Establishment of Par-Polarized Cortical Domains via Phosphoregulated Membrane Motifs

Abstract

The Par polarity complex creates mutually exclusive cortical domains in diverse animal cells. Activity of the atypical protein kinase C (aPKC) is a key output of the Par complex as phosphorylation removes substrates from the Par domain. Here, we investigate how diverse, apparently unrelated Par substrates couple phosphorylation to cortical displacement. Each protein contains a basic and hydrophobic (BH) motif that interacts directly with phospholipids and also overlaps with aPKC phosphorylation sites. Phosphorylation alters the electrostatic character of the sequence, inhibiting interaction with phospholipids and the cell cortex. We searched for overlapping BH and aPKC phosphorylation site motifs (i.e., putative phosphoregulated BH motifs) in several animal proteomes. Candidate proteins with strong PRBH signals associated with the cell cortex but were displaced into the cytoplasm by aPKC. These findings demonstrate a potentially general mechanism for exclusion of proteins from the Par cortical domain in polarized cells.

Figure 1. Lgl, Mira, and Numb localize to the cell cortex through BH motifs

A–A″. Basic-Hydrophobic motif signal in Lgl, Mira, and Numb. Previous work identified domains with multiple potential cortical targeting domains in Lgl, Mira, and Numb. The BH-search algorithm was used to analyze the sequences of Drosophila Lgl, Mira, and Numb. The BH signal is shown across each protein’s domain architecture. BH motifs are defined by peaks with a BH signal greater than 0.6 (denoted by blue line) and the peak area indicates the strength of the peak’s BH character. Each protein has a single BH motif that contains at least one aPKC phosphorylation site. Abbreviations are as follow: THR, tomosyn homology region; CLD, cortical localization domain; Cargo BD, cargo binding domain; PTB, phosphotyrosine binding domain; DPF, α-adaptin-binding motif; NPF, Eps15-binding motif. B. BH motifs are necessary and sufficient for cortical targeting of Lgl, Mira, and Numb. Deletion constructs were used to determine if the BH motif is necessary and sufficient for localization to the S2 cell cortex. Representative images of the localization are presented for each protein when transiently transfected in S2 cells and characterized by either immunofluorescence for the HA epitope tag or by EGFP fluorescence. Localization to the cell cortex was quantified as an intensity ratio for the cortex versus the cytoplasm. At least 17 cells were quantified and the mean and the standard error of the mean (SEM) are shown. Asterisks indicate p < 0.0001 as assessed by a t-test to compare each ΔBH construct to its respective full-length control or each EGFP-fusion to its EGFP control. Scale bar, 10 μm.

Figure 2. Phospholipid binding mediates localization to the cell cortex

A. BH motif phospholipid binding was tested using a cosedimentation phospholipid binding assay. B. The BH motifs of Lgl, Mira, and Numb bind directly to phospholipid vesicles. A representative Coomassie-stained SDS-PAGE gel from a lipid vesicle-binding cosedimentation assays is presented. Each protein was characterized with an NH₂-terminal Maltose Binding Protein (MBP) fusion. The BH motif from Neuralized (Neur residues 68–88) was used as a positive control. The supernatant (S) and pellet (P) contain the unbound and bound fractions, respectively. All vesicles contained a 4-1 mixture of 1,2-dioleoyl-sn-glycero-3-phosphocholine (PC) to 1,2-Dioleoyl-sn-Glycero-3-[Phospho-L-Serine] (PS). Lanes marked by +X contained a third lipid that was added to be 10% of the total lipid by mass; abbreviations for these lipids are as follows. PA, L-α-Phosphatidic acid; PI, L-α-phosphatidylinositol; PI4P, L-α-Phosphatidylinositol-4-phosphate; PIP₂, Phosphatidlyinositol-4,5-bisphosphate; PIP₃, 1,2-dioleoyl-sn-glycero-3-[phosphoinositol-3,4,5-trisphosphate]. The fraction bound expressed as a percentage is shown in the bottom panel. Error bars represent SEM from three independent measurements. Asterisks (*, **, ***) represent significance levels of p < 0.05, 0.01, and 0.001, respectively. See also Supplemental Figure 1 for sequence analysis of Neur BH motif.

Figure 3. Electrostatic interactions mediate Lgl, Mira, and Numb interactions with phospholipids and the cell cortex

A. BH motif point mutants test the role for charged residues in cortical localization. B–B″. Charged residues contribute to BH motif of Lgl, Mira, and Numb. The sequence of the BH motifs from Lgl, Mira, and Numb are displayed, highlighting basic residues (blue) and aPKC phosphosites (red). Terminology and location of the point mutants are displayed with pink aspartic acids. The inset shows the effect of each mutation on the BH signal and the area of the BH peak, as computed by BH-search. The BH threshold is displayed by a blue line with a BH signal of 0.6. C. S2 cell localization assay tested if BH motif charge mediates localization to the cell cortex. D–D″. Mutations to acidic residues reduce localization to the cell cortex in transiently transfected S2 cells. The localization of each full length protein with point mutations was characterized by immunostaining for the HA epitope tag (located on each protein’s NH₂-terminus), and the localization was quantified as a cortical to cytoplasmic signal intensity ratio for at least 16 cells (mean ± SEM). Scale bar, 10 μm. See also Supplemental Figure 3A for additional point mutants. E. Lipid binding assays tested if BH motif charge disrupts phospholipid binding. F–F″. Acidic mutations reduce phospholipid binding. A Coomassie-stained gel from lipid vesicle binding sedimentation assays with vesicles of 4:1 PC:PS plus 10% PIP₂ is shown. S indicates supernatant and P the pellet fraction. Arrowheads mark the full MBP-BH protein while other bands are truncation products (e.g. MBP alone). The fraction of protein bound, quantified as the amount of protein in the pellet over total protein, was quantified in triplicate for each vesicle composition. The mean and SEM of the fraction bound are shown. Asterisks indicate p < 0.05 and 0.0001 for one and three asterisks, respectively. Significance was evaluated using a non-parametric t-test relative to the unmutated BH motif. See also Supplemental Figure 2.

Figure 4. Phosphorylation directly inhibits Par substrate BH motif interactions with phospholipids and the cell cortex

A. Schematic of aPKC regulation of BH motif cortical localization assay. B–B″. Phosphorylation by aPKC disrupts Lgl, Mira, and Numb BH motif cortical enrichment. The localization of EGFP fused BH motifs both in the absence and presence (+aPKC) of aPKC is shown, along with that for BH motifs with phosphorylation sites mutated to alanine (3A, S96A, S48AS52A). Expression of aPKC was verified by immunostaining. Representative images from each transfection are shown. Expression of aPKC reduces cortical enrichment, and this reduction is dependent on the presence of the phosphorylation site. The mean and SEM are shown for measurements from at least 16 cells. Asterisks indicate p < 0.0001 as assessed by a non-parametric t-test to compare each aPKC-expressing cell to its respective control without aPKC expression. Scale bar, 10 μm. See also Supplemental Figure 2A. C. The effect of aPKC phosphorylation on phospholipid binding was tested with lipid cosedimentation assays. D–D″. Effect of aPKC on Par substrate BH motif binding to PC/PS/PIP₂ vesicles. Lipid-binding cosedimentation assays were performed in the presence of aPKC and ATP. Samples were analyzed by immunoblotting for the MBP tag. S marks the supernatant and P marks the pellet. The “S-to-A” mutations are: Lgl3A, MiraS96A and NumbS48AS52A. All vesicles were composed of 4:1 PC:PS plus 10% PIP₂. The fraction of protein bound, quantified as the amount of protein in the pellet over total protein, was quantified in triplicate for each vesicle composition. The mean and SEM of the fraction bound are shown. Asterisks indicate p < 0.0001. Significance was evaluated using a non-parametric t-test relative to binding in the absence of aPKC. See also Supplemental Figure 3B–D.

Figure 5. PRBHscreen identifies candidate aPKC-regulated PRBH motifs

A. Schematic representation of bioinformatics identification of PRBH motifs from proteomic files. The BH motif scoring algorithm identifies sequences with strong BH character. Vertical position on a scale from −2 to 2 shows each residues’ BH score. A representative BH plot has been included to depict the maximum BH score. aPKC consensus sites were identified and scored by scanning for S/T residues with preferred residues in the specific NH₂- and COOH-terminal sites. The scale shows the aPKC site score (ranging from −0.1 to 0.35) for each residue in the specified site. Unlisted residues have no effect on the overall phosphosite score. An overall PRBH score was assigned to sequences identified in both the BH and aPKC site identifying algorithm. B–C. PRBH score distribution from in the Drosophila and human proteome. Select PRBH motifs from the fly and human proteome are highlighted in B and C, respectively. Gray points display each identified putative PRBH motif with its respective PRBH max BH score and max aPKC site score. Blue, purple and red denote the number of identified aPKC sites. Blue lines mark the PRBH threshold values, sequences with scores less than these values are not candidate PRBH motifs. See also Supplemental Figure 4 and the Supplemental Data Tables for analysis of the C. elegans and sponge proteomes. D. aPKC phosphorylates candidate PRBH motifs. An autoradiograph for P-32 emission shows that Amer1^BH1-2, CKIγ-2^BH, MPP7^PRBH-GK, and PIP82^PRBH were phosphorylated by aPKC. A Coomassie-stained loading control gel is displayed. E–H. aPKC inhibits localization to the cell cortex for several candidate PRBH motifs. Representative images are shown with a quantification of the cortical to cytoplasmic signal intensity ratio as the mean ± SEM. Cartoons show the constructs characterized with candidate aPKC phosphosites listed. Each protein was characterized as an EGFP-fusion protein. Statistical significance testing was performed using the t-test of aPKC cotransfected cells to singly transfected cells. Asterisks indicate p < 0.001 and 0.0001 for two and three asterisks, respectively. Significance was evaluated using a non-parametric t-test relative of the singly transfected cells to aPKC cotransfected cells. ns, not significant. See also Supplemental Figure 4.

Figure 6. Model for Par polarization by phosphoregulated BH motifs

PRBH motifs mediate localization to regions of the cell cortex opposite the Par complex. The box detail shows the molecular interactions occurring at the interface between the Par and substrate domains. Par substrates will localize to the cortex by electrostatic interactions with phospholipids until they encounter aPKC and become phosphorylated. Phosphorylation reduces their membrane affinity. Weak accessory interactions mediate localization to the cell cortex but they must allow the substrate to dissociate from the cell cortex, when the BH motif is phosphorylated.