Integrated control of formin-mediated actin assembly by a stationary inhibitor and a mobile activator

Figure 1.. Hof1-FL is a potent inhibitor of Bnr1-mediated actin nucleation. (A) Schematic of Hof1 constructs purified. CC2, coiled-coil 2 domain; F-BAR, FCH Bar-Amphiphysin-Rvs domain; MBP, maltose-binding protein; SH3, Src homology 3 domain. (B) Coomassie-stained SDS-PAGE gel of purified MBP-tagged Hof1 polypeptides (100 ng each). (C) Inhibitory effects of Hof1-FL on C-Bnr1–mediated actin assembly. Reactions contain 2 µM actin monomers (5% pyrene labeled), 2 nM C-Bnr1 (FH1-FH2-C), and indicated concentrations of Hof1-FL (0–100 nM). (D) Concentration-dependent effects of Hof1-FL or Hof1-NT on 2 nM C-Bnr1 or Bnr1 (FH2), as indicated; data from assays as in C. (E) TIRF microscopy analysis showing the effects of different Hof1 polypeptides on C-Bnr1. Reactions contained 0.5 µM actin monomers (10% OG labeled; 0.5% biotin-actin), 3 µM yeast profilin, 0.2 nM C-Bnr1, and concentrations of Hof1 polypeptides indicated in F. Images are from 300 s after initiation of actin assembly. Bar, 50 µm. (F) Quantification of actin filament nucleation from TIRF assays as in E. Data are averaged from four FOVs in each of two independent experiments (eight FOVs total). Error bars represent SEM. Statistical…

Figure 2.. EM structures of Hof1–Bnr1 complexes. (A) Representative 2D class averages of C-Bnr1 alone and Hof1-FL alone. Bars, 10 nm. (B) Most abundant 2D class averages of complexes formed by C-Bnr1 + Hof1-FL (percentage of total particles that produced 2D average displayed). Bar, 10 nm. (C) 3D reconstruction of Hof1-FL bound to C-Bnr1 at 10 Å resolution. Crystal structures of Hof1 F-BAR domain and Bni1 FH2 domains docked into the EM density using UCSF Chimera. Note that we used the crystal structure of the Bni1 FH2 domain, because no structure is available for Bnr1 FH2. (D) 3D reconstruction of Hof1 F-BAR bound to Bnr1 (FH2) at 24-Å resolution, with crystal structures docked as above. (E) Close-up views of the Hof1-FL + C-Bnr1 structure, showing approximate positions of contact between the Hof1 F-BAR domain (green) and each of the two FH2 domains it binds, color-coded in gold and purple as in C. The tips of the F-BAR domains sit close to key actin-binding residues in the FH2 domain (K1601 and I1431).

Figure 3.. Deletion of the Hof1 F-BAR domain in vivo gives rise to actin cable defects. (A) SIM of F-actin organization in fixed wild-type, hof1Δ, and hof1ΔFBAR cells. (B) Automated analysis of actin cable organization from SIM images as in A with SOAX, showing cable segments (purple) and cable intersections (green). (C) Polarity map of actin cable segments in representative cells from the SIM images, with color-coding according to azimuthal orientation of cable segments relative to mother-bud axis. Azimuthal angle is defined such that the angle is zero when a cable segment is parallel with the x axis with a range of ±90 degrees. Cells analyzed marked with a yellow arrowhead in B. (D) Average number of cable segments per cell analyzed by SOAX, quantified for 20 cells per strain. (E) The same cells were quantified for number of cable intersections per cell by SOAX. (F) CVs of F-actin staining in mother cells after treatment with 100 µM CK666. Data were averaged for 20 cells in each strain. (G) Representative traces (18 for each strain) of GFP-Sec4 paths, tracking vesicle movement from its starting position in the mother cell until it reached the bud neck. Convergence point of traces…

Figure 4.. Bud6 overcomes Hof1 inhibition of Bnr1-mediated actin assembly in vitro. (A) Schematic of Bud6 domains and C-Bud6 polypeptides (wild type and mutant) used for in vitro assays. (B) Representative images from TIRF microscopy assays showing the effects of wild-type C-Bud6 and mutant C-Bud6-35 and C-Bud6-8 polypeptides on actin assembly in the presence of C-Bnr1 and Hof1-FL. All reactions contain 0.5 µM monomeric actin (10% OG labeled; 0.5% biotin-actin) and 3 µM yeast profilin and as indicated 0.2 nM C-Bnr1 (FH1-FH2-C), 20 nM Hof1-FL, and/or 100 nM C-Bud6 polypeptides. Images show representative FOVs 600 s after initiation of actin assembly. (C) Quantification of actin assembly rates from TIRF assays as in B. Rates of OG-labeled F-actin accumulation averaged from four FOVs in each of two independent experiments (eight FOVs total). Error bars represent SEM. Statistical significance was calculated by one-way ANOVA (n.s., not significant; ****, P ≤ 0.0001). (D) Representative TIRF images from binding assays. 500 nM soluble MBP-Hof1-NT-SNAP-549 incubated with beads coated with immobilized 649-biotin-SNAP-C-Bnr1 in the presence and absence of 200 nM unlabeled C-Bud6. Bar, 5 µm. (E)…

Figure 5.. Deletion of BUD6 suppresses hof1Δ defects. (A) Yeast strains (fivefold serial dilutions) grown at 25°C or 34°C on YEPD plates for 2 d. (B) Growth rates of the same strains measured by optical density (OD600) at 34°C in liquid culture (YEPD) in a shaking microplate absorbance reader. Data were averaged from eight independent trials. Lighter shading represents SEM. (C) F-actin organization in the same strains imaged by SIM before and after treating cells with the Arp2/3 complex inhibitor CK666 to remove actin patches. (D) Number of actin cable segments per cell determined by SOAX analysis, averaged from 10 cells per strain. (E) Number of cable intersections determined by SOAX analysis for same cells as above. (F) CV analysis on cable fluorescence after treatment with CK666, analyzed for 20 mother cells per strain. (G) Tortuosity of GFP-Sec4 secretory vesicle paths (n = 50 vesicles per strain). Error bars represent SEM. Statistical significance in all panels calculated by one-way ANOVA (n.s., not significant; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001).

Figure 6.. Bud6 localizes to secretory vesicles and is transported on actin cables to the bud neck. (A) Colocalization of Bud6-mCherry and GFP-Sec4 in live yeast cells. Bar, 5 µm. Merge highlights two puncta used in line scan analysis below. (B) Pearson’s correlation analysis of Bud6-mCherry signal with GFP in strains containing GFP tags on different membrane compartments (n = 50 cells per strain). Error bars represent SEM. Statistical significance calculated by one-way ANOVA (****, P ≤ 0.0001). (C) Line-scan analysis shows overlap between Bud6-mCherry and GFP-Sec4 puncta, corresponding to puncta highlighted in A. (D) Time-lapse dual-color live imaging of Bud6-mCherry and GFP-Sec4. Arrow highlights the movement of a vesicle. (E) Kymograph showing Bud6-mCherry and GFP-Sec4 comigration. (F) Total Bud6-mCherry fluorescence level in cells of variable bud size from an asynchronous population (n = 100 cells). (G) Representative time-lapse images of Bnr1-GFP and Hof1-GFP strains after α-factor arrest/release. Bar, 3 µm. (H) Quantification of Bnr1-GFP and Hof1-GFP levels at the bud neck for each strain, from time lapse imaging in G (n = 5 cells per strain).

Figure 7.. Working model for Hof1-Bud6 positive feedback loop controlling Bnr1-mediated actin cable nucleation. (A) Model showing myosin-mediated transport of secretory vesicles carrying Bud6 on actin cables to the bud neck, where Bnr1 and Hof1 reside. Close up on the right shows proposed interplay among Hof1, Bnr1, and Bud6 at the bud neck. (B) Representative SIM images of actin cable staining in wild-type and myo2-66 cells 5 min after shift to 35°C. (C) Quantification of actin cable fluorescence levels in mother cells, from experiments as in B, averaged from 10 cells per strain. Error bars represent SEM. Statistical significance was calculated by unpaired t test (*, P ≤ 0.05).