BPAG1e and β4 integrin colocalize at the leading edge of migrating epithelial sheets. A scratch wound was introduced into a confluent monolayer of HEKs. The monolayer was allowed to reclose for 6 h and then was fixed and prepared for indirect immunofluorescence microscopy with antibodies against BPAG1e and β4 integrin as indicated. Top panels, cells migrating into the scratched area; bottom panels, cells distant from the wound site. W and arrow in the phase image in the top panel mark the wound site and direction of migration, respectively. Bar, 20 μm.

Generation of keratinocytes deficient in BPAG1e expression. (A) HEKs were infected with a lentivirus encoding shRNA targeting BPAG1e. Total cell lysates from a number of different stable clones were prepared for SDS-PAGE/immunoblotting and then probed with BPAG1e and mTOR antibodies as shown. The percentage expression of BPAG1e in clones 2, 5, 12, 17, and 18 relative to HEKs (Control) was calculated after densitometric scanning of the blots and is indicated underneath each lane. (B) Cell lysates from HEKs and HEK clone 2 cells (HEKCn2s) were probed in immunoblots with antibodies against BPAG1e, BP180, plectin, and β4 integrin as marked. In the bottom immunoblot in B, cell surface proteins that had been biotinylated and subsequently immunoprecipitated by streptavidin-coated beads were probed with antibodies against β4 integrin. (C) Extracellular matrix preparations of the HEKs and HEKCn2s were probed in immunoblots with antibodies specific for the α3, β3, and γ2 laminin subunits. (D) Cell-surface expression of β4 integrin and α3 integrin was compared in HEKs and HEKCn2s by FACS as indicated (black tracing; secondary antibody alone). (E) HEKs and HEKCn2s were plated onto glass coverslips. Forty-eight hours later the cells were fixed and prepared for double-label immunofluorescence microscopy with a mix of antibodies against BPAG1e and β4 integrin, as indicated. In the right-hand column, the separate green and red fluorescence images are overlaid. Bars, 20 μm.

Plectin and β4 integrin colocalization in keratinocytes is not dependent on BPAG1e. HEKs and HEKCn2s were plated onto glass coverslips and 48 (A) or 24 h (B) later were fixed and then imaged by confocal immunofluorescence microscopy after staining with antibodies against plectin and β4 integrin, as indicated. Red and green images are shown overlaid in the right-hand column. Bars, 20 μm.

Cell surface area and cytoskeleton localization are perturbed in keratinocytes deficient in BPAG1e. (A) HEKs and HEKCn2s were plated onto glass coverslips and 24 h later were fixed and stained with for double-labeling with rhodamine-conjugated phalloidin and antibodies against paxillin. Arrows in A, D, and E indicate lamellipodia. More that 50 live cells in each of three trials were scored for the number of lamellipodia that each displays. Lamellipodia scores are graphed in B (graph represents mean ± SEM). The percentage of HEKs exhibiting a single lamellipodium is higher than in HEKCn2s (* p < 0.019). (C) HEKs and HEKCn2s were plated onto uncoated glass coverslips, and 24 h later cell surface area was measured using Metamorph software. More than 50 cells were evaluated per trial in three trials. HEKCn2s are slightly smaller than HEKs (* p <0.01). (D and E) HEKs and HEKCn2s were plated onto glass coverslips and 24 h later were fixed and stained with antibodies against γ-tubulin (D) or cytokeratin 5 (E). The cells were then imaged by confocal microscopy. Bars, 20 μm.

Rac and cofilin activities are lowered in keratinocytes deficient in BPAG1e. (A) Levels of active Rac in subconfluent HEKs, HEKCn12s and HEKCn2s were determined by G-LISA through binding of GTP bound Rac1 to an affinity plate followed by antibody detection. Activity levels of Rac1 in HEKCn2s and HEKCn12s are plotted relative to that in HEKs and is significantly lower (significance relative to HEKs, * p < 0.05). Data are from three independent trials. (B) Total cell lysates were immunoblotted for total Rac1 and actin. (C) β4 integrin was immunoprecipitated from cell lysates of HEKs (left) and HEKCn2s (right), and precipitates were then immunoblotted for Rac1 (top panel) and β4 integrin (bottom panel). Blots were densitometrically scanned, the level of Rac immunoprecipitated was normalized to β4 integrin in three independent trials, and the results were plotted in the graph below relative to that in HEKs (* p < 0.012). (D) Cell lysates from HEKs, HEKCn2s, and HEKCn12s were prepared for immunoblotting using antibodies specific for Ser-3–phosphorylated cofilin or F-actin. Blots were densitometrically scanned, the levels were normalized to actin, and the fold increase of phosphorylated cofilin, relative to the level in HEKs, was quantified. The graph beneath the blot displays the mean ± SEM in three independent trials (significance relative to HEKs; * p < 0.05).

Constitutively active Rac1 and cofilin rescue motility defects of keratinocytes deficient in BPAG1e. HEKs (A, Supplemental Video 1) HEKCn2s (B, Supplemental Video 2), HEKCn12s (C), and HEKCn2s infected with adenovirus encoding constitutively active forms of Rac1 (D), RhoA (E), Cdc42 (F), or cofilin (G) were plated onto uncoated glass-bottomed dishes, and 16 h later motility patterns of individual cells were tracked over a 2 h period. Virally infected cells were identified by visualizing GFP. Vector diagrams show the migration pattern of 10 cells, with each line representing the migration of a single cell. (H) Cells from a minimum of four trials per treatment, with a minimum of 10 cells tracked per trial, were scored as trails (■) if they migrated in a predominantly linear and/or backtracking manner or circles (□) if they moved over a approximately circular track and were plotted as percentage of total. The percentage of HEKCn2s moving in a trail under each condition was compared with the percentage of HEKs moving in a similar manner (* p < 0.01). (I) Migration rates for cells visualized in A–G were calculated by summing the displacement of cell bodies between each time point in order to calculate total displacement over time. Differences between cell lines/infections were below significance.

Dynamics of β4 integrin in keratinocytes deficient in BPAG1e. HEKs or HEKCn2s were infected with retrovirus encoding GFP-tagged β4 integrin. FRAP of the arrowed boxes (images of these bleached areas are shown at higher magnification in B) was measured either at the leading edge of cells migrating into a scratch wound (A and B, two top panels; C, black/red lines; arrows indicate direction of migration) or in stationary cells close to or at confluence (A and B, two bottom panels; C, gray/orange lines). Fluorescence recovery is plotted as a percentage of the prebleach levels, allowing for background fading. The graph in C represents mean ± SEM; n = 8.