 2.

The Virtual Cell predicts experimental actin comet tail characteristics at steady state if aggregate size and velocity are known. We adapted a previously published Virtual Cell model to our experimental system to test the theoretical effect of aggregate size and velocity on actin polymerization. Using this model, we show that when aggregate size and velocity are known, actin polymerization downstream of Nck SH3 aggregates can be predicted. (A) Virtual Cell reaction diagram of the simplified actin dendritic nucleation model in which N-WASp is activated on the membrane by Nck. In the Virtual Cell reaction diagram, green circles represent species that participate in reaction, yellow ovals represent biochemical reactions, lines connecting species and reactions indicate the reactants and products of each reaction, and the vertical line in the left schematic designates the separation between membrane and cytosol. The reaction scheme on the left describes membrane reactions, whereas the reaction scheme on the right describes cytosolic reactions. The detailed schematic can be found in the Virtual Cell database under the user JDitlev, model Nck Induces Actin Comet Tail Formation Single NWASP Activation of Arp2/3 Simplified Actin Dendritic Nucleation. (B) Example of an Nck-induced actin comet tail from Virtual Cell simulations at steady state. Nck SH3 domains were localized to a patch on the membrane (red membrane mesh units) inducing actin polymerization in the adjacent mesh units in the cytosol (scaled from blue ∼195 µM to red ∼1,500 µM). Actin then flowed away from the Nck SH3 patch, as indicated by the advection arrow, at a constant velocity to mimic actin propulsion of the Nck SH3 aggregate. (C) Predicted versus experimental actin comet tails induced by aggregates traveling at similar velocities (0.060–0.090 µm/s). Experimental data were divided into five groups by aggregate size (n = 2–9). The mean experimentally measured actin concentrations from linescans were compared with Virtual Cell predictions. Virtual Cell simulations qualitatively predicted experimental results at a given aggregate velocity. Error bars indicate the SEM of averaged experimental data from each group at each measured point. (D) Predicted versus experimental actin comets tails induced by aggregates with a similar number of Nck SH3 molecules (12,000–22,000 molecules). Experimental data were divided into five groups by velocity (n = 2–11). The mean experimentally measured actin concentrations from linescans were compared with Virtual Cell predictions. Virtual Cell simulations predicted experimental results at a given aggregate size. Error bars indicate the SEM of averaged experimental data from each group at each measured point. (E) Four representative comparisons chosen from 145 experimentally measured actin comet tails with corresponding simulations. The following parameters were used for simulations run with the experimentally measured size (Exp.) and velocity of individual Nck SH3 aggregates (VCell): 17,000 molecules, 0.04 µm/s (top left); 10,000 molecules, 0.050 µm/s (top right); 17,000 molecules, 0.057 µm/s (bottom left); and 17,000 molecules, 0.120 µm/s (bottom right).

Figure 3.

Representative experimental images of aggregates containing different densities of Nck SH3 molecules. (A–E) To test the effect of decreasing Nck SH3 density on localized actin polymerization, Nck SH3 domains with dummy fusion proteins were cotransfected and aggregated, resulting in aggregates of varying density. In individual images, CD16-7–mCherry–Nck SH3–HA is red, CD16-7–mCerulean-HA is cyan, and YFP-actin is green. For the mCherry-Nck/YFP-actin merged images, brightness and contrast have been adjusted to allow for comparison of Nck SH3 density-dependent actin polymerization associated with Nck SH3 aggregates. Higher magnifications of clusters are shown in the insets. Bars, 10 µm. (A) Aggregates containing 80–100% dummy and 0–20% Nck SH3 (0–5,000 Nck SH3/µm²) induce no detectable actin polymerization. (B) Aggregates containing 60–80% dummy and 20–40% Nck SH3 (5,000–10,000 Nck SH3/µm²) induce minimal actin polymerization that results in the formation of sparse actin spots. (C) Aggregates containing 40–60% dummy and 40–60% Nck SH3 (10,000–15,000 Nck SH3/µm²) induce actin polymerization that results in the formation of dense actin spots. (D) Aggregates containing 20–40% dummy and 60–80% Nck SH3 (15,000–20,000 Nck SH3/µm²) induce actin polymerization that results in the formation of dense actin spots or short actin comet tails. (E) Aggregates containing 0–20% dummy and 80–100% Nck SH3 (20,000–25,000 Nck SH3/µm²) induce actin polymerization that results in the formation of a robust actin comet tail.

Figure 4.

Quantitative analysis of aggregates of various Nck SH3 density. (A–E) Aggregates were divided into different groups based on their Nck SH3 density. Measured characteristics of individual aggregates were averaged and plotted against Nck SH3 density. Error bars indicate SEM of each density group. Significance is indicated by asterisks above data points (P < 0.05), with each notation indicating a significant difference between each set of data. The plots show that increasing Nck SH3 density results in a nonlinear increase in the velocity of aggregates (A), actin comet tail length (B), number of actin molecules in the comet tail core (C), the peak actin concentration in actin foci or comet tails (D), and the distance between the aggregate centroid and the peak actin concentration in aggregate-induced actin structures (E). These data are also presented as the experimental dataset in Fig. 5 to allow for comparison with model predictions.

Figure 5.

Comparison between experimental results and Virtual Cell biochemical model predictions from simulations using different mechanisms of Arp2/3 activation. (A–D) Quantitative Virtual Cell biochemical simulations were performed using three different mechanisms of N-WASp (NW) and Arp2/3 complex activation, five different aggregate sizes, six different aggregate velocities, and seven different Nck SH3 densities. Models were simulated using experimentally measured aggregate sizes, velocities, and densities. Simulation results were weighted based on experimentally measured aggregate size and velocity distributions to allow for an accurate comparison of predicted and experimental results. The plots show comparisons of simulation results from models using one Nck/one N-WASp/one Arp2/3, two Nck/two N-WASp/one Arp2/3, or four Nck/two N-WASp/one Arp2/3 with experimentally measured actin comet tail length (A), number of actin molecules in the comet tail core (B), peak actin concentration (C), and distance between aggregate centroid and peak actin concentration (D). For each measured actin comet tail characteristic, the 4:2:1 Nck/N-WASp/Arp2/3 activation scheme best predicted experimental results, suggesting that this mechanism, and not the 1:1:1 or 2:2:1 mechanism, was responsible for the activation of Arp2/3 downstream of Nck SH3 domains. For graphs on right, data have been normalized to the maximum values within each group to allow for another comparison of predicted and experimental results. The data for the experimental datasets are duplicated from Fig. 4 for comparison with model predictions.

Figure 6.

WIP is an essential component of the Nck–N-WASp–Arp2/3 pathway. (A–D) Confocal images of either WIP WT and KO or N-WASp WT and KO MEFs transfected with a combination of mCherry-actin, membrane-bound Nck SH3 fusion, and GFP-WIP or GFP–N-WASp demonstrate the necessity of WIP for Nck-induced, N-WASp–dependent actin polymerization. Higher magnifications of clusters are shown in the insets. Bars, 10 µm. (A) Antibody-induced aggregation of Nck SH3 domains (green) induces the formation of actin comet tails (red) in WIP WT MEFs (top) but does not induce actin polymerization in WIP KO MEFs (bottom). (B) Aggregation of Nck SH3 domains (cyan) in WIP KO MEFs rescued with GFP-WIP (green) induces actin comet tails (red) similar to those seen in WIP WT MEFs. (C) Nck SH3 aggregates (cyan) neither recruit GFP–N-WASp (green) nor induce actin polymerization (red) in WIP KO MEFs (top), whereas Nck SH3 aggregates in WIP WT MEFs both recruit GFP–N-WASp and induce actin polymerization (bottom). (D) Nck SH3 aggregates (cyan) recruit GFP-WIP (green) in both N-WASp KO MEFs (top) and N-WASp WT MEFs (bottom) but only induce actin polymerization (red) in N-WASp WT MEFs.

Figure 7.

Proposed 4:2:1 Nck/N-WASp/Arp2/3 mechanism leading to actin polymerization. In our proposed mechanism, two Nck molecules bind each N-WASp; one Nck molecule binds directly to the proline-rich region of N-WASp, whereas a second Nck molecule binds the proline-rich region of WIP, which in turn binds the WH1 domain of N-WASp. Two activated N-WASp molecules then bind and activate Arp2/3, resulting in actin filament nucleation and polymerization. Although not drawn precisely to scale, the molecular mechanism presented here is intended to demonstrate the spatial arrangement of all proteins near the membrane in our experimental system. B, basic region; GBD, GTPase binding domain; WBD, WASp binding domain.