The mammalian Scribble polarity protein regulates epithelial cell adhesion and migration through E-cadherin

Abstract

Scribble (Scrib) is a conserved polarity protein required in Drosophila melanogaster for synaptic function, neuroblast differentiation, and epithelial polarization. It is also a tumor suppressor. In rodents, Scrib has been implicated in receptor recycling and planar polarity but not in apical/basal polarity. We now show that knockdown of Scrib disrupts adhesion between Madin-Darby canine kidney epithelial cells. As a consequence, the cells acquire a mesenchymal appearance, migrate more rapidly, and lose directionality. Although tight junction assembly is delayed, confluent monolayers remain polarized. These effects are independent of Rac activation or Scrib binding to betaPIX. Rather, Scrib depletion disrupts E-cadherin-mediated cell-cell adhesion. The changes in morphology and migration are phenocopied by E-cadherin knockdown. Adhesion is partially rescued by expression of an E-cadherin-alpha-catenin fusion protein but not by E-cadherin-green fluorescent protein. These results suggest that Scrib stabilizes the coupling between E-cadherin and the catenins and are consistent with the idea that mammalian Scrib could behave as a tumor suppressor by regulating epithelial cell adhesion and migration.

Figure 1.

Suppression of Scrib expression in MDCK cells causes a delay in tight junction assembly. (A) MDCK II cells were transiently transfected with either pS-Luciferase vector (Luc) as a control or three pSUPER constructs targeting different sequences of canine Scrib mRNA (ScrbKD). Equal amounts of proteins were analyzed by SDS-PAGE and immunoblot. (B) Control and ScrbKD MDCK cells grown at high density were fixed and stained for Scrib and occludin. (C) Control and ScrbKD cells were subjected to a calcium switch and then fixed and stained for ZO-1 at the indicated times after readdition of calcium. Magnified details of the junctions are shown in inverted black-and-white scale to highlight the defects caused by loss of Scrib.

Figure 2.

Cells lacking Scrib lose their epithelial morphology at low density. (A) Control (Luc) and ScrbKD cells were grown on 6-well plates at low density for 3 d. Images were obtained by phase-contrast microscopy using a 10× objective. (B) Control and ScrbKD cells were fixed and stained with phalloidin to visualize F-actin. Typical colonies are shown. (C) Surface areas per cell (μm²) were measured for 80–90 frames (as shown in A) and sorted into bins 200 μm² wide. The percentage of cells in each bin is shown.

Figure 3.

Suppression of Scrib expression increases cell motility and attenuates orientated migration. (A) Motility of cells transfected with Luc, ScrbKD1 shRNA, or ScrbKD1 plus GFP-tagged human Scrib was measured using a Boyden chamber assay. Images were captured after crystal violet staining of the cells at the bottom side of the filter. (B) Quantification of cell migration by measuring A595 of eluted crystal violet. Error bars represent mean ± SD. (C) Immunoblot analysis of transfected cells with anti-Scrib antibody. Equal protein concentrations were loaded and normalized using anti-tubulin. (D) Rose plot of individual cell tracks from time-lapse movies of the wounding assay.

Figure 4.

βPIX is not involved in Scrib function at adherens junctions. (A) MDCK II cells were transiently transfected with either pS-Luciferase control vector (Luc) or three pSUPER constructs targeting different sequences of canine βPIX mRNA (PixKD). Proteins were analyzed by SDS-PAGE and immunoblot. (B) Quantification of Boyden chamber migration assay with Luc, ScrbKD, and ScrbKD plus PixKD cells. Error bars represent mean ± SD. (C) Immunoblot showing Scrib single KD and Scrib and βPIX double KD. (D) Rac pull-down assays with control and ScrbKD cells in both normal medium (top) and low-calcium medium (bottom). (E) Immunoblot analysis of phospho-ERK in control and ScrbKD cells with and without stimulation by 15 ng/ml HGF.

Figure 5.

Decreased adhesiveness of cells lacking Scrib. (A) 3 × 10⁴ control and ScrbKD cells were seeded into hanging drop cultures and allowed to aggregate overnight. After trituration by passing the cell cluster 10 times through a 200-μl pipette tip, images were captured by phase-contrast microscopy using a 10× objective. (B) E-cadherin is not preferentially lost from the ScrbKD cells in suspension culture. (C) Quantification of the degree of aggregation shown in A. Data are presented as the area of the aggregated cells/number of individual nonaggregated cells and represent means of 10–12 images from triplicates of each sample ± SD. (D) Effect of βPIX silencing on cell aggregation.

Figure 6.

Suppression of Scrib expression reduces E-cadherin–mediated adhesion. (A) In vitro E-cadherin binding assay. Different concentrations of recombinant E-cadherin extracellular domain were coated on plates. Control (Luc) and ScrbKD cells were collected in suspension and allowed to adhere to the plates for 60 min. Images were captured for nonwashed plates (to give total cell numbers), and plates were subjected to washing (to determine attached cell numbers). (B) Quantification of cell attachment to E-cadherin ectodomain. Data represent means of 10–12 images from triplicates of each condition ± SD.

Figure 7.

Loss of Scrib causes a defect in adherens junction structure. (A) Control and Scrib-depleted cells were plated at subconfluence and allowed to form islands of cells. They were then fixed and stained for E-cadherin (red) and Na/K-ATPase (green). Image stacks were collected using confocal microscopy with 0.95-μm z steps. (B) Control and ScrbKD cells were fixed and stained for Scrib and β-catenin. (C) Immunoblot of junctional proteins in control and ScrbKD cells. Whole cell lysates were blotted for E-cadherin, α- and β-catenin, and α-tubulin. (D) Cells were treated with a nonpermeable biotin linker, sulfo-NHS-SS-biotin, and biotinylated proteins were captured onto streptavidin-agarose and detected by immunoblot. (E) Control and ScrbKD cells plated at low density were lysed with buffer containing 0.5% Triton X-100. After centrifugation, supernatants and pellets were resolved by SDS-PAGE. Distributions of α- and β-catenin were detected with immunoblot. 25% of the soluble and 50% of the insoluble fractions are shown. (F) Quantification by Odyssey image system. Data are presented as the ratio of insoluble to soluble fractions. Error bars represent the SD from four independent experiments.

Figure 8.

Effects of E-cadherin depletion and expression of an E-cadherin–α-catenin fusion protein. (A) Cells were transfected with an shRNA targeted against canine E-cadherin. Lysates were immunoblotted for E-cadherin and Scrib. Equal protein concentrations were loaded and normalized using anti-tubulin. (B) Quantification of the motility of cells transfected with Luc or EcadKD shRNAs using Boyden chamber assay as described in Fig. 3. (C) Morphology of cells depleted of E-cadherin. Control and EcadKD cells were grown on six-well plates at low density for 3 d. Images were obtained by phase-contrast microscopy using a 10× objective. (D) Expression of an Ecad–αcat and Ecad–GFP fusion in MDCK cells. The fusion proteins and endogenous E-cadherin were detected by immunoblot with anti–E-cadherin antibodies. (E) Effects of the Ecad–αcat and Ecad–GFP fusions on motility of cells depleted of Scrib. Migration through filters was quantified as in Fig. 3. (F) Aggregation assay. 3 × 10⁴ cells were seeded into hanging drop cultures and allowed to aggregate for overnight. After trituration with a 200-μl pipet tip, images were captured by phase-contrast microscopy using a 10× objective. (G) Quantification of the aggregation assay. Data are presented as the area of the aggregated cells/number of individual nonaggregated cells per field and represent means of 10–12 fields from triplicates of each sample ± SD.