Dsh expression activates Dlg-mediated spindle orientation through the GK interacting protein GukH.

<p>(<b>A</b>) <i>Drosophila</i> S2 cells were transfected with the SH3-Hook-GK domains of Dlg fused to the cell adhesion protein Ed (Ed:Dlg) alone or in combination with full-length myc-tagged Dsh (wild-type or the D316A mutant). In additional experiments, transfected cells were treated with RNAi directed against GukH. ‘Induced polarization’ of cortical Ed:Dlg was achieved through Ed-mediated cell contacts, and spindle orientation angles of mitotic cells were measured relative to the center of the induced Ed:Dlg crescent (<i>white dashed lines in Merge images</i>). Shown below cell images are western blots of S2 lysates (20 µg total protein loaded) transfected with either wild-type or D316A Dsh (as pMT:myc fusion constructs), which demonstrate equal protein expression (α-myc). An antibody against β-actin was used as a loading control. (<b>B</b>) Cumulative percentage plots reflect fraction of cells with spindles oriented at or below a given angle for all acquired images (n≥30 for all conditions). Ed alone (<i>black; circles</i>) and Ed fused to the entire SH3-Hook-GK (<i>orange; squares</i>) measurements populate a diagonal expected for completely random orientations between 0-90°. Deletion of the Hook domain (ΔHook; <i>blue; upward triangles</i>) or co-expression of Dsh (<i>green; downward triangles</i>) results in a leftward shift expected for cell populations with increased percent of lower angle measurements. Co-expression of the non-binding Dsh-D316A mutant (<i>red; diamonds</i>) did not improve spindle orientation. *, p <0.05 compared to Ed alone, One-way ANOVA with <i>Dunnett</i>'<i>s post-hoc</i> test. (<b>C</b>) GukH RNAi treatment blocks Dlg-mediated spindle orientation induced by Hook truncation (<i>yellow; squares</i>) or Dsh expression (<i>cyan; upward triangles</i>) to a level statistically equivalent to Ed alone (<i>black; circles</i>). (<b>D</b>) Three additional RNAi sequences were generated against distinct sequence elements of GukH and examined for confirmation of specificity. Each alternative (‘Alt #1-3’) target produced a similar loss-of-function phenotype as that in panel C. (<b>E</b>) Schematic representation of the GukH transcript and location of each RNAi target (see <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0114235#s2" target="_blank">Methods</a> for primer sequence information). Diverse targets include 5′ and 3′ untranslated regions along with two distinct targets within the coding region. All targets were against sequences shared universally among GukH isoforms. The average spindle orientation angle for each condition is shown in parentheses (compare to 26.3° for Ed:Dlg+Dsh in panel B).