AlphaE-catenin regulates actin dynamics independently of cadherin-mediated cell-cell adhesion

Abstract

alphaE-catenin binds the cell-cell adhesion complex of E-cadherin and beta-catenin (beta-cat) and regulates filamentous actin (F-actin) dynamics. In vitro, binding of alphaE-catenin to the E-cadherin-beta-cat complex lowers alphaE-catenin affinity for F-actin, and alphaE-catenin alone can bind F-actin and inhibit Arp2/3 complex-mediated actin polymerization. In cells, to test whether alphaE-catenin regulates actin dynamics independently of the cadherin complex, the cytosolic alphaE-catenin pool was sequestered to mitochondria without affecting overall levels of alphaE-catenin or the cadherin-catenin complex. Sequestering cytosolic alphaE-catenin to mitochondria alters lamellipodia architecture and increases membrane dynamics and cell migration without affecting cell-cell adhesion. In contrast, sequestration of cytosolic alphaE-catenin to the plasma membrane reduces membrane dynamics. These results demonstrate that the cytosolic pool of alphaE-catenin regulates actin dynamics independently of cell-cell adhesion.

Figure 1.

αE-catenin exists as a monomer and homodimer in MDCK cytosol. (A) Recombinant αE-catenin fractions collected from S200 gel filtration. The two peaks correspond to αE-catenin monomer (α^M) and homodimer (α^D). (B) Coomassie-stained native PAGE. Purified recombinant αE-catenin homodimer and monomer (α^D/α^M) served as markers for MDCK cytosol. (C) Recombinant αE-catenin homodimer and monomer and MDCK cell cytosol were run on a native PAGE gel and blotted for αE-catenin (left) and β-cat (right). αE-catenin monomer (α^M), homodimer (α^D), and heterodimer with β-cat (α/β) are marked. Note that recombinant αE-catenin and MDCK cytosol were run on the same gel but shown at different exposures, which are separated by a dotted line. (D, top) MDCK cytosol from cyclohexamide (CHX)-treated cells was separated by native PAGE and Western blotted for αE-catenin. (D, bottom) Total cell lysates were also separated by SDS-PAGE and blotted for αE-catenin and GAPDH. (E) Ratio of αE-catenin homodimer to monomer from experiment shown in D was determined by measuring band immunofluorescence intensity. (F) Native PAGE of MDCK cytosol blotted for vinculin and actin. Positions of α^M, α^D, and α/β complexes are marked for reference. Error bars represent SEM from two independent experiments.

Figure 2.

shRNA-mediated knockdown of αE-catenin depletes membrane and cytosolic pools. (A) Control MDCK and stable αE-catenin knockdown MDCK cells (αE-cat shRNA) were stained for αE-catenin, β-cat, E-cadherin, and F-actin. Bar, 10 µm. (B) Whole cell lysates from MDCK and αE-cat shRNA cells separated by SDS-PAGE and blotted for αE-catenin. (C) Band intensities shown in B were measured, normalized to MDCK control levels, and plotted. (D) Cytosol (cyto) and membrane (mem) fractions from MDCK and αE-cat shRNA cells were separated by SDS-PAGE and blotted for αE-catenin and β-cat. (E) Individual band intensities shown in D were measured, and percent distribution was plotted. (F) Native PAGE of MDCK and αE-catenin knockdown cytosol blotted for αE-catenin. αE-catenin monomer (α^M), homodimer (α^D), and heterodimer with β-cat (α/β) are marked. IB, immunoblot. Error bars represent SEM from at least three independent experiments.

Figure 3.

Targeting endogenous αE-catenin to mitochondria depletes the cytosolic αE-catenin pool. (A) Schematic of αE-catenin mitochondrial-targeting constructs. β-Cat–ActA, minimal αE-catenin–binding domain of β-cat (aa 92–179) fused to mRFP and the mitochondrial targeting region (aa 436–637) of ActA. ActA, control construct containing mRFP fused to aa 436–637 of ActA. (B) RFP and αE-catenin staining of β-cat–ActA and ActA MDCK cell lines. Yellow boxes highlight magnified regions shown in the bottom row. (C) Coimmunoprecipitation of αE-catenin from β-cat–ActA and ActA lysates using either RFP (lanes 4–6) or E-cadherin (lanes 7–9) antibodies. Blots probed for E-cadherin and RFP (top) and αE-catenin (bottom) are shown. Note that ActA migrates slightly faster than the IgG heavy chain (arrow). Asterisk marks cross reacting IgG heavy chain. β-Cat–ActA_1 and β-cat–ActA_2 were two independent, stable cell lines. (D) ActA- or β-cat–ActA-expressing MDCK cells (asterisks) mixed with wild-type MDCK cells and stained for αE-catenin. (E) The mean level of αE-catenin at cell–cell contacts was quantified between two expressing cells (E/E), an expressing and nonexpressing MDCK cell (E/N), and two nonexpressing MDCK cells (N/N). Results are presented in a box and whisker format. The ends of the box mark the upper and lower quartiles, the horizontal line in the box indicates the median, and the whiskers outside the box extend to the highest and lowest value within 1.5 times the interquartile range. Outliers are represented as dots. About 30 cell–cell contacts for each condition were measured. (F) Cytosol (C) and membrane (M) fractions from β-cat–ActA and ActA stable cells were separated by SDS-PAGE and blotted for αE-catenin, mtHsp70, and GAPDH. (G) αE-catenin band intensities shown in F were measured, and the percentage of αE-catenin in the cytosol fraction is graphed. (H) Native PAGE of ActA and β-cat–ActA cytosol blotted for αE-catenin (left) and β-cat (right). An additional slow-migrating band (asterisk) present in β-cat–ActA cytosol cross reacted with RFP (not depicted) and is presumed to be an αE-catenin/β-cat–ActA heterodimer synthesized in the cytoplasm that binds posttranslationally to mitochondria. (I) Immunofluorescence of EPLIN, ZO-1, α-actinin, β-cat, F-actin, vinculin (all shown in green), and RFP (red) in β-cat–ActA cells. (J) Immunofluorescence of αE-catenin (red) and afadin and mDia2 (green) in β-cat–ActA cells. Bars, 10 µm. Error bars represent SEM from three independent experiments.

Figure 4.

Sequestration of cytosolic αE-catenin to mitochondria increases cell migration but does not disrupt cell–cell adhesion. (A) ActA, β-cat–ActA, and αE-cat shRNA cell suspensions were triturated, fixed, and imaged at the indicated times. (B) Cell clusters were binned into the following classes: 1–11, 12–20, 21–50, 51–100, or >100 cells, and the percentage of cell clusters in a bin class at a given time point is shown. Data shown are a representative example from two independent experiments. (C) Confluent monolayers were scratched (0 h) and imaged over time to track wound closure. (D) Rate of wound closure was measured and plotted as the mean width of the wound over time and defined in arbitrary units (au). Data shown are a representative example from two independent experiments. (E) 250–400 individual cells from each cell type were tracked for 2 h during wound closure. Velocity is defined as length of track/time. Coordination is defined as 1/radius (the inverse of the mean difference of angle between a target cell and neighbors; the lowest possible value is 2/π [∼0.64], and the highest is infinity). Persistence is defined as deviation/track length (the lowest possible value is 0, and the highest is 1). Error bars indicate SEM. **, P < 0.002 (Mann-Whitney test). Bars, 100 µm.

Figure 5.

Cytosolic αE-catenin regulates actin-dependent membrane dynamics. (A and B) Representative kymographs of EGFP-actin in membrane protrusions from ActA and β-cat–ActA cells. (A) 2-pixel-wide kymographs were compiled parallel to protrusion direction over 10 min. (B) Number of protrusions per 10-min window (left) and the mean speed of protrusions (right) were measured in ActA and β-cat–ActA cells and shown using a box and whisker plot. 11 cells with 30 protrusions in ActA and 11 cells with 50 protrusions in β-cat–ActA cells were quantified. (C and D) Representative kymographs of membrane protrusions from MDCK and αE-cat shRNA cells (C) and quantification (D). 75 protrusions from 12 MDCK cells and 116 protrusions from 15 αE-cat shRNA cells were measured. Results are presented in a box and whisker format. The ends of the box mark the upper and lower quartiles, the horizontal line in the box indicates the median, and the whiskers outside the box extend to the highest and lowest value within 1.5 times the interquartile range. Outliers are represented as dots. *, P < 0.02; ***, P < 0.0002 (Mann-Whitney test). Bars, 5 µm.

Figure 6.

Depletion of αE-catenin cytosolic pool by cadherin-independent recruitment to membranes. (A) Schematic of αE-catenin membrane-targeting constructs. Lyn–β-cat, minimal αE-catenin–binding domain of β-cat (aa 92–179) fused to mCherry and a 10-aa palmitoylation and myristoylation (PM) sequence from Lyn. Lyn–β-cat mutant, similar to Lyn–β-cat but containing two mutated residues in the β-cat fragment (asterisks), shown to eliminate αE-catenin binding. (B) Cells expressing Lyn–β-cat or Lyn–β-cat mutant (asterisks) were mixed with wild-type MDCK cells and stained for αE-catenin. (C) The mean level of endogenous αE-catenin at cell–cell contacts was quantified as described in Fig. 3 E and graphed. 30–50 cell–cell contacts for each of the three conditions were measured. ***, P < 0.0002 (Mann-Whitney test). Results are presented in a box and whisker format. The ends of the box mark the upper and lower quartiles, the horizontal line in the box indicates the median, and the whiskers outside the box extend to the highest and lowest value within 1.5 times the interquartile range. Outliers are represented as dots. (D) Cytosol (C) and membrane (M) fractions from Lyn–β-cat and Lyn–β-cat mutant cells separated by SDS-PAGE and blotted for αE-catenin, mCherry, and GAPDH. (E) Percentage of αE-catenin in cytosol fraction from experiment shown in D was measured and plotted. (F) E-cadherin immunoprecipitates from Lyn–β-cat and Lyn–β-cat mutant cell lysates were blotted for E-cadherin and αE-catenin. Cyto, cytosol; mem, membrane. (G) Fc immunoprecipitates (IP) from Lyn-Fc–β-cat and Lyn-Fc–β-cat mutant cell lysates were blotted for αE-catenin and Fc. Sup, supernatant. (H) Native PAGE of Lyn–β-cat mutant and Lyn–β-cat cytosol blotted for αE-catenin (left), mCherry (middle), and β-cat (right). An additional band (asterisks) in Lyn–β-cat cytosol cross reacted with mCherry and αE-catenin and is likely an αE-catenin/Lyn–β-cat heterodimer. (I and J) Representative time-lapse montage of mCherry-labeled membrane protrusions in Lyn–β-cat mutant and Lyn–β-cat cells. (I) Arrows, membrane extensions; arrowheads, membrane retractions. (J) Quantification of membrane dynamics in 29 Lyn–β-cat mutant and 24 Lyn–β-cat cells. Change in membrane area was calculated as the mean difference in area between two 10-s frames normalized for cell area (left, schematic; right, quantification). **, P < 0.002 (Mann-Whitney test). Error bars indicate SEM from three independent experiments. Bars: (B) 10 µm; (I) 5 µm.

Figure 7.

Redistribution of cytosolic and membrane-associated αE-catenin pools affects Arp2/3 complex enrichment in lamellipodia. (A) Representative images from two ActA, β-cat–ActA, and Lyn–β-cat cells fixed and stained with anti-p34 antibody (Arp2/3 complex) and Alexa Fluor–labeled phalloidin (F-actin). Bar, 10 µm. (B) Fluorescence intensity of p34 and F-actin signals in lamellipodia was measured by line scan analysis. Mean fluorescence <3 µm extending from the cell edge (0) in the cell cortex was plotted. 40–50 protrusions from each cell type were measured at three separate points and averaged.

Figure 8.

Perturbation of cytosolic and membrane-associated αE-catenin pools alters actin ultrastructure. (A–I) Platinum replica electron microscopy of membrane protrusions from control ActA (A–C), β-cat–ActA (D–F), and Lyn–β-cat (G–I) cells. Red boxes denote magnified regions shown in the indicated panel. Pseudo coloring in C, F, and I highlight lamellipodia.