Generation and characterization of conditional yeast mutants affecting each of the 2 essential functions of the scaffolding proteins Boi1/2 and Bem1

Abstract

Boi1 and Boi2 are closely related yeast scaffolding proteins, either of which can perform an essential function. Previous studies have suggested a role in cell polarity, interacting with lipids, components of the late secretory pathway, and actin nucleators. We report detailed studies of their localization, dynamics, and the generation and characterization of conditional mutants. Boi1/2 are present on the plasma membrane in dynamic patches, then at the bud neck during cytokinesis. These distributions are unaffected by perturbation of the actin cytoskeleton or the secretory pathway. We identify 2 critical aromatic residues, present in both Boi1 and Boi2, in the essential C-terminal Pleckstrin-Homology domain, that cause temperature-sensitive growth resulting in defects in polarized growth leading to cell lysis. The scaffolding protein, Bem1, colocalizes with Boi1 in patches at the growing bud, and at the bud neck, the latter requiring the N-terminal SH3 domain of Boi1p. Loss of function of Boi1-SH3 domain renders Bem1 essential, which can be fully replaced by a fusion of the SH3b and PB1 domains of Bem1. Thus, the 2 essential functions of the Boi1/2/Bem1 proteins can be satisfied by Bem1-SH3b-PB1 and Boi1-Pleckstrin-Homology. Generation and characterization of conditional mutations in the essential function of Bem1 reveal a slow onset of defects in polarized growth, which is difficult to define a specific initial defect. This study provides more details into the functions of Boi1/2 and their relationship with Bem1 and presents the generation of conditional mutants that will be useful for future genetic analysis.

Introduction

The mechanism by which cells polarize is a fundamental question in cell biology that has been especially well explored in the budding yeast (Chiou et al. 2017). For example, to generate a bud, the cell has to select an appropriate site, build a polarized actin cytoskeleton, and deliver post-Golgi vesicles to sites of growth. In order to coordinate the process, there are many essential components involved.

The first step in generating a new bud is to select a site to which cell growth will occur. This involves the local activation of the small RhoGTPase, Cdc42 that is activated by its GEF, Cdc24 (Ziman and Johnson 1994; Zhao et al. 1995; Pruyne and Bretscher 2000a, 2000b; Bi and Park 2012). Active Cdc42 has many roles, initially directing the nucleation and assembly of actin cables through the formin protein Bni as well as participating in the assembly of septin proteins at the bud neck (Brown et al. 1997; Chen et al. 1997; Evangelista et al. 1997; Adamo et al. 2001; Zhang et al. 2008). The actin cables are then used as the tracks for the myosin-V Myo2 to deliver post-Golgi secretory vesicles, marked by the Rab Sec4, to sites of exocytosis (Salminen and Novick 1987; Kabcenell et al. 1990; Pruyne et al. 1998). During transport, the secretory vesicles acquire the exocyst, a multiprotein complex that facilitates the tethering of secretory vesicles to the plasma membrane and is essential for fusion with the plasma membrane in a process also involving Cdc42 (TerBush et al. 1996; Wu et al. 2010; Gingras et al. 2022).

Considerable interest has focused on the initial single-site activation of Cdc42. Extensive studies have shown that this involves a local positive feed-back loop coordinated by the scaffolding protein Bem1 (Leeuw et al. 1995; Park et al. 1997; Bose et al. 2001; Yamaguchi et al. 2007; Takaku et al. 2010; Liao et al. 2013; Grinhagens et al. 2020). Bem1 is a modular protein containing 2 SH3 domains (SH3a and SH3b), a PX domain, and a C-terminal PB1 domain. A 2-hybrid screen for interactors of Bem1’s second SH3 domain yielded a protein called Boi1 (Bem One Interactor 1), and its homolog, Boi2. Loss of either Boi1 or Boi2 has no effect on cell growth, whereas loss of both is lethal or highly deleterious (Bender et al. 1996; Matsui et al. 1996). Boi1 and Boi2 are large modular scaffolding proteins consisting of an SH3 domain, an SAM domain, followed by a proline-rich region that binds the Bem1SH3b domain and terminating in a Pleckstrin-Homology (PH) domain. Remarkably, the essential function of the Boi1/2 proteins can be supplied by just the Boi1 or Boi2 PH domain that is reported to bind Cdc42 (Bender et al. 1996; Matsui et al. 1996; Kustermann et al. 2017). Studies have shown that when the Boi proteins are depleted from the cell, secretory vesicles accumulate within the bud, and that the N-terminal Boi1 domain binds several components of the exocyst, and the C-terminal PH domain binds Sec1 (Kustermann et al. 2017; Masgrau et al. 2017). It was also found that the cells lacking Boi1/2 were capable of survival with a gain-of-function mutant of the exocyst complex, EXO70* (Masgrau et al. 2017). The collective results suggest that Boi1/2 proteins coordinate cell polarization with secretory vesicle exocytosis. Additionally, Boi1/2 have been reported to bind through their N-terminal region directly to the formin Bni1 and accessory protein Bud6 that are involved in the nucleation of actin filaments (Glomb et al. 2020).

Early work has also shown that an inactivating mutation in the SH3 domain of Boi1 in boi2Δ cells (boi1W53K boi2Δ) renders BEM1 essential (Bender et al. 1996). Thus, the Boi1/2 proteins have 1 essential function requiring the PH domain, and a second essential function overlapping with Bem1. In our previous work, we have made extensive use of fast-acting conditional mutants to TPM1, MYO2, and BNI1 that have been very useful in pinpointing their essential functions (Evangelista et al. 1997; Pruyne et al. 1998; Schott et al. 1999; Pruyne and Bretscher 2000b;Chernyakov et al. 2013). In this study, we set out to examine the distribution of Boi1/2 and Bem1 in more detail and generate conditional mutations in each of the 2 essential functions as an unbiased approach to explore their precise roles.

Materials and methods

Development of yeast strains and growth in different media types

Yeast strains used in this study are listed in Table 1. All strains used were haploid BY4741/2 unless otherwise noted. Yeast were grown in either selectable synthetic complete liquid media or on plates. Plates supplemented with 5-fluoroorotic acid (5FOA) were used to shuffle out yeast containing a plasmid with URA (Boeke et al. 1984). Sporulation of yeast was conducted by growing strains at 26°C in sporulation media (1% yeast extract, 1% potassium acetate, and 0.05% glucose). Tetrad dissections were performed using a Zeiss Standard 14 dissection scope.

Table 1.

List of yeast strains.

Strain genotype	Name	Reference
Mat-a BY4741 Wildtype	ABY1655	C. Boone
Mat-α BY4742 Wildtype	ABY1656	C. Boone
Mat-a boi1Δ::URA3	ABY4643	This study
Mat-α boi1Δ::URA3	ABY4644	This study
Mat-a boi2Δ::HISMX	ABY4645	This study
Mat-α boi2Δ::HISMX	ABY4646	This study
Mat-a Boi1-mNG::URA3	ABY4647	This study
Mat-a Boi2-mNG::HISMX	ABY4642	This study
Mat-a Boi1-mNG::URA3 Myo1-mScarlet::NATMX	ABY5021	This study
Mat-a Boi1-mNG::URA3 Boi2-mScarlet::NATMX	ABY5031	This study
Mat-a tpm1-2::LEU2 tpm2Δ::HISMX	ABY944	D. Pruyne
Mat-α tpm1-2::LEU2 tpm2Δ::KANMX Boi2-mNG::HISMX	ABY5029	This study
Mat-a Boi2-mNG::HISMX mScarlet-Sec4::NATMX	ABY5028	This study
Mat-a boi1Δ::URA3 Boi2-mNG::HISMX mScarlet-Sec4::NATMX	ABY5040	This study
Mat-a sec4-8::HISMX	ABY9429	P. Novick
Mat-a sec4-8::HISMX Boi1-mNG::URA3	ABY5034	This study
Mat-a sec4-8::HISMX Boi2-mNG::HISMX	ABY5036	This study
Mat-a sec6-4::HISMX	ABY9035	P. Novick
Mat-a sec6-4::HISMX Boi1-mNG::URA3	ABY5035	This study
Mat-a sec6-4::HISMX Boi2-mNG::HISMX	ABY5037	This study
Mat-α Boi1PH::LEU2 boi2Δ::HISMX	ABY4770	This study
Mat-α Boi1PH-mNG::NATMX boi2Δ::HISMX	ABY4772	This study
Mat-a Boi1SH3-mNG::LEU2	ABY4774	This study
Mat-a Boi1SH3-SAM-mNG::LEU2	ABY4775	This study
Mat-a Boi1SH3-SAM-PxxP-mNG::LEU2	ABY4776	This study
Mat-a boi1Δ::HISMX boi2Δ::KANMX pRS315-boi1-1	ABY4734	This study
Mat-α boi1Δ::HISMX boi2Δ::KANMX pRS315-boi1-2	ABY4735	This study
Mat-α boi1Δ::HISMX boi2Δ::KANMX pRS315-boi1-3	ABY4702	This study
Mat-a boi1Δ::HISMX boi2Δ::KANMX pRS315-BOI1	ABY4737	This study
Mat-a boi1Δ::HISMX boi2Δ::KANMX pRS315-Boi1PH	ABY4738	This study
Mat-a boi1Δ::KANMX boi2-1::LEU2	ABY3208	W. Liu
Mat-a boi1Δ::KANMX boi2-4::LEU2	ABY3211	W. Liu
Mat-a boi1Δ::KANMX boi2-5::LEU2	ABY3212	W. Liu
Mat-a boi1Δ::KANMX boi2-6::LEU2	ABY3213	W. Liu
Mat-a boi1Δ::KANMX boi2-7::LEU2	ABY3214	W. Liu
Mat-a boi1Δ::HISMX boi2Δ::KANMX pRS315-boi1W794R	ABY4766	This study
Mat-a boi1Δ::HISMX boi2Δ::KANMX pRS315-boi1F799S	ABY4764	This study
Mat-a boi1Δ::HISMX boi2Δ::KANMX pRS315-boi1PHW794R	ABY4905	This study
Mat-a boi1Δ::HISMX boi2Δ::KANMX pRS315-boi1PHF799S	ABY4909	This study
pck1Δ::LEU2 yEP50-pck1-2 SEY6210	AAY603	S. Emr
Mat-a boi1W53K::LEU2 boi2Δ::KANMX	ABY4110	This study
Mat-a bem1Δ::URA3	ABY4102	This study
Mat-α bem1Δ::URA3 boi2Δ::HIS	ABY4107	This study
Mat-a boi1-mNG::URA3 Bem1-mScarlet::NATMX	ABY4942	This study
Mat-a bem1-mNG::URA3	ABY4140	This study
Mat-a bem1-mNG::URA3 Myo1-mScarlet::NATMX	ABY4175	This study
Mat-α boi1W53K::LEU2 boi2Δ::HISMX Bem1-mNG::URA	ABY4139	This study
Mat-α Boi1PH::LEU2 boi2Δ::HISMX Bem1-mNG::URA	ABY4148	This study
Mat-a boi1W53K::LEU2 boi2Δ::KANMX bem1Δ::NATMX pRS313-BEM1	ABY4114	This study
Mat-α boi1W53K::LEU2 boi2Δ::HIS bem1-3::NATMX	ABY4116	This study
Mat-α boi1W53K::LEU2 boi2Δ::HISMX bem1-11::NATMX	ABY4119	This study
Mat-α boi1W53K::LEU2 boi2Δ::HISMX bem1-41::NATMX	ABY4122	This study
Mat-α boi1W53K::LEU2 boi2Δ::HISMX bem1-89::NATMX	ABY4128	This study
Mat-α boi1W53K::LEU2 boi2Δ::HISMX bem1-404::NATMX	ABY4134	This study
Mat-α boi1W53K::LEU2 boi2Δ::HISMX bem1-461::NATMX	ABY4136	This study
Mat-a/α boi1W53K::LEU2 boi2Δ::HISMX bem1-461::NATMX GFP-Sec4::URA3/SEC4	ABY4904	This study

Molecular cloning

Deletion of genes or tagging was performed using PCR-constructed DNA fragments with 30 bp of homology to the chromosome (Longtine et al. 1998). Truncations of Boi1 N-terminal region was also made through homologous recombination of mNG fragment to the end of each domain. Plasmids used in this study are listed in Table 2. Truncated Boi1PH(730-980) was made by cloning Boi1PH into integrating vector: pRS305 with Boi1 downstream region (+150–350 bp) cloned ahead of the Boi1 promoter (–370 bp) with a SalI endonuclease cut site separating the 2 fragments. Upon digestion with SalI, the linearized vector completely replaces the endogenous gene with LEU selection (Fig. 5a) (Chernyakov et al. 2013). Other integration plasmids were used for tagging proteins for fluorescent microscopy: pRS30Nat-mScarlet-Sec4 adapted from Donovan and Bretscher (2012), pRS306-GFP-Sec4 (same study), pRS306-Bni1-GFP (Pruyne and Bretscher 2000a, 2000b). Point mutants on plasmids were made by developing back-to-back 5′ phosphorylated primers with 1 containing the mutation in the middle of the sequence. Protocol was adapted from Thermo Scientific Phusion Site-Directed Mutagenesis Kit. The point mutant: boi1W53K was made by integration of pRS305-boi1W53K which has a 300-bp homologous region to upstream of Boi1’s promoter (685–385 bp upstream of start codon) cloned after the 3′UTR with a XbaI endonuclease cut site. Upon linearization of the plasmid, the boi1W53K mutant replaced the endogenous gene. Yeast transformations were conducted using a standard Lithium Acetate protocol. Bacterial cloning techniques used either Top10 or DH5alpha competent cells.

Table 2.

List of plasmids.

Plasmid name	Bacterial strain	Name	Reference
pRS315-BOI1	Top10	AB4538	This study
pRS316-BOI1	Top10	AB4452	This study
pRS316-BOI2	Top10	AB4446	This study
pRS315-boi1-1	Top10	AB4590	This study
pRS315-boi1-2	Top10	AB4591	This study
pRS315-boi1-3	DH5α	AB4565	This study
pRS315-boi1W794R/F799S	DH5α	AB4729	This study
pRS315-boi1PH	Top10	AB4576	This study
pRS315-boi1PHW794R	DH5α	AB4730	This study
pRS315-boi1PHF799S	DH5α	AB4732	This study
pRS315-boi1PHW794R/F799S	DH5α	AB4731	This study
yEP352-Cdc42	DH5α	AB3929	R. Gingras
yEP352-Rho1	DH5α	AB3920	R. Gingras
yEP352-Rho2	DH5α	AB3922	R. Gingras
yEP352-Rho3	DH5α	AB3865	R. Gingras
yEP352-Rho4	DH5α	AB3925	R. Gingras
yEP352-Rho5	DH5α	AB3927	R. Gingras
pRS426-MSS4	DH5α	AB4618	S. Emr
pRS313-BEM1	Top10	AB4762	This study
pRS313-bem1-Nterm	Top10	AB4682	This study
pRS313-bem1-Cterm	Top10	AB4652	This study
pRS313-Bem1-SH3aΔ	Top10	AB4653	This study
pRS313-Bem1-SH3bΔ	Top10	AB4655	This study
pRS313-Bem1-PXΔ	Top10	AB4654	This study
pRS313-Bem1-PB1Δ	Top10	AB4683	This study
pRS313-Bem1-SH3a-PB1	Top10	AB4745	This study
pRS313-Bem1-SH3b-PB1	Top10	AB4746	This study
pRS316-Bem1	Top10	AB4578	This study

Fig. 5.

Generation of conditional mutants in BOI1. a) Strategies of the screen to develop the temperature-sensitive mutants in BOI1 in a boi2Δ background. Strategy 1 highlights integration of the mutated gene to replace the endogenous gene, and strategy 2 illustrates a plasmid shuffle assay in a boi1Δ boi2Δ background. Each screen was done 3+ times and about 15,000 clones were analyzed. b) Growth of 3 boi1 conditional mutants. Each mutant is recessive as the temperature sensitivity was suppressed by incorporation of either BOI1 or BOI2 on a centromeric plasmid. Dilution assays were conducted on SC -LEU/-URA plates and grown for 2 days at the respective temperatures. c) Domain maps highlighting mutation variance across BOI1 and BOI2, purple/blue stars indicate single nucleotide mutations changing the amino acid. Highlighted mutations show similarities across screens with F791/F799 and W786/W794 as the 2 recurrent mutations found in each candidate. d) Single W794R or F799S mutations introduced into full-length BOI1 in boi1Δ boi2Δ cells show temperature-sensitive growth. Cells were grown on SC -LEU media for 2 days at the respective temperatures. e) Single W794R or F799S mutations introduced into Boi1PH in boi1Δ boi2Δ cells show temperature-sensitive growth. Cells were grown on -LEU plates for 2 days. f) Homology modeling shows where the 2 mutations potentially fall in the protein structure. Model was made with human DAPP1 PH domain bound to inositiol to represent PIP₂ binding.

Construction of temperature-sensitive mutants

Temperature-sensitive mutants were made through error-prone PCR mutagenesis as adapted (Cadwell and Joyce 1992; Wilson and Keefe 2001). The mutated fragment of BOI1 was cloned into either a pRS315 or pRS305 vector for further experiments. At least 3 separate libraries were constructed for each strategy collecting bacterial clones off of agar plates and Maxi prepping the libraries. Using strategy 1 (and also described above) the boi1-ts fragment was transformed into cells (Fig. 5a). Using strategy 2, boi1-ts was transformed into cell types with boi1Δ boi2Δ and pRS316-BOI1. The fragment of boi1-ts had differing homologous regions to the chromosome and not to the BOI1 fragment on the pRS316 plasmid so as not to recombine with the plasmid. Transformations were plated on -LEU plates (strategy 1) or -LEU/-URA plates (strategy 2) and then placed at 26°C for 2 days. For strategy 1: plates were replicated using velvet squares and replicated plates were placed at 35°C for 2 days, while the original plate was placed back at 26°C. The 2 plates were then screened by eye to find colonies capable of growth at 26°C and that were dead at 35°C. Over 15,000 colonies were screened and candidates were selected for further dilution assays on selectable media. For strategy 2: plates were replicated onto -LEU 5FOA media to eliminate the pRS316-BOI1 plasmid and were grown for 3 days at 26°C. After sufficient growth, they were then replicated again onto 2 -LEU plates, 1 placed at 26°C and the other at 35°C for 2 days. They were then screened by eye to find temperature-sensitive colonies. Bem1-ts libraries were made using the same strategy (Fig. 7f).

Fig. 7.

Bem1 and Boi1 have a secondary essential overlapping function that requires Boi1’s SH3 domain. a) Growth assays of wildtype, boi2Δ, boi1W53KΔ, boi1W53K boi2Δ, bem1Δ, and boi2Δ bem1Δ. Cells were grown on YPD plates for 2 days at the respective temperatures. b) Tetrad dissections of boi2Δ boi1W53K crossed with boi2Δ bem1Δ. Dissections revealed that boi1W53K boi2Δ bem1Δ is inviable. Spores are representative of 3 plates of dissections. Spores were grown on respective media for 2 days. c) Microscopy images of Boi1-mNG and Bem1-mScarlet at 5 different stages of bud growth including cytokinetic bud. Microscopy was max projections of 15 planes covering ∼4.2 µm of the cell. Three hundred milliseconds exposures were used to visualize Boi1-mNG and Bem1-mScarlet. Scale bars are 2 µm. d) Individual slices from 2 of the panels from (c) showing a single plane in the middle of the cell. Images were taken the same as in (c), and scale bars are 2 µm. e) Domain map of Bem1 truncations and their growth in boi1W53K boi2Δ cells. Growth is indicated as follows: none (−), temperature sensitive (+), or wildtype-like (+++). f) Growth assays of each of the strains diluted 10×. g–i) Images of Bem1-mNG, boi1W53K boi2Δ Bem1-mNG, and Boi1PH boi2Δ Bem1-mNG, respectively. Images are max projections of 15 z-planes throughout the cell ∼4.2 µm. Timelapses were taken with 5-min intervals, with 100 ms exposure. All scale bars are 2 µm.

BOI2 conditional mutants were made through PCR-mutagenized DNA encoding full-length BOI2. The mutagenic PCR was performed using the following mixture: 1 mM dCTP, 1 mM DTTP, 0.2 mM dATP, 0.2 mM dGTP, 75 ng template DNA (pWL106), 5.5 mM MgCl₂, 0.5 mM primers, 1X Taq buffer (Roche), 10 µg BSA, and 0.05% Tween-20 per 100 µl reaction. The mutagenic PCR was subjected to 8 cycles of amplification and the BamHI-NotI fragment was subcloned into pWL105 to make the libraries. Transformants were grown at the permissive temperature of 26°C and then replica-plated to 37°C to identify temperature-sensitive clones. A total of 14,000 colonies were screen and 5 temperature-sensitive strains were saved (Fig. 5, Supplementary Fig. 4.1 and Table 1).

Microscopy techniques and imaging

For live-cell imaging, yeast were grown in liquid culture to mid-log phase (0.5–0.8 OD) and were prepared by adhering 200 µl of cells to 4-well glass-bottom dishes (CellVis) coated with concanavalin-A submerged with proper media for cell survival. For fixed imaging, yeast were grown in liquid culture to mid-log phase and 4% paraformaldehyde was added to the culture for 10 min. Cells then followed appropriate protocol (see Actin staining) and then were adhered to coverslips, again by using concanavalin-A. Coverslips were then mounted using soft mounting media (Invitrogen) and secured by using clear nail polish. Some fixed cells were also imaged using agarose pads, where 4 µl of fixed cells were added to a 1% agarose pad and media was allowed to diffuse into the agarose (Pringle et al. 1989; Shin et al. 2018). All fluorescent microscopy was conducted on an inverted microscope (Leica DMI6000B) which had a spinning disk confocal unit (Yokogawa CSU-X1) with 100× 1.45 NA objective (Leica) with either a Evolve 512Delta EMCCD or with a Flash 4.0v2 CMOS camera. DIC imaging was conducted on this microscope as well as on a Leica De-Convolution Microscope (DMi8). Slidebook 6.0 or FIJI software were used for image analysis.

Cell death assay

Yeast cells were analyzed for death rate in Supplementary Fig. 5.1 during live-cell imaging. Cells were marked “dead” when lysing took place and was measured to verify death using Live-or-Dye (Biotium). Dye was in 10% DMSO and was added 1:100 into yeast cells and allowed to incubate for at least 30 min before shifting the temperature. After confirming that the yeast lysing was indeed killing the cells, we used DIC to measure the rest of the cultures.

Fluorescence recovery after photobleaching

For the fluorescence recovery after photobleaching (FRAP) experiment, specific regions of the cell, either a Boi1/2 patch or half of the bud neck, were photobleached using a single point with the 3i Vector photobleaching system after 1 s of capturing frames were taken every 160 ms for 32 s. Photobleaching control was taken as the region of a neighboring mother cell. In analysis graphs background control was subtracted and normalized to max intensity.

Mutant microscopy

Imaging the temperature-sensitive mutants using traditional methods is highly variable and difficult to control. In these studies, we used a microfluidic device temperature control system fitted to either the 3i Confocal or the Leica microscopes (CherryBiotech). Cells were adhered to a glass slide using concanavalin-A, a plastic spacer with a cutout hole was added on top of the cells, and 50 µl of selectable media was placed on top of the cells. The ThermoChip Immersion was placed on top of the cells with the plastic spacer so the yeast were close to the microfluidic chamber and still had space to continue growth. The CherryTemp was set to 26°C and shifted to 35°C depending on the experiment. The temperature shift takes only seconds, but we waited 5 min for equilibration and then started imaging.

Actin staining

The actin staining was conducted using 2 different protocols (noted in figure legends). The first staining protocol washed the fixed cells with PBS ×2 and then with PBS + 0.1%TritonX. Cells were spun down after each wash on a very low centrifugation speed—500g. The cells were incubated with PBS + 0.1%TritonX for 20 min, and then spun down and were incubated with 4% phalloidin AlexaFlour-488 in PBS + 0.1%TritonX. Prior to incubation, the phalloidin was dried and resuspended in equal PBS + 0.1%TritonX. After phalloidin incubation cells were spun down and washed with PBS + 0.1%TritonX ×2, and once with PBS. 1:1,000 DAPI incubation was also conducted for 10 min prior to adhering to the concanavalin-A cover slips and mounted using soft mounting prodiamond gold (Invitrogen) onto glass slides (Fig. 3a). The second phalloidin stain protocol used dried phalloidin AlexaFlour-568 resuspended in PBS + 0.1%TritonX. Four microliters of fixed cells were added 3× to 1% agarose pads until the media diffused into the agarose. Phalloidin was added on top of the cells and sufficient time was allotted for the PBS to diffuse (2–3 min). Once diffused, cover-slips were added on top of the samples and then they were imaged.

Fig. 3.

Boi1-mNG and Boi2-mNG localization on the plasma membrane and at the bud neck is independent of actin cortical patches, cables, and secretory vesicles. a) Fixed cell imaging of Boi1-mNG (200 ms exposure), phalloidin (200 ms exposure), and DAPI staining (100 ms exposure) at 4 different stages during budding, small bud, medium, large, and during cytokinesis. All images are max-projection z-stacks of 15 planes covering ∼4.2 µm. b) Images of Boi2-mNG in 4 different stages of the cell cycle with the same parameters as in (a). c) Boi1-mNG and Boi2-mNG after CK-666 treatment, images taken within 10–20 min after treatment. Both images are max-projection z-stacks of 15 planes covering ∼4.2 µm. Exposure 200 ms for both cell types. Second panel is actin phalloidin staining (75 ms exposure) fixed after 5 min of CK-666 treatment. d) Boi2-mNG with either treatment with a 2% DMSO control or LatA (200 µM) 2 examples of LatA treatment show localization during mid and late bud cycle. Images were taken with 100 ms exposure, and max-projections of 15 planes covering ∼4.2 µm. e) Boi2-mNG in tpm1-2 tpm2Δ cells. Max projections of 15 planes ∼4.2 µm. Exposure at 100 ms. Second panel is actin phalloidin staining at 25 ms exposure. f) Boi2-mNG and mScarlet-Sec4 colocalization. Fifty milliseconds exposure for each. Images taken in a single plane at the bottom of the cell. Arrows indicate individual Boi2-mNG localization, individual mScarlet-Sec4 localization, or colocalization. See Supplementary Video 4 for more detail on mScarlet-Sec4 localization compared with Boi2-mNG. All scale bars are 2 µm.

Latrunculin-A and CK-666 treatment

Treatment of the cells with latrunculin-A (LatA) was conducted prior to fixation for actin staining. Cells were treated to a final concentration of LatA of 250 µM in 2% DMSO (ThermoFisher) for 5 min before addition of 4% PFA. For live-cell imaging with LatA cells were adhered to the glass-bottom dishes and then LatA was spiked into the media with a final concentration of 250 µM in 2% DMSO. Control cells were treated with just 2% DMSO (Ayscough et al. 1997). For the CK-666 treatment the cells were incubated with CK-666 for 5 min prior to fixation or prior to adhering for live-cell imaging.

Results

Boi1/Boi2 are essential proteins localized to short-lived polarized patches on the plasma membrane

To study the cellular functions of Boi1/2, we confirmed that deletion of either of the 2 genes confers no apparent growth defect, yet loss of both is lethal indicating that Boi1 and Boi2 share an essential function in our genetic background (BY4741/2) (Supplementary Figs. 1.1 and 1.2). Previously, Boi1 and Boi2 have been shown to be generally polarized throughout the cell cycle and at the bud neck during later growth (Hallett et al. 2002; Kustermann et al. 2017; Masgrau et al. 2017). Here, we localize Boi1 and Boi2 C-terminally tagged endogenously with mNeonGreen (mNG) at a higher spatial and temporal resolution than prior studies. During polarized growth, both proteins localized to the nascent bud and in patches associated with the cell cortex (Fig. 1, a and b). We confirmed that this localization was only at the plasma membrane, and not on internal structures, as single confocal slices through the middle of the cell exclusively show cortical association (Fig. 1, c and d). During late bud growth, the proteins shift localization to the bud neck. By imaging alongside the class-II myosin, Myo1, tagged with mScarlet, Boi1 was found to localize at the bud neck for 7–8 min before completion of cytokinesis marked by the loss of Myo1-mScarlet and then remains there (Fig. 1e) (Bi et al. 1998; Okada et al. 2021).

Fig. 1.

Boi1-mNG and Boi2-mNG localize to short-lived stable polarized patches on the plasma membrane and at the bud neck during late bud growth. a and b) Images of endogenously tagged Boi1 or Boi2 with mNeonGreen taken at 5 stages of cell growth. Captures were at 200 ms exposure with max-projection of 15 stacks covering 4.2 µm of the cell in the z-plane. c and d) Single slices of Boi1-mNG and Boi2-mNG showing localization is only at the plasma membrane and not on cytoplasmic vesicles. e) Images of Boi1-mNG with Myo1-mScarlet. Images are z-projections of 15 max-projections planes covering 4 µm of the cell. Arbitrary timepoint “0 min” is 3 min before arrival of Boi1-mNG to the bud neck. Figures are representative of at least 5 examples. Scale bars are 2 µm. f and g) Stills of Supplementary Videos 1 and 2. Single-plane timelapse of Boi1-mNG or Boi2-mNG captured every 160 ms with 50 ms exposure. h) Graph showing patch lifetimes of Boi1-mNG and Boi2-mNG in the growing cell. The mean patch longevity for Boi1-mNG is 9 s, and Boi2-mNG is 14 s. Timing of patch longevity is significantly different P < 0.001 through a Mann–Whitney statistical test. i) Colocalization of Boi1-mNG and Boi2-mScarlet, max-projections of 15 stacks covering 4.2 µm of the cell. Boi1-mNG is exposed for 100 ms, and Boi2-mScarlet for 300 ms. Arrows indicate instances of colocalization, Boi1-mNG alone, and Boi2-mScarlet alone. All scale bars are 2 µm.

To assess relative dynamics of Boi1/2 patches on the membrane, the lifetime of single patches was measured (Fig. 1, f and g; Supplementary Videos 1 and 2). Boi1 patches were found to be shorter-lived with a mean lifetime of about 9 s, whereas Boi2 patches persisted longer, with a lifetime of about 14 s, a timing consistent with earlier measurements (Fig. 1h) (Gingras et al. 2022). However, these dynamics may be influenced by the number of molecules of Boi2 patches compared with Boi1. To explore whether Boi1 and Boi2 patches colocalize, cells were imaged expressing Boi1-mNG together with Boi2-mScarlet. Although both proteins colocalized, there was not perfect overlap with clear instances of Boi1 and Boi2 localizing to separate patches (Fig. 11, example video: Supplementary Video 3). This lack of colocalization is not due to the increased abundance of Boi2-mScarlet or the longer lifetime since Boi1-mNG is seen to be in independent patches lacking Boi2-mScarlet. Thus, Boi1 and Boi2 are similarly localized to short-lived patches that presumably participate in their overlapping essential function, yet each may also play additional distinct, subtle, roles.

Boi1 puncta are more dynamic than Boi2 puncta

As the Boi1-mNG patch life on the membrane is shorter than Boi2-mNG, we next examined the protein turnover within single patches using FRAP. Individual cortical patches were bleached and recovery was monitored for Boi1-mNG and Boi2-mNG (Fig. 2, a–d). Boi1-mNG fluorescence appeared to recover much faster than Boi2-mNG, with most of the Boi2-mNG patches never recovering after photobleaching (Fig. 2e). This suggests that Boi1’s shorter patch lifetime may reflect a higher protein turnover relative to Boi2.

Fig. 2.

Boi1 is more dynamic than Boi2 at patches within the bud and at the bud neck. a) Boi1-mNG FRAP images before, during, and seconds after photobleaching (depicted with lightning bolt) a region in the growing bud (outlined with box). Images are single z-plane slices in the middle of the cell at a 50-ms exposure. Timelapse was taken with 160 ms between each frame. b) Boi2-mNG FRAP images of a mid-budding cell taken with the same parameters. c and d) FRAP imaging of Boi1-mNG and Boi2-mNG at the bud neck taken with the same parameters as A and C with photobleaching of half the signal at the neck. e and f) Subsequent traces of intensity. g) Average normalized intensity of traces of Boi1-mNG (n = 4) and Boi2-mNG (n = 8) at the bud neck. Images were normalized to the highest average intensity. k) Examples of laterally mobility in Boi1-mNG and Boi2-mNG patches on the cortex. Arrows indicate a specific patch that moves from the first timepoint on the membrane. All scale bars are 2 µm.

To assess the dynamics of Boi1 and Boi2 at the bud neck, half of the signal at the bud neck was bleached and fluorescence recovery recorded (Fig. 2, f–i). While Boi1 fluorescence appeared to recover more readily than Boi2, neither protein recovered particularly well (Fig. 2j). Furthermore, patches hardly move laterally on the cortex (Supplementary Fig. 2.1). These experiments illustrate that Boi1 is more dynamic than Boi2 in both protein turnover at the patches and slightly at the bud neck. Boi1 is also more dynamic in patch longevity, highlighting a difference between the 2 proteins’ dynamics.

Boi1-mNG and Boi2-mNG’s localization is independent of actin and secretory vesicle transport

In order to examine the relationship between Boi1 and Boi2 patches and other polarized structures, we initially assessed their localization relative to actin. Previous studies have reported that Boi1, and the S. pombe homolog Pob1, are capable of redirecting actin filament formation and interacting with actin-nucleator formins (Toya et al. 1999; Rincón et al. 2009; Glomb et al. 2020). Neither Boi1-mNG nor Boi2-mNG colocalize with actin cortical patches (Fig. 3, a and b) indicating they are not associated with the endocytic machinery. Consistent with this lack of colocalization, treatment of cells for 5 min with the Arp2/3 inhibitor CK-666, which results in the complete disassembly of actin patches (Nolen et al. 2009; Hetrick et al. 2013), has no impact on the polarized localization of Boi1 or Boi2 (Fig. 3c).

To determine if Boi1/2 patches are dependent on actin cables, we examined Boi2-mNG localization in the tpm1-2 tpm2Δ mutant which rapidly disassembles cables upon shift to the restrictive temperature (Pruyne et al. 1998). After shifting tpm1-2 tpm2Δ cells expressing Boi2-mNG to 35°C for 5 min, Boi2 polarization remained unchanged (Fig. 3e), indicating that the polarization of Boi2 patches is independent of the presence of actin cables. Consistent with this, treatment of cells for 10 min with the drug LatA, which results in disassembly of the entire actin cytoskeleton (Spector et al. 1989; Ayscough et al. 1997), had no impact on the presence of polarized Boi2 patches at the cell cortex or the bud neck (Fig. 3d). Thus, the localization of Boi2 is independent of the actin cytoskeleton, consistent with earlier observations (Kustermann et al. 2017).

Previous studies have implicated Boi1/2 in exocytosis, showing that after depletion of Boi1/2, the cell accumulates secretory vesicles (Kustermann et al. 2017; Masgrau et al. 2017). To examine the relationship of Boi2 to secretory vesicles, we coimaged Boi2-mNG with the secretory vesicle marker, mScarlet-Sec4. Strikingly, the majority of Boi2-mNG patches do not colocalize with secretory vesicles, or vice versa, although some colocalization of tethered vesicles and Boi2 patches does occur (Fig. 3f;  Supplementary Video 4). The majority of noncolocalization is at least partially due to the relatively short tether-to-fusion time of secretory vesicles and the number of untethered vesicles in the bud (Gingras et al. 2022). Furthermore, examination of videos shows that secretory vesicles tether in places apparently devoid of Boi2-mNG patches. This led us to question whether these vesicles are tethering with regions which have no Boi1/2 present, or if they are merely tethering at Boi1 patches. To answer this, we looked at Boi2-mNG mScarlet-Sec4 in a boi1Δ. In this cell type, the mScarlet-Sec4 vesicles had a similar localization and movement as the cells in an otherwise wildtype background (Supplementary Video 5), indicating that mScarlet-Sec4 vesicles are capable of tethering on the membrane at sites lacking Boi1/2.

Given their reported role in exocytosis of secretory vesicles, we examined whether the localization of Boi1 or Boi2 were influenced in mutants with conditional defects in the late secretory pathway. Localization of Boi1-mNG or Boi2-mNG to polarized patches is unaffected by shifting the conditional sec4-8 mutant to the restrictive temperature for 30 min (Supplementary Fig. 3.1), a condition that compromises both vesicle transport and overall secretion (Salminen and Novick 1987; Harsay and Bretscher 1995). Thus, the polarized patch localization of Boi1 and Boi2 is independent of polarized secretory vesicle delivery and the regulation of this localization is upstream of exocytosis.

Earlier studies have reported that Boi2 is localized to the cytosol in the conditional exocytic mutant sec6-8 after shifting to the restrictive temperature for 1–2 h (Masgrau et al. 2017). To explore this phenomenon, we shifted the conditional exocytic mutant sec6-4 to the restrictive temperature for 30 min. Unlike the sec4-8 conditional mutant, the sec6-4 mutation only compromises vesicle tethering, not transport, at the restrictive temperature. Interestingly, even prior to shifting, Boi1-mNG appeared to localize to depolarized cortical patches in the mother—a localization never observed in wildtype cells—which became more prominent at the restrictive temperature (Supplementary Figs. 3.2 and 3.3). Together, these results indicate that the presence of Boi1/2 in polarized cortical patches is not directly influenced by the delivery of secretory vesicles nor a requirement for exocytosis; however, patch regulation may be influenced by exocyst function.

Boi1PH is responsible for Boi1 cortical bud localization while Boi1SH3 is important for Boi1 and Bem1 localization at the bud neck

After our initial characterization of Boi1/2, we then wanted to understand the contributions of the individual domains. It has previously been shown that the Boi1/2 PH domain including the C-terminal tail represents the minimal essential domain which facilitates interactions with both the plasma membrane and Cdc42. Boi1 and Boi2 also each contain an SH3 domain, an SAM motif, and a proline-rich region necessary for interaction with Bem1 (Fig. 4a) (Bender et al. 1996; Matsui et al. 1996; Hallett et al. 2002; Kustermann et al. 2017). We focused on the PH and SH3 domains by first replacing expression of the endogenous BOI1 with just the Boi1 PH domain [boi1(730-980)], in a boi2Δ background to generate Boi1PH boi2Δ. This strain grew normally at all temperatures (Fig. 4b), confirming that just the PH domain can provide the Boi1/2 essential function. Upon tagging Boi1PH with the mNeonGreen, the patch-like distribution is diminished and Boi1PH-mNG is only polarized to the bud and not found at the bud neck (Fig. 4c) (Hallett et al. 2002; Yu et al. 2004). Examining a few optical sections revealed that Boi1PH-mNG is restricted to the cortex and not found on internal structures (Fig. 4d and Supplementary Video 6). These results confirm that the shared essential function mediated by the PH domain is at the cell cortex, and another part of the protein is responsible for the patch-like distribution that the wildtype-protein exhibits. It has been suggested by previous work that the protein’s SAM domain might help facilitate homodimerization (Jia et al. 2020). This may explain why the patch formation is eliminated in the Boi1PH-mNG cells.

Fig. 4.

Boi1PH is the essential domain important for stable Boi1 cortical membrane localization, and the Boi1SH3 domain is important for Boi1 and Bem1’s localization at the neck. a) Schematic showing the truncations of Boi1 used in this figure. b) Dilution assays of wildtype, boi2Δ, and boi1PH(730-980) boi2Δ cells at 26, 30, 35, and 37°C all on YPD plates grown for 2 days. c) Max projections of 15 z-planes of boi1PH-mNG boi2Δ cells at the nascent, small bud, medium bud, large, and precytokinetic bud. Images were exposed for 200 ms. Images cover ∼4.2 µm of the cell. d) Stills of Supplementary Video 6 showing boi1PH-mNG boi2Δ cells. Images were taken with 50 ms exposure max projections of 6 planes at the bottom of the cell. e) Images of 3 different tagged truncations of Boi1 in a wildtype cell. Images were taken at 500 ms exposure, max projections with 13 z-planes covering ∼4 µm. All scale bars are 2 µm.

After initial characterization of the Boi1PH domain, we explored the localization of the N-terminal part of the protein. Tagged versions of 3 different truncations of Boi1 (Boi1SH3, Boi1SH3-SAM, Boi1SH3-SAM-PxxP) all localized exclusively to the bud neck and were never found at the bud cortex (Fig. 4e). Thus, the cortical bud localization of Boi1 is mediated by the PH domain, whereas the bud neck localization is mediated by the Boi1 SH3 domain.

Generation of temperature-sensitive mutations in Boi1 highlight 2 aromatic residues in the PH domain

To investigate the essential function of Boi1/2 further, we used PCR mutagenesis to generate temperature-sensitive alleles of BOI1 in a boi2Δ background. Two different strategies were employed to incorporate the mutated gene. The first was direct integration of the mutated gene into the genome using an integrating plasmid, and the second employed a plasmid shuffle screen (Fig. 5a). Using each strategy in 3 separate screens for mutants that could grow at 26°C but not at 35°C, we were unable to recover conditional mutants using the integration approach, but recovered several from the plasmid shuffle screens. Three alleles, boi1-1, boi1-2, and boi1-3, were selected for further study. Interestingly, they failed to grow at 35°C, but growth was restored at 37°C. All 3 mutants were recessive, as expression of the wildtype BOI1 or BOI2 genes suppressed their temperature-sensitive phenotype (Fig. 5b). Multiple attempts were made to introduce these mutations into the chromosome, but all failed. We conclude that they are lethal at the chromosomal locus, so all further studies were performed with them expressed from a centromeric plasmid.

Interestingly, although multiple mutations were found in the temperature-sensitive alleles, several of them had mutations in one of 2 aromatic residues in the PH domain of BOI1. Remarkably, in a similar screen for conditional mutations in BOI2 in a boi1Δ background we again found the corresponding residues affected (Fig. 5c and Supplementary Fig. 4.1). These mutations were either the W794R or the F799S in BOI1 or W786R and F791S in BOI2. These individual mutations were found to confer temperature-sensitive growth when individually placed in the full-length protein or the Boi1PH version (Fig. 5, d and e). Although single mutations caused temperature sensitivity, the double W794R F799S mutant was not viable (Supplementary Fig. 4.2). Both W794 and F799 are adjacent to the PIP₂ binding pocket of Boi1, potentially explaining the temperature sensitivity (Fig. 5f). To further explore what residues caused temperature sensitivity at Boi1 positions 794 and 799, we generated libraries of Boi1 with random codons at these 2 positions and asked which of these were temperature sensitive when introduced into a boi1Δ boi2Δ background. Among several colonies, we found: W794R, W794G, F799S, F799L, and F799G (Supplementary Fig. 4.3). These results establish the critical and restrictive nature of residues FW794 and F799 in Boi1.

The boi1 conditional mutants exhibit defects in cell morphology, secretion, and actin patch localization

To better characterize which cellular process Boi1/2’s essential function supports, we observed the growth phenotypes of the boi1-ts. The boi1 conditional mutants did not cease growth immediately upon shifting to the restrictive temperature (Supplementary Fig. 5.1) and showed variable viability after shifting to 35°C for 4 h, with about 50% loss for boi1-1 and boi1-2, and 100% for boi1-3 (Supplementary Fig. 5.2). As boi1-1 exhibits the least clear temperature sensitivity, our characterizations focused on boi1-2 and boi1-3 (both containing the F799S mutation).

To explore Boi1’s essential function, we compared the growth morphology of boi1-2 boi2Δ cells with well characterized mutants affecting either the actin cytoskeleton, or exocytosis. Upon shifting to 35°C, boi1-2 cells showed a variable phenotype of diminished cell growth and then eventually lysing (Fig. 6a). In contrast, tpm1-2 tpm2Δ cells, which lose actin cables at the restrictive temperature (Pruyne et al. 1998), exhibit altered growth and also eventually lyse (Fig. 6b). Conditional defects in secretion as exhibited by a sec4-8 mutant, simply halt cell growth at the restrictive temperature and do not lyse (Fig. 6c). This phenotypic analysis shows that the essential role of Boi1 is not restricted to organizing actin, or to secretion, but has a more complex function. The mutant most like the boi1-2 was the conditional pkc1-2 mutant affecting the essential kinase activity of the cell wall integrity pathway (Levin et al. 1990; Paravicini et al. 1992). This mutant appears to have a similar lysing pattern which shows lysing after about 80 min at the restrictive temperature (Fig. 6d).

Fig. 6.

Characterization of the boi1 conditional mutants. a) Morphology of boi1Δ boi2Δ pRS315-boi1-2 at 26°C and then shifted to 35°C for several minutes. Microscopy images are stills taken with DIC microscopy to look at overall cellular structure. Scale bars are 5 µm. b–d) Morphology of different mutants related to secretion and cell wall integrity for comparison: tpm1-2 tpm2Δ, sec4-8, and pkc1-2, respectively. Images were taken with the same parameters as (a). e) Actin morphology of cells grown at 26 and 35°C for 1 h and fixed with PFA. Cells were stained with phalloidin-568 (see Materials and Methods) and microscopy is max-projections of 15 z-stack planes with 100 ms exposure. Scale bars are 2 µm. f) Graphs showing the number of patches in the mother in cells at 26°C and shifted to 35°C for 1 h: 26°C BOI1 (n = 9), boi1-2 (n = 10), and boi1-3 (n = 12); at 35°C, BOI1 (n = 17), boi1-2 (n = 9), and boi1-3 (n = 12). Statistical analysis was conducted as 1-way ANOVO tests between columns. P-value <0.01 for the significant values.

The actin cytoskeleton was partially disrupted in the boi1 conditional mutants. Whereas wildtype cells have cortical patches almost exclusively in the bud and robust actin cables whether grown at 26°C or shifted to 35°C for 1 h, boi1-2 cells lost much of the patch polarity even at 26°C with diminished cables, with even greater loss of patch polarity and cables at 35°C (Fig. 6e). Similarly, boi1-3 lost polarity of patches and had reduced cables after shifting to 35°C (Fig. 6e). Quantitation of patch localization showed significant differences for patch polarity in both the boi1-2 and boi1-3 cells (Fig. 6f). Our phenotypic characterization of the conditional boi1 mutants did not reveal a clear role in either secretion or cytoskeletal polarity making its essential function difficult to pinpoint.

BOI1 participates in the Cdc42 pathway and the full length is influenced by PIP₂ concentrations

Earlier work has suggested that the Boi1/2 proteins function in the Cdc42 and potentially Rho3 pathways (Bender et al. 1996; Matsui et al. 1996; Liao et al. 2013; Jia et al. 2020). Cdc42 is an essential conserved small GTPase that regulates cell polarity from yeast to vertebrates (Johnson 1999; Chiou et al. 2017) and its GEF, Cdc24 is also essential in yeast. Overexpression of either CDC42 or CDC24 on a 2-µm vector suppressed the growth defect of boi1-3 (Supplementary Fig. 6.1a). To see how specific this suppression was to Cdc42, each of yeast’s 6 small GTPases of the Rho family were expressed from 2 µm vectors. CDC42 suppressed well, with only RHO3 overexpression also showing some suppression (Supplementary Fig. 6.1, b and c).

Next, we examined whether this suppression was seen in the full-length protein containing a single mutation in the PH domain, or in expression of just the PH domain with the single mutation (Supplementary Fig. 6.1, d–f). CDC42 overexpression was able to suppress the temperature sensitivity of both the boi1PH-W794R and the boi1PH-F799S mutants, and only RHO3 overexpression showed some suppression of the boi1PH-W794R.

We were surprised to find that the original conditional boi1 mutants show temperature-sensitive growth at 35°C, yet grew quite well at 37°C (see Fig. 5b). It has been shown that shifting yeast cells from 26 to 37°C elevates the level of the plasma membrane signaling lipid, PI(4,5)P₂ (Stefan et al. 2002). Moreover, the mutations in the PH domain that confer temperature sensitivity likely affect the binding of PI(4,5)P₂ (see Fig. 5f). If this is the reason why the conditional mutants grew better at 37°C, we predicted that overexpression of the enzyme that synthesizes PI(4,5)P₂ should suppress growth at 35°C. Accordingly, we overexpressed the PI-4P 5-kinase MSS4 in each of the 3 conditional mutants, and found that all were partially suppressed, but to different degrees (Supplementary Section 6.2). Next, we examined cells expressing full-length Boi1 with either the single W794R or F799S mutations, or these mutations in the context of the Boi1 PH domain (Supplementary Fig. 6.2). In both cases, MSS4 overexpression suppressed the constructs harboring the W794R mutation, but not those containing the more severe F799S mutation (Supplementary Fig. 6.2). Interestingly, at 37°C, MSS4 overexpression suppressed the mutations in the context of the full-length protein, but not in the context of the PH domain. This finding is addressed later in the Discussion.

Boi1 and Bem1 have a secondary overlapping essential function that requires Boi1’s SH3 domain

The redundant Boi1 and Boi2 proteins were initially identified in a screen for interactors of Bem1’s second SH3 domain (SH3b) (Bender et al. 1996; Matsui et al. 1996). Disrupting the function of Boi1’s SH3 domain with the boi1W53K mutation in a boi2Δ background rendered BEM1 essential. Therefore, in boi2Δ cells there is an essential function requiring either Boi1-SH3 or Bem1 (Matsui et al. 1996). We confirmed that the boi1W53K mutation does not confer a growth defect in an otherwise wildtype background, or in boi2Δ cells, and that boi2Δ boi1-W53K bem1Δ is lethal (Fig. 7, a and b). We also confirmed that bem1Δ and bem1Δ boi2Δ are temperature sensitive (Fig. 7a).

While it is known that Boi1/2 and Bem1 each localize to sites of polarization, it is not clear how well they co-localize. We therefore imaged Boi1-mNG and Bem1-mScarlet and found that they colocalize at every stage in the cell cycle (Fig. 7, c and d). However, Boi1 arrives at the bud neck about a minute before Bem1 when compared with the cytokinesis marker Myo1 (compare Fig. 1e to Supplementary Fig. 7.1). We also observed Bem1 arriving after Boi1 in dual-color microscopy videos (Supplementary Video 7). Furthermore, in boi1W53K boi2Δ or boi1PH boi2Δ cells, Bem1-mNG is not recruited to the bud neck (Fig. 7, f–h). These results indicate that Bem1 is recruited to the bud neck by Boi1 through its SH3 domain. Since boi1W53K boi2Δ has no growth or cytokinesis defect, this implies that Bem1’s localization at the bud neck is not important for cytokinesis.

To define the minimal Bem1 domains necessary for growth in a boi1W53K boi2Δ background, we conducted a plasmid shuffle screen with several truncated versions of Bem1. A construct lacking both SH3 domains was not viable, and lacking just the SH3a grew well, but lacking SH3b grew poorly (Fig. 8e and Supplementary Fig. 7.2). Further analysis showed that in the boi1W53K boi2Δ background, robust growth required the presence of both the SH3b and PB1 domains. Indeed, a fusion of SH3b-PB1 restored growth to the wildtype levels, whereas a fusion of SH3a-PB1 did not. Moreover, the SH3b-PB1 construct fully suppressed the temperature sensitivity at 37°C of bem1Δ (Fig. 7e and Supplementary Fig. 7.2).

Fig. 8.

BEM1 conditional mutants exhibit unusual cellular morphology, budding pattern, and actin morphology. a) Schematic for the integration of mutated BEM1 to replace the endogenous BEM1. b) Domain map of Bem1 showing where several of the mutations from different candidates fall. c) Growth assay of bem1 conditional mutants with empty pRS316 or a pRS316-BEM1 plasmid grown at 26, 30, 35, or 37°C for 3 days. d) DIC images of bem1-461 boi1W53K boi2Δ at 26 and 37°C over the course of a timelapse up to 7 h of growth. Scale bars are 5 µm. (e) Upper panel is boi1W53K boi2Δ GFP-Sec4/SEC4 diploid cells at 26°C and then shifted to 35°C for ∼22 min. Lower panel shows the bem1-461 boi1W53K boi2Δ GFP-SEC4/SEC4 diploid strain. Scale bars are 5 µm. (f) Actin phalloidin staining of the cells grown at 26°C or after shifting to 35°C for 1 h. Images are max projections of 15 z-planes with 100 ms exposure. Scale bars are 2 µm.

We conclude that the SH3b and PB1 domains of Bem1 can fully substitute for the full-length protein, and that the SH3b domain performs a redundant function with the Boi1 SH3 domain. Additionally, in the Boi1PH boi2Δ bem1Δ strain, the Bem1-SH3b-PB1 construct could restore growth comparable to full-length BEM1 (Supplementary Fig. 7.2). Thus, the 3 genes BOI1, BOI2, and BEM1 can be functionally replaced by the Boi1PH and Bem1-SH3b-PB1 constructs.

Conditional bem1 mutants generated in the boi1W53K boi2Δ background have significant morphological and budding defects

To facilitate investigation of the essential function of Bem1 in boi1W53K boi2Δ cells, we targeted mutated BEM1 to its chromosomal locus and screened for temperature-sensitive mutations (Fig. 8a). Several candidates were recovered and all were found to be recessive to BEM1 (Fig. 8, b and c). Sequence analysis showed many mutations in some of the conditional bem1 alleles, but notably several of them were stop codons that eliminated the important PB1 domain (Supplementary Fig. 8.1). Two of the mutants that had truncations in the PB1 domain also had mutations in the PX domain which caused a more severe temperature sensitivity compared with the truncation alone (Fig. 8c and Supplementary Fig. 7.2).

One of these mutants, bem1-461 was selected for further study. Upon shifting to 37°C, the mutant continued to grow but in a largely isotropic manner, and then lysed. The lysing pattern is similar to what we saw in the conditional boi1 mutants; however, growth is not initially impaired. Additional other phenotypes result including bud-on-bud pattern similar to mutants of pkc1-2 (Paravicini et al. 1992) (Fig. 8c and Supplementary Fig. 8.2). This mutant also showed a largely depolarized actin cytoskeletal structure accounting for the severe defects in polarized growth (Fig. 8f). Replacement of SEC4 with GFP-SEC4 suppressed the temperature sensitivity of the bem1-ts, so we examined its distribution in SEC4/GFP-SEC4 diploids (Fig. 8e). The secretory vesicles marked by GFP-SEC4 do not initially depolarize upon shifting to the restrictive temperature, but eventually do. This is consistent with the depolarized actin cytoskeleton and depolarized growth seen after shifting for an extended period. Thus, the onset of defects appears to be quite slow in this mutant, making pinpointing a specific process difficult.

Discussion

Boi1/2 have been implicated in functioning during cell polarity by interacting with Cdc42, PIP₂, the exocyst and Sec1 during exocytosis, and actin nucleators. Genetic analysis has also clearly shown that Boi1/2 share an overlapping and essential function with Bem1. With this broad set of proposed roles, we aimed to probe the essential functions of Boi1/2 and Bem1 by analyzing each protein’s localization throughout the cell cycle, and by generating and characterizing conditional mutants in the essential domains.

Careful characterization showed that both Boi1 and Boi2 are present in dynamic cortical patches where Boi2 patches are more stable. Surprisingly, although Boi1 and Boi2 are functionally redundant, they frequently showed distinct patches in the bud, suggesting that they may have divergent as well as overlapping functions. It is unclear whether Boi1 or Boi2 may be arriving in a sequential order, as examples of the 2 proteins arriving one before the other have been observed, and it is unclear whether it is to the same site on the membrane, or rather nearby. Both proteins localize at the bud neck prior to cytokinesis. This highlights a role at the neck to potentially aid in the organization of cytokinetic proteins or to allow for exocytic vesicles to accumulate at the neck to ensure the closure of the plasma membrane and cell wall during separation.

Given the potential interplay of Boi1/2 in actin structure and secretory vesicle exocytosis, we examined the effect of disrupting both actin and exocytosis on Boi1/2 localization. Neither had any affect, showing that if Boi1/2 play roles in these processes, they are upstream in the pathway. Our analysis found that Boi1/2 remains at the cell cortex upon shifting sec6-4 to the restrictive temperature, again suggesting that they may function upstream of exocytosis. We also asked if Boi1/2 patches correlate with sites of secretory vesicle exocytosis especially given that Boi1/2 patches have a lifetime about double that of secretory vesicle tethering (Gingras et al. 2022). Interestingly, there was very little colocalization between Boi1/2 and Sec4, and since Boi2 patches do not move extensively on the membrane, it is difficult to reconcile with a model that invokes a direct and essential role for Boi1/2 in the tethering/exocytosis of secretory vesicles. This result suggests that Boi1 may not be interacting with the exocyst or Sec1 during exocytosis for the tethering and secretion of vesicles, but may be acting as a coincidence detector between the exocytic machinery and the Rho-GTPase given previous data reporting their interaction.

Moreover, secretory mutants show immediate cessation of growth upon shifting to the restrictive temperature, which was not seen in our boi1 conditional mutants, again shedding doubt that they play an essential role in exocytosis. Nevertheless, 2 studies have made a clear case for a role in exocytosis (Kustermann et al. 2017; Masgrau et al. 2017). In both studies, the level of Boi1 proteins was depleted by either dilution or targeted degradation, both of which take significant time. It is possible that secretory vesicle accumulation in both these studies is a secondary and indirect effect of inhibiting another important process. As the essential region of Boi1/2 lies in the C-terminal PH domain, we turned to examining the properties of cells that express just the Boi1-PH domain in boi2Δ cells.

The Boi1-PH boi2Δ cells grew normally at all temperatures and localized tightly to the cortex of the bud. The localization was more diffuse than Boi1-mNG, indicating that the part of the protein lacking in this construct is responsible for patch localization. Boi1-PH-mNG did not relocate to the bud neck for cytokinesis, whereas the nonessential regions containing the SH3 domain did, indicating that the role of Boi1/2 localization at the bud neck is not important for cell viability.

It is clear from these studies that the essential function of the Boi1/2 proteins is not simple, and therefore we set out to generate conditional mutations in BOI1 in a boi2Δ strain. Generating conditional mutations in BOI1 turned out to be challenging, and the construction of the boi1-ts library and mutants were the first published attempts. These conditional mutations were only viable on a plasmid, presumably due to a slightly higher level of expression than when placed in the chromosome. These mutants had an interesting phenotype in that they grew at 26°C, but not at 35°C, but did again at 37°C. Moreover, due to the delicate nature of the mutants, they would not tolerate tagging of key proteins involved in secretion, such as Sec4 and subunits of the exocyst. This condition made it difficult to characterize the mutants to understand which cellular process the boi1-ts were involved in. Furthermore, the fact that the conditional mutants continued to exhibit cell growth at the restrictive temperature, unlike exocytic mutants sec6-4 or sec4-8, indicates that the essential function is unlikely to be directly involved in exocytosis. Thus, the characterization did not give a clear phenotype as we had hoped—having apparent defects in growth and eventually lysing after extended periods at the restrictive temperature.

In previous unpublished work, we isolated conditional mutations in BOI2 in a boi1Δ background. Many of the BOI1 and BOI2 conditional mutants had mutations in either of 2 homologous residues: W794R or F799S in BOI1, or W786R or 791S in BOI2. Replacing the BOI1 gene in a boi2Δ strain with the single W794R or F799S mutation into full-length Boi1, or Boi1-PH, resulted in temperature-sensitive growth. Moreover, a screen in which any amino acid was placed at position 794 or 799 revealed the same amino acid changes already recovered that conferred temperature sensitivity. Thus, our mutational analysis reveals the critical nature of these 2 residues, which structurally are localized near the predicted PIP₂ binding region of the PH domain. Consistent with this, overexpression of the kinase that generates PIP₂ at the plasma membrane, Mss4, partially suppresses the conditional mutants at 35°C.

As noted above, our temperature-sensitive boi1 mutants, failed to grow at 35°C, but did grow at 37°C. We suspected that this phenotype might be due to a defect in PIP₂ binding, as PIP₂ levels rise upon shifting to 37°C (Stefan et al. 2002). Interestingly, the boi1-PHW794R allele in a boi2Δ background is temperature sensitive for growth at both 35 and 37°C. While MSS4 overexpression partially restores growth at 35°C, it does not do so at 37°C. This suggests that while elevating PIP₂ may help growth in the presence of boi1-PHW794R, it is not the only essential function of the PH domain. Since the same mutation in the context of the full-length Boi1 protein does grow at 37°C, this essential function of the PH domain at 37°C must be at least partially redundant as it can be satisfied by some other part of the full-length molecule.

As far as we are aware, this is the only incidence where a single PH domain has been shown to be essential. Simply binding a lipid cannot be essential itself, so it is necessary to predict that the Boi1PH construct must bind another factor, perhaps acting as a coincidence detector. An obvious candidate is Cdc42 that has been shown to bind the PH domain (Bender et al. 1996) and CDC42 overexpression, or its activator CDC24, can suppress our conditional mutants. However, arguing against this model is the generation of the triple R827A, L829A, T894A mutant in the Boi1PH domain that no longer binds Cdc42 and that has no effect on cell growth (Kustermann et al. 2017). Thus, the essential function of the PH domain remains open, and perhaps genetic approaches using the new conditional mutants will be more illuminating.

We next turned our attention to the overlapping function between Boi1/2 and Bem1. We confirmed that Bem1 becomes essential in a boi1W53K boi2Δ strain. By deletion analysis, we found that fusing the second SH3 domain of Bem1 to its C-terminal PB1 domain could provide growth equivalent to wildtype cells. Thus, the 2 essential functions of Bio1/2/Bem1 can be supplied by just the Bem1-SH3-PB1 construct replacing BEM1, and the Boi1PH replacing BOI1 and BOI2.

To analyze the essential function of Bem1 in the boi1W53K boi2Δ cells, we once again turned to generating conditional mutations in BEM1. In contrast to the case described above, generating conditional BEM1 mutations by direct insertion into the chromosome was relatively straightforward. Upon shifting to the restrictive temperature, one of these mutants, bem1-461, continued to grow isotropically and then lysed. Consistent with depolarized growth, the actin cytoskeleton was highly depolarized after shifting to the restrictive temperature for an hour. However, the onset of these phenotypes was slow—GFP-Sec4 marked secretory vesicles did not depolarize immediately upon shifting to the restrictive temperature, so is difficult to assess if the depolarized phenotype is a direct or indirect consequence of affecting the essential function of Bem1.

In this study, we have defined more clearly the localization and dynamics of the Boi1/2 scaffolding proteins and identified 2 critical residues that impact the essential function of the Boi1/2 PH domains. We have also clarified the regions, the SH3 and PB1 domain, important in Bem1 that provide full function in a boi1W53K boi2Δ cell. It is somewhat disappointing that the phenotypes of the 2 sets of conditional mutations that we generated in BOI1 and BEM1 were not more illuminating, but future genetic approaches using these mutants may help in pinpointing the precise essential functions of these scaffolding proteins.

Data Availability

All strains and plasmids are available upon request. The authors affirm that all data important and necessary for the conclusions within this article are present within the article, figures, and Tables 1 and 2.

Supplemental material is available at G3 online.