Figure 1.
HSlo expression on HEK293 cell surface is concentrated at adhesion junctions and is co-localized with β-catenin.
(A) Projected 3-D reconstruction from a Z-series of fluorescence images of HSlo-HEK cells surface-labeled with Ibtx-biotin and streptavidin-Alexa488 conjugates. HSlo has a punctate distribution on the cell surface, and is clearly concentrated at cell adhesion junctions (shown by the arrows). (B) Orthogonal view of confocal images of the surface-labeled HSlo-HEK cells. Note the relative abundance of HSlo near cell-cell contacts. (C) Cells were first surface labeled for HSlo (red) as above, and then permeabilized for labeling with mouse anti- β-catenin antibody (green). Co-localized spots are shown in orange (circle as an example) in the overlay image from 3D reconstructed projection. There is strong co-localization of HSlo and β-catenin at the cell surface. Co-localization analysis using the JACoP plugin in Image J confirm a strong co-localization between the β-catenin and HSlo; Pearson's correlation was 0.748, Li's intensity correlation quotient was 0.38 and the Manders M1 (fraction of catenin overlapping with HSlo) and M2 (fraction of HSlo overlapping with catenin) coefficients were 0.991 and 0.947 respectively. Scale bars are 10 µm.
Figure 2.
Mutant channel with deletion of the S10 region has decreased surface expression in a stable HEK293 cell line.
(A) Sequence of the Slo channel near the S10 region, which follows closely after the Ca2+ bowl region. (B, E) Immuno-fluorescence images showing HSlo wt expression detected by anti-FLAG antibody. Permeabilized cells show total expression (B), while non-permeabilized cells (E) show only surface expression. (C, F) mutant with the S10 region deleted (HSloΔS10) has reduced surface expression of HSlo. Here the exposure conditions for this pair of micrographs were chosen to give (part C) approximately the same brightness as part B, to normalize for total HSlo expression. Arrows show similar intracellular distribution of channel proteins in both wt and deletion mutant. Each panel also includes the corresponding DIC image for comparison. (D, G) negative controls using wt-HSlo without primary antibodies. Scale bars = 10 µm. (H, I) FACS quantification of the surface to total ratio of channel expression in wt HSlo and HSloΔS10 mutant (here abbreviated as DS10). Surface to total ratios were obtained by integration of the histograms. There is a 40% decrease in surface protein expression for the HSloΔS10 mutant determined by the surface to total HSlo ratio (+/− SEM, p = 0.032 in two sample t-test .n = 5).
Figure 3.
Transfection of siRNA against β-catenin into HSlo-HEK cells decreases surface expression of HSlo.
(A–D) HSlo surface expression was detected by anti-FLAG in HSlo- HEK cells transfected with different siRNAs. Compared to cells transfected with either no DNA (A), or with siRNA against the chick β4 subunit of Slo that has no comparable sequence identity in mammals (B), cells transfected with siRNA against β-catenin have decreased surface labeling (C). As expected, transfection of siRNA against HSlo eliminates nearly all surface HSlo signal (D). Scale bars = 20 µm. (E) Western blot analysis showed the efficiency and specificity of HSlo and β-catenin siRNA knockdown of protein expression BsloC, a rabbit polyclonal antibody against the very C-terminus of bovine Slo (BsloC) that shows high sequence homology with hSlo, was used for detection of Slo. Detection of h-vinculin was used to confirm equivalent loading of wells. (F, G). FACS experiments show the decrease in the ratio of surface to total protein expression. This ratio was obtained from the integration of the histograms. Figure 3G shows that relative surface HSlo expression decreases 29% in β-catenin siRNA transfected cells compared to control transfected cells (+/− SEM). One way ANOVA p = 0.0179. In a Student-Newman-Keuls Multiple Comparisons Test between 1. siRNA beta catenin and siRNA chick beta-4 p<0.05; 2. mock transfected cells and beta-catenin siRNA treated cells p<0.05; and 3. mock transfected cells and beta-4 siRNA treated cells p>0.05.).
Figure 4.
Total Slo expression and more significantly, surface clusters of Slo on the surface of hair cells are decreased with β-catenin knockdown.
Shown are confocal images in the X-Y plane of control tall hair cells (A, B and C) and hair cells transfected with siRNA to β-catenin (D, E, and F) stained with antibodies to β-catenin (A,D) and Slo (B,E). The corresponding pseudo colored (β-catenin in red and Slo in green) merged images are shown in C and F. Slo clusters of varying intensities on the surface of the cell are indicated (yellow arrows). Note that the thresholds for the Slo images have been changed to show Slo clusters only. These images were obtained at approximately the same z plane depth. There is a 20% reduction in β-catenin staining after knockdown (mean fluorescence intensity 812+/−31 S.E.M. control vs. 646+/−30 S.E.M., p = 0.003 on a t test, n = 3). There is a reduction in Slo staining (G) and Slo clusters (E,H) after β-catenin knockdown. Mean fluorescence intensity of Slo per µm2 in control hair cells in a 2500 µm2 area was 366 A.U. (+/−21 S.E.M, n = 3 cochlea) vs. 266 A.U. (+/−8) in β-catenin siRNA treated hair cells (p = 0.018 on a t test, n = 3 cochlea). There is an even greater reduction (70%) in Slo clusters after β-catenin knockdown (E) in a similar 50 µ×50 µ area. The number of Slo clusters in control tall hair cells were 2695 (+/−675 S.E.M., n = 3 cochlea) and contrasts with the 790 (+/−132 S.E.M., n = 3) Slo clusters in tall hair cells from β-catenin siRNA treated cochlea (p = 0.025 on a t test). Also note that we were unable to quantify co-localization between Slo and β-catenin in hair cells, since Slo forms clusters in these cells, unlike β-catenin, which has a more uniform distribution. In contrast, in HEK cells both proteins have a more uniform distribution with wide overlap. Manders overlap co-efficient (M1) in hair cells was 0.015, effectively ruling out meaningful co-localization analysis.
Figure 5.
HSlo surface expression in β-catenin -null H28 cells occurs only with co-expression of β-catenin.
HSlo was surface-labeled with anti-Flag antibody and Alexa594-conjugated 2nd antibody (red), while β-catenin was N-terminally tagged with EGFP (green). HSlo surface expression was not detected when cells were transfected with empty vector (A), or EGFP- β-catenin alone (B). There was minimal surface expression of Slo with transfection with HSlo alone (C). However, when H28 cells were co-transfected with both HSlo and EGFP- β-catenin, surface expression of HSlo can be seen (arrows) in the cells having a high level of β-catenin expression (D1, D2). All scale bars = 20 µm.
Figure 6.
β-catenin co-immunoprecipitates with HSlo from HSlo-HEK cells.
(A) β-catenin co-purifies with HSlo. N-terminal FLAG-tagged HSlo was purified from HSlo-HEK cells using an anti-FLAG agarose column. Protein eluted with FLAG peptide was separated by SDS-PAGE. Coomassie staining revealed a single band that corresponds in size to HSlo (120 kDa). A western blot from the same gel, probed with mouse anti β-catenin antibody revealed a single band that corresponds in size to β-catenin (∼90 kDa). (B) Reciprocal immunoprecipitation experiments using protein A/G agarose and a rabbit antibody against the very C-terminus of bovine Slo (BsloC) revealed that HSlo was present in β-catenin immunoprecipitates (upper panel) and vice versa (lower panel). On the other hand, the S10 deletion mutant HSloΔS10 has weaker Slo signal than the wild type HSlo in β-catenin immunoprecipitates. The ratio of HSlo in immunoprecipitates from wild type HSlo∶ HSloΔS10 was 4∶1. The ratio of HSlo from lysates of WT-HSlo and HSloΔS10 was 2∶1. Thus even when corrected for reduced expression of HSlo in the HSloΔS10 cell line, there was a reduction in the amount of HSlo immunoprecipitated by β-catenin. When immunoprecipitated with BsloC, β-catenin was not detectable in HSloΔS10 pulldown, indicating a weaker interaction with β-catenin.
Figure 7.
Mutation of the putative GSK3_1 phosphorylation sites in the S10 region has clear effects on HSlo surface expression.
(A) Construct for the mutants at the GSK3_1 site (SXXXS) on S10. The S10AA phosphonull mutant has very weak surface expression (B) with prominent intracellular ER-like retention (arrows in C), while the S10DD phosphomimetic mutant has excellent cell surface expression (D) with minimal ER-like retention (E). Since the total expression of the S10AA mutant was lower than that of S10DD according to western blotting results, we acquired panels (B) and (C) with identical exposure times that were longer than those for panels (D) and (E), so that the fluorescence intensity in the images of permeabilized cells C and E is comparable. Scale bars = 20 µm. FACS experiments gave similar results (F). Each sample represents the fluorescence distribution of 20,000 cells. (G) Further quantification of the ratio of surface to total protein expression using FACS histogram integration on sample sets with surface labeling and total labeling (+/− SEM). A one-way ANOVA yielded a p value of 0.036. A multiple comparison test revealed significant differences between HSlo and S10DD (p<0.05), and S10AA and S10DD (p<0.05).
Figure 8.
Phosphorylation-mutation effects on HSlo kinetics.
(A) Representative current traces from inside-out patch recordings of the wild type and mutant HSlo channels in nominally 0 µM Ca2+ demonstrating faster activation of S10DD channels and slower activation of S10AA channels compared to the wild type HSlo. (B) Corresponding activation time constants, obtained from monoexponential fits to the activation time course at the potentials given. Error bars represent SEM from 11–16 patches. The purpose of fitting the time course with a single exponential decay is to make it easier to distinguish the groups between each other. Dashed lines are for 0 µM Ca2+, solid lines for 10 µM Ca2+, (C) G-V curves. Solid symbols are with 10 µM internal Ca2+, while open symbols are with zero Ca2+. The V1/2 of activation in 10 µM Ca2+ was 20 mV for HSlo, 46 mV for S10DD (DD), and 67 mV for S10AA (AA). In zero Ca2+ the wild type channel had V1/2 = 133 mV, while the S10DD and S10AA mutants were 107 and 146 mV.
Figure 9.
HSlo expression inhibits Wnt signaling in 293T cells as assayed with the TOP-FLASH assay.
This assay measures the activity of a luciferase reporter whose expression is under the control of the TCF/LEF promoter to provide an assessment of β-catenin or Wnt ligand-stimulated activation of TCF/LEF mediated transcription. (A) Dose dependent titration of Wnt signaling intensity by increasing amounts of HSlo co-transfection. The normalized relative Wnt signaling intensity decreases when HSlo was co-expressed with either Wnt3a ligand or exogenous β-catenin. The amount of either Wnt3a DNA (on the left side in the bar graph) or β-catenin DNA (on the right side) was fixed and predetermined to give maximum TOP-FLASH responses when HSlo was absent. p200 was used as a positive control. p200 is part of the C-terminus of polycystin-1 protein, and is known to specifically inhibit the binding between TCF-LEF and β- catenin, thus inhibiting luciferase signal. (B) Transfection of HSlo S10 deletion mutant or phosphorylation mutants inhibits Wnt signaling in TOP-FLASH assays in a manner similar to wt HSlo. The data represent means from 3 samples each (+/− SD, n = 3). A total amount of 1.6 µg DNA was transfected into each well with a fixed ratio of HSlo/ HSlo-mutants and Wnt3a or HSlo/ HSlo-mutants and β-catenin. In the Wnt3a experiments we used 1.55 µg of HSlo or HSlo mutants with 0.05 µg of Wnt3a plasmid DNA. In the β-catenin experiments we used 0.9 µg of HSlo or HSlo mutant plasmid DNA and 0.7 µg of β-catenin plasmid DNA.
Figure 10.
HSlo interaction with β-catenin involves multiple sites on HSlo, including regions outside S10.
(A) Three HSlo peptides were chosen for further study from several peptide segments with high binding scores. Shown are the sequences of these peptides. (B, C) Mapping peptide regions to the crystal structure of HSlo C-terminus [36]. The 3D filled –in structure of the C-terminus of Slo (B) viewed at a tangent down the four-fold axis. The black vertical line indicates the four-fold axis. Shown in pink is the peptide sequence beginning at residue E562 (named EDT here), in blue the S10 region and in orange the Ca2+ bowl. As is evident the S10 region and the peptide beginning at E562 are adjacent in the tertiary structure of the protein (although separated in the primary structure). C shows the identical view with the ribbon diagram of its secondary structure. As is evident these two regions are on the surface of the protein where they are accessible for protein-protein interactions with β-catenin. (D) Western blots of eluted HSlo protein from an in vitro competitive binding assay, showing reduced binding of whole HSlo molecule to immobilized β-catenin in the presence of HSlo peptides. Eluate from β-catenin immobilized to columns incubated with cell lysates of HEK cells expressing HSlo in the presence of each peptide was separated on SDS-PAGE in duplicate. A charged peptide fragment from the extracellular protein acetylcholinesterase (AchE) served as a negative control (single lane on the right). (E) Averaged densitometric analysis of competitive binding measurements (+/− SEM, n = 3). In the presence of the peptide beginning at residue E562 HSlo binding to immobilized β-catenin is reduced 43% (p = 0.039, one way ANOVA).