Nucleocytoplasmic shuttling of the Golgi phosphatidylinositol 4-kinase Pik1 is regulated by 14-3-3 proteins and coordinates Golgi function with cell growth

Abstract

The yeast phosphatidylinositol 4-kinase Pik1p is essential for proliferation, and it controls Golgi homeostasis and transport of newly synthesized proteins from this compartment. At the Golgi, phosphatidylinositol 4-phosphate recruits multiple cytosolic effectors involved in formation of post-Golgi transport vesicles. A second pool of catalytically active Pik1p localizes to the nucleus. The physiological significance and regulation of this dual localization of the lipid kinase remains unknown. Here, we show that Pik1p binds to the redundant 14-3-3 proteins Bmh1p and Bmh2p. We provide evidence that nucleocytoplasmic shuttling of Pik1p involves phosphorylation and that 14-3-3 proteins bind Pik1p in the cytoplasm. Nutrient deprivation results in relocation of Pik1p from the Golgi to the nucleus and increases the amount of Pik1p-14-3-3 complex, a process reversed upon restored nutrient supply. These data suggest a role of Pik1p nucleocytoplasmic shuttling in coordination of biosynthetic transport from the Golgi with nutrient signaling.

Figure 1.

Interaction of Pik1p with the 14-3-3 proteins Bmh1/2p. (A) Pik1p and Bmh1/2p interact in the yeast two-hybrid system. Pik1p full-length (aa 2-1066; pMB138) and a C-terminal truncation (aa 2-769; pMB105) were expressed as LexA fusion proteins and targets as GAL4 DNA activation domain fusions. Growth is shown on −His reporter plates. (B) Quantification of Pik1p-Bmh1p and Bmh2p interaction. β-Galactosidase activity was determined in the ONPG liquid culture assay (n = 5; mean ± SD). Frq1p was used as control. (C) GST-14-3-3 pull down of Pik1p from yeast lysates. GST-Bmh1p full length (pMB186) or truncated (aa 1-198; pMB241) fusion proteins were purified and incubated with a cleared yeast lysate (CSY513) in the presence of phosphatase inhibitors. Western blots were probed with anti-Pik1p polyclonal antibody. SDS-PAGE loads were normalized to equal amounts of GST-fusion proteins and Western blots probed with peroxidase anti-peroxidase (PAP) complex. (D) GST-14-3-3 pull down of in vitro-translated Pik1p. Pik1p was translated in vitro (pLD103) (TNT Quick-Coupled Transcription/Translation System) and incubated with GST-Bmh1p (pMB186) or GST-Bmh2p (pMB187) immobilized on glutathione-Sepharose. Radiolabeled Pik1p was detected by autoradiography. (E) Characterization of the Pik1p complex. TAP purification using a C-terminal Pik1p TAP tag (CSY370) was performed in the presence of phosphatase inhibitors, purified material was fractionated on a 4–16% gradient SDS-PAGE gel, and proteins were identified by MALDI-TOF and nanoES MS/MS. The analysis also identified common contaminants in TAP purification experiments: Sse1p, Ssa1/2p, Ssb1/2p, and human keratin. (n.d., not defined).

Figure 2.

14-3-3 proteins are present in a complex with Frq1p and Pik1p. (A) TAP purification using a C-terminal Frq1p TAP tag (Open Biosystems) was performed in the presence of phosphatase inhibitors and starting cell lysate, and purified material was analyzed by SDS-PAGE and Western blotting with anti-Pik1p, anti-Bmh1/2p, and PAP antibody. (B) Frq1p-TAP purified material was fractionated on a 4–16% gradient SDS-PAGE gel, and proteins were identified by nanoES MS/MS. For full list of identified proteins, see Supplemental Table 4.

Figure 3.

14-3-3 overproduction interferes with late secretory pathway function. (A) Synthetic enhancement of pik1-101 temperature-sensitive growth defect by high copy BMH1/2. The pik1-101 mutant (CSY212) was transformed with 2μ BMH1 (pMB180), 2μ BMH2 (pMB181), 2μ PIK1 (pCS8), or control vector (pRS426), and serial dilutions of cells were tested for growth at 25, 32, or 34°C. (B) Gap1*p-GFP surface transport is similarly impaired in pik1-101 and wild-type cells overexpressing BMH2 (pMB263). Wild-type (NY604) and pik1-101 (YMB064) cells were transformed with Gap1*p-GFP (TPQ99). Cells were grown overnight at 25°C in SCRaf-URA and induced with 3% galactose for 3 h. The temperature shift was performed during the last hour of induction. Cells were analyzed by fluorescence microscopy. Gap1*p-GFP accumulation is indicated by arrowheads. (C) Gap1*p-GFP accumulating structures were counted for pik1-101 (n = 68), wild-type (pRS425; n = 256), wild-type cells overexpressing BMH1 (pMB262; n = 54) or overexpressing BMH2 (pMB263; n = 243). Data are mean ± SEM. (D) Gap1*p-GFP surface activity is reduced in cells overexpressing BMH2. Cells were grown, and Gap1*p-GFP expression was induced as described in B. Gap1*p-GFP activity was determined by measuring the uptake of [¹⁴C]citrulline. Data are mean ± SD (n = 3).

Figure 4.

Localization of the Pik1p–14-3-3 interaction. (A) Subcellular localization of Pik1p and Bmh1/2p. BMH1-13myc– (YMB054) and BMH2-13myc (YMB058)–expressing cells were subjected to fixation, and Bmh proteins were labeled using anti-myc antibody (9E10) and Pik1p with anti-Pik1p antibody. DNA was stained with DAPI. Cells were viewed by fluorescence microscopy. (B) Pik1p is localized to the nucleolus. Wild-type cells (NY13) were labeled with anti-Pik1p antibody, mAb clone 2.3b for nucleoli, and DAPI. (C) Pik1p and Bmh1/2p distribution in subcellular fractions. Cells containing PIK1-TAP and BMH2-13myc (YMB119) were subjected to 13,000 × g and 100,000 × g centrifugation, and fractions examined by Western blotting. Pik1p-TAP was detected with PAP complex and Bmh1p-13myc or Bmh2p-13myc with anti-myc antibody (9E10). (D) Pik1p and 14-3-3 interact in the cytoplasm. Cells containing PIK1-13myc and BMH1-3ha (YMB148) were subjected to 100,000 × g centrifugation. After detergent lysis of the cytosolic or the membrane fraction, Pik1p-13myc was immunoprecipitated using anti-myc antibody (9E10). Coprecipitated Bmh1p-3ha was detected using anti-HA antibody (Y11). Loading controls represent 0.5% of the fractions. (E) PI 4-kinase activity of immunoprecipitated Pik1p from cytosol or membrane fractions prepared as described in D. Pik1p-13myc was precipitated with anti-myc antibody (9E10) and subjected to lipid-kinase assay by using phosphatidylinositol as substrate. Shown are results of two representative experiments. The amount of radioactive labeled PI(4)P was normalized to Pik1p-13myc protein levels.

Figure 5.

Functional 14-3-3 is required for normal Pik1p localization. Wild-type (NY1211), bmh1Δ (GG3001), and bmh1Δ bmh2ts (GG3000) mutant strains expressing GFP-PIK1 from the endogenous promoter on a CEN-based plasmid were grown overnight, shifted to 37°C for 1 h when indicated, and visualized using a laser confocal microscope. (A) Laser confocal sections through the middle of representative cells in each treatment are shown. (B) Cells with nuclear GFP-Pik1p were counted (n = 80–90). (C) Cells with GFP-Pik1p at Golgi dots were counted (n = 80–90). (D) GFP-Pik1p fluorescence decreases in bmh mutants. Total intensity of small dotted structures (<0.95 μm in diameter) per cell for wild-type and mutant cells (mean ± SEM). For illustration of the quantification procedure, see Supplemental Figure 3.

Figure 6.

Pik1p interaction with 14-3-3 and localization is regulated by phosphorylation. (A) Pik1p binding to 14-3-3 requires phosphorylation. Extracts from cells containing Pik1p-TAP (CSY513) were either treated with λ-PPase or with phosphatase inhibitors followed by incubation with GST alone, GST-Bmh1p, or GST-Bmh2p immobilized on glutathione-Sepharose. Immunoblot analysis was with PAP complex. The input represents 5% of lysate. (B) Liquid β-galactosidase assay and yeast two-hybrid interactions of Pik1p phospho-mutants (pik1-145 to 148, pMB365-368, respectively) with Bmh1p (pMB116) and Bmh2p (pMB125) were quantified. Interaction between Pik1p (pMB138) and Bmh1p or Bmh2p serves as positive control, and interactions with pGADT7 as negative control. Values were from nine independent experiments. (C) Pik1p phosphorylation upon nucleocytoplasmic relocation. Wild-type (NY1211), sec6-4 (NY778) and pik1-130 (YMB087) cells were grown at 25°C, incubated for 30 min at 25 or 37°C as indicated, and lysed in the presence or absence of phosphatase inhibitors. Lysates without phosphatase inhibitors were incubated with λ-PPase. Samples were analyzed by SDS-PAGE (acrylamide:bisacrylamide ratio, 500:1) to resolve phosphorylated and nonphosphorylated Pik1p and immunoblotting with anti-Pik1p antibody. (D) The pik1-130 mutant cells lack nuclear localization at the restrictive temperature. Cells carrying the pik1-130 mutation (YMB087) were processed for immunofluorescence microscopy by using anti-Pik1p antibody and DAPI after incubation for 60 min at 25 or 37°C as indicated. Arrowheads indicate nuclear exclusion of Pik1-130p.

Figure 7.

Pik1p relocalizes from the TGN to the cytoplasm and nucleus under growth-limiting conditions. (A) Pik1p is lost from the TGN in late log phase. Cells (NY1211) expressing GFP-PIK1 from the endogenous promoter on a CEN-based plasmid were observed by laser confocal microscopy. (B) Depletion of fermentable carbon sources results in shuttling of Pik1p from the TGN to the nucleus. Early log phase cultures (NY1211) were either depleted of glucose or glycerol for 45 min. (C) Relocalization of Pik1p from the TGN to the nucleus is reversed upon readdition of glucose. Cultures (NY1211) were depleted of glucose as described in B. After indicated times, glucose was added back to a final concentration of 2%. (D) The TGN marked by Sec7p-GFP is present in early log phase, late log phase, and glucose-depleted cells (YGY84). Cultures were treated as described in A or B. (E and F) Percentage of cells with nuclear GFP-Pik1p in A or B. (G and H) Percentage of cells with Golgi GFP-Pik1p in A or B. Values are mean ± SD (n = 100).

Figure 8.

Relocation of Pik1p from the TGN to the cytoplasm is mediated by Pik1p phosphorylation and 14-3-3 binding. Early log phase cultures (YMB148) were depleted of glucose for 45 min. If indicated, glucose was added back to a final concentration of 2%. Pik1p-13myc was immunoprecipitated using anti-myc antibody (9E10). Coprecipitated Bmh1p-3ha was detected using anti-HA antibody (Y11). Samples were analyzed by SDS-PAGE (acrylamide:bisacrylamide ratio, 500:1) to resolve phosphorylated and nonphosphorylated Pik1p and immunoblotting with anti-myc antibody (9E10).

Figure 9.

Co-overexpression of BMH1 and BMH2 prevents relocation of Pik1p to the nucleus upon glucose deprivation. (A) Early log phase cultures of cells (NY1211) expressing GFP-PIK1 from the endogenous promoter on a CEN-based plasmid, and BMH1 (pMB180), BMH2 (pMB263), or BMH1 and BMH2 on a 2μ-based plasmid were depleted of glucose for 45 min. Cells were observed by confocal microscopy. (B and C) Number of cells with nuclear GFP-Pik1p (A) and with GFP-Pik1p at the Golgi (C) were plotted (mean ± SD).

Figure 10.

Pik1p phosphorylation site mutations affect nucleocytoplasmic shuttling but not Golgi association. (A) Cells (YMB003 containing pMB384) expressing GFP-PIK1 (pMB422), GFP-PIK1(S396A) (pMB423), or GFP-PIK1(S396D) (pMB424) from the ADH promoter on a CEN-based plasmid were observed by laser confocal microscopy. Depletion of glucose was performed as in Figure 7. (B) Percentage of cells with nuclear GFP-Pik1p. (C) Percentage of cells with Golgi GFP-Pik1p. Values are mean ± SD (n = 100). Similar results were obtained in absence of pMB384. (D) PI 4-kinase activity of S396 phosphorylation site mutants GFP-Pik1p, GFP-Pik1p^S396A, and GFP-Pik1p^S396D. GFP-Pik1p was precipitated with polyclonal anti-GFP antibody and subjected to lipid-kinase assay by using phosphatidylinositol as substrate. Results are from two representative experiments (mean ± SEM). (E) Representative immunoblot (monoclonal anti-GFP antibody) of immunoprecipitated GFP-Pik1p, GFP-Pik1p^S396A, and GFP-Pik1p^S396D for experiment (D).

Figure 11.

Model of the role of 14-3-3 interaction in the regulation of cellular Pik1p pools. A cytoplasmic Pik1p–14-3-3 complex maintains a pool of Pik1p that can be recruited to the TGN, where Pik1p binds to the membrane through the action of Frq1p and produces PI(4)P. Cytoplasmic Pik1p is protected from dephosphorylation and nuclear import through the action of 14-3-3. Glucose deprivation releases Golgi Pik1p, increases Pik1p–14-3-3 complex formation, and, possibly due to an effective increase in Pik1p concentration, increases also the nuclear pool. This regulation might provide a very rapid mode of adjustment of PI(4)P levels and thus transport capacity of the Golgi to coordinate nutrient signaling with growth. The k-values represent hypothetical reaction constants (see Discussion and Supplemental Material).