PIKfyve Regulation of Endosome-Linked Pathways

PIKfyve inhibition creates swollen vacuoles inaccessible to fluid phase marker

Knockdown of PIKfyve in HeLa cells creates swollen vacuoles visible by phase contrast light microscopy in ∼30% of cells as previously reported (23). We could obtain highly efficient knockdown of the PIKfyve activator protein Vac14 but this only produced the vacuole phenotype at very low penetrance (∼3%) and did not augment the effect of PIKfyve knockdown on vacuole formation (not shown). MF4 is chemically similar to the recently described specific PIKfyve inhibitor YM201636 by Jefferies et al., with the only difference being that MF4 lacks an amino group on the pyridine ring (24), (Figure 1E). MF4 inhibited PIKfyve with an IC₅₀ of 23 nm, whereas an inactive analogue MF2 showed no activity even at 5 μm. Corresponding MF4 values for class I PtdIns 3-kinases which we determined are 0.25 μm (p110α), 1 μm (p110β), 0.9 μm (p110γ) and 0.8 μM (p110δ). Application of MF4 gives a vacuolar phenotype in all cells within 4 h. Electron microscopic analysis indicates that the large phase lucent vacuoles are inaccessible to internalized horseradish peroxidase (HRP), but they do become surrounded by a second class of smaller (but still swollen) HRP-containing vacuoles, positive for early (EEA1) or late endosomal markers [lysosome-associated membrane protein (LAMP-2)] (Figure 1A and B), which we will henceforth refer to as 'swollen endosomes’. The retromer components Vps26 and SNX1 also associate with swollen endosomal structures (Figure 1C). MF4 does not appreciably reduce cellular PtdIns3P levels or its cellular distribution, as assessed by immunofluorescence labelling of cells for this lipid with a GST-2xFYVE probe (Figure 1D).

Disruption of endosome to TGN trafficking

PIKfyve has been implicated in the control of the retromer-mediated pathway between endosomes and the TGN through siRNA studies of CI-M6PR distribution with respect to the marker protein TGN-46 (23). Certain aspects of these studies may be complicated by the fact that TGN-46 itself undergoes a cycling itinerary, which partially overlaps with ciM6PR (25). We have now analyzed the steady-state distribution of TGN-46, CI-M6PR and golgin-245 following individual knockdown of PIKfyve, hVac14, a combination of the two and the PIKfyve inhibitor. Knockdown of PIKfyve or its inhibition leads to a more dispersed distribution of CI-M6PR but also of TGN-46, whilst knockdown of Vac14 has little effect (Figure 2A and B). In cells where the TGN46 has been highly dispersed, there is little co-localization with CI-M6PR or with an alternative TGN marker golgin-245 (26), which retains its characteristic perinuclear localization (Figure 2C). Under conditions where TGN-46 is highly dispersed, the cis-Golgi marker GM130 retains its ribbon-like disposition (Figure 2D), characteristic of intact Golgi structure. No changes in levels of CI-M6PR or of retromer components are seen as a consequence of PIKfyve inhibition (Figure 2E).

Cell surface exposed CI-M6PR can be retrieved to the TGN following endocytosis, and Shiga toxin follows an overlapping retromer-dependent route (27). We have followed trafficking of a chimeric CD8-CI-M6PR along this route following application of an anti-CD8 antibody and of fluorescently labelled Shiga toxin B-subunit. Furthermore, we have followed the trafficking of a CD8-Furin chimeric protein, which also traffics from the cell surface to the TGN, but in a retromer-independent fashion (28). Following PIKfyve inhibition, internalized CD8-CI-M6PR and CD8-Furin both can reach a TGN compartment, but show slower rates of accumulation at this perinuclear destination (Figure 3A). In cells which show highly dispersed TGN-46 punctae, these do not co-localize with CD8-CI-M6PR at the final time-point (Figure 3B). Shiga toxin can reach the perinuclear TGN as judged by co-localization with golgin-245, (Figure 4A and B) following PIKfyve inhibition, but its accumulation there, is impeded and co-localization with dispersed TGN-46 positive punctae can also be observed (Figure 4C).

Downregulation of RTKs

PIKfyve siRNA studies of EGF-induced EGFR downregulation have not revealed any defect in EGFR downregulation (23). Our data concur with these previous studies, but we noticed that when combined with Vac14 knockdown a modest defect in EGFR downregulation is evident (Figure 5A and B). A far more effective block to EGFR and also c-Met downregulation was seen following application of PIKfyve inhibitor (Figure 5C, D and E). Fluorescence examination of EGFR distribution shows that residual receptor following PIKfyve inhibition is concentrated in punctate structures which frequently co-localize with the early endosomal marker EEA1 (29). In many cases EGFR can be visualized in the interior of endosomes decorated with EEA1 on their limiting membrane (Figure 5G). Thus acute pharmacological inhibition has revealed an important function of PIKfyve in mammalian cells that provides some reconciliation with data obtained with Drosophilia knock-outs (21), which indicated defects in receptor downregulation. If activated receptor was retained at the limiting membrane of endosomes, one may expect more prolonged signalling following acute activation (30). We do not see this when we examine activation of mitogen-activated protein kinase (MAPK) or AKT/protein kinase B (PKB). Rather we see a dampening of AKT signalling at later time-points, e.g. 30 min (Figure 5F). Although IC50 values of MF4 (23 nm) and YM210636 (33 nm) are similar towards PIKfyve, we obtain less selectivity over PtdIns 3-kinase p110α than that reported for YM210636 (24). To be certain that the observed effects on EGFR degradation were not as a result of p110α inhibition, specific inhibitors for both p110α and β were used. Both inhibitors caused a significant reduction in phosphorylation of AKT/PKB but had no effect on EGFR degradation (data not shown). These results in combination with the observation that MF4 does not cause a reduction in initial EGF-dependent phosphorylation of AKT/PKB, suggest that the effects on EGFR degradation are not due to a non-specific inhibition of class 1 PtdIns 3-kinases.

Role of PIKfyve in autophagy

PtdIns3P generated by the PtdIns 3-kinase Vps34 is required for initial steps of autophagosome formation, but the role of PtdIns(3,5)P₂ is less well established. The Atg18 gene in yeast binds both 3-phosphoinositides with a preference for PtdIns(3,5)P₂(16). One signature for the accumulation of autophagosomes is the enrichment of a faster migrating form (LC3 II) of the autophagy protein LC3, following its conjugation to phosphatidyl-ethanolamine (31). Whilst, neither Vac14 nor PIKfyve knockdown had any significant influence on starvation-induced accumulation of GFP-LC3-II (data not shown), inhibitor treatment led to a significant increase, as judged biochemically (Figure 6A and B) and by the visualization of GFP-labelled punctae (Figure 6C). Addition of the PIKfyve inhibitor, MF4 did not significantly increase the accumulation of GFP-LC3 in the presence of leupeptin, a condition in which the degradation of GFP in the late endosome/autolysosome should be reduced. This indicates that MF4 treatment does not increase the rate of formation of autophagosomes but instead interferes with their consumption.

Antibodies and other reagents

Rabbit polyclonal human Vac14 antibodies were raised by immunizing rabbits with the N-terminal amino acids of Vac14 (HLEVRHQRSGRGDHLDRR), conjugated via a cysteine to Limulus polyphemus haemocyanin (Covance) and affinity purified on peptide columns. Sheep anti-TGN46, sheep anti-GM130 antibodies and sheep anti-GFP were generous gifts of Vas Ponnambalam (University of Leeds, UK) Francis Barr and Ian Prior (both University of Liverpool, UK) respectively. Rabbit polyclonal anti-PIKfyve and anti-CI-M6PR antibodies were generous gifts from Lois Weisman (University of Michigan, USA) and Paul Luzio (Cambridge Institute of Medical Research, UK). Mouse monoclonal anti-Met (25H2) antibody, rabbit monoclonal phosphorylated AKT (pAkt) (Ser 473) and mouse monoclonal phosphorylated MAPK (pMAPK) (Thr 202, Tyr 204) antibodies were from Cell Signalling. Monoclonal anti-EGFR R1 was from Cancer Research UK Laboratories (CRUK) and goat polyclonal anti-EGFR 1005 were from Santa Cruz. Mouse monoclonal anti-tubulin was from Sigma. Mouse monoclonal anti-golgin 245 (p230) antibody was from BD Biosciences. Rabbit polyclonal anti-EEA-1 was previously described (4). Secondary antibodies were from Molecular Probes and Licor Biosciences. HeLa M cell lines stably expressing CD8-CI-M6PR and CD8-Furin along with mouse monoclonal CD8 antibody were generous gifts from Matthew Seaman (Cambridge Institute of Medical Research, UK). Purified mouse EGF was obtained from J. Smith, Liverpool, UK. 2GL9 cells, HEK293A cells stably transfected with GFP-LC3 have been previously described (34). GST-2xFYVE (Hrs) vector was a gift from Harald Stenmark (35) Following production in bacteria it was isolated with Glutathione Sepharose and then biotinylated with Sulfo-NHS-LC-Biotin (Pierce, UK). Purified PIKfyve protein was a kind gift from Frank Cooke (UCL, UK).

Synthesis and characterization of MF4 and MF2

Synthesis of the PIKfyve inhibitor (hereafter designated MF4) and its inactive analogue, MF2 is described below. PI-103 (36) (200 mg, 0.57 mmol) and 2,6-di-tert-butyl-4-methylpyridine (136 mg, 0.66 mmol) were dissolved in CH₂Cl₂ (20 mL) and cooled on ice. Triflic anhydride (0.3 mL, 0.63 mmol) was added drop-wise and the reaction was warmed to room temperature. Further CH₂Cl₂ was added (10 mL) and the reaction was allowed to proceed overnight. The solvent was removed in vacuo and the product was purified by silica gel flash chromatography with 25% EtOAc:Hexanes to yield 67 mg (24% yield). Liquid chromatography–electrospray ionisation–mass spectrometry (LC-ESI-MS) [MH]⁺ m/z calculated for C₂₀H₁₅F₃N₄O₅S 481.1, found 480.9. This product (67 mg, 0.14 mmol), cesium carbonate (64 mg, 0.20 mmol) and nicotinamide (21 mg, 0.17 mmol) were added to dioxane (4 mL). Xantphos (10 mg, 0.016 mmol) and Pd₂(dba)₃ (5 mg, 0.006 mmol) were added and the reaction was heated to 100°C for 24 h (37). The reaction was diluted with CH₂Cl₂, filtered and the solvent was removed in vacuo. The product was purified by silica gel chromatography using a gradient of 2–3% MeOH:CH₂Cl₂ and by reverse phase high performance liquid chromotography (HPLC) using a MeCN/H₂O/0.1% trifluoroacetate (TFA) solvent system to give a white solid. LC-ESI-MS [MH]⁺ m/z calculated for C₂₅H₂₀N₆O₃453.2, found 453.0. MF2 was prepared similarly using PIK-112 (36) in place of PI-103. LC-ESI-MS [MH]⁺ m/z calculated for C₂₆ H₂₂N₆O₂451.2, found 451.4.

In vitro kinase assay

Purified GST-PIKfyve bound to 5 μL of glutatione beads was incubated with inhibitors at fourfold dilutions over a concentration range of 5 μm to 0.001 μm in a total reaction volume of 25 μL. The assay buffer contained 10 μm ATP, 2 μCi of (γ-32P) ATP, 100 μm PI-3P, 400 μm phosphatidyl-ethanolamine, 25 mm HEPES (pH 7.4), 120 mm NaCl, 1.5 mm MgCl₂, 1 mm DTT, 0.5 mg/mL BSA and 2% dimethyl sulphoxide (DMSO). Reactions were terminated by spotting onto nitrocellulose membranes, which were then washed five or six times to remove unbound radioactivity and dried. Transferred radioactivity was quantitated using phosphorimaging, and IC50 values were calculated by fitting the data to a sigmoidal dose-response using Prism (GraphPad).

Cell culture, PIKfyve inhibitor treatment and siRNA transfection

HeLa and HEK293A cells were cultured in 5% CO₂ in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% foetal calf serum and 1% non-essential amino acids. To retain expression of stably transfected constructs, culture medium for CD8 cell lines was supplemented with 0.5 mg/mL G418. On-Target Plus siRNA oligos against PIKfyve and Vac14 were purchased from Dharmacon. A 2-hit protocol over 72 h was used. Briefly, HeLa cells were seeded in six-well plates at a density of 2.4 × 10⁵ cells/well. Cells were transfected with 40 nm of the corresponding oligo using oligofectamine (Invitrogen), 24 h later. After a further 24 h, cells were reseeded onto six-well plates at the same density or onto cover-lips at a density of 0.12 × 10⁶ cells/well and transfected again with 40 nM oligos for 24 h. Cells were then processed for western blotting and immunofluorescence as described below.

Growth factor stimulation and cell lysis

Where indicated, cells were serum-starved for 16 h and then stimulated with 100 ng/mL EGF or hepatocyte growth factor (HGF). Cells were washed twice in ice-cold phosphate buffered saline solution and lysed in Nonidet P-40 lysis buffer (0.5% Nonidet P-40, 25 mm Tris/Cl, pH 7.5, 100 mm NaCl). Lysates were cleared by centrifugation, resuspended in 5× SDS-sample buffer and resolved by SDS-PAGE followed by immunoblotting for detection with a Licor Odyssey.

Immunofluorescence

Cells were processed for immunofluorescence by fixation in 3% paraformaldehyde (PFA) in PBS, permeabilized with 0.2% Triton-X-100, blocked in 0.2% fish skin gelatin in PBS and incubated with antibodies in blocking buffer. Secondary antibodies were conjugated to either AlexaFluor 488, 594 or 633. Cover-slips were imaged using a Leica confocal SP2 AOBS (HCX PL APO CS 63.0 × 1.40 oil objective).

HRP Internalization and Electron microscopy (EM)

Cells were incubated in 10 mg/mL HRP in serum-free medium at 37°C for 10 min, 1 h or 4 h. Subsequently, cells were washed three times in ice-cold PBS and fixed in 0.5% glutaraldehyde for 30 min. Cells were washed three times at room temperature with 0.1 m Tris-Cl pH 7.6, followed by incubation for 10 min in 0.1% diaminobenzidine (DAB, Sigma) in Tris buffer and then for 10 min in 0.1% DAB and 0.01% H₂0₂ in Tris buffer at room temperature in the dark. The reaction was terminated, by washing 3 × 5 min in Tris buffer, and then 2 × 5 min in PBS. Cells were then post-fixed in the dark on ice with 1% OsO₄ in 0.1 m phosphate buffer pH 7.4 for 1 h. Subsequently, the cells were washed 3 × 30 mi in PBS, followed by 2 × 30 min in distilled water and then incubated in 5% uranyl acetate in 30% ethanol for 1 h. Samples were dehydrated by sequential 10-min ethanol washes in 30%, 60%, 70%, 80% and then twice in 100% ethanol, then infiltrated with a 1:1 ratio of resin to 100% ethanol for 30 min. The infiltration resin was then replaced with 100% resin and the flat-embedded samples were polymerized at 60°C for 2 to 3 days.

Anti-CD8 uptake experiments

HeLaM CD8-CIM6PR or CD8-Furin cells were seeded onto cover-slips and treated with either siRNA or PIKfyve inhibitor. Cover-slips were then transferred to a well containing 3 mL ice-cold dihydroxybenzoate (DHB) (DMEM containing 25 mm Hepes and 0.2% fatty-acid free BSA) for 15 min. CD8-CIM6PR and CD8-Furin receptors at the cell surface were labelled by placing cover-slips, cell side down, upon a 100 μL drop of DHB containing 1 μg monoclonal anti-CD8 antibody at 4°C, for 1 h. Cover-slips were washed twice in ice-cold PBS, and then transferred to pre-warmed growth medium for the indicated time periods to allow uptake of the anti-CD8 antibody to occur. At the assay endpoint, cells were fixed and processed for immunofluorescence as described earlier.

Shiga toxin uptake assay

The assay has been described previously (38). Briefly, HeLa cells were seeded onto cover-slips and treated with either siRNA or PIKfyve inhibitor as described earlier. Subsequently, cells were incubated on ice for 30 min with 2 μg/μL Cy3-STxB in DHB. The 0-min time-point was immediately fixed and the remaining cover-slips washed twice in ice-cold PBS, transferred to pre-warmed growth medium and returned to the incubator for the indicated time periods to allow uptake to proceed. At the assay endpoint, cells were fixed and processed for immunofluorescence as described earlier.

Induction of autophagy and GFP-LC3 shift assay

HEK293A cells stably expressing GFP-LC3 were seeded onto 12-well plates and treated with siRNA or PIKfyve inhibitor as described above. To induce autophagy the growth medium was either replaced with fresh medium (fed) or cells were washed four times with Earl's buffered saline solution and then starved for 2 h in the final wash. Where indicated, Leupeptin was added to the starvation medium at a concentration of 0.25 mg/mL. Following starvation, cells were either fixed and processed for immunofluorescence or lysed in 1× SDS-sample buffer. After protein determination lysates were supplemented with 1 mm DTT and 0.1% bromophenol blue and resolved by SDS-PAGE for western blotting with polyclonal sheep anti-GFP antibodies, which were quantified using a Licor Odyssey instrument.