Figure 1.
Morphological phenotypes of KAI1/CD82-overexpressing cells.
(A) KAI1/CD82-expressing cells displayed altered morphology. Du145-Mock and -KAI1/CD82 transfectant cells were plated on FN-coated (10 µg/ml) glass coverslips at 37°C, 5% CO2 for 4 h or overnight in serum-free DMEM medium. (B) Diminished lamellipodia in KAI1/CD82-overexpressing cells. Du145-Mock and -KAI1/CD82 cells were spread on FN-coated (10 µg/ml) plates in DMEM containing 1% FCS at 37°C overnight. The DIC images of cells were captured on an Axiovert inverted microscope using DIC optics with a 40 x F Fluor oil immersion objective. The arrows indicate lamellipodia. (C) KAI1/CD82-overexpressing cells frequently exhibit elongated cellular extensions. Du145-Mock and -KAI1/CD82 cells were spread on FN-coated (10 µg/ml) plates in DMEM containing 10% FCS and HGF (100 ng/ml) at 37°C for 6 h. The DIC images of cells were captured using DIC optics with a 40 x F Fluor oil immersion objective. Scale bar, 50 µm.
Figure 2.
KAI1/CD82 attenuates the formation of lamellipodia and retraction of cellular tails.
(A) KAI1/CD82 inhibits lamellipodia and protrusion formations. Du145-Mock and -KAI1/CD82 transfectant cells were plated on FN-coated glass coverslips from 3 to 6 h. DIC images were acquired using time-lapse vidoemicroscopy. Arrows indicate lamellipodia. (B). KAI1/CD82 inhibits the retraction of the rear tail. Du145-Mock and -KAI1/CD82 cells were placed on FN-coated (10 µg/ml) coverslips from 3 to 6 h and treated with 100 ng/ml HGF from 4 to 6 h. Cell morphology was photographed for 3 h using time-lapse videomicroscopy. Arrows indicate the retraction processes in Du145-Mock cells and arrowheads indicate the rear tail in Du145-KAI1/CD82 cells. The time-lapse intervals are labeled inside images. Scale bar, 20 µm.
Figure 3.
The actin cortical meshwork and stress fiber were disrupted upon KAI1/CD82 expression.
(A) F-actin distribution. After being spread on FN (50 µg/ml)-, LN1 (50 µg/ml)-, or LN5 (2 µg/ml)-coated coverslips at 37°C, 5% CO2 for 6 h, Du145 transfectant cells were fixed, permeabilized, and then stained with TRITC-conjugated α-phalloidin. The fluorescent images were captured under an Axiophot fluorescent microscope equipped with an Optronics digital camera at magnification 63X. (B) Live imaging of actin polymerization in Du145-Mock and -KAI1/CD82 cells. The cells that were transiently transfected with pEGFP-actin construct were spread on FN-coated coverslips in complete DMEM and photographed using time-lapse confocal videomicroscopy. Arrowheads and arrows indicate actin polymerization during the development of peripheral meshwork and stress fiber, respectively. (C) Less F-actin in Du145-KAI1/CD82 cells. Left panel, Du145-Mock and -KAI1/CD82 cells spread in serum-free DMEM were detached, fixed, permeabilized, incubated with Alexa 488-conjugated phalloidin or mouse IgG2b, and then analyzed with flow cytometry. Right top panel, quantitation of the mean fluorescence intensity (MFI) of F-actin. Data are expressed as the mean MFI of three independent experiments (P<0.05 between Mock and KAI1/CD82 cells). Right bottom panel, the cell lysates from the experiment were analyzed by Western blot for total cellular actin proteins. Cofilin blot was used as the protein loading control. (D) The presence of KAI1/CD82 proteins is inversely correlated with actin cytoskeletal assembly. HT29 colon cancer cells and prostate epithelial cells were fixed, permeabilized, and incubated with Alexa 488-conjugated phalloidin and Alexa 594-conjugated KAI1/CD82 mAb. The images were acquired with confocal microscopy. For HT29 cells, the photographs were taken from the basal sections of the cells. Shown in insets is the magnified staining from the frame-selected areas. White dashed lines outline the cell boundary. Scale bars: 10 µM. (E) KAI1/CD82 silencing leads to stress fiber formation. HT29 cells were transiently transfected with KAI1/CD82-specific siRNA duplexes, seeded sparsely, and stained with Alexa 488-conjugated phalloidin and Alexa 594-conjugated KAI1/CD82 mAb as described above. Scale bar: 10 µM. White dotted lines outline the cell periphery. Arrows indicate the cells exhibit robust stress fiber formation but no KAI1/CD82 staining.
Figure 4.
Regulation of KAI1/CD82-induced actin cytoskeletal changes.
(A) Effects of integrin, growth factor, and chemokine on KAI1/CD82-induced actin cytoskeletal changes. Du145-Mock or -KAI1/CD82 transfectant cells were spread on FN-coated plates in complete DMEM and treated with integrin β1 activating mAb A1A5 (8 µg/ml), EGF (100 ng/ml), HGF (100 ng/ml), or SDF-1 (100 ng/ml) at 37°C from 4 to 6 h. Cells were then fixed, permeabilized, and incubated with Texas Red-conjugated α-phalloidin. Digital images were captured under a fluorescent microscope. Scale bar, 20 µm. Quantitative analysis (bottom panel): the cortical F-actin intensity was quantitated as described in Experimental Procedures. Bars denote the average intensity of three individual experiments. In each group, 50∼95 cells were quantified. The differences between Mock and all CD82 groups are statistically significant (P<0.05), while the differences between untreated CD82 and each treated CD82 group are not statistically significant (P>0.05). (B) Effect of the confluent cell culture on KAI1/CD82-induced actin cytoskeletal changes. Actin polymerization was analyzed in the Du145-Mock and -KAI1/CD82 cells cultured in complete DMEM at confluent stage. Scale bars, 100 µm in Du145-Mock cells and 50 µm in Du145-CD82 cells. Arrows indicate the well-developed actin cortical meshwork seen in the KAI1/CD82-expressing cells with cell-cell contacts. Arrowheads indicate that no well-developed cortical network was found in KAI1/CD82-expressing cells without cell-cell contacts. Quantitative analysis (bottom panel): the cortical F-actin intensity was quantitated as described in Experimental Procedures. Bars denote the average intensity of 32–40 cells from three individual experiments. The differences between the Mock and KAI1/CD82 cells without cell-cell contacts and between the KAI1/CD82 cells with and without cell-cell contacts are statistically significant (P<0.05), while the differences between the groups of Mock and KAI1/CD82 cells containing cell-cell contacts are not statistically significant (P>0.05). (C) The effect of cell-cell contacts on actin polymerization in Du145-KAI1/CD82 cells. Du145-KAI1/CD82 cells expressing EGFP-actin were photographed using time-lapse confocal and DIC videomicroscopy. For each group, a DIC image is included to display whether the green fluorescent cell is in cell-cell contact. Arrows and arrowheads indicate actin polymerization during the development of peripheral actin fiber. Scale bar: 10 µm.
Figure 5.
KAI1/CD82 regulates the activities of Rho GTPases.
(A) KAI1/CD82 inhibits Rac1 activity. Du145-Mock or -KAI1/CD82 transfectant cells were lysed in a lysis buffer containing 1% NP-40 and 0.2% SDS detergents. Cell lysates were subjected to affinity precipitation with GST-PAK1, which binds to only the activated or GTP-bound Rac. The co-precipitated, GTP-bound Rac GTPase was detected by Rac mAb. The intact cell lysates were blotted with Rac mAb to demonstrate equivalent levels of total Rac proteins between Mock and KAI1/CD82 transfectant cells. Blots show the result from a representative experiment; the graph represents the relative density of the Rac band (mean±SD, n = 4), based on densitometric analysis. *: P<0.05. (B) KAI1/CD82 does not significantly alter RhoA activity. Du145 transfectants were pretreated as described above. GTP-bound RhoA was pulled down by GST-Rhotekin and detected by RhoA mAb. The blot shows the result from a representative experiment; the graph represents the relative density of the RhoA bands (mean±SD, n = 9), based on densitometric analysis. P>0.05 between Mock and KAI1/CD82. (C) KAI1/CD82 does not significantly alter Cdc42 activity. GTP-bound Cdc42 was pulled down by GST-PAK1 and detected by Cdc42 mAb. The blot shows the results from a representative experiment; the graph represents the relative density of the Cdc42 bands (mean±SD, n = 4), based on densitometric analysis. P>0.05 between Mock and KAI1/CD82. In all experiments, tubulin protein levels in cell lysates were detected via Western blot and served as protein loading controls.
Figure 6.
KAI1/CD82 blocks the enrichment of cofilin at the cell periphery.
(A) The levels of total and phosphorylated cofilin proteins in Du145-Mock and -KAI1/CD82 cells were assessed by Western blot using pAbs against cofilin and phosphorylated cofilin, respectively, as described in Materials and Methods. Tubulin blot is used as a control for protein loading. (B) KAI1/CD82 prevents cofilin from being targeted to the cell periphery. Du145 transfectant cells were spread on FN-coated coverslips in complete DMEM from 3 to 6 h. The cells were fixed, permeabilized, and incubated with cofilin pAb and TRITC-conjugated α-phalloidin, followed by the FITC-conjugated second Ab staining. Digital images were captured under a confocal microscope, and each image represents a single XY section. The arrow indicates the translocation of cofilin into lamellipodia, while the arrowhead indicates the relatively transparent zone beneath the actin cortical meshwork and within the cytoplasm. Scale bar, 20 µm. (C) Comparison of the subcellular distribution of phosphorylated cofilin between Du145-Mock and -KAI1/CD82 cells. The experiment was performed as described in (B) except that the pAb against phosphorylated cofilin was used as the primary Ab. Scale bar, 20 µm.
Figure 7.
KAI1/CD82 inhibits ROCK activity and the enrichment of RhoA at the cell periphery.
(A). Du145-Mock and -KAI1/CD82 transfectant cells were cultured in either DMEM containing 0.5% FBS for 40 h or DMEM containing 10% FBS and HGF (100 ng/ml) for overnight. The cells were then lysed with 1% Triton X-100 lysis buffer. Equal amounts of cell lysate were assayed for ROCK activity with MBL Rho Kinase Assay kit. In each experiment, ROCK activities were measured in duplicate for each transfectant. The results shown in histogram are the average values of three individual experiments±SD. *: P<0.05, **: P<0.01. (B). The cells were pretreated with HGF (100 ng/ml) overnight and then assayed for RhoA activity as aforementioned. Histogram represents the relative density of the RhoA-GTP bands (mean±SD, n = 3). *: P<0.05. (C). The cells spread on LN1 (20 µg/ml)-coated plates were treated with or without HGF (100 ng/ml) overnight and then fixed, permeabilized, and incubated sequentially with RhoA mAb and Alexa594-conjugated second Ab. Images were acquired as described above under a fluorescent microscope. Arrows indicate the enrichment of RhoA at the cell periphery and in the retraction tail. The bottom panel shows the fluorescent profiles of line scan from Mock and KAI1/CD82 cells.
Figure 8.
The effect of KAI1/CD82 on the PIP2 in the plasma membrane.
(A). Du145-Mock and -KAI1/CD82 transfectant cells were spread on either FN- or LN1-coated plate in complete DMEM at 37°C overnight, fixed, permebilized, and then incubated with PIP2 mAb, followed with the incubation of Alexa 594-conjugated second Ab and FITC-conjugated phalloidin. Confocal images of X-Y sections were collected, scale bar: 20 µm. (B). The pEGFP-PLCδ PH domain construct was transiently transfected into Du145-Mock and -KAI1/CD82 transfectants. At 48 h after transfection, the cells were spread on an FN (10 µg/ml)-coated plate, fixed, permeabilized, stained with Alexa 594-conjugated phalloidin, and analyzed with confocal microscopy. The X-Y section images were captured as described above. Arrowhead: the localization of EGFP-PLCδ PH domain at the cell periphery. Scale bar: 20 µm. The bottom panel shows F-actin (red) and PLCδ PH domain (green) fluorescent profiles of line scan (white lines in the Merge images) from Mock and KAI1/CD82 cells. Green arrow: the enrichment of PLCδ PH domain at the cell periphery.