Actin-dependent recruitment of AGO2 to the zonula adherens

Abstract

Adherens junctions are cadherin-based structures critical for cellular architecture. E-cadherin junctions in mature epithelial cell monolayers tether to an apical actomyosin ring to form the zonula adherens (ZA). We have previously shown that the adherens junction protein PLEKHA7 associates with and regulates the function of the core RNA interference (RNAi) component AGO2 specifically at the ZA. However, the mechanism mediating AGO2 recruitment to the ZA remained unexplored. Here, we reveal that this ZA-specific recruitment of AGO2 depends on both the structural and tensile integrity of the actomyosin cytoskeleton. We found that depletion of not only PLEKHA7, but also either of the three PLEKHA7-interacting, LIM-domain family proteins, namely LMO7, LIMCH1, and PDLIM1, results in disruption of actomyosin organization and tension, as well as disruption of AGO2 junctional localization and of its miRNA-binding ability. We also show that AGO2 binds Myosin IIB and that PLEKHA7, LMO7, LIMCH1, and PDLIM1 all disrupt interaction of AGO2 with Myosin IIB at the ZA. These results demonstrate that recruitment of AGO2 to the ZA is sensitive to actomyosin perturbations, introducing the concept of mechanosensitive RNAi machinery, with potential implications in tissue remodeling and in disease.

FIGURE 1:

PLEKHA7 interacts with AGO2 and actin-related, LIM domain-containing proteins at apical AJs. (A and B) GO enrichment analyses of the PLEKHA7-interacting protein dataset (www.imexconsortium.org/; identifier: IM-25739). The GO-slim Molecular Function and GO-slim Biological Process datasets were interrogated using Fisher’s exact test with FDR correction to assess enrichment significance. The 10 most significantly enriched GO terms are shown; GO term accession numbers are displayed in parentheses. (C) LIM domain-containing proteins identified in the same dataset of PLEKHA7 interactors; the unique peptide counts recorded from mass spectrometry analysis following PLEKHA7 immunoprecipitation are shown. (D and E) Immunoprecipitation (IP) of PLEKHA7, LMO7, LIMCH1, PDLIM1, and AGO2 from Caco2 cells, immunoblotted (IB) for the same markers; IgG is the negative control. Molecular masses (kD) are indicated on the right. (F–G) Immunofluorescence of Ecad, PLEKHA7, LMO7, LIMCH1, PDLIM1, and AGO2 in confluent Caco2 cell monolayers. Images were obtained by confocal microscopy, and Z-series stacks were acquired through the entire plane; representative apical and basal Z-slices are shown. Fluorescence intensity of 6-µm line scans drawn perpendicular to cell–cell junctions was measured from n = 30 cell–cell junctions (10 junctions/field) representative of three independent experiments; statistical analyses were performed using two-way ANOVA tests; error bars represent mean ± SD; ****P < 0.0001. Scale bar = 20 µm.

FIGURE 2:

LMO7, LIMCH1, PDLIM1, and PLEKHA7 loss each disrupt junctional localization of AGO2. (A) Immunoblotting of PLEKHA7 KO Caco2 cells, compared with control wild-type cells (WT). (B) Immunofluorescence of control (WT) or PLEKHA7 KO cells for AGO2 and Ecad. AGO2 junctional fluorescence intensity normalized to cytoplasmic was quantified from n = 30 cell–cell junctions (10 junctions/field) representative of three independent experiments; statistical analyses were performed using unpaired two-way t test; ****P < 0.0001. (C) Immunoblotting of LMO7, LIMCH1, and PDLIM1 shRNA-mediated knockdown (shLMO7, shLIMCH1, shPDLIM1, respectively) Caco2 cells, compared with non-target (NT) shRNA control cells; asterisk indicates the specific LMO7 band lost by shRNA targeting. (D) Immunofluorescence of NT, shLMO7, shLIMCH1, and shPDLIM1 Caco2 cells for AGO2, PLEKHA7, and Ecad. AGO2 and PLEKHA7 junctional fluorescence intensity normalized to cytoplasmic was quantified from n = 30 cell–cell junctions (10 junctions/field) representative of three independent experiments; statistical analyses were performed using one-way ANOVA test; ****P < 0.0001; *P < 0.05; ns, non-significant. Insets in all cases are marked by white rectangles and are 3× magnification of the original image. In all immunoblots, β-actin is the loading control; molecular masses (kD) are indicated on the right. Top-view immunofluorescence images were obtained by confocal microscopy and are single apical Z-slices. Scale bars = 20 µm. (E) AGO2 RNA immunoprecipitation (RNA IP) followed by qRT-PCR analysis for miR-24, miR-200c, and miR-203a miRNAs in NT, shLMO7, shLIMCH1, and shPDLIM1 Caco2 cells. Error bars represent mean ± SD from n = 3 independent experiments; statistical analysis was performed using one-way ANOVA ***P < 0.005; **P < 0.01.

FIGURE 3:

PLEKHA7, LMO7, LIMCH1, or PDLIM1 loss does not affect each other’s localization to the junctions. (A) Immunofluorescence of control (WT) or PLEKHA7 KO cells for LMO7, LIMCH1, and PDLIM1. (B–D) Immuno­fluorescence of each LMO7, LIMCH1, and PDLIM1 in control (NT) and shLMO7, shLIMCH1, and shPDLIM1 Caco2 cells. For all figures, top-view immunofluorescence images were obtained by confocal microscopy and are single apical Z-slices. Junctional fluorescence intensity normalized to cytoplasmic was quantified from n = 30 cell–cell junctions (10 junctions/field) representative of three independent experiments; statistical analyses were performed using unpaired two-way t test; *P < 0.05; ns, nonsignificant. Scale bars = 20 µm.

FIGURE 4:

PLEKHA7, LMO7, LIMCH1, and PDLIM1 each distinctly influence apical actomyosin organization. (A) Immunofluorescence and super resolution microscopy of PLEKHA7 KO Caco2 cells compared with control wild-type (WT) Caco2 cells. (B) Immunofluorescence and super resolution microscopy of LMO7, LIMCH1, and PDLIM1 knockdown (shLMO7, shLIMCH1, and shPDLIM1) Caco2 cells, compared with control (NT) shRNA cells. Top-view immunofluorescence images are maximum intensity projections of the two most apical Z-slices. Insets marked by white rectangles are 6× magnification of the original image. Insets for shPDLIM1 cells noted with asterisks are magnified 5× from areas marked with yellow dashed boxes. Asterisk marked insets are from full maximum intensity projection images to indicate basolateral actin filaments linked to the apical actin ring (marked by white arrows). In all cases, Ecad is used as a co-stain denoting AJs. Fluorescence intensity of 3-µm line scans drawn perpendicular to cell–cell junctions was measured from n = 30 cell–cell junctions (10 junctions/field) representative of three independent experiments; statistical analyses were performed using two-way ANOVA tests; error bars represent mean ± SD; ****P < 0.0001. Scale bars = 20 µm.

FIGURE 5:

AGO2 recruitment to the ZA depends on filamentous actin integrity. (A–G) Immunofluorescence of Caco2 cells for F-actin, PLEKHA7, AGO2, E-cadherin, LMO7, LIMCH1, and PDLIM1, fixed at indicated time points throughout a calcium switch assay. Images at 0 min indicate Ca²⁺ depleted cells immediately before Ca²⁺ reintroduction and at 30- and 60-min time points are following Ca²⁺ re-addition. Cells were pretreated and maintained in either DMSO vehicle (control) or 10 µM LatA throughout the calcium switch assay. Insets are marked by white rectangles and are 3× magnification of the original image. Scale bars = 20 µm. (H) Fluorescence intensity of 6-µm line scans drawn perpendicular to cell–cell junctions was measured for each marker at the 60-min time point upon recovery from n = 30 cell–cell junctions (10 junctions/field) representative of three independent experiments; statistical analyses were performed using two-way ANOVA tests; error bars represent mean ± SD; ****P < 0.0001; **P < 0.01; *P < 0.05; ns, non-significant.

FIGURE 6:

Contractile tension at the ZA is disrupted in LMO7, LIMCH1, PDLIM1 knockdown and PLEKHA7 KO cells. Immunofluorescence of (A–E) total, as well as of tensile α-catenin using the α18 antibody that recognizes α-catenin upon tension, as well as of (F–I) phosphorylated myosin light chain at S19 (pS19-MRLC) that also indicates tension, in wild-type (WT) or PLEKHA7 KO Caco2 cells and in control (NT) or LMO7, LIMCH1, and PDLIM1 knockdown (shLMO7, shLIMCH1, and shPDLIM1) cells. Ecad is used as a co-stain denoting AJs. α18 fluorescence intensity measurements were normalized to total α-catenin (B and D) and pS19-MRLC were normalized to Ecad (G and I). All quantifications are from n = 30 cell–cell junctions (10 junctions/field) representative of three independent experiments. Statistical analyses were performed using unpaired two-way t tests (B and G) or one-way ANOVA tests (D, E, and I). ****P < 0.0001; **P < 0.01; *P < 0.05; ns, non-significant. Scale bars = 20 µm.

FIGURE 7:

Contractile actin tension at the ZA is required for AGO2 junctional recruitment. (A–F) Immunofluorescence of Caco2 cells during a calcium switch assay in which DMSO or Blebbistatin (Blebb) were included in the calcium-containing recovery medium. Cells fixed at 30 and 60 min post Ca²⁺ reintroduction were stained for AGO2 (A), PLEKHA7 (B), E-cadherin (C), LMO7 (D), LIMCH1 (E), and PDLIM1 (F). Insets are marked by white rectangles and are 3× magnification of the original image. Fluorescence intensity of 6-µm line scans drawn perpendicular to cell–cell junctions was measured for each marker at the 60-min time point upon recovery from n = 30 cell–cell junctions (10 junctions/field) representative of three independent experiments; statistical analyses were performed using two-way ANOVA tests. Error bars represent mean ± SD. ****P < 0.0001; *P < 0.05; ns, non-significant. Scale bars = 20 µm.

FIGURE 8:

PLEKHA7, LMO7, LIMCH1, or PDLIM1 depletion each disrupt AGO2-Myosin IIB interaction at the ZA. (A) Immunoprecipitation (IP) of AGO2 and Myosin IIB from Caco2 cells, immunoblotted (IB) for the same markers; IgG is the negative control. Molecular masses (kD) are indicated on the right. (B–E) Wild type (WT) or PLEKHA7 KO Caco2 cells and control (NT) or LMO7, LIMCH1, and PDLIM1 knockdown (shLMO7, shLIMCH1, and shPDLIM1) cells stably expressing an AGO2-Flag construct (see Supplemental Figure S1, A–D) were subjected to PLA for AGO2-Flag, using an anti-Flag antibody, and Myosin IIB, followed by immunofluorescence staining for Ecad and confocal microscopy; DAPI was used to stain nuclei. Caco2 cells stably transduced with an empty vector were used as negative PLA control. Insets are marked by white rectangles and are 3× magnification of the original image. The ratio of junctional vs cytoplasmic PLA signals was quantified for each condition and from n = 30 cells (10 cells/field) representative of three independent experiments; statistical analyses were performed using either unpaired two-way t test (B–C) or one-way ANOVA test (D–E). Error bars represent mean ± SD. **P < 0.01; *P < 0.05. Scale bars = 20 µm.