Figure 1.
Loss of Scrb causes E-cadherin accumulation into perinuclear vesicles.
(A) MDCK cells stably expressing E-cadherin-GFP (Ecad-GFP) were nucleofected with either a control shRNA (LucKD) as a control or with an shRNA that specifically targets canine Scrb (ScrbKD), along with mRFP to mark transfected cells. Cells were plated on chambered coverslips and visualized live. Scale bars are 20 µm. (B) MDCK cells were nucleofected as in A, but using GFP as a transfection marker, then plated on slides for 2 d before fixing and staining for E-cadherin and GFP. (C) Normal localization of Ecad-GFP to the lateral membranes is rescued by expression of human, Flag-tagged Scrb with shRNA against the canine Scrb. (D) Cells were lysed in hypotonic buffer with no detergent, and centrifuged at 12,000×g for 10 min. Supernatant fractions were analyzed by immunoblot for adherens junction components. Tubulin was used as a loading control. The amounts of E-cadherin in the supernatant fractions were quantified and normalized to the intensities of the α-tubulin bands. Error bars represent mean +/− SEM (n = 3). (E) FRAP analysis of E-cadherin dynamics in GFP-E-cadherin MDCK stable cell line. Confocal sections of representative cells are shown before and after photobleaching. Curves show fluorescence recovery in control and knockdown cells (n = 5) (mean +/−1 SD). Mean intensities were adjusted for bleaching during imaging, normalized by subtraction of the residual fluorescence immediately after bleaching (∼35%) and analyzed using Prism software to fit a double-exponential model for recovery. Fluorescence intensity, I = A*(1-exp(−K1*t))+B*(1−K2*t)), where t = time after bleaching; K1 and K2 are rate constants, and A and B are the maximal fractional recoveries of Ecad-GFP in each component. For control cells, regression coefficient R2 = 0.55, for K1 = 0.05+/−0.025 sec−1, K2 = 0.0018+/−0.0034 sec−1, A = 0.15+/−0.039, and B = 0.31+/−0.37. For Scrb-depleted cells, R2 = 0.68, for K1 = 0.158+/−0.074 sec−1, K2 = 0.008+/−0.002 sec−1, A = 0.25+/−0.036, and B = 0.28+/−0.03. Equal amounts of whole cell lysates from control and ScrbKD MDCK II cells were analyzed by immunoblot for Scrb expression.
Figure 2.
E-cadherin endocytosis is promoted in the absence of Scrb.
(A) Schematic of experimental protocol to distinguish effects of Scrb depletion on delivery of E-cadherin to the plasma membrane versus on internalization of E-cadherin from the membrane. The rr1 antibody, which recognizes the extracellular domain of E-cadherin, will be retained at the plasma membrane at 4°C. Internalization is slow in control cells, so most of the antibody will still be at the plasma membrane after 2 h at 37°C. However, if internalization is accelerated by loss of Scrb, the rr1 antibody will be recruited along with E-cadherin into intracellular vesicles. (B) Control and Scrb-depleted Ecad-GFP cells were plated on 0.4 µm filters for 18 h. The E-cadherin extracellular domain-specific antibody (rr1) was added to the bottom wells and allowed to bind to cells for 1 h at 4°C. Cells were then washed to remove unbound antibody, incubated at either 4°C or 37°C for 2 h, then fixed and stained for the antibody (red) and for Scrb (blue). All images are confocal sections (Zeiss LSM 510; 40× oil immersion lens, NA 1.4). Scale bars are 20 µm. (C) Quantification of the overlap coefficient for colocalization of Ecad-GFP and rr1 (mean +/−1 SEM; n = 5 fields).
Figure 3.
Internalized E-cadherin is retrieve to the Golgi in Scrb-depleted cells, is associated with retromer components, and requires retromer for Golgi accumulation.
(A) Staining of Ecad-GFP expressing MDCK cells for mannose-6 phosphate receptor. Scale bars are 20 µm. (B) Co-localization of Ecad-GFP with the Golgi marker GM130. All images are confocal sections (Zeiss LSM 510; 40× oil immersion lens, NA 1.4). Overlap coefficients were determined using Openlab software (+/−1 SEM; n = 3). (C) Co-localization of Ecad-GFP with Golgi 58K. Overlap coefficients were determined using Openlab software (+/−1 SEM; n = 3).
Figure 4.
(A) Depletion of retromer component Vps29 blocks accumulation of E-cadherin in the Golgi.
Ecad-GFP cells were nucleofected with shRNAs to silence Scrb and/or Vps29. After 3 d cells were fixed, permeabilized and stained for GM130 and GFP. A second shRNA against Vps29 gave identical results. Confocal images were all captured using the same channel settings. (B) Regions of interest were drawn around the Golgi area for individual cells and the mean pixel intensities were measured for GFP fluorescence. Histogram shows mean GFP fluorescence +/−1 SEM (n = 5). (C) Silencing of Scrb increases the association of E-cadherin with retromer component Vps29. Cells were nucleofected with shRNAs. After 48 h cell lysates were immunoprecipitated with anti-E-cadherin and blotted for associated Vps29. Band intensities were quantified using Image J (+/−1 SEM, n = 3).
Figure 5.
The E-cadherin/p120 interaction is disrupted in Scrb-depleted cells.
(A) E-cadherin was immunoprecipitated from control and ScrbKD lysates as described in Methods, and the samples were blotted for p120. Band intensities were measured as described in Methods. Error bars = mean +/−1 SEM (n = 4). (B) Equal amounts of LucKD and ScrbKD total cell lysates were blotted for p120. Error bars indicate mean +/− SEM (n = 3). (C) β-catenin was immunoprecipitated from control and ScrbKD MDCK cell lysates and the samples were blotted for E-cadherin (mouse antibody) and β-catenin (rabbit antibody). Error bars indicate mean +/− SEM (n = 3). (D) Control and Scrb-depleted cells stably expressing Ecad-GFP were grown on slides for 48 h, then fixed and immunostained for p120 or β-catenin. Scale bars are 20 µm. Pixel intensities were measured across the lines shown and are displayed beneath the images. Quantification of co-localization between Ecad-GFP and the catenins was performed for multiple fields using Openlab. Error bars show mean +/−SEM (n = 5 fields per condition).
Figure 6.
Diversion of E-cadherin to the Golgi is independent of p120.
(A) Silencing Scrb reverses the loss of E-cadherin caused by depletion of p120. MDCK cells were transfected with shRNAs as shown, and incubated for 3d. Lysates were blotted for Scrb, p120 and endogenous E-cadherin. Tubulin was used as a loading control. Note the rescue of E-cadherin expression in the p120/Scrb double knockdown, as compared to the p120 single knockdown. (B) Silencing of Scrb diverts E-cadherin to the Golgi in cells lacking p120. Silencing of p120 alone results in the internalization and destruction of E-cadherin. Cells were stained for GM130 as a Golgi marker. Scale bars are 20 µm. (C) Schematic for Scrb function in regulating E-cadherin localization and recycling. Scrb is proposed to have two separate functions, first to stabilize p120 association with E-cadherin, and second to block retromer association with internalized E-cadherin, preventing its diversion to the Golgi.