Phosphorylation at Ser7 and 295 is required for lateral membrane localization of Numb. GFP-Numb2A localization overlaps with aPKC at the apical region of the cell. Polarized MDCK cells transfected with GFP-Numb (A) or GFP-Numb2A (B) were fixed and co-stained with anti-aPKC (red) and anti-ZO-1 (blue). Confocal X–Y sections were taken as indicated in Figure 5 to show the subapical localization of aPKC (I) and the basolateral accumulation of Numb (II). Optical sections taken in the Z-axis are shown above. (C) MDCK cells transfected with GFP-Numb or GFP-Numb2A were stained anti-gp135 to mark the apical membrane. Merged images of GFP-Numb or GFP-Numb2A (green), and gp135 (red) staining in optical sections taken in the Z-axis are shown. Size bars indicate 10 μm.

Numb phosphorylation at Ser7 and Ser295 regulates cortical membrane localization. (A) HeLa cells were transfected with either GFP-Numb (left panels) or GFP-Numb2A (right panels), and immunostained with anti-α-adaptin. GFP-Numb is localized at the basal substratum-associated membrane with α-adaptin (left panels, PM) and to cytoplasmic vesicles in optical sections taken at the midpoint through the cell (left panels, Mid). GFP-Numb2A was predominantly localized at the cortical membrane (right panels, PM) and little cytosolic/vesicle-associated localization is observed (right panels, Mid). The relative association of GFP-Numb and GFP-Numb2A at the plasma membrane and in the cytosol was determined by calculating a ratio of the mean GFP fluorescence intensity using Improvision Volocity 3.7 software program. Values are given in the text. Bottom panels show merged images of the boxed regions that have been digitally enlarged. (B) Coexpression of CA PKCζ (HA-CA-PKCζ) resulted in loss of GFP-Numb at the cortical membrane (left panels, PM), and increased cytosolic staining (left panels, mid). HA-CA-PKCζ expression has no effect on GFP-Numb2A plasma membrane localization (right panels). Boxed regions have been digitally enlarged and represented as panels below. Size bars represent 10 μm.

Phosphorylation is required for Numb localization to the basolateral membrane in response to cell polarization. Polarized MDCK cells transfected with GFP-Numb (A) or GFP-Numb2A (B) were switched to low-calcium media (0 Ca²⁺) for 20 h followed by Ca²⁺ re-addition for 24 h to allow for repolarization. Changes in the localization of GFP-Numb, aPKC, and ZO-1 were determined by immunostaining as in Figure 6 (A, B). (C) Colocalization of Numb with the apical membrane marker gp135 in a calcium switch experiment was determined as in Figure 6. Merged images of GFP-Numb or GFP-Numb2A (green), and gp135 (red) staining in optical sections taken in the Z-axis are shown. Colocalization of GFP-Numb (left panels) and GFP-Numb2A (right panels) with gp135 after calcium readdition was quantified using the colocalization module of Improvision Volocity 3.7 software program. Values are given in the text. Size bars indicate 10 μm.

Numb is phosphorylated in vivo. (A) Autoradiograph of Numb immunoprecipitates from ³²P-labeled orthophosphate-labeled HEK293T cells treated with the phorbol ester TPA. Equal loading confirmed by Numb immunoblotting (lower panel). (B) Numb immunoprecipitates were incubated with [γ-³²P]ATP and either no PKC, or with the indicated recombinant PKC isozymes. Ten percent of the kinase reaction was subjected to autoradiography to confirm activity (upper panel). Ninety percent of each kinase reaction was boiled, diluted, and Numb was re-immunoprecipitated and subjected to autoradiography (middle panel). Immunoblotting with anti-Numb antibody confirmed the presence of Numb protein in each reaction (lower panel). Note that the band migrating just above Numb in the PKCη reaction likely represents ³²P-labeled autophosphorylated PKCη that was not removed by the denaturing immunoprecipitation step. (C) GST-Numb PTBi and not GST-Numb PTBo or GST alone bound to endogenous PKCζ from MDCK cells lysates (upper panel). Equivalent input of GST-PTB fusion proteins was confirmed by Coomassie blue staining of membrane.

Expression of CA PKCζ causes redistribution of membrane-associated Numb to the cytoplasm. (A) Cortical membrane localization of endogenous Numb is lost in cells expressing CA PKCζ (HA-CA-PKCζ) compared to untransfected cells (upper left panels). Boxed regions have been digitally enlarged to show plasma membrane puncta staining of Numb in an untransfected cell (inset box 1) compared to a HA-CA-PKCζ-transfected cell (inset box 2). Numb cytosolic staining is increased in the presence of HA-CA-PKCζ (lower left panel). (B) Transfection of HeLa cells with DN PKCζ (HA-DN-PKCζ) had no effect on plasma membrane Numb localization. Size bars represent 10 μm. (C) Biochemical fractionation of HeLa cells transfected with HA-CA-PKCζ or HA vector. Numb immunoprecipitations were performed on the triton soluble fraction (S) and triton insoluble pellet (P) and immunoblotted with anti-Numb antibody (upper panel). Twenty μg of protein from each fraction was separated by SDS–PAGE and blotted with anti-Numb, anti-PKCζ to confirm expression levels, and anti-actin to confirm equal protein loading (lower panels).

Cell polarity determines Numb localization in MDCK cells. (A) Polarized MDCK cells were fixed and co-stained with anti-NumbC (green) and either anti-aPKC or anti-Par3 (red). Shown are confocal X–Y sections taken at the subapical region of the cells (I) and 2–4 μm below at the basolateral region (II) as depicted in the diagramatic representation of the cells to the left of the images. Accumulation of Numb at the lateral membranes, below the subapical aPKC and Par3 staining, is shown in the Z sections shown above. (B) Polarized MDCK cells were grown in low calcium medium for 20 h (calcium switch) followed by re-addition of normal media for 24 h. Fixed cells were co-stained with anti-NumbC (green) and anti-α-adaptin (red). (C) MDCK cells were treated with 200 pmol aPKCλ or PKCα siRNA oligonucleotides and grown on filters for 72 h. Following fixation and permeabilization, the monolayers were co-stained with anti-aPKC (which also recognizes aPKCλ; blue), anti-PKCα (red), and anti-NumbC (green). Size bars indicate 10 μm. In lower panel MDCK cells were transfected with 40, 120, or 200 pmol of aPKCλ or PKCα siRNA oligonucleotides, or scrambled RNA control as indicated. In total, 20 μg of protein lysate for each condition was separated by SDS–PAGE and blotted with anti-aPKC or anti-PKCα. Membranes were reprobed with anti-Transferrin receptor as a loading control.

Identification of mammalian Numb phosphorylation sites by phosphopeptide mapping and mass spectrometry. (A) Numb was immunoprecipitated from ³²P orthophosphate-labeled, untreated or TPA-treated HEK293T cells and resolved by SDS–PAGE and autoradiography. Bands corresponding to radiolabeled Numb (confirmed by parallel immunoblot) were excised from the membrane and tryptic fragments were purified and resolved by two-dimensional thin layer electrophoresis and ascending chromatography. Five phosphopeptides were recovered in samples from untreated (left panel) or TPA-treated cells (right panel; spots a–e). (B) Mass spectrometry analysis of Numb proteins purified by immunoprecipitation. MS/MS spectra of peptides LRQSFRR, and QLSLRINELPSTMQR indicating phosphorylated residues Ser7 (left) and Ser295 (right) are shown. (C) Sequence conservation surrounding Ser7 and Ser295 in Numb protein sequence alignment from: C. elegans (ceCKA1; Q9XTY6), Drosophilia melanogaster (dNumb; P16554), mouse (mNumb; AAD47836), human (hNumb; NP001005744), rat (rNumb; ABC69734), chicken (cNumb; AAD49434), and related mouse Numblike protein (mNbl; NP035080). (D) HeLa cells were transfected with HA-tagged Numb or Numb2A, and immunoprecipitated Numb subjected to in vitro kinase reactions with PKCζ as described in Figure 1B. Immunoblotting with anti-Numb antibodies revealed equivalent loading in all lanes (lower panel). Storm Phosphoimager analysis was used to detect ³²P-labeled incorporation, and analysis performed using ImageQuant 5.0 software. A histogram of ³²P-labeled incorporation from a representative experiment of four independent repeats is shown. (E) Immunoprecipitated Numb or Numb2A subjected to kinase assays with PKCζ as described in Figure 1B (in the presence of cold ATP), resolved by SDS–PAGE and immunoblotted with affinity-purified anti-pS7 antibody (upper panel) or anti-Numb (lower panel).

Phosphorylation regulates asymmetric localization of Drosophila Numb. (A) Interaction of Drosophila Numb with PKCζ. Cell lysates obtained from HEK293 cells transfected with myc-tagged Numb or an empty control vector were immunoprecipitated with an anti-myc antibody (IP) and immunoprecipitates were immunoblotted (IB) with PKCζ antibody. Whole-cell lysates were analyzed by Western blotting (IB) to control for expression. (B) Autoradiography demonstrating that Drosophila Numb is an in vitro substrate for PKC isozymes alpha (α) and zeta (ζ) (upper panel). Immunoblot with anti-Numb antibodies confirms presence of Numb protein in each reaction (lower panel). (C) PKCζ phosphorylates Drosophila Numb on conserved serine residues in vitro. Right panel is a Coomassie-stained SDS–PAGE gel showing the purified protein samples used in the kinase assay. Autoradiography (left panel) of the same gel showing phosphorylation of GST-Numb but not GST alone. PKCζ also efficiently phosphorylates GST-Numb S52A and GST-Numb 4A, but only poorly phosphorylates GST-Numb 5A. The compared phosphorylation of GST-Numb 5A and GST-Numb 4A suggests that Ser52 is one of the site phosphorylated in vitro and that Ser52 is not the sole acceptor site. The asterisk marks autophosphorylated PKCζ. (D) Mass spectrometry analysis. Ion chromatograms for the doubly charged peptide SFRDSFR (m/z 457.7) and its corresponding phosphopeptide at Ser52 (m/z 497.7) obtained from the aPKC-treated (top left) and control (top right) GST-Numb tryptic digests. Middle panel shows the MS–MS spectrum of m/z 497.7 for the phosphopeptide SFRDpSFR with characteristic b- and y-type fragment ions, indicating the position of the modified residue at Ser52. (E) The full sequence of Drosophila Numb. The residues that are underlined correspond to the sequenced peptides (65% coverage). The phosphorylated sites are indicated in red. Indirect evidence suggest that Ser304 (in blue) is also phosphorylated (see text). (F) Localization of aPKC (blue) and myc-tagged Numb (green) in dividing pI cells (senseless in red). Numb localized at the anterior cortex of pI cells, whereas aPKC localized at the posterior cortex opposite to Numb at prometaphase and metaphase. Numb4A and NumbS52A also localized opposite to aPKC at the anterior cortex. In contrast, the distribution of Numb5A was not restricted to the anterior cortex.