Minimal functional domains of the core polarity regulator Dlg

Abstract

The compartmentalized domains of polarized epithelial cells arise from mutually antagonistic actions between the apical Par complex and the basolateral Scrib module. In Drosophila, the Scrib module proteins Scribble (Scrib) and Discs-large (Dlg) are required to limit Lgl phosphorylation at the basolateral cortex, but how Scrib and Dlg could carry out such a 'protection' activity is not clear. We tested Protein Phosphatase 1α (PP1) as a potential mediator of this activity, but demonstrate that a significant component of Scrib and Dlg regulation of Lgl is PP1 independent, and found no evidence for a Scrib-Dlg-PP1 protein complex. However, the Dlg SH3 domain plays a role in Lgl protection and, in combination with the N-terminal region of the Dlg HOOK domain, in recruitment of Scrib to the membrane. We identify a 'minimal Dlg' comprised of the SH3 and HOOK domains that is both necessary and sufficient for Scrib localization and epithelial polarity function in vivo. This article has an associated First Person interview with the first author of the paper.

Keywords: Cell; Dlg; Polarity; Scrib; Scrib module.

Fig. 1.

The Dlg RVxF motif is critical for function. (A) Cartoon showing location of the RVxF motif in the Dlg protein and conservation of the motif across species. The resides mutated in Dlg^ASAKA are highlighted in red. The RVxF consensus as defined by Wakula et al. (2003) is shown. (B,C) Like Dlg^WT (B), Dlg^ASAKA localizes to the basolateral membrane and is enriched at the cell cortex (C). (D-F) Dlg^WT localization (D) as well as Dlg^ASAKA localization (E) is sensitive to scrib depletion, quantified in F. (G,H) Compared to WT (G), dlg null mutant wing discs form disorganized tumors (H). (I,J) Expression of Dlg^WT rescues this phenotype (I), while expression of Dlg^ASAKA does not rescue (J). (K,L) In the follicle epithelium, dlg^m52 null mutants (K) lose polarity, characterized by lateral aPKC spread, and this is rescued by expressing Dlg^WT (L). (M) In contrast, polarity loss is not rescued by Dlg^ASAKA expression. Scale bars: 10 µm (B,K), 100 µm (G). Magenta lines in B-E and white lines in K-M indicate clones of given genotype. Clones in B-E are flip-out GAL4 clones and those in K-M are MARCM clones. (F) One-way ANOVA with Tukey's multiple comparisons test. Error bars represent s.d. Data points are PM Index measurements in single cells. PM Index=cortical/cytoplasmic intensity. **P<0.01, ****P<0.0001.

Fig. 2.

Scrib and Dlg regulate Lgl independently of PP1. (A) Co-IP of transgenic Scrib or Dlg and PP1 from follicle cells fails to detect an interaction, although interaction between PP1 and Sds22 is robustly captured. (B) Co-IP of overexpressed Dlg or Scrib and PP1 from S2 cells following in situ crosslinking also failed to reliably detect interaction between these proteins. Asterisks in A and B indicate relevant bands. (C,D) Lgl^WT membrane localization (C) is severely disrupted by dlg RNAi (D). (E,F) Lgl^KAFA (E) has increased cytoplasmic localization compared to Lgl^WT and is further decreased by dlg RNAi (F). (G) Quantification of Lgl membrane localization. Scale bar: 10 µm. Clones in C-F are flip-out GAL4 clones. (G) One-way ANOVA with Tukey's multiple comparisons test. Error bars represent s.d. PM Index=cortical/cytoplasmic intensity. Data points are measurements from individual cells. n.s. (not significant) P>0.05, ****P<0.0001.

Fig. 3.

Dlg^SH3-HOOK is sufficient for Scrib localization in an induced polarity system. (A) Cartoon of S2 induced polarity assay. Polarizing Dlg by fusion to Ed enables testing of Scrib recruitment in a minimal synthetic system. (B) S2 cells expressing Ed-GFP can be clustered by adhesion between Ed molecules, but this does not alter Scrib localization. (C) When Dlg^{PDZ3-SH3-HOOK-GUK} is fused to Ed, it creates a polarity crescent at the contact site that is able to recruit Scrib. (D) A minimal fragment, Dlg^SH3-HOOK retains the ability to recruit Scrib to the contact site. (E) Quantification of Scrib recruitment to the polarity site in various Ed-Dlg constructs schematized below. Dlg^{PDZ3-SH3-HOOK-GUK} is able to enrich Scrib, while the Ed-GFP negative control cannot. The HOOK and SH3 domains are necessary and, when in combination, sufficient to recruit Scrib. Statistical tests are versus the Ed-GFP negative control construct. Red line denotes the average for the Ed-GFP negative control and indicates no Scrib contact site enrichment. Scale bar: 10 µm. (E) One-way ANOVA with Tukey's multiple comparisons test. Error bars indicate s.d. Data points are individual cell clusters. Enrichment index=contact site/non-contact site intensity. n.s. (not significant) P>0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Fig. 4.

Dlg SH3 and HOOK domains are sufficient for function in vivo. (A-C) Like WT Dlg (A), Dlg^SH3-HOOK-GUK (B) and Dlg^SH3-HOOK (C) localize to the basolateral membrane in follicle cells. All constructs contain an HA epitope tag used for detection. (D) Quantification of cortical localization in A-C. (E-H) Compared to WT (E) and dlg null mutants (F), Dlg^SH3-HOOK-GUK (G) and Dlg^SH3-HOOK (H) fully rescue polarity and epithelial architecture in wing imaginal discs. (I-L) In monolayered dlg-depleted follicle cells (I), both Dlg^SH3-HOOK-GUK (J) and Dlg^SH3-HOOK (K) provide polarity-rescuing activity, quantitated in L. Full restoration of monolayering is more efficient by Dlg^SH3-HOOK-GUK than Dlg^SH3-HOOK: 89.5% (n=38) of Dlg^SH3-HOOK-GUK show no regions of multilayering in rescued follicles, compared to 20.5% (n=39) of Dlg^SH3-HOOK rescued follicles and 0% (n=37) of follicles with dlg-depleted clones alone. (M-P) Both Dlg^SH3-HOOK-GUK (N) and Dlg^SH3-HOOK (O) fully rescue loss of cortical Scrib seen in dlg-depleted cells (M), quantified in P. Scale bars: 10 µm (A,I,M), 100 µm (E). White or magenta lines indicate clones of given genotypes; Clones in I-K and M-O are flip-out GAL4 clones. (D,L,P) One-way ANOVA with Tukey's multiple comparisons test. Error bars indicate s.d. PM Index=cortical/cytoplasmic intensity. aPKC spread is a ratio of lateral:apical fluorescence intensity. Data points are individual cell measurements. n.s. (not significant) P>0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Fig. 5.

Dlg SH3-HOOK is primarily required to regulate Scrib localization. (A) dlg^m52 null mutant cells show reduced cortical localization of Scrib. (B) Scrib mislocalization is rescued by expression of Dlg^WT. (C) Scrib mislocalization is not rescued by expression of Dlg^ASAKA. (D) Quantification of Scrib localization in A-C. (E,F) Both dlg^m52 null mutants (E) and dlg^m30 SH3 point mutant cells (F) lose polarity and mislocalize aPKC. (G,H) Preventing Scrib mislocalization by cortical tethering (myr-Scrib) partially suppresses the polarity loss phenotypes of dlg^m30 SH3 mutant cells (H) but not dlg^m52 null mutant cells (G). (G′,H′) myr-Scrib contains a V5 epitope tag used for detection. (I) Quantification of aPKC mislocalization in (E-H). Scale bars: 10 µm. Magenta or white lines indicate MARCM clones of given genotypes. (D,I) One-way ANOVA with Tukey's multiple comparisons test. Error bars indicate s.d. PM Index=cortical/cytoplasmic intensity. aPKC spread is a ratio of lateral:apical fluorescence intensity. Data points are individual cell measurements. n.s. (not significant) P>0.05, ***P<0.001, ****P<0.0001.