Cdc42 and Par6-PKCzeta regulate the spatially localized association of Dlg1 and APC to control cell polarization

Abstract

Cell polarization is essential in a wide range of biological processes such as morphogenesis, asymmetric division, and directed migration. In this study, we show that two tumor suppressor proteins, adenomatous polyposis coli (APC) and Dlg1-SAP97, are required for the polarization of migrating astrocytes. Activation of the Par6-PKCzeta complex by Cdc42 at the leading edge of migrating cells promotes both the localized association of APC with microtubule plus ends and the assembly of Dlg-containing puncta in the plasma membrane. Biochemical analysis and total internal reflection fluorescence microscopy reveal that the subsequent physical interaction between APC and Dlg1 is required for polarization of the microtubule cytoskeleton.

Figure 1.

Dlg1 interacts with APC at the leading edge of migrating astrocytes. (A) Cells were lysed immediately (0 h), 1 h, or 4 h after scratching. Immunoprecipitations were performed with anti-Dlg antibodies (IP Dlg1) or control rabbit anti–mouse IgG (Ctl) and were analyzed on Western blots (WB) using anti-APC or anti-Dlg1 antibodies. Bottom panel (input) shows equal APC and Dlg1 content in different cell lysates. (B) Cells were fixed and stained before (confluent), just after (0 h), and 4 h after scratching. Localization of Dlg1 was visualized with conventional epifluorescence. (C) Costaining of Dlg1 and APC visualized by confocal microscopy 0 and 4 h after scratching. Yellow in the merged image reflects colocalization of APC (red) and Dlg1 (green) clusters (similar results were observed 8 h after wounding). Bkgd, background. Higher magnification of the boxed area (solid lines) are shown on the right. Bars, 10 μm. (D) Quantification of APC and Dlg1 colocalization. Colocalization was measured in a central region of the cell in front of the nucleus (cell center) or at the leading edge (cell edge) in migrating or nonmigrating cells. Random background colocalization was measured in front of the leading edge (background; see Materials and methods). As an example, regions that were used for quantification are shown in C (dotted lines). Error bars represent SEM.

Figure 2.

Spatial organization of APC and Dlg1 at the leading edge. (A) Organization of the microtubule cytoskeleton in a scratched monolayer of astrocytes visualized by TIRF microscopy at low (left) and high (right) resolution. Conventional epifluorescence (red) and TIRF (green) images are superimposed on the high resolution image. (B) TIRF microscopy of EB1 (red) and microtubules (green). (C) TIRF microscopy of APC (red) and microtubules (green; left) and APC (red) and EB1 (green; right). Note that APC clusters are in proximity of microtubule plus ends only at the leading edge in contrast to EB1 clusters. (D) Low and high (E) magnification TIRF images (red or white, Dlg1; green, tubulin). (E) Two pools of Dlg1 can be distinguished: one pool that is associated with microtubule tips (arrowheads) and a membrane-associated pool that is localized in front of microtubules (arrows). Bars, 10 μm; (left panels in A and D), 20 μm. (F) TIRF microscopy of Dlg1 (red; left) and actin (green; right). (C, E, and F) Higher magnifications of the boxed areas are shown on the right.

Figure 3.

A Cdc42-PKCζ–dependent, GSK-3β/APC-independent pathway controls Dlg1 localization. (A) APC constructs that were used in this study. Astrocyte monolayers were scratched, and leading edge cells were immediately microinjected with the indicated constructs or incubated in the presence of PKCζ pseudosubstrate (PKCζ-PS; 10 μM for 1 h). Numbers correspond to the amino acid sequences of APC. (B) APC localization visualized with epifluorescence (green, tubulin; red, APC). (C) Percentage of cells with APC clusters at the leading edge. (D) Dlg1 localization visualized with epifluorescence. (B and D) 4 h after wounding, cells were fixed and stained with antibodies recognizing the microinjected constructs. Cells expressing the injected constructs are indicated by an arrow. Similar results were observed 8 h after wounding. Bars, 10 μm. (E) Percentage of cells with Dlg1 recruitment at the leading edge. (C and E) Results are means ± SEM of three independent experiments scoring at least 150 cells.

Figure 4.

APC–Dlg1 interaction is required for astrocyte polarization. When indicated, cells were nucleofected with pEGFP and siRNA and incubated for 3 d. Monolayers were scratched, microinjected with the indicated constructs, or incubated in the presence of PKCζ pseudosubstrate (PKCζ-PS; 10 μM for 1 h). (A) Polarized microtubule anchoring at the plasma membrane was assessed in astrocytes expressing the indicated constructs. (B) 8 h after wounding, cells were fixed and stained with antibodies recognizing microinjected constructs (green cells expressing the injected constructs are indicated by asterisks), antipericentin (red), and Hoechst (blue). Red lines indicate the directions of the scratch. Bars, 10 μm. (C) Centrosome polarization was assessed in astrocytes expressing the indicated constructs. As described in Materials and methods, 25% of polarized centrosome corresponds to a random orientation. Results are means ± SEM of three independent experiments scoring at least 100 (A) or 300 (C) cells. (D) Schematic diagram showing molecular pathways occurring at the leading edge of migrating astrocytes that control APC and Dlg1 localization and subsequent microtubule polarization and cell migration.