Bil2 Is a Novel Inhibitor of the Yeast Formin Bnr1 Required for Proper Actin Cable Organization and Polarized Secretion

Abstract

Cell growth in budding yeast depends on rapid and on-going assembly and turnover of polarized actin cables, which direct intracellular transport of post-Golgi vesicles to the bud tip. Saccharomyces cerevisiae actin cables are polymerized by two formins, Bni1 and Bnr1. Bni1 assembles cables in the bud, while Bnr1 is anchored to the bud neck and assembles cables that specifically extend filling the mother cell. Here, we report a formin regulatory role for YGL015c, a previously uncharacterized open reading frame, which we have named Bud6 Interacting Ligand 2 (BIL2). bil2Δ cells display defects in actin cable architecture and partially-impaired secretory vesicle transport. Bil2 inhibits Bnr1-mediated actin filament nucleation in vitro, yet has no effect on the rate of Bnr1-mediated filament elongation. This activity profile for Bil2 resembles that of another yeast formin regulator, the F-BAR protein Hof1, and we find that bil2Δ with hof1Δ are synthetic lethal. Unlike Hof1, which localizes exclusively to the bud neck, GFP-Bil2 localizes to the cytosol, secretory vesicles, and sites of polarized cell growth. Further, we provide evidence that Hof1 and Bil2 inhibitory effects on Bnr1 are overcome by distinct mechanisms. Together, our results suggest that Bil2 and Hof1 perform distinct yet genetically complementary roles in inhibiting the actin nucleation activity of Bnr1 to control actin cable assembly and polarized secretion.

Keywords: Bil2; Bni1; Bnr1; Bud6; actin; cable; formin; secretion.

FIGURE 1

BIL2 is required for efficient polarized delivery of secretory vesicles. (A) Representative time-lapse images showing the transport path of a secretory vesicle (GFP^Envy-Sec4) moving from the mother to the bud. At each time point shown, the vesicle being tracked is highlighted by a green circle. To the right is a sum of vesicle positions over time, with a line (red) marking the transport path. The sum of the transport paths are isolated and expanded on the right. (B) Bouquets of representative transport paths for secretory vesicles (15 each) in wildtype and bil2Δ cells. Vesicle traces are organized such that they start at the periphery of the bouquets and terminate at the central dot (corresponding to the bud neck). (C) Quantification of GFP^Envy-Sec4 path lengths from traces as in (B) (n = 25 vesicles per experiment, 2 independent experiments. Number of cells: experiment 1: 17 for each strain; experiment 2: 18 for each strain). (D) Tortuosity of transport paths (ratio of path length to distance traveled) for the same vesicles in (C). (E) Fraction of vesicles successfully transported from the mother compartment to the bud during a 30 s observation window (n = 20 cells per condition per experiment, 2 independent experiments). Each dot represents the fraction of vesicles successfully transported to the bud in one cell. In all panels, bars show mean and SD. Statistical significance calculated by 2-way student T-test in all panels (n.s., no significance, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001).

FIGURE 2

Loss of BIL2 disrupts the spatial organization of Bnr1-polymerized actin cable networks. (A) Representative structured illumination microscopy (SIM) images of F-actin organization in CK666 treated, phalloidin stained wildtype and bil2Δ cells at different stages of bud growth. (B) Automated traces of actin cables from SIM images as in (A), created using SOAX. Left, phalloidin stained cell. Right, purple cable segments generated by SOAX. (C) Average number of actin cable segments per cell analyzed by SOAX (n = 20 cells per condition from two independent experiments). (D) Coefficient of variation (CoV) of phalloidin staining within the mother compartment of cells treated with CK666 (n = 20 cells per condition from two independent experiments). (E) Representative images of endogenously-expressed Bnr1-GFP in wildtype (WT) and bil2Δ cells, with quantification of signals at the bud neck (mean and SD) below each image. In all panels, statistical significance calculated by 2-way student T-test (n.s., no significance, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001).

FIGURE 3

Purified Bil2 inhibits Bnr1-mediated actin nucleation but not elongation. (A) Bulk pyrene-actin assembly assays showing that Bil2 inhibits Bnr1-dependent actin nucleation, both in the presence and absence of Bud6. Bil2 with Bud6 and Bil1 present did not inhibit Bnr1. Reactions contain 2 μM actin monomers (5% pyrene labeled) and 5 μM profilin, with 2 nM C-Bnr1 (FH1-FH2-C; 758–1,375), 100 nM Bud6(L) (489–788), 100 nM Bil1, and/or 100 nM Bil2, as indicated. (B) Representative images from TIRF microscopy experiments showing the effects of Bil2 on Bnr1-mediated actin assembly. Reactions contain 1 μM actin monomers (10% Oregon green labeled) and 3 μM profilin, with 0.1 nM C-Bnr1 and/or 100 nM Bil2, as indicated. Images shown are from 200 s after the initiation of actin assembly. (C) Quantification of the number of actin filaments nucleated per field of view (FOV) at 200 s into TIRF reactions as in (B) (four FOVs per condition). Shown are the mean and SEMs. (D) Quantification of filament elongation rates for TIRF reactions as in (B), except that 100 nM Bil2 was flowed into reactions 5 min after initiation of actin assembly (n = 20 filaments per condition). (E) Bil2 inhibits Bnr1 (FH2)-mediated actin filament assembly. Quantification of number of filaments nucleated per field of view (FOV) at 200 s into TIRF reactions as in B (four FOVs per condition). Reactions contain 1 μM actin monomers (10% Oregon green labeled) and 0.1 nM Bnr1 (FH2), with and without 100 nM Bil2. (F) TIRF fields showing that Bil2 produces latrunculin-resistant actin puncta. Reactions contain 1 μM actin monomers (10% Oregon green labeled), 3 μM profilin, and 100 nM Bil2, with or without 100 nM Latrunculin B. Images shown are from 200 s after the initiation of actin assembly. Shown are the mean and SEMs. Statistical significance calculated by 2-way student T-test in all panels (n.s., no significance, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001).

FIGURE 4

Synthetic genetic interactions between BIL2 and HOF1. (A) bil2Δ haploids were crossed to haploids carrying deletions in other Bnr1 regulatory genes, including HOF1, BUD6, and BUD14. Diploids were sporulated, and tetrads dissected and genotyped (n = 144, 64, and 104 tetrads from crosses with hof1Δ, bud6Δ, and bud14Δ, respectively). Resulting wildtype, single mutant, and double mutant haploids were analyzed for viability at 25°C. (B) The indicated haploid strains were compared for growth in synthetic complete media at 25°C in a shaking microplate reader for 50 h, monitoring growth (OD₆₀₀) every 5 min. Lines represent the average of three independent cultures per strain. (C) Cell size was determined from DIC images in ImageJ by outlining each cell and calculating its area (n = 20 cells per condition). Shown are the mean and SD. (D) Representative max projection Z-stacks of phalloidin stained cells imaged by structured illumination microscopy (SIM). Note cell size is to scale, i.e., hof1Δ bil2Δ cells are enlarged compared to wildtype, hof1Δ, and bil2Δ cells, as indicated in (C).

FIGURE 5

GFP-Bil2 localization to polarity sites and association with secretory vesicles. (A) Representative images of live cells expressing GFP-Bil2 (from a low copy plasmid under the control of the ACT1 promoter) and either integrated Bud6-mCherry or mCherry-Sec4 (expressed from a low copy plasmid under the control of its own promoter). (B) GFP-Bil2 colocalization in live cells with mCherry-Sec4 (secretory vesicle marker) or Sec7-mCherry (trans-Golgi marker) quantified by Pearson correlation. (C) Comparison of GFP-Sec4 vesicle transport paths (ratio of path length to distance traveled) in wildtype (WT) and bil2Δ cells with or without the pACT1-GFP-Bil2 plasmid (n = 25 vesicles per condition). (D) Representative fields of view of secretory vesicles isolated from cells expressing GFP-Bil2 (plasmid, as in A) along with Bud6-mCherry (integrated) or mCherry-Sec4 (plasmid, as in A). (E) Quantification of GFP-Bil2 colocalization with Bud6- and Sec4-positive secretory vesicles. Statistical significance in all panels calculated by 2-way student T-test (n.s., no significance, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001).

FIGURE 6

Working model for the regulation of Bnr1-mediated actin cable nucleation. Bnr1 is anchored at the bud neck and assembles cables in the mother cell. Bud6 functions as an NPF, promoting actin nucleation by Bnr1. Bil2 and Hof1 specifically inhibit actin nucleation by Bnr1. Bud6 is delivered on secretory vesicles to the bud neck and overcomes Hof1 inhibition of Bnr1 as part of a positive feedback loop promoting cable assembly (Garabedian et al., 2018). Bud6 is not sufficient to overcome Bil2 inhibition of Bnr1. However, bud6 together with its binding partner Bil1 overcomes Bil2 inhibition to promote Bnr1-mediated actin nucleation. Bil2 and Bud6 are found together on many secretory vesicles, suggesting that Bil2 may help keep Bud6 inactive until it reaches the bud neck where Bil1 is found (Graziano et al., 2013). Bnr1 is also predicted to be autoinhibited via interactions of its N-terminal diaphanous inhibitory domain (DID) with its C-terminal diaphanous autoregulatory domain (DAD) (Li et al., 2003). However, it is not yet clear what mechanisms trigger the release of Bnr1 from autoinhibition. After a filament is nucleated, Bnr1 remains processively attached to the growing barbed end, where the duration and rate of filament elongation are controlled by other cellular factors, including Smy1 and Bud14 (Eskin et al., 2016). Model created using BioRender.com.