A pik1-101 gga2Δ double mutant exhibits synthetic genetic and exocytosis defects. (A) pik1-101 mutant cells were crossed to gga2Δ cells, and tetrads were isolated at 25°C. Growth of resulting haploids was monitored after 3 d at 25°C. Squares indicate pik1-101 gga2Δ double mutants, showing a synthetic growth defect. (B) The expression of invertase was derepressed by shifting cells for 1 h into low glucose (0.1%). Left panel, wild-type (YAB200897; invertase secretion set to 100%), gga1Δ (YAB531; 95.8 ± 1.2%), gga2Δ (YAB532; 103.4 ± 1.2%), gga1Δ gga2Δ (YAB538; 97.2 ± 1.6%), and sec6-4 (NY778; 4.6 ± 0.8%); right panel, wild-type (CSY209; invertase secretion set to 100%), gga2Δ (CSY567; 99.9 ± 1.9%), pik1-101(CSY544; 76.7 ± 2.5%), pik1-101gga2Δ (CSY545; 64.4 ± 2.2%), chs6Δ (CSY566; 98.2 ± 1.7%), pik1-101chs6Δ (CSY561; 72.8 ± 1.4%), and sec6-4 (NY778; 5.3% ± 0.5). The exocyst mutant sec6-4 was used as positive control. The experiment was performed at the restrictive temperature of 37°C. Values indicate percentages of secreted invertase (n = 6–15, mean ± SEM). The difference in invertase secretion between the pik1-101 and pik1-101gga2Δ mutants is highly significant (p = 0.003). There is no significant difference (p = 0.18) between pik1-101 and pik1-101chs6Δ. Note that mutants in the same strain background are shown within the same panel with their respective wild type.

Pik1p regulates the Vps pathway at the exit from the Golgi. (A) Vps10p transport to endosomes is impaired in pik1-101. Wild-type (CSY392), pik1-101 (CSY391), pik1-101 gga2Δ (CSY398), vps28Δ (CSY900), and pik1-101 vps28Δ (CSY399) containing the CPY receptor Vps10p genomically tagged with 3xHA were grown overnight in YPD, and the next day were shifted for 3–4 h to minimal medium lacking methionine. Proteins were pulse-labeled with ³⁵[S]methionine for 10 min and then chased for 0, 30, and 60 min. Proteins were precipitated with TCA and washed with acetone, and Vps10p-3xHA was immunoprecipitated using an α-HA antibody. Samples were subjected to SDS-PAGE and autoradiography. Vps10–3xHA* indicates the lower protease-resistant form. (B) GGA overexpression suppresses growth defect of pik1-101 at 34°C. Wild-type (NY1211) or pik1-101 (CSY712) cells containing p415 GALs and p416 GALs, control vectors, p415 carrying GALs-GGA1 or p416 carrying GALs-GGA2 as indicated were grown on selective media with 2% raffinose to midlog phase and plated onto selective media with 2% galactose at 25 or 34°C, respectively, for 3 d.

Modeling of a PI-binding motif in the N-terminal VHS domain of Gga2p. The VHS domains of the GGA proteins from different species were analyzed in terms of their potential binding to PI(4)P. (A) Multiple sequence alignment of the GGA VHS domains from fungi, arthropods, and vertebrates with the VHS domain of Target of Myb1 (Tom1) and the ENTH and ANTH domains from Epsin and CALM, respectively. Conserved residues are highlighted in yellow, and secondary structural elements are indicated on top of the one family member containing structural information. Residues that are described to contact Ins(1,4,5)P₃ in Epsin and PI(4,5)P₂ in CALM are indicated by dark green triangles. The Ins(1,4,5)P₃-binding site in Epsin resides mostly within the very N-terminus, which only adopts a helical fold when bound to the ligand (not shown; for details see Ford et al. 2002). Residues thought to be involved in membrane binding of Tom1 are indicated by light green triangles. Residues that are in contact with BACE phosphopeptide in human GGA1 are highlighted by blue circles. The two potential PI(4)P-binding sites in fungal GGAs, which show a strong basic charge on the surface area, are located in helix α1 or helix α8 and are indicated by orange or yellow triangles, respectively. For a list of GenBank accession numbers, see Table S3. (B) Structural display of the Epsin ENTH domain and the CALM ANTH domain complexed with PI(1,4,5)P₃ and PI(4,5)P₂, respectively. The surface of the protein domains is shown in gray with basic amino acid, including histidine, highlighted in light and dark blue, respectively. The dot surface of the lipid is shown in red. (C) Surface representation of the modeled structure of the VHS domain of yeast Gga2p using the Phyre-server for fold recognition. The two potential sites of interaction of Gga2p^VHS with PI(4)P are circled in orange or yellow, respectively. (D) Alignment of the putative phosphoinositide interaction sites of yeast GGAs with vertebrate ANTH and the clathrin adaptors AP1γ and AP2α. The positively charged and aromatic residues contacting the phosphoinositide head group in AP1γ and AP2α are highlighted in red. Note: This site overlaps with the predicted PI-interacting region preceding helix α8 (orange triangles).

PI(4)P and Arf1p cooperate in the recruitment of the VHS-GAT domain to the TGN. (A) Liposome pulldown assay was performed as in Figure 5D using GST-Gga2p^VHS-GAT (1.5 μg). (B) Wild-type (CSY349) and pik1-101 cells (CSY906) were transformed with a plasmid encoding for GFP-Gga2p^VHS-GAT (pLD216). Cells were grown in selective media to midlog phase and subsequently analyzed by fluorescence microscopy. In wild-type cells, the GFP-Gga2p^VHS-GAT domain was able to localize to the Golgi, whereas in pik1-101 cells GFP-Gga2p^VHS-GAT was mostly cytosolic. (C) Liposome pulldown assay was performed as in Figure 5D using GST-Gga2p and GST-Gga2p^KKHR-EEAE (pMG10; 1.5 μg). (D) Localization of GFP-Gga2 and GFP-Gga2p^KKHR-EEAE. Wild-type cells (NY1211) expressing GFP-Gga2p (pCS136) or GFP-Gga2p^KKHR-EEAE (pMG9) were grown in selective media to midlog phase and subsequently analyzed by fluorescence microscopy. (E) Mutation of the Gga2p Arf-binding domain and VHS domain leads to decreased membrane association of GFP-Gga2p. Wild-type cells (NY604) were transformed with plasmids encoding GFP-Gga2p (pCS136), GFP-Gga2p^I207N (pMB430), GFP-Gga2p^KKHR-EEAE (pMB432), or GFP-Gga2p^{KKHR-EEAE,I207N} (pMB433). GFP-GGA2 constructs express Gga2p under the control of the CPY promoter. Subcellular fractionations, and quantifications of Western blots were performed as described in Figure 5B. Shown is the result of one representative experiment.

Results of genome-wide synthetic lethality screen of the pik1-101 mutant allele. Genes that have been found in SGA analysis to interact genetically with pik1-101 (synthetic lethal or sick) are represented as nodes. All genetic interactions shown in this scheme have been confirmed by tetrad analysis. Several genetic interactors of CLA4 were found to exhibit synthetic interactions (URM1, ELP2, NCS2, YNL120C, and UBA4). The synthetic genetic interaction with UBA4 was not identified in the genome-wide screen but was found by tetrad analysis in the further course of this study. For a full list of genes isolated in the screen, see Table S1.

Ultrastructural analysis of pik1-101 and gga2Δ mutant cells. The pik1-101 and gga2Δ mutant cells were grown at permissive temperature and processed for electron microscopy. (A–E) pik1-101. (F) gga2Δ. Immunolabeling was with α-Tlg1 antibody. Secondary antibody was coupled to 12-nm colloidal gold. Bar, (A and F) 1 μm; (D and E) 500 nm; (B and C) 250 nm; and (A, C, and F, insets) 100 nm. Note the appearance of ring-like (arrows) and tubular (arrowhead) structures. For quantitation of the results, see Table. S2.

Binding of Gga2p to PIs and Arf1p is required for proper localization to the Golgi. (A) Wild-type (CSY349) and vps34Δ cells (CSY911) expressing an additional copy of Sec7p-DsRed under the control of a strong promoter as well as pik1-101 cells (CSY906) containing genomically tagged Sec7p-DsRed were transformed with a plasmid encoding for GFP-Gga2p (pCS136). Cells were grown in selective media to midlog phase, incubated at 25 or 37°C for 1 h, and subsequently analyzed by fluorescence microscopy. In wild-type and vps34Δ cells, Gga2p colocalizes with Sec7p-DsRed, whereas in pik1-101 cells GFP-Gga2p is mostly cytosolic. (B) Subcellular distribution of endogenous Gga2p was evaluated by subcellular fractionation from wild-type (CSY210), pik1-101 (CSY544), and arf1Δ (CSY704) cells. Wild-type, arf1Δ, and pik1-101 mutant cells were grown to midlog phase, homogenized, and then soluble and membrane fractions were separated by 100,000 × g centrifugation of a postnuclear supernatant. Temperature shift to 37°C was performed for 1 h before fractionation. Equal volumes of fractions were loaded and analyzed by SDS-PAGE and immunoblotting. The Image Quant Software was applied for quantification (n = 3, mean ± SD). Less endogenous Gga2p is membrane-bound in pik1-101 cells at both permissive and restrictive temperatures compared with wild type. (C) Equal amounts of subcellular fractions prepared from wild-type (CSY210) and pik1-101 (CSY544) cells as in B were analyzed by SDS-PAGE and Western blots using α-ADH and α-Arf antibodies. (D) Overexpression of gga mutants unable to bind Arf1p does not rescue the growth defect of pik1-101 at 34°C. Wild-type (NY1211) or pik1-101 (CSY712) cells containing p415 GALs and p416 GALs control vectors, or expressing Gga1p^L203Q (pLD232) or Gga2p^I207N (pLD231) as indicated were grown on selective media with 2% raffinose to midlog phase and plated onto selective media with 2% galactose at 25 or 34°C, respectively, for 3 d. (E) Liposomes containing either 1% PI(3)P, PI(4)P or PtdIns were incubated with purified GST-Gga2p (1.5 μg) or GST (0.5 μg) and purified myristoylated ARF1 in the absence or presence of GTPγS. The liposome-bound fraction of GST-Gga2p was detected using a α-GST antibody. The recruitment of GST-Gga2p to PI(4)P containing liposomes is increased in the presence of active ARF1, indicating a synergistic function of PI(4)P and Arf1p in the recruitment of Gga2p to membranes.

Binding of Gga2p^VHS (A), Gga2p^VHS-GAT (C) and Gga2p^{VHS(KKHR-EEAE)} (E) to PI(4)P liposomes determined by surface plasmon resonance. Gga2p^VHS and Gga2p^VHS-GAT were injected at concentrations from 250 nM to 4 μM (bottom to top curves) over immobilized PI(4)P liposomes and over control PC liposomes. The values of control surface were subtracted from the signal. The change in SPR signal during association and dissociation is shown in colored curves. Black bars are report points set on the sensograms in the steady-state region of the curve. (B and D) Plot of steady-state binding levels (Req) against concentrations of Gga2p^VHS or Gga2p^VHS-GAT and fit to steady-state affinity model. Gga2p^{VHS(KKHR-EEAE)} was used at concentrations from 250 nM to 2 μM and revealed no binding to PI(4)P liposomes.