SGEF forms a complex with Scribble and Dlg1 and regulates epithelial junctions and contractility

Abstract

The canonical Scribble polarity complex is implicated in regulation of epithelial junctions and apical polarity. Here, we show that SGEF, a RhoG-specific GEF, forms a ternary complex with Scribble and Dlg1, two members of the Scribble complex. SGEF targets to apical junctions in a Scribble-dependent fashion and functions in the regulation of actomyosin-based contractility and barrier function at tight junctions as well as E-cadherin-mediated formation of adherens junctions. Surprisingly, SGEF does not control the establishment of polarity. However, in 3D cysts, SGEF regulates the formation of a single open lumen. Interestingly, SGEF's nucleotide exchange activity regulates the formation and maintenance of adherens junctions, and in cysts the number of lumens formed, whereas SGEF's scaffolding activity is critical for regulation of actomyosin contractility and lumen opening. We propose that SGEF plays a key role in coordinating junctional assembly and actomyosin contractility by bringing together Scribble and Dlg1 and targeting RhoG activation to cell-cell junctions.

Figure 1.

SGEF interacts with Scribble PDZ domain through a novel PBM. (A) Schematic representation of SGEF and Scribble constructs used in this figure. LRR, leucine-rich repeat. (B–G) Lysates from HEK293FT cells expressing the indicated constructs were immunoprecipitated (IP) using GFP antibodies (GFP-trap nanobodies). In all experiments, the precipitates were immunoblotted with anti-GFP antibodies to detect the immunoprecipitated protein and with anti-myc or anti-His to detect the potential interacting partner. (H) Sequence alignment comprising the Scribble binding domain of SGEF in vertebrates (aa 32–55 in human). (I) The indicated boxed regions within the Scribble binding domain in SGEF were mutagenized to Ala in full-length SGEF. Lysates from HEK293FT cells expressing the myc-tagged WT-SGEF or the indicated Ala mutants and GFP-Scribble were immunoprecipitated using GFP antibodies (GFP-trap nanobodies). The precipitates were immunoblotted with anti-GFP antibodies to detect the immunoprecipitated protein and with anti-myc to detect my-tagged SGEF mutants. Green boxes show the sequence that when mutated still maintain an interaction with Scribble whereas the red boxes denote sequence that when mutated lead to loss of interaction with Scribble. (J) Crystal structure of the Scribble PDZ1 domain in complex with an SGEF iPBM peptide. The crystal structure of the apo Scribble PDZ1 domain is shown in gray, while the complex with the SGEF_peptide (KSPNGLLITDFP) is shown in red (left panel). The right panel is a surface representation of the PDZ1/SGEF-PDZ_peptide complex.

Figure 2.

SGEF’s N terminus interacts with Dlg1 GUK domain. (A) Schematic representation of SGEF and Dlg1 constructs used in this study. (B–F) Lysates from HEK293FT cells expressing the indicated constructs were immunoprecipitated using GFP antibodies (GFP-trap nanobodies). In all experiments, the precipitates were immunoblotted with anti-GFP antibodies to detect the immunoprecipitated protein and with anti-myc to detect the interacting partner. (G) Sequence alignment comprising the Dlg1 binding domain of SGEF in vertebrates (aa 301–350 for human).

Figure 3.

SGEF forms a ternary complex with Scribble and Dlg1. (A) Lysates from HEK293FT cells expressing GFP-Dlg1 and HA-Scribble in the presence or absence of myc-SGEF were immunoprecipitated using GFP antibodies (GFP-trap nanobodies). The precipitates were immunoblotted with anti-myc, anti-GFP, and anti-HA antibodies as indicated. (B) Endogenous Dlg1 was immunoprecipitated from Caco2 cell lysates and immunoblotted for Dlg1, SGEF, and Scribble. (C) Endogenous Dlg1 was immunoprecipitated from CTRL and SGEF KD MDCK cell lysates and immunoblotted for Dlg1, SGEF, and Scribble. (D) Lysates from HEK293FT cells expressing the indicated constructs were immunoprecipitated using GFP antibodies (GFP-trap nanobodies). The precipitates were immunoblotted with anti-GFP antibodies to detect the immunoprecipitated protein and with anti-myc and anti-HA to detect the interacting partners. (E) Summary of results from C. (F) Cartoon representation of the ternary complex between Scribble, SGEF, and Dlg1. DBM, Dlg binding motif; iPBM, internal PDZ-binding motif.

Figure 4.

SGEF localizes at cell–cell junctions in epithelial cells. (A and B) IF of MDCK cells showing colocalization of mNeon-SGEF WT with Scribble and Dlg1, ZO-1 (TJ marker), and β-catenin (AJ marker). The right panels are single Z-planes along the length of the dotted yellow line. a, apical, b, basal. Scale bar, 10 µm, and XZ, 3 µm. (C) Gastrula-stage Xenopus embryos expressing mNeon-SGEF WT (green), TagBFP-ZO-1 (TJ marker), and PLEKHA7-mCherry (AJ marker) were live-imaged using confocal microscopy. En face views (left) are brightest point projections across multiple Z-planes. Side views (right) are average intensity projections along the length of the highlighted junction (see Materials and methods). Scale bars, XY, 10 µm, and XZ, 1 µm. (C′) Intensity profiles of SGEF (green solid line) relative to AJs (red dotted line) and TJs (blue dotted line) along the z axis in Xenopus gastrula-stage epithelial cells. Note that SGEF’s peak intensity is close to ZO-1’s, but it tapers away more slowly than ZO-1 along the lateral membrane. The graph shows normalized averaged intensities fitted with a smoothed curve; error bars indicate SD; n = 47 junctions, 18 embryos, five experiments. (D) Gastrula-stage Xenopus embryos expressing mNeon-RhoG (green), BFP-ZO-1 (TJ marker, blue), and PLEKHA7-mCherry (AJ marker, red) were live imaged using confocal microscopy. Brightest point projections of en face views and averaged side views of the highlighted junction (as in C) are shown. Scale bars, XY, 10 µm, and XZ, 1 µm. (D′) Intensity profiles of RhoG (green solid line) relative to AJs (red dotted line) and TJs (blue dotted line). Note that the RhoG signal is more basolateral compared with SGEF. The graph shows normalized averaged intensities fitted with a smoothed curve; error bars indicate SD. n = 19 junctions, 10 embryos, three experiments. (E) Gastrula-stage Xenopus embryos expressing mRFP-ZO-1 (TJ marker, magenta) and mNeon-SGEF WT (top, green), SGEF 1–227 (middle, green), or SGEF 228–871 (bottom, green) were live-imaged using confocal microscopy. En face views are brightest point projections across multiple Z-planes. Side views (right) are single Z-planes at the locations marked by yellow dotted lines. Note that WT SGEF and SGEF 1–227 appear junctional, whereas SGEF 228–871 appears diffusely localized. Scale bars, 20 µm, and XZ, 5 µm. (F) Quantification of the ratio of junctional to cytosolic intensities of mNeon-tagged SGEF WT, SGEF 1–227, and SGEF 228–871 in Xenopus embryos. n = SGEF WT: 234 junctions, 12 embryos, six experiments; SGEF 1–227: 214, 11, 5; SGEF 228–871: 180, 9, 4. Error bars represent SEM. ****, P < 0.00005 using the Mann–Whitney U test.

Figure 5.

SGEF regulates AJ properties of epithelial cells. (A) Cell lysates from confluent CTRL and SGEF KD MDCK cells were analyzed by WB using anti-SGEF antibodies. Tubulin was used as a loading control. (B) Cell lysates from confluent CTRL, SGEF KD, and Rescue WT MDCK cells were probed with E-cadherin, Pan-cadherin, cadherin-6, β-catenin, p120-catenin, Scribble, and Dlg1 antibodies. Tubulin was used as a loading control. (C, C′, and D–F) IF showing the distribution of endogenous E-cadherin, p120-catenin, Scribble, Dlg1, β-catenin, and mNeon-SGEF (green) in CTRL, SGEF KD, and Rescue WT MDCK cells. The bottom panel in each set of images shows a zoomed image of the selected regions (dotted yellow line). Note that panels C and C′ show images from same field. Confocal images are maximum projections of apical Z-planes. Scale bars, top panels: 30 µm; bottom panels: 10 µm. (G) Linescan (6-µm line drawn perpendicular to center of junctions) of IF images in panels C to F. At least two fields from two independent experiments were used for quantification (≥200 junctions). The intensity profiles were manually centered around the highest peak for each condition. (H) XZ view of MDCK cells from CTRL, SGEF KD, and Rescue WT cells stained for E-cadherin (red), β-catenin (magenta), nucleus (blue) and mNeon-SGEF WT (green in merge panel). Scale bar, 10 µm. (I) Quantification of height in CTRL, SGEF KD, and Rescue MDCK cells. n = 50 cells for each condition from three independent experiments. Error bars represent min to max with all points. ****, P < 0.00005; ns, nonsignificant using Student’s t test (two-tailed, unpaired).

Figure 6.

SGEF KD regulates TJ architecture and permeability. (A) Confocal images showing maximum projection of apical Z-planes in CTRL, SGEF KD, and Rescue WT MDCK cells stained for endogenous ZO-1 and mNeon-SGEF (green). The bottom panels show a zoomed image of the selected regions (dotted yellow line). Scale bars, top panel: 20 µm; bottom panels: 10 µm. (B–D) Quantification of zigzag index, apical cell area, and axial ratio in CTRL, SGEF KD, and Rescue WT cells. Two fields from two independent experiments were used for quantification. (n = at least 75 cells for zigzag index, n = 100 for area and n = 150 for axial ratio). Error bars represent min to max values. (E) TEER of CTRL, SGEF KD, and Rescue WT cells is plotted. Data represent the average of three experiments performed in duplicates. CTRL was normalized to 1, and data were plotted relative to CTRL. Error bars represent SEM. *, P < 0.05; **, P < 0.005; ***, P < 0.0005; ****, P < 0.00005; ns, nonsignificant using the Mann–Whitney U test (B–D) or Student’s t test (two-tailed, unpaired; E).

Figure 7.

SGEF KD stimulates actomyosin contractility. (A) Confocal images showing maximum projection of apical Z-sections in CTRL, SGEF KD, and Rescue WT MDCK cells stained for endogenous ZO-1 (green), myosin IIB (red), and mNeon-SGEF (magenta, in Rescue). mNeon signal is shown in magenta in Rescue panel to maintain color consistency. Scale bar, 5 µm. (B) Confocal images showing CTRL, SGEF KD, and Rescue WT MDCK cells stained for endogenous F-actin using phalloidin (green) and myosin IIB (red). Left panel: Maximum projection of apical Z-planes; right panel: maximum projection of basal Z-planes. Images were processed using the HyVolution deconvolution package (see Materials and methods). Scale bars, 0.5 µm. (C) Quantification of intensities of ZO-1 at junctions measured using a rectangle of 2 × 3 µm placed along BCJs. At least two fields from two independent experiments were used for quantification (≥100 junctions). Error bars represent SEM. ****, P < 0.00005; ns, nonsignificant using the Mann–Whitney U test. (D) Total cell lysates from confluent CTRL, SGEF KD, and Rescue WT MDCK cells were immunoblotted with ZO-1, myosin IIB, and afadin antibodies. Tubulin was used as a loading control. (E) Maximum projection of confocal images showing the localization of endogenous afadin in CTRL, SGEF KD, and Rescue WT cells. Scale bar, 5 µm. (F) Quantification of the ratio of TCJ over BCJ intensity of afadin was measured as described in Materials and methods. At least three fields from two independent experiments were used for quantification (≥200 junctions). Error bars represent SEM. ****, P < 0.00005; ns, nonsignificant using the Mann–Whitney U test.

Figure 8.

SGEF regulates apical constriction in epithelial cells. (A) Gastrula-stage Xenopus embryos expressing mRFP-ZO-1 (TJ marker) with 3xGFP-SGEF overexpressed (OE) at high levels (bottom). Yellow arrows point to apically constricted cells. Scale bar, 20 µm. (B) Time lapse of CTRL and 3xGFP-SGEF overexpression of a single cell over a period of 24 min. Note that the SGEF OE cell constricts apically whereas CTRL cell retains the same apical area. Scale bars, 10 µm. (C) Time projection of ZO-1 signal over a 203-s interval shows that junctions in SGEF OE cells are more dynamic than in CTRLs. Scale bar, 20 µm. (D) Graph showing the average apical surface area of SGEF OE cells is significantly smaller than CTRL cells, and some SGEF OE cells exhibit severe apical constriction. CTRL, n = 132 cells, three embryos, two experiments; SGEF OE, n = 147 cells, three embryos, two experiments. (E) CTRL and SGEF OE gastrula-stage Xenopus embryos were fixed and stained with Alexa Fluor 568–phalloidin to reveal F-actin. Images in the top row were taken with lower laser power optimized for viewing cell–cell junctions, and images in the bottom row were taken with higher laser power optimized for viewing medial-apical actin. Scale bar, 10 µm. (E′) F-actin intensity at BCJ was quantified from fixed phalloidin stained embryos. n = control: 288 junctions, 11 embryos, three experiments; SGEF OE: 304 junctions, 13 embryos, two experiments. (E′′) Medial-apical F-actin intensity was quantified from fixed phalloidin stained embryos. n = control: 50 cells, seven embryos, three experiments; SGEF OE: 50 cells, eight embryos, three experiments. (F) CTRL and SGEF OE embryos expressing an F-actin probe (Lifeact-mRFP, magenta in merge) and a myosin II intrabody (SF9-mNeon, green in merge) were live imaged by confocal microscopy. The control image shown is from a control region of a mosaic SGEF OE embryo. Scale bar, 10 µm. (G) CTRL and SGEF OE embryos coexpressing mNeon-Vinculin, mCherry-α-catenin, and BFP-membrane. Scale bar, 10 µm. (H) Graph comparing junctional intensities of vinculin (normalized to membrane probe intensity) in CTRL and SGEF OE embryos. n = control: 63 junctions, eight embryos, three experiments; SGEF OE: control: 63 junctions, eight embryos, three experiments. Confocal images in A, B, and E–G are brightest point of apical sections. All graphs show mean ± SEM. ****, P < 0.00005 using the Mann–Whitney U test.

Figure 9.

The guanine nucleotide exchange activity of SGEF is required for junctional maintenance, whereas scaffolding activity of SGEF is required for apical contractility. (A) Active RhoG was precipitated from total lysates of CTRL, SGEF KD, Rescue mNeon-SGEF WT, and Rescue CD mNeon-SGEF using GST–ELMO and immunoblotted with anti-RhoG antibodies. (B) For quantification, active RhoG levels were normalized to total RhoG levels. Data are mean ± SEM of three independent experiments. **, P < 0.005; ns, nonsignificant using Student’s t test (two-tailed, unpaired). (C) Lysates from CTRL, SGEF KD, and SGEF KD cells rescued with mNeon-SGEF 1–227, mNeon-SGEF 1–400, or mNeon-SGEF CD were probed for E-cadherin, β-catenin, ZO-1, and myosin IIB antibodies. Tubulin was used as a loading control. (D and E) Confluent MDCK CTRL, SGEF KD, and SGEF KD cells rescued with mNeon-SGEF 1–227, mNeon-SGEF 1–400, or mNeon-SGEF CD were stained for endogenous E-cadherin, β-catenin, and mNeon-SGEF (green). Confocal images are maximum projections of apical Z-planes. Scale bar, 5 µm. (F) XZ view of MDCK cells from CTRL, SGEF KD, and SGEF KD cells rescued with mNeon-SGEF 1–227, mNeon-SGEF 1–400, or mNeon-SGEF CD stained for F-actin (magenta), nucleus (Hoechst), and mNeon-SGEF (green). Scale bar, 10 µm. (G) Confluent MDCK CTRL, SGEF KD, and SGEF KD cells rescued with m-Neon-SGEF 1-227, mNeon-SGEF 1-400, or mNeon-SGEF CD were steined for endogenous ZO-1, myosin IIB, and mNeoen-SGEF (green). Confocal images are maximum projections of apical Z-planes. Scale bar, 5 μm. (H) Linescan (6-µm line drawn perpendicular to center of junctions) of IF images in D. At least two fields from two independent experiments (≥150 junctions) were used for quantification. The intensity profiles from were manually centered around the highest peak for each condition. (I) Quantification of height in CTRL, SGEF KD, and SGEF KD cells rescued with mNeon-SGEF 1–227, mNeon-SGEF 1–400, or mNeon-SGEF CD cells. n = 50 cells for each condition. Error bars represent min to max values with all points. Error bars represent SEM. ****, P < 0.00005, using the Mann–Whitney U test. (J) Quantification of zigzag index of CTRL, SGEF KD, and SGEF KD cells rescued with mNeon-SGEF 1–227, mNeon-SGEF 1–400, or mNeon-SGEF CD. At least two fields from two independent experiments (≥200 junctions) were used for quantification. ****, P < 0.00005; ns, nonsignificant using Student’s t test (two-tailed, unpaired).

Figure 10.

SGEF does not affect polarity but regulates lumen formation in 3D MDCK cysts. (A) IF of MDCK CTRL and SGEF KD cells using gp135 (green), actin (magenta), and nucleus (Hoechst). Scale bar, 10 µm. (B and C) MDCK CTRL, SGEF KD, and Rescue WT cells were plated on matrigel to form 3D cysts. Cysts were fixed and stained for β-catenin (red), gp135 (green), and nuclei (blue) in B and E-cadherin (green), phalloidin (red), and nuclei (blue) in C. For detailed protocol of growing and staining cyst, see Materials and methods. Scale bars, 5 µm. (D) Cysts from CTRL, SGEF KD, Rescue WT, and Rescue CD were classified based on the number of cysts (single or multiple) and the phenotype of the lumen (open or closed). Three independent experiments were counted for each condition (≥200 cysts/condition). Images in B and C are single Z-sections corresponding to the center of the cyst.