PtdIns-specific MPR pathway association of a novel WD40 repeat protein, WIPI49

Abstract

WIPI49 is a member of a previously undescribed family of WD40-repeat proteins that we demonstrate binds 3-phosphorylated phosphoinositides. Immunofluorescent imaging indicates that WIPI49 is localized to both trans-Golgi and endosomal membranes, organelles between which it traffics in a microtubule-dependent manner. Live cell imaging establishes that WIPI49 traffics through the same set of endosomal membranes as that followed by the mannose-6-phosphate receptor (MPR), and consistent with this, WIPI49 is enriched in clathrin-coated vesicles. Ectopic expression of wild-type WIPI49 disrupts the proper functioning of this MPR pathway, whereas expression of a double point mutant (R221,222AWIPI49) unable to bind phosphoinositides does not disrupt this pathway. Finally, suppression of WIPI49 expression through RNAi, demonstrates that its presence is required for normal endosomal organization and distribution of the CI-MPR. We conclude that WIPI49 is a novel regulatory component of the endosomal and MPR pathway and that this role is dependent upon the PI-binding properties of its WD40 domain.

Figure 1.

Identification of a novel family of mammalian WD-40 repeat containing proteins. (A), Sequence alignment of the four mammalian Aut10p homologues. Sequences were aligned using ClustalX and displayed using MacBoxShade. Normal residues are black text on white background, conserved residues are white text on black background, similar residues are white text on gray background, and identical residues are yellow text on black background. The predicted location of the seven WD40 motifs inferred by 3D modeling of the protein sequences are indicated by both the blue and yellow bars, whereas the location of WD40 repeats identified through primary sequence analyses are represented by yellow bars only. Double asterisks (^**) indicate the position of a RR motif conserved throughout the protein family shown (B). Phylogenetic analysis of the evolutionary relationship between NP_009006, NP_062559, NP_056425, NP_060453, and Aut10p indicates that the human homologues of Aut10p comprise two separate lineages. (C). Similarity and identity matrix derived from protein sequences using MacBoxShade algorithms. Both the phylogenetic tree shown in B and the similarity/identity matrix presented in C indicate that all four human Aut10p homologues are approximately equally distant in terms of divergence from Aut10p.

Figure 2.

WIPI49 binds to PtdIns3P, PtdIns3,5P₂, and PtdIns5P in vitro. (A) Liposomes of defined phosphoinositide content were generated as described in MATERIALS AND METHODS and subsequently incubated with purified GST-WIPI49 or GST-SNX3-PX (∼250 ng) for 10 min at 37°C. Resultant lipid complexes were pelleted by centrifugation, separated from supernatants, and resolved by SDS-PAGE and Western blotted using GST-specific antisera. Lane 1, base lipids; lane 2, PtdIns3P; lane 3, PtdIns4P; lane 4, PtdIns5P; lane 5, PtdIns3;4P₂; lane 6, PtdIns4,5P₂; 7, PtdIns3,5P₂; lane 8, PtdIns3,4,5P₃. (B) Data from seven independent liposome-binding experiments with GST-WIPI49 was quantified using NIH Image 1.6 and is presented using arbitrary binding values where material bound to PtdIns3P is equivalent to one unit. Error bars, SE. Lane 1, base lipids; lane 2, PtdIns3P; lane 3, PtdIns4P; lane 4, PtdIns5P; lane 5, PtdIns3,4P₂; lane 6, PtdIns4,5P₂; 7, PtdIns3,5P₂; lane 8, PtdIns3,4,5P₃.

Figure 3.

WIPI49 localizes to trans-Golgi and endosomal membranes. An antibody directed against the C-terminus of WIPI49 was raised in rabbits and affinity-purified. (A) Western blot analysis demonstrates anti-WIPI49 recognizes a single band of immunoreactivity 49 kDa in wild-type Cos7 lysates resolved by SDS-PAGE (lane 1). In COS7 cells transiently transfected with a plasmid encoding GFP-WIPI49, anti-WIPI49 detects both endogenous WIPI49 and a stronger second band of immunoreactivity at approximately 77 kDa that corresponds to GFP-WIPI49 (lane 2). (B) Cos7 cells were fixed in paraformaldehyde and processed for immunofluorescence using monoclonal anti-GM130 (a and d) and anti-WIPI49 (b and e). Merged images are shown in c and f. The cell shown in panels d, e, and f has been treated with nocodazole for 30 min before fixation to induce formation of Golgi ministacks. Cos7 cells transiently expressing Hrs-myc (g-i) were processed for immunofluorescence using murine anti-Myc (g and red in i) and rabbit anti-WIPI49 (h and green in i). The overexpression of Hrs results in WIPI49 accumulating in the Hrs-induced swollen endosomes. In addition, in Cos7 cells transfected with GFP-Rab5 (j-l) the Golgi localized pool of WIPI49 (k and red in l) is apparently redistributed to Rab5-positive endosomal regions. Asterisk (^*) indicates position of the nucleus. Scale bar, 5 μm.

Figure 4.

Transfected GFP- and DsRED2-tagged WIPI49 localize to the Golgi and endosomal membranes. Cos7 cells transiently expressing GFP-WIPI49 (a-c and d-f) were processed for immunofluroescence and stained with either murine anti-p230 to detect trans-Golgi elements (a and red in c) or goat anti-EEA1 to detect early endosomes (d and red in f). GFP-WIPI49 is observed to partially colocalize with p230 as well as a subpopulation of EEA1-positive membranes. To examine the relationship of WIPI49 to endosomes in more detail Cos7 cells were cotransfected with either Hrs-myc and GFP-WIPI49 (g-i) or with GFP-Rab5 and DsRED2-WIPI49 (j-l). Hrs-myc was detected by staining with murine anti-Myc (g and red in i) and observed to colocalize substantially with GFP-WIPI49 (h and green in i). In those cells expressing GFP-Rab5 (j and green in l) the DsRED2-WIPI49 (k and red in l) was seen to be localized predominately to Rab5-positive membranes with a significant reduction in the Golgi localized pool. To determine the relationship between the peripheral WIPI49 membranes and PtdIns3P-enriched regions Cos7 cells were cotransfected to express FYVEx2-GFP (m and green in o) in conjunction with DsRED2-WIPI49 (n and red in o). Although the perinuclear WIPI49 fluorescence is separated from the FYVEx2-GFP, there is some colocalization between the peripheral elements.

Figure 5.

Dynamic relationship between DsRED2-WIPI49- and GFP-Rab5-positive membranes. See also Supplementary Video 4. Cos7 cells transfected with plasmids encoding DsRED2-WIPI49 (red) and GFP-Rab5 (blue) were analyzed by fluorescence microscopy. (A) DsRED2-WIPI49 (left), GFP-Rab5 (middle), and the merged image (right) in transfected cells. The arrow indicates a membrane compartment positive for both Rab9 and WIPI49 that undergoes coordinate movement and then segregation of the Rab9 and WIPI49 signals about two thirds of the way through the film. The boxed region in c relates to Figure 5B. (B) Enlarged selected video images from a time-lapse series corresponding to the boxed region from the cell represented in A. Overlay images show prolonged association of GFP-Rab5 with DsRED2-WIPI49 membranes before segregational movement of GFP-Rab5 away from the DsRED2-WIPI49 (arrowhead). Subsequent to this segregation, a second GFP-Rab5 structure transiently associates with the DsRED2-WIPI49 membrane and then trafficks away (arrow). ^*, nucleus. Bar, 3 μm. Frame time indicated relates to seconds.

Figure 6.

Migration of DsRED2-WIPI49 membranes to, and association with, Rab9-positive structures. See also Supplementary Video 5. Cos7 cells transfected with plasmids encoding DsRED2-WIPI49 (red) and YFP-Rab9 (blue) were analyzed by fluorescence microscopy. (A) DsRED2-WIPI49 (left), YFP-Rab9 (middle), and the merged image (right) in transfected cells. (B) Enlarged selected video images from a timelapse series corresponding to the boxed region from the cell represented in A. Overlay images show directional trafficking of a DsRED2-WIPI49-positive vesicle (arrowhead) to, in addition to association and potential fusion with, a YFP-Rab9/DsRED2-WIPI49 endosome (arrow). Bar, 3 μm. Frame time indicated relates to seconds since commencement of filming.

Figure 7.

Perturbation of the mannose-6-phosphate pathway in Cos7 cells overexpressing WIPI49. (A) The subcellular distribution of the CI-MPR and γ-adaptin was examined in Cos7 cells transfected with either GFP-WIPI49 (a-c) or DsRED2-WIPI49 (d-f and g-i). In Cos7 cells transfected with GFP-WIPI49 (b and green in c) the CI-MPR (a and red in c) is observed to be localized in compartments positive for GFP-WIPI49. In Cos7 cells transiently expressing DsRED2-WIPI49 (e and red in f) there is again colocalization between the recombinant protein and CI-MPR, predominately in the perinuclear Golgi compartment. Examination of the γ-adaptin subunit of AP1 (g and green in i) in Cos7 cells expressing DsRED2-WIPI49 demonstrates substantial colocalization between the two proteins. (B) Overexpression of DsRED2-WIPI49 causes an increased secretion of cathepsin D precursor forms compared with cells transfected with DsRED2 only. Cos7 cells were transfected with either DsRED2 or DsRED2-WIPI49 were pulse-chased with ³⁵S-Met, and then cell-associated (C) and secreted (S) forms of cathepsin D were immunoprecipitated and resolved by SDS-PAGE before gel drying and exposure to a phosphoimager cassette.

Figure 8.

WIPI49 is enriched in clathrin-coated vesicles. An enriched CCV fraction was prepared from rat brain as described in the MATERIALS AND METHODS. After purification the relative amounts of clathrin, γ-adaptin, and WIPI49 were determined from equivalent levels of starting brain homogenate, and the initial pellet was retrieved after clarification of the homogenate and the enriched CCV fraction. Both clathrin and γ-adaptin are enriched in CCVs over the starting material, whereas WIPI49 has a lower degree of enrichment.

Figure 9.

The PtdIns binding ability of WIPI49 is required for perturbation of the MPR pathway. (A) The ability of both GST-WIPI49 and GST-WIPI49 R226A, R227A to bind to base liposomes or liposomes that had incorporated PtdIns3P, PtdIns5P, or PtdIns3,5P₂ was compared. Material pelleting with liposomes (P) was resolved against the supernatant remaining after liposome pelleting (S). GST-WIPI49 binds to PtdIns3P containing liposomes (and to a lesser extent PtdIns5P/PtdIns3,5P₂ containing liposomes), whereas GST-WIPI49 R226A, R227A does not. (B) Ectopic expression of WIPI49 R226A, R227A does not disrupt the MPR pathway. When Cos7 cells are transfected with DsRED2-WIPI49 R226A, R227A (b and e and red in c, f) both the CI-MPR (a and green in c) and γ-Adaptin (d and green in f) have a normal subcellular distribution. Thirty-six hours posttransfection cells were fixed and stained with either rabbit anti-CI-MPR or mouse anti-γ-adaptin. (C) The requirement of WIPI49 interacting with PtdIns3P for its recruitment to membranes is supported by inhibition of PtdIns3P generation by the 3-kinase inhibitor LY294002. Cos7 cells expressing GFP-WIPI49 were treated with LY294002 for either 0, 5, or 15 min. After treating, cells were fixed and observed by immunofluroescence. Over time there is a loss of membrane-associated WIPI49.

Figure 10.

Suppression of WIPI49 expression by RNAi. (A) Cos7 cells transfected with either pSUPERpuro, pSUPERpuro-WIPI49I, pSUPERpuro-WIPI49II, or pSUPERpuro-WIPI49III were grown in media supplemented with 10 μg/ml of puromycin for 3 weeks before harvesting in sample buffer and resolution by SDS-PAGE. Subsequent to transfer to PVDF, the samples were probed with rabbit anti-WIPI49. It is clear that although neither pSUPERpuro nor pSUPERpuro-WIPI49I caused a reduction in the protein levels of WIPI49, both pSUPERpuro-WIPI49II and pSUPERpuro-WIPI49III do. (B) Cos7 cells transfected with pSUPERpuro-WIPI49II and selected for 3 weeks in puromycin were fixed in paraformaldehyde and processed for immunofluorescence using goat anti-EEA1 (a and red in c) and murine anti-GM130 (b and green in c). Although the morphology of the Golgi is normal, the early endosomes are more elaborate and larger than usual. The inset in “a” highlights the endosomal morphology. Scale bar, 5 μm. (C) Wild-type Cos7 cells (WT) or Cos7 cells with knocked down levels of WIPI49 (KD) were lysed and fractionated across a continuous sucrose gradient extending from 21 to 54% sucrose as described in MATERIALS AND METHODS. Subsequent to fractionation 24 fractions were collected (where fraction 1 is 21% sucrose and fraction 24 54% sucrose) and analyzed by Western blotting for the presence of EEA1 or CI-MPR.