Characterization of VPS34-IN1, a selective inhibitor of Vps34, reveals that the phosphatidylinositol 3-phosphate-binding SGK3 protein kinase is a downstream target of class III phosphoinositide 3-kinase

Abstract

The Vps34 (vacuolar protein sorting 34) class III PI3K (phosphoinositide 3-kinase) phosphorylates PtdIns (phosphatidylinositol) at endosomal membranes to generate PtdIns(3)P that regulates membrane trafficking processes via its ability to recruit a subset of proteins possessing PtdIns(3)P-binding PX (phox homology) and FYVE domains. In the present study, we describe a highly selective and potent inhibitor of Vps34, termed VPS34-IN1, that inhibits Vps34 with 25 nM IC50 in vitro, but does not significantly inhibit the activity of 340 protein kinases or 25 lipid kinases tested that include all isoforms of class I as well as class II PI3Ks. Administration of VPS34-IN1 to cells induces a rapid dose-dependent dispersal of a specific PtdIns(3)P-binding probe from endosome membranes, within 1 min, without affecting the ability of class I PI3K to regulate Akt. Moreover, we explored whether SGK3 (serum- and glucocorticoid-regulated kinase-3), the only protein kinase known to interact specifically with PtdIns(3)P via its N-terminal PX domain, might be controlled by Vps34. Mutations disrupting PtdIns(3)P binding ablated SGK3 kinase activity by suppressing phosphorylation of the T-loop [PDK1 (phosphoinositide-dependent kinase 1) site] and hydrophobic motif (mammalian target of rapamycin site) residues. VPS34-IN1 induced a rapid ~50-60% loss of SGK3 phosphorylation within 1 min. VPS34-IN1 did not inhibit activity of the SGK2 isoform that does not possess a PtdIns(3)P-binding PX domain. Furthermore, class I PI3K inhibitors (GDC-0941 and BKM120) that do not inhibit Vps34 suppressed SGK3 activity by ~40%. Combining VPS34-IN1 and GDC-0941 reduced SGK3 activity ~80-90%. These data suggest SGK3 phosphorylation and hence activity is controlled by two pools of PtdIns(3)P. The first is produced through phosphorylation of PtdIns by Vps34 at the endosome. The second is due to the conversion of class I PI3K product, PtdIns(3,4,5)P3 into PtdIns(3)P, via the sequential actions of the PtdIns 5-phosphatases [SHIP1/2 (Src homology 2-domain-containing inositol phosphatase 1/2)] and PtdIns 4-phosphatase [INPP4B (inositol polyphosphate 4-phosphatase type II)]. VPS34-IN1 will be a useful probe to delineate physiological roles of the Vps34. Monitoring SGK3 phosphorylation and activity could be employed as a biomarker of Vps34 activity, in an analogous manner by which Akt is used to probe cellular class I PI3K activity. Combining class I (GDC-0941) and class III (VPS34-IN1) PI3K inhibitors could be used as a strategy to better analyse the roles and regulation of the elusive class II PI3K.

Figure 1. VPS34-IN1, a selective inhibitor of Vps34 kinase

(A) Chemical structure of the VPS34-IN1. (B) Insect cell expressed recombinant human Vps34–Vps15 complex was assayed by measuring phosphorylation of PtdIns in a ³²P-radioactive kinase assay in the absence or presence of the indicated concentrations of VPS34-IN1. Reactions were chromatographed on a Silica 60 TLC plate and ³²P-radioactivity associated with the spot comprising PtdIns(3)P was visualized (top panel) and quantified (bottom panel) by phosphoimager analysis on a Fujifilm Image reader FLA-2000 employing the AIDA image analysis software. Data are shown as the mean kinase activity±S.D. for three independent experiments, relative to DMSO-treated sample. The IC₅₀ histogram was generated using Prism Software with non-linear regression analysis. (C) Protein kinase profiling of the VPS34-IN1 at a single concentration of 1 μM carried out against the Dundee panel of 140 protein kinases at the International Centre for Protein Kinase Profiling. Results for each kinase are presented as the mean kinase activity±S.D. for an assay undertaken in triplicate relative to a control kinase assay in which the inhibitor was omitted. Abbreviations and assay conditions used for each kinase are defined at http://www.kinase-screen.mrc.ac.uk.

Figure 2. VPS34-IN1 selectively inhibit class III PI3K

(A) Lipid kinase profiling of the VPS34-IN1 at 1 μM was carried out against a panel of 19 lipid kinases at the International Centre for Protein Kinase Profiling. (B) Summary of the measurements of VPS34-IN1 for the indicated kinases and IP1 levels in cells (see the Materials and methods section for assay conditions). (C) Lipid kinase profiling of VPS34-IN1 inhibitor was carried out against ProQinase panel of 13 lipid kinases. Summary of IC₅₀ measurements for each kinase is presented in the Table. Abbreviations used for each kinase and assay conditions used are defined at http://www.proqinase.com. Dose–response curves for data marked with an asterisk are presented in Supplementary Figure S3.

Figure 3. VPS34-IN1 reduces GFP–2×FYVE_Hrs probe localization on endosomes in a dose-dependent manner

U2OS cell line stably expressing GFP–2×FYVE_Hrs was treated with indicated concentration of VPS34-IN1 inhibitor for 1 h. The cells were permeabilized by freeze–thaw in liquid nitrogen and fixed in 4% paraformaldehyde. The GFP signal was enhanced by using mouse anti-GFP primary and anti-mouse Alexa Fluor® 488 secondary antibody. The top panel shows representative cell images of GFP–2×FYVE_Hrs localization and the bottom panel shows corresponding DAPI staining for each condition. The histogram displays average cell fluorescence±S.D. compared with DMSO-treated control. Similar result was obtained in at least one other experiment. The average cell fluorescence was calculated by dividing total fluorescence of each field by the number of the cells in the field. For each condition, ten random fields were chosen containing 20–25 cells/field. Images were taken using Nikon Eclipse Ti-S microscope using ×40 objective. Analysis was performed using NIS-Elements BR3.1 program. Scale bar, 20 μm. *P≤0.05, ***P≤0.001. ns, not significant.

Figure 4. VPS34-IN1 rapidly reduces GFP–2×FYVE_Hrs probe localization on endosomes

(A) The U2OS cell line stably expressing GFP–2×FYVE_Hrs was recorded for approximately 1 min, before adding either no inhibitor (top panel), 1 μM VPS34-IN1 (second panel), 0.5 μM GDC-0941 (third panel) or a combination of 1 μM VPS34-IN1 and 0.5 μM GDC-0941 (bottom panel). Images were taken starting at 1.5 min from the time that the inhibitor was added (the first time point we could reliably measure) and subsequently at 0.5 min intervals up to a period of 1 h. Time-lapse microscopy was performed on Zeiss 710 microscope using ×63 objective. Similar results were obtained in at least two separate experiments. Scale bar, 20 μm. (B) As in (A) except the U2OS cell line stably expressing GFP–PH_Akt1 was starved of serum overnight and treated with either no inhibitor (left panel), 1 μM VPS34-IN1 (middle panel) or 0.5 μM GDC-0941 (right panel) for 1 h before stimulation with IGF (100 ng/ml for 15 min). The cells were fixed with 4% paraformaldehyde and images were taken using Zeiss 710 microscope at ×63 objective. Representative images are shown and similar results were obtained in two experiments. Scale bar, 20 μm.

Figure 5. SGK3 binds to PtdIns(3)P via its PX domain

(A) ITC was performed by gradually titrating 0.5 mM PtdIns(3)P into the reaction chamber containing wild-type 3×FLAG–SGK3-(1–162) PX domain (left panel) or mutant 3×FLAG–SGK3-(1–162) R50A PX domain (middle panel) and 3×FLAG–SGK3-(1–162) R90A PX domain (right panel). The top panel illustrates enthalpic heat released during titration at 30°C and the bottom panel presents integrated binding isotherms and the best fit curves. K_d are indicated on the graphs. The measurement was performed on a VP-ITC MicroCalorimeter MicroCal machine. K_d were calculated using Origin 7.0 with ITC data analysis disc program. (B) The ability of the indicated GST fusion proteins to bind various phosphoinositides was analysed. The indicated phosphoinositides (500 pmol) were spotted on to nitrocellulose membranes, which were then incubated with 10 nM of the wild-type and indicated mutants of 3×FLAG–SGK3-(1–162) PX domain or GST-fusion PH domains [PLCδ-(1–178), GRP1-(241–399), TAPP1-(195–315) proteins]. The membranes were washed and the PX domain protein bound to the membrane by virtue of its interaction with lipid was detected using by using HRP-conjugated anti-FLAG antibodies (SGK3 PX domain) or anti-GST (PH domains).

Figure 6. PtdIns(3)P binding localizes SGK3 to endosomes

(A) U2OS stably expressing wild-type or indicated mutant of full-length SGK3 with a C-terminal GFP-tagged were fixed with 4% (v/v) paraformaldehyde and GFP distribution in cell was visualized. Images were taken using a Nicon Eclipse Ti-S microscope with ×40 objective (upper panel). SGK3 co-localization with early endosomal marker EEA1 marker was visualized using rabbit anti-EEA1 primary and anti-rabbit Alexa Fluor® 594 secondary antibody (lower panel). Pictures were taken with Zeiss 710 microscope using ×100 objective. Mander's correlation coefficient in 112 cells was calculated using the Volocity 6.3 program. Scale bar, 20 μm. (B) The U2OS cell line stably expressing full-length SGK3 with an C-terminal GFP-tagged were recorded for approximately 1 min, before adding either no inhibitor (top panel), 1 μM VPS34-IN1 (second panel), 0.5 μM GDC-0941 (third panel) or a combination of 1 μM VPS34-IN1 and 0.5 μM GDC-0941 (bottom panel). Images were taken starting at 1.5 min from the time that the inhibitor was added (the first time point we could reliably measure) and subsequently at 0.5 min intervals up to a period of 1 h. Time-lapse microscopy was performed on Zeiss 710 microscope using ×63 objective. Representative images are shown and similar results were obtained in two experiments. Scale bar, 20 μm.

Figure 7. SGK3 phosphorylation and kinase activity is controlled by its ability to bind PtdIns(3)P

(A) U2OS stably expressing wild-type or indicated mutant of full-length SGK3 with a C-terminal FLAG-tag were treated in the absence or presence of the indicated PDK1 inhibitor (1 μM GSK2334470) for 1 h. The cells were lysed and SGK3 immunoprecipitated with anti-FLAG antibody. The immunoprecipitates were subjected to immunoblot analysis with the indicated antibodies (lower panel) after being assayed for SGK3 kinase activity by measuring phosphorylation of the Crosstide substrate peptide in the presence of 0.1 mM [γ-³²P]ATP in a 30 min reaction (upper panel). Kinase reactions are presented as means±S.D. for triplicate reaction. The similar result was obtained in at two separate experiments. ****P≤0.0001.

Figure 8. Evidence that SGK3 phosphorylation and kinase activity is controlled by Vps34 and class I PI3K

(A) U2OS stably expressing wild-type or indicated mutant of full-length SGK3 with a C-terminal FLAG-tag were treated in the absence or presence of the indicated concentrations of VPS34-IN1 inhibitor or class I PI3K inhibitor (0.5 μM GDC-0941 [34]) PDK1 inhibitor (1 μM GSK2334470 [43]), mTOR inhibitor (0.1 μM AZD8055 [45]) for 1 h. The cells were lysed and SGK3 immunoprecipitated with anti-FLAG antibody. The immunoprecipitates were subjected to immunoblot analysis with the indicated antibodies (lower panel) after being assayed for SGK3 kinase activity by measuring phosphorylation of the Crosstide substrate peptide in the presence of 0.1 mM [γ-³²P]ATP in a 30 min reaction (upper panel). Kinase reactions are presented as means±S.D. for triplicate reaction. (B and C) As in (A) except cells were treated with the indicated doses of the BKM120 class I PI3K inhibitor. Similar results were obtained in at least two separate experiments for all data shown in this Figure. ***P≤0.001, **P≤0.01 and *P≤0.05. ns, not significant.

Figure 9. Class I and class III PI3K inhibitors rapidly inactivate SGK3 and evidence that SGK2 is not regulated by Vps34

(A and B) U2OS stably expressing wild-type full-length SGK3 with a C-terminal FLAG-tag were treated in the absence or presence of the 1 μM VPS34-IN1 (A) or 0.5 μM GDC-0941 (B) for the indicated times. The cells were lysed and SGK3 immunoprecipitated with anti-FLAG antibody. The immunoprecipitates were subjected to immunoblot analysis with the indicated antibodies (top panel) after being assayed for SGK3 kinase activity by measuring phosphorylation of the Crosstide substrate peptide in the presence of 0.1 mM [γ-³²P]ATP in a 30 min reaction (bottom panel). Kinase reactions are presented as means±S.D. for triplicate reaction. (C) As above except U2OS stably expressing wild-type full-length SGK2 with a C-terminal FLAG-tag were treated in the absence or presence of the indicated inhibitors for 1 h. Similar results was obtained in at least two separate experiments for all data shown in this Figure. ***P≤0.001, **P≤0.01 and *P≤0.05. ns, not significant.

Figure 10. Model of how SGK3 is regulated by two distinct pools of PtdIns(3)P

Our data suggest that SGK3 is controlled by two pools of PtdIns(3)P. One pool is produced via phosphorylation of PtdIns by class III PI3K termed Vps34 at the endosome and the other pool of PtdIns(3)P is generated as a result of dephosphorylation of PtdIns(3,4,5)P₃ to PtdIns(3)P through the sequential actions of SHIP1/2 and INPP4B PtdIns phosphatases. Our findings suggest that binding of PtdIns(3)P to the PX domain triggers SGK3 activation by promoting phosphorylation of the T-loop by PDK1 and the hydrophobic motif by mTOR. Further work is required to understand how binding of PtdIns(3)P promotes phosphorylation of SGK3 by PDK1 and mTOR.