Regulation of intrinsic polarity establishment by a differentiation-type MAPK pathway in S. cerevisiae

All cells establish an axis of polarity, which is critical for cell shape, the organization of cellular compartments and progression through the cell cycle (Doerr and Ragkousi, 2019). Cell polarity can be reorganized in response to intrinsic and extrinsic cues and is required during development and for other processes that require dynamic changes in cell shape such as cell migration and differentiation (Henderson et al., 2018; Pei et al., 2019; Piroli et al., 2019). Defects in cell polarity are commonly associated with human disease. For example, the mis-regulation of cell polarity leads to metastasis in many types of cancers (Fomicheva et al., 2020; Noguchi et al., 2018; Taciak et al., 2018).

Cdc42p and other members of the Rho family of GTPases are central regulators of cell polarity (Arkowitz and Iglesias, 2008; Irazoqui and Lew, 2004; Jimeno and Santos, 2017; Pringle et al., 1995; Pruyne and Bretscher, 2000a,b; Woods and Lew, 2019; Zihni and Terry, 2015). In the yeast Saccharomyces cerevisiae, Cdc42p regulates bud emergence, which is an essential process where daughter cells or buds are produced from mother cells. The pathway that regulates bud emergence has been extensively studied and represents one of the best examples for how cells establish an axis of polarity (Fig. 1A, green pathway). Activation of Cdc42p by the guanine nucleotide exchange factor (GEF) Cdc24p at bud sites – or at random sites in bud-site-selection mutants – results in symmetry breaking, which commits the cell to grow at a particular site. Positive and negative feedback loops promote polarity establishment by amplifying the levels of active or GTP-bound Cdc42p at the incipient bud site (Chiou et al., 2018; Irazoqui et al., 2003; Kozubowski et al., 2008; Woods and Lew, 2019). The GTP-bound conformation of Cdc42p binds effector proteins (Bendezú and Martin, 2013; Guo et al., 2010; McCormack et al., 2013; Wu and Jiang, 2005; Wu et al., 2006; Yang et al., 2007), including Gic1p and Gic2p (Brown et al., 1997; Chen et al., 1997; Iwase et al., 2006; Kawasaki et al., 2003; Liu et al., 2019), p21 activated kinases (PAKs) Ste20p and Cla4p (Cvrckova et al., 1995; Gulli et al., 2000; Hofmann et al., 2004; Simon et al., 1995; Takahashi and Pryciak, 2007), and the formin Bni1p (Evangelista et al., 1997; Sherer et al., 2018). In a highly choreographed manner that is coordinated with the cell cycle (Moran et al., 2019), effector proteins regulate the assembly of the septin ring, which forms a diffusion barrier between mother and daughter cells, and the formation of actin cables, which direct the delivery of vesicles to the nascent bud site. In this way, buds are produced from mother cells in a precisely spatially and temporally regulated manner.

Cdc42p also regulates MAPK pathways (Brown et al., 1996; Martin, 2019; Simon et al., 1995). MAPK pathways are evolutionarily conserved signaling modules that control cell differentiation and the response to stress in eukaryotes (Dinsmore and Soriano, 2018; Papa et al., 2019; Raman et al., 2007; Yoon and Seger, 2006). In yeast, nutrient limitation induces a Cdc42p-dependent MAPK pathway that regulates filamentous/invasive/pseudohyphal growth [Fig. 1A, blue pathway, fMAPK pathway (Adhikari et al., 2015; Borneman et al., 2007; Cullen, 2015; Gancedo, 2001; Gimeno et al., 1992; Leberer et al., 1997; Mösch et al., 1999; Mosch et al., 1996; Pan et al., 2000; Peter et al., 1996; Roberts and Fink, 1994)]. Filamentous growth occurs in many fungal species, and in some plant and animal pathogens, filamentous growth is required for virulence (Desai et al., 2014; Lagree and Mitchell, 2017; Lo et al., 1997). At the head of the fMAPK pathway, plasma membrane proteins Msb2p and Sho1p regulate the Cdc42p module, which leads to the recruitment and activation of Ste20p, the specific Cdc42p effector that regulates the fMAPK pathway. Ste20p in turn induces a MAPK cascade composed of Ste11p (MAPKKK), Ste7p (MAPKK) and Kss1p (MAPK) (Madhani et al., 1997; Roberts and Fink, 1994). MAPK Kss1p regulates the activity of transcription factors Ste12p and Tec1p, as well as transcriptional repressors and other factors (Bardwell et al., 1998; Madhani and Fink, 1997; van der Felden et al., 2014) to regulate expression of target genes that bring about the filamentous cell type (Roberts et al., 2000; Rupp et al., 1999).

Despite the fact that bud emergence is extensively regulated by positive and negative feedback loops and is coordinated with the cell cycle, bud emergence is not known to be regulated by extrinsic cues or signaling pathways. The fMAPK and polarity pathways both require Cdc42p, which suggests that functional cross-talk may occur between the two pathways. In support of this possibility, we have recently found that bud-site-selection proteins regulate the fMAPK pathway [Fig. 1A, green arrow (Basu et al., 2016)]. We tested the hypothesis that reciprocal cross-talk might occur from the fMAPK pathway to the bud emergence pathway, and discovered a role for the fMAPK pathway in regulating bud emergence. An active fMAPK pathway rescued the bud emergence defect of a polarity mutant (cdc24-4), and was required for proper bud emergence during filamentous growth. The fMAPK pathway accomplished this role by inducing expression of a direct effector of Cdc42p that functions in budding (GIC2) and by directly stimulating GTP-Cdc42p levels. The activity of the fMAPK pathway was regulated during the cell cycle to peak prior to and during bud emergence and when mis-regulated induced growth at multiple sites. Our study therefore identifies a new regulatory facet of bud emergence that is subject to regulation by a MAPK pathway and external cues. Given that Rho GTPases are functionally connected to MAPK pathways in higher eukaryotes, polarity establishment might be generally regulated by MAPK and other pathways.

Multicopy Suppressor of Budding Defect 2 (MSB2) was initially characterized as a high-copy suppressor of the budding and growth defects of the cdc24-4 mutant (Bender and Pringle, 1989, 1992). Msb2p was subsequently identified as the mucin-type glycoprotein that regulates the fMAPK pathway [Fig. 1A, blue pathway (Cullen et al., 2004)]. To determine whether Msb2p regulates budding through the fMAPK pathway, plasmids carrying alleles that hyperactivate the fMAPK pathway were introduced into the cdc24-4 mutant and examined for growth at 37°C, in synthetic medium lacking uracil to maintain selection for the plasmids, and supplemented with 0.5 M sorbitol, which stabilizes cell integrity (Bender and Pringle, 1989, 1992). Insertion of Green Fluorescent Protein (GFP) into the extracellular domain of Msb2p results in a version of the protein that hyperactivates the fMAPK pathway (Adhikari et al., 2015; Vadaie et al., 2008). pGFP-MSB2 weakly suppressed the growth defect of the cdc24-4 mutant (Fig. 1B). Hyperactive versions of Sho1p (pSHO1^P120L) and Ste11p (pSTE11-4) more robustly rescued the growth defect of the cdc24-4 mutant (Fig. 1B). pSTE11-4 also rescued the polarity defect of the cdc24-4 mutant (Fig. S1A) and was used for most of the subsequent experiments. Therefore, activated versions of MAPK pathway components can rescue the growth and polarity defects of a bud emergence mutant.

Msb2p, Sho1p and Ste11p regulate two MAPK pathways [fMAPK and high-osmolarity glycerol response (HOG) (Saito, 2010)]. A pathway-specific regulator of the HOG pathway, Pbs2p (Brewster et al., 1993; Nishimura et al., 2016; Zarrinpar et al., 2004), was not required for growth of the cdc24-4 mutant (Fig. 1C). In fact, the cdc24-4Δ pbs2Δ double mutant grew better than the cdc24-4 single mutant, perhaps because the HOG pathway negatively regulates the fMAPK pathway and its loss results in hyperactive fMAPK pathway activity (Davenport et al., 1999). By comparison, a pathway-specific regulator of the fMAPK pathway, Tec1p (Madhani and Fink, 1997), was required for viability of the cdc24-4 mutant (Fig. 1B). Tec1p functions with another transcription factor, Ste12p (Liu et al., 1993), which was also required for growth of cdc24-4 (Fig. 1C). Ste11p and Ste12p also regulate the mating pathway, but these experiments were carried out in a mating defective strain (ste4Δ). Tec1p was required for suppression of the growth defect of the cdc24-4 mutant by pSTE11-4 (Fig. 1B). pSTE11-4 also induced phosphorylation of the MAP kinase for the fMAPK pathway, Kss1p (P∼Kss1p) in the cdc24-4 mutant (Fig. S1B). Therefore, the MAPK pathway that controls filamentous and invasive growth regulates bud emergence in a defined genetic context.

The fMAPK pathway controls filamentous growth by regulating target genes that control cell adhesion [FLO11 (Rupp et al., 1999)], bud-site selection [BUD8 (Cullen and Sprague, 2002; Harkins et al., 2001; Taheri et al., 2000)] and cell elongation [CLN1 (Kron et al., 1994; Loeb et al., 1999; Madhani et al., 1999)]. None of these genes was required for viability of the cdc24-4 mutant (Fig. S1C). To identify relevant polarity targets of the fMAPK pathway, gene ontology (GO) term analysis was performed on data generated in Chow et al. (2019). Polarity targets of the fMAPK pathway included genes that regulate bud-site selection (Fig. S1D; AXL2, BUD8, RSR1 and RAX2), polarity establishment (MSB2, GIC2 and RGA1), and septin ring organization (GIC2, AXL2 and RGA1). We explored the roles of many of these genes in this study. We started with GIC2, which encodes a direct effector of Cdc42p that along with GIC1 functions in bud emergence [Fig. 1A, green (Brown et al., 1997; Chen et al., 1997)]. The GIC2 gene was a target of the fMAPK pathway (Fig. S1E; see also MacIsaac et al., 2006).

Gic2p was required for viability of the cdc24-4 mutant (Fig. 1B; Costanzo et al., 2010) and to suppress the growth defect of cdc24-4 by pSTE11-4 (Fig. 1B). Expression of GIC2 from a fMAPK pathway-independent promoter partly rescued the growth defect of the cdc24-4 mutant (Fig. 1B; pTEF2-GIC2) even in cells lacking an intact fMAPK pathway (pTEF2-GIC2 ste12Δ). GIC1 was not a target of the fMAPK pathway (Fig. S1E) and pSTE11-4 suppressed the cdc24-4 gic1Δ mutant better than cdc24-4 gic2Δ (Fig. S1F). Thus, Gic2p is a target of the fMAPK pathway that is required for fMAPK pathway-dependent rescue of the cdc24-4 mutant.

Gic proteins regulate clustering or polarization of Cdc42p at bud sites (Daniels et al., 2018; Kang et al., 2018), which is a critical step in polarity establishment. Time-lapse fluorescence microscopy was performed to examine the viability of the cdc24-4 mutant on SD-URA media. This approach showed that GFP-Cdc42p clustered at a specific site on the cell cortex (Fig. 1D; wild-type, asterisk), from which site new buds emerged. The cdc24-4 mutant was defective for GFP-Cdc42p clustering and bud emergence (Fig. 1D; cdc24-4). Analysis of 30 cells showed this difference to be statistically significant (Fig. 1E). pSTE11-4 partially restored GFP-Cdc42p clustering and bud emergence to the cdc24-4 mutant (Fig. 1D,E). Rescue of bud emergence by pSTE11-4 was dependent on Gic2p (Fig. 1D,E).

During bud emergence, Gic proteins also regulate formation of the septin ring (Bi and Park, 2012; Iwase et al., 2006; Kang et al., 2018; Okada et al., 2013; Sadian et al., 2013). In wild-type cells, the septin Cdc12p-GFP was localized in a ring prior to bud emergence (Fig. 1F; black arrows) and to the mother-bud neck in budding cells (Fig. 1F; asterisk). In the cdc24-4 mutant, Cdc12p-GFP showed a punctate pattern on the cell cortex (Fig. 1F; red arrowheads). Based on the analysis of over 150 cells, the localization defect was statistically significant (Fig. 1G). pSTE11-4 partially rescued the septin localization defect of the cdc24-4 mutant (Fig. 1F,G), which was dependent on Gic2p (Fig. 1F,G). The fMAPK pathway did not suppress the growth and morphological defects of the cdc12-6 mutant and septin kinase mutants, elm1Δ and gin4^G377T, which indicates that the pathway functions upstream of the septins themselves (Fig. S2). Gic2p also promotes the interaction between Cdc42p and Bni1p (Chen et al., 2012). Bni1p was not required for GFP-Cdc42p clustering or bud emergence (Fig. 1E,F) but was required for septin localization in the cdc24-4 mutant (Fig. 1F,G). Bni1p is required for septin ring assembly (Gladfelter et al., 2005; Kadota et al., 2004). In wild-type cells, Gic2p was required for invasive growth (plate-washing assay; Fig. S3A), colony ruffling (Fig. S3B) and septin organization (Fig. S3C–E) of filamentous cells. Gic2p did not regulate the fMAPK pathway (Fig. S3F). These results establish a new role for the fMAPK pathway in regulating bud emergence by a mechanism that involves the regulated expression of Gic2p. In the next section, we examine the role of the fMAPK pathway in regulating bud emergence in wild-type cells, under conditions of filamentous growth.

Like other monomeric GTPases, the exchange of GDP for GTP alters the conformation of Cdc42p and allows the protein to bind effector proteins. During budding, the increase in GTP-Cdc42p levels is critical for symmetry breaking and polarity establishment (Irazoqui et al., 2003; Kozubowski et al., 2008). Msb2p interacts with GTP-Cdc42p to activate the fMAPK pathway (Cullen et al., 2004). During filamentous growth, the level and activity of Msb2p is stimulated by positive feedback (Fig. S1D). To test whether Msb2p and the fMAPK pathway affect the levels of GTP-Cdc42p in the cell, a fluorescent reporter that measures GTP-Cdc42p levels was examined [Gic2p-PBD-tdTomato (Okada et al., 2013)]. Compared with cells undergoing yeast-form growth (YEPD), cells undergoing filamentous growth showed elevated levels of GTP-Cdc42p (Fig. 2A; YEP-GAL). The increase in GTP-Cdc42p levels was dependent on the fMAPK pathway (Fig. 2A) in a manner that was statistically significant (Fig. 2B; YEP-GAL; P<0.00001). As previously shown (Okada et al., 2013), the level of GTP-Cdc42p was also higher in cells responding to the mating pheromone α-factor (Fig. 2A,B). In this case, the increase was independent of the fMAPK pathway.

One caveat in interpreting the above results is that GIC2-PBD-tdTomato is driven by its endogenous promoter, which, as shown above, is regulated by the fMAPK pathway. As a separate test, the activity of a Cdc42p biosensor (Smith et al., 2013) was examined by fluorescence lifetime imaging [FLIM-FRET (Sun et al., 2011)]. In wild-type cells, the Cdc42p biosensor exhibited shorter lifetimes with a version that mimics the GTP-bound conformation (Fig. 2C; Cdc42p^Q61L) and longer lifetimes with a version that mimics the GDP-bound conformation (Fig. 2C; Cdc42p^D57Y). Activation of the fMAPK pathway reduced the lifetime of the biosensor (Fig. 2C; MSB2^Δ100-818). This change corresponded to a 25% increase in the total levels of GTP-Cdc42p, which has previously been shown to reflect changes in Cdc42p activity during bud emergence (Smith et al., 2013). Cdc42p activity can also be affected by the Gic proteins (Daniels et al., 2018; Kang et al., 2018; Kawasaki et al., 2003) but the lifetime of the biosensor was not altered in the gic1Δ gic2Δ double mutant (Fig. 2C). Hyperactive versions of Msb2p, Sho1p and Ste11p also partially suppressed the growth defect of the gic1Δ gic2Δ double mutant at 37°C (Fig. S3G; Gandhi et al., 2006), which supports a role for fMAPK in regulating bud emergence that is separate from Gic protein function.

Bud emergence occurs in the G₁ phase of the cell cycle (Gulli et al., 2000; Hartwell et al., 1970; Howell and Lew, 2012; Lew and Reed, 1993; Moran et al., 2019; Pringle et al., 1995). Whether the fMAPK pathway is active in the G₁ phase of the cell cycle has not been explored. To address this question, cells were synchronized in the G₁ phase of the cell cycle by α-factor (Breeden, 1997), and fMAPK pathway activity was assessed under basal (YEPD) and activating conditions (YEP-GAL) as cells progressed through the cell cycle. An epitope-tagged cyclin, Clb2p-HA, showed the expected pattern of cell-cycle regulation (Cepeda-García, 2017; Cross et al., 2005; Eluere et al., 2007; Irniger et al., 1995; Kuczera et al., 2010; Richardson et al., 1992; Wäsch and Cross, 2002), increasing after α-factor release by 60 min and decreasing by 90 min as cells entered anaphase (Fig. 3A; α-HA). P∼Kss1p levels also increased after α-factor release by 80 min and peaked by 100 min (Fig. 3A; P∼Kss1p). By comparison, the level of P∼Fus3p, which is the MAP kinase that regulates the mating pathway, decreased after release from α-factor (Fig. 3A; P∼Fus3p). Conditions that activate the fMAPK pathway (YEP-GAL) led to a delay in Clb2p-HA accumulation (Leitao and Kellogg, 2017) and also showed an increase in P∼Kss1p levels after the G₂/M phase of the cell cycle, prior to the next round of bud emergence (Fig. 3B). These results demonstrate that the activity of the fMAPK pathway is cell-cycle regulated and increases prior to and during bud emergence.

Bud emergence is spatially regulated in that it results from Cdc42p activation at bud sites (Caviston et al., 2003; Das et al., 2007; Freisinger et al., 2013; Okada et al., 2013). We previously showed that bud-site-selection proteins regulate the fMAPK pathway (Basu et al., 2016), which indicates that the fMAPK pathway is active at sites where bud emergence occurs. In line with this possibility, key regulators of the fMAPK pathway, including Sho1p (see below) and Ste20p are recruited to bud sites prior to bud initiation (Moran et al., 2019). Moreover, polarity targets of the fMAPK pathway included proteins that promote budding at bud sites (Fig. S1D). Polarity targets also included proteins that prevent budding at previous division sites or cytokinesis remnants. These included Rga1p (Fig. S1D), which along with other GTPase activating proteins (GAPs) prevents budding within the existing growth site (Kadota et al., 2004; Miller et al., 2017; Tong et al., 2007), and Rax2p (Fig. S1D), which restricts budding at previous division sites in a complex with Rax1p, Gps1p, Nba1p and Nis1p (Meitinger et al., 2014). Rga1p has been shown to negatively regulate the fMAPK pathway (Smith et al., 2002). Because these were negative regulators of the fMAPK pathway, cells were examined under pathway basal conditions (YEPD). Cells lacking Rax2p, Nba1p or Gps1p also showed elevated fMAPK pathway activity (Fig. 3C; Fig. S4A), which indicates that the negative polarity complex negatively regulates the fMAPK pathway. The elevated activity of the fMAPK pathway in negative polarity complex mutants also required Msb2p (Fig. 3D; Fig. S4A,B). Cells lacking the negative polarity complex did not have an impact on mating (Fig. S4C). Therefore, proteins that spatially promote budding promote fMAPK activity, and proteins that spatially restrict budding also restrict fMAPK pathway activity. Cells lacking other Cdc42p-interacting proteins, Boi1p and Boi2p (Bender et al., 1996; Glomb et al., 2020; Liao et al., 2013; Masgrau et al., 2017) and Msb1p (Bender and Pringle, 1991; Bi et al., 2000; Liao et al., 2013) were examined but did not show altered activity of fMAPK pathway or a difference in invasive growth compared to wild-type cells (Fig. S4D–I). Collectively, the data indicate that fMAPK pathway activity is temporally and spatially regulated to coincide with bud emergence.

To further define how the fMAPK pathway regulates bud emergence, the timing of bud emergence was examined using a septin marker (Cdc3p-mCherry) that shows a characteristic localization pattern throughout the cell cycle (Kim et al., 1991; Lippincott et al., 2001). In yeast-form cells grown in SD-URA (GLU), the timing of GFP-Cdc42p clustering, septin recruitment by Cdc3p-mCherry and bud emergence were similar between wild-type cells and an fMAPK pathway mutant (Fig. S5A). In filamentous wild-type cells grown in S-GAL-URA (GAL), GFP-Cdc42p clustering occurred at the incipient site by 10 min [Fig. 4A (green arrowheads) and 4B (Movie 1)] from the preceding round of cytokinesis marked by the septin hourglass split into a double ring (Fig. 4A; set as t=0; Cid et al., 2001; Lippincott et al., 2001). The recruitment of Cdc3p-mCherry also occurred by 10 min [Fig. 4A (red arrowheads) and 4B (Movie 1)]. The average time for bud emergence for 17 wild-type cells was 30 min from the preceding round of cytokinesis (Fig. 4B). Strikingly, in an fMAPK pathway mutant grown under filamentous conditions (GAL), a delay in GFP-Cdc42p clustering, Cdc3p-mCherry recruitment and bud emergence were observed (Fig. 4B). Although initial GFP-Cdc42p clustering and Cdc3p-mCherry recruitment occurred within the same time span (Fig. 4B,C; example 1, green and red arrowheads), some cells showed disappearance and reappearance of the GFP-Cdc42p cluster at the polarity site [Fig. 4C,D (oscillation seen in the pixel intensity graphs); Movies 2–4]. The disappearance and reappearance of the GFP-Cdc42p cluster, referred to here as transient disappearance of the polarity complex, accounted for a significant delay in bud emergence in the fMAPK pathway mutant (Fig. 4B,C). Some cells failed to make buds within the time of the experiment (Fig. 4D; Movies 3,4). In other examples, the initial polarity site disappeared and reappeared at a new site from where the bud emerged [Fig. 4E (in the mother cell, compare the location of green and red arrowheads with the black arrow that marks bud emergence); Movies 5,6]. The transient disappearance of the polarity complex, seen at some level in wild-type cells (although they always made a bud), was increased by 3-fold in the fMAPK mutant (Fig. 4F). Representing GFP-Cdc42p intensity in each cell as the coefficient of variation (CV) of pixel intensity (Lai et al., 2018) over time brought out the polarity defect in the fMAPK mutant relative to the wild-type (Fig. 4G). The CV of pixel intensity also showed a larger change in wild-type (0.16 to 0.23: 7) compared to the ste12Δ mutant (0.14 to 0.18: 4), which was statistically significant (P<0.01). These results define a function for the fMAPK pathway in stimulating the rate of bud emergence under conditions that promote filamentous growth.

To look at active Cdc42p during bud emergence in filamentous conditions, Gic2p-PBD-tdTomato reporter was co-localized with Sho1p-GFP, an fMAPK pathway regulator which is a direct effector of Msb2p (Cullen et al., 2004; Tatebayashi et al., 2007) that interacts with Cdc24p (Vadaie et al., 2008), Ste20p and Ste11p (Tatebayashi et al., 2006; Zarrinpar et al., 2004). Although Sho1p is known to localize to polarized sites (Pitoniak et al., 2009), whether it localizes to presumptive bud sites is not known. In wild-type cells, Sho1p-GFP was localized to incipient bud sites prior to bud emergence together with GTP-Cdc42p (Fig. S5B,C; wild-type, red arrow). After bud emergence, Sho1p-GFP was found at the bud tips as has been reported previously (Pitoniak et al., 2009). In tec1Δ mutant cells, Sho1p-GFP initially localized to the incipient bud site but failed to become enriched, and instead migrated back and forth along the distal pole (Fig. S5C; tec1Δ, black arrows). The low level of active Cdc42p in these cells rapidly disappeared. Bud formation was delayed and, in the cell shown, bud emergence did not occur within the time of the experiment. Thus, although active Cdc42p clusters at incipient bud sites, an intact fMAPK pathway is required to promote bud emergence under nutrient-limiting conditions. Cells lacking the fMAPK pathway showed a growth defect under filamentous conditions (Fig. S5D; GAL) and a defect in the rate of bud formation (Fig. S5E,F). In particular, at 5, 10 and 18 h, the ste12Δ mutant formed buds at a slower rate than wild-type cells. Therefore, the fMAPK pathway stimulates the rate of budding during filamentous growth.

Wild-type cells normally grow at a single site due to a regulatory phenomenon known as singularity in budding. Cells containing active versions of Cdc42p bypass this regulation and grow at multiple sites (Caviston et al., 2002; Howell et al., 2012, 2009; Richman and Johnson, 2000; Wedlich-Soldner et al., 2003, 2004). Cells containing hyperactive versions of Msb2p also had this property (Fig. 5A; Basu et al., 2016). Further inspection showed that 16% of cells carrying MSB2^Δ100-818 showed multiple growth sites (Fig. 5B; MSB2^Δ100-818). Given that the fMAPK pathway regulates bud emergence, the ability of Msb2p to induce multiple growth sites might be mediated by the fMAPK pathway. We found that the formation of multiple growth sites by MSB2^Δ100-818 required the fMAPK pathway (Fig. 5B; MSB2^Δ100-818 ste12Δ). Growth at multiple sites was also induced by STE11-4 or over-expression of SHO1 (Fig. 5B). Generally speaking, the formation of multiple growth sites correlated with fMAPK pathway activity (Fig. 5B,C). One exception was GAL-SHO1, which showed high levels of multiple growth site formation (Fig. 5B), yet modestly induced fMAPK pathway activity (Fig. 5C). This may be because over-expression of SHO1 induces a unique cell morphology where cells have hyper-elongated buds (Fig. 5D). The activity of the fMAPK pathway is stimulated by positive feedback, which is evident by immunoblot of the Kss1p protein (Fig. 5C, middle blot), whose levels are controlled by the fMAPK pathway (Roberts et al., 2000). In some fMAPK hyperactive mutants (MSB2^Δ100-818 and STE11-4) the levels of total Kss1p were lower than would be expected by positive feedback. Although the reason for this is not known, it could be due to the presence of negative feedback that acts to attenuate the activated pathway. Thus, Msb2p might itself induce multiple growth sites but require the fMAPK pathway for positive feedback. However, this was not the case. MSB2^Δ100-818 expressed from an inducible promoter (pGAL) that did not depend on the fMAPK pathway also required Ste12p to form multiple growth sites (Fig. 5B,C; compare GAL-MSB2^Δ100-818 with GAL-MSB2^Δ100-818 ste12Δ). Therefore, Msb2p activation itself is not sufficient to induce growth at multiple sites. The fact that MSB2^Δ100-818 is more effective at inducing growth at multiple sites than STE11-4 can be explained by fMAPK pathway activity.

Transcriptional targets of the fMAPK pathway may be required for the formation of multiple growth sites. Bni1p and Gic2p were required for multiple growth site formation by the fMAPK pathway (Fig. 5E). High-copy expression of GIC2 itself induced multiple growth sites (Fig. 5E), in line with a previous report (Jaquenoud et al., 1998). The negative polarity complex did not have an impact on growth at multiple sites (Fig. 5B; pSTE11-4 nba1Δ), although it did stimulate the fMAPK pathway activity (Fig. 5C). Collectively, these results support a role for the fMAPK pathway in the regulation of bud emergence, because activation of the fMAPK pathway can induce growth at multiple sites.

The multiple growth sites produced by the fMAPK pathway differed in shape from previous reports on multiple bud formation by hyperactivation of Cdc42p (Caviston et al., 2002; Wedlich-Soldner et al., 2003, 2004). Hyperactive Cdc42p induces symmetry breaking in the mother cell to trigger formation of a second bud. By comparison, the hyperactive fMAPK pathway did not affect budding in mother cells. The multiple growth sites were formed inside the growing bud (Fig. 6A). To our knowledge, this new phenotype has not been previously characterized. Growth at a second site within the growing bud was not due to abortive growth at the initial site, because polarity proteins including Cdc24p-GFP (Fig. 5D) localized to multiple sites. Actin dynamics in live cells using Abp140p-YFP, a marker for the actin cables (Asakura et al., 1998; Riedl et al., 2008; Yang and Pon, 2002), showed multiple growth sites in the hyperactive fMAPK mutants compared with wild-type cells (Fig. 6B). Actin staining in filamentous cells using rhodamine phalloidin also showed polarized actin cytoskeleton at more than one site (Fig. S6A–C). Cells expressing MSB2^Δ100-818 showed asymmetric clustering of GFP-Cdc42p (Fig. 6C; Movie 8). Over time, GFP-Cdc42p migrated along the bud cortex (Fig. 6C), which was evident by kymograph analysis (Fig. 6D), which allows tracking of changes in protein localization over time (Kaksonen et al., 2003). Sec3p-GFP localization corroborated that these sites were actively growing (Fig. S6C). Some wandering of GFP-Cdc42p occurred in wild-type cells (Movie 7), which might be due to the off-center delivery of vesicles that dilute the polarity complex (Chiou et al., 2017; Dyer et al., 2013).

The hyperactive fMAPK pathway may alter the dynamics of the interaction between Cdc42p and its effectors in a manner that makes the Cdc42p polarity axis ‘forget’ its orientation. This may mimic symmetry breaking occurring but within the bud cortex. In line with this possibility, versions of Cdc42p that contain point mutations in the effector-binding domain (such as Cdc42p^V36T), which impair its interaction with effector proteins (Gladfelter et al., 2002, 2005, 2001), showed a similar phenotype (Fig. 6E). Here multiple protrusions occurred adjacent to the previous polarization site. Similarly, cells expressing MSB2^Δ100-818 formed multiple growth projections (Fig. 6A; three, four or five sites). Other genes can induce multiple buds when over-expressed (Sopko et al., 2006), but these did not rescue the growth defect of the cdc24-4 mutant and in most cases made it worse (Fig. S6D). These results demonstrate the importance of a properly regulated fMAPK pathway in bud morphogenesis.

Here we show that bud emergence, which is one of the most intensively studied and well understood polarity establishment processes in eukaryotes, is regulated by a MAPK pathway. The fMAPK pathway regulates transcription of polarity target genes and GTP-Cdc42p levels to increase the rate of bud emergence during filamentous growth. In this way, bud emergence is regulated by a pathway whose activity is sensitive to extrinsic cues. Importantly, the ability of the fMAPK pathway to induce the expression of polarity target genes allows the pathway to tailor bud emergence by altering the levels of proteins that act at multiple steps in the polarity pathway.

The fMAPK pathway may also regulate bud emergence during vegetative growth. Ste20p is the first effector recruited by Cdc42p at bud sites (Moran et al., 2019), which is known to activate the fMAPK pathway at bud sites to regulate bud emergence. Ste12p and Tec1p are also required for viability of the cdc24-4 mutant under vegetative growth conditions. A function for the fMAPK pathway in regulating bud emergence might be masked by genetic buffering under normal growth conditions. Indeed, cells lacking TEC1 are synthetically lethal with a diverse class of cytoskeletal and cell-cycle regulatory genes (Costanzo et al., 2010). More recently, Rsr1p in its GDP-locked state has been shown to regulate the timing of Cdc42p polarization in early G₁ by interaction with Bem1p (Miller et al., 2019). Rsr1p (Basu et al., 2016) and Bem1p (Basu et al., 2020) both regulate the fMAPK pathway.

A key polarity target of the fMAPK pathway is Gic2p, which is a direct effector of Cdc42p that controls multiple steps in bud emergence. The regulation of GIC2 expression may be a key step in regulating polarity establishment in general. GIC2 expression is regulated by the fMAPK pathway and other proteins that regulate filamentous growth, including Phd1p (Gimeno and Fink, 1994; MacIsaac et al., 2006), Rim101p (Hu et al., 2007; Lamb and Mitchell, 2003), SAGA (Venters et al., 2011) and Rpd3p(L) (Hu et al., 2007; Venters et al., 2011). GIC2 expression is also regulated during mating (Roberts et al., 2000) and the Gic proteins are required for shmoo formation (Brown et al., 1997). Gic2p may function in other contexts as well, and has been implicated as an effector of the protein kinase C pathway (Zanelli and Valentini, 2005). Generally speaking, changes in gene expression may affect the regulation of bud emergence during cell differentiation and the response to stress.

Here and in Basu et al. (2016), we identify cross-talk between the polarity pathway and the fMAPK pathway. Cross-talk from the fMAPK pathway to the polarity pathway impacts bud emergence in several ways: (1) the activity of the fMAPK pathway is amplified by positive feedback to stimulate GTP-Cdc42p levels, (2) the fMAPK pathway induces target genes that encode polarity pathway components, and (3) the fMAPK pathway induces target genes that prevent growth at dormant growth sites (Meitinger et al., 2013). We also show that the activity of the fMAPK pathway is critical for proper bud emergence. Too little fMAPK pathway activity causes a delay in bud emergence during filamentous growth, and too much leads to growth at multiple sites. Precise regulation of the fMAPK pathway comes from activation at the right place (incipient bud sites) and time (at M/G₁). The fact that the fMAPK pathway is cell-cycle regulated is a novel finding. Cell-cycle regulation of the fMAPK pathway might occur through TEC1, which like other G₁ specific genes in the ‘SIC’ cluster (Cho et al., 1998; Spellman et al., 1998; Wittenberg and Reed, 2005) is induced at the M/G₁ boundary by the cell-cycle regulated transcription factor Swi5p (Spellman et al., 1998).

During vegetative growth in S. cerevisiae, Cdc42p-dependent budding is dictated by bud-site-selection proteins (Chiou et al., 2017). In other fungal species, diverse mechanisms of polarity control predominate. In Schizosaccharomyces pombe, growth at the two poles is maintained by regulated oscillations in Cdc42p activity (Das et al., 2012). Changes in the orientation of the polarity complex occur during mating (Dyer et al., 2013; Nern and Arkowitz, 2000) and filamentous growth (this study), and may be a common feature of fungal cell differentiation. In the filamentous fungus Ashbya gossypii, Rsr1p regulates symmetric growth cone formation at the hyphal tip (Bauer et al., 2004). In the major human fungal pathogen Candida albicans, Rsr1p and Cdc42p are also required for proper hyphal growth (Si et al., 2016). MAPK pathways may be critical regulators of Cdc42p-dependent morphogenesis during hyphal/pseudohyphal growth. Parallels can also be drawn to higher eukaryotes. In mammals, the ERK pathway is directly involved in breaking radial symmetry of spreading RAT2 fibroblast cells (Klímová et al., 2016). In this case, ERK spatially restricts p190A-RhoGAP activity to limit growth at the cell rear. Functional cross-talk between MAPK pathways and Rho GTPases may constitute an under-explored regulatory circuit in many systems.

Yeast strains are described in Table S1. Plasmids are listed in Table S2. Gene disruptions and GAL1 promoter fusions were made by polymerase chain reaction (PCR)-based methods (Baudin et al., 1993; Longtine et al., 1998). Some gene disruptions were made with antibiotic resistance markers KanMX6 (Longtine et al., 1998), HYG and NAT (Goldstein and McCusker, 1999). Internal epitope fusions were made by the pop-in, pop-out strategy (Schneider et al., 1995). Some strains were made ura3- by selection on 5-fluoroorotic acid (5-FOA). Gene disruptions were confirmed by PCR Southern analysis and confirmed by phenotype when applicable.

The pRS series of plasmids (pRS315 and pRS316) have been described (Sikorski and Hieter, 1989). To construct cdc24-4 mutant (PC6936), NAT cassette in PC6077 was replaced with URA3 using pop-in, pop-out strategy (Schneider et al., 1995). Following complementary primers were used: 5′- CAGAAGAGTACCATTGCTGTTATCATTTGTTGCCTAGCCCTATCAACTAAAGGGAACAAAAGCTGG-3′ and 5′-CAATATAATGGGTATAGCTTGAACATCTGCCCCTCTCTATCTAATGTCACTATAGGGCGAATTGG-3′. The strain was subsequently made ura3Δ by selection on 5-FOA. cdc24Δ was confirmed by phenotypic analysis. Msb2p (pGFP-MSB2; Adhikari et al., 2015), Sho1p (pRS316-SHO1^D16H,P120L-GFP or SHO1^P120L; Vadaie et al., 2008) and Ste11p (YCp50-STE11-4; Stevenson et al., 1992) have been described. Plasmid pRS316-SHO1-GFP has been described (Marles et al., 2004) and was provided by Dr Alan Davidson (University of Toronto, Ontario, Canada). In SHO1^D16H, the D16H mutation alone does not hyperactivate the fMAPK pathway, but together with P120L mutation results in a hyperactive version of the protein. Over-expression constructs came from an ordered collection (Gelperin et al., 2005). YEp352-TEF2-GIC2 was created by amplifying the GIC2 open reading frame (ORF) by PCR from genomic DNA using complementary primer pairs 5′-GACTCTAGAATGACTAGTGCAAGTATTACCAATACTGGAAACG-3′ and 5′-CAGTGTCGACTTAAGTTTGCAGGGGCTCGAGC-3′. The PCR product was digested with XbaI and SalI and inserted into the YEp352-TEF2 plasmid (Pitoniak et al., 2015) digested with the same enzymes. Positive clones were confirmed by DNA sequencing.

Plasmid YCplac181-CDC12-GFP::KanMX6 was constructed from YCplac181-CDC12-GFP (PC1365), provided by John Pringle (Stanford University) (Fares et al., 1996), by homologous recombination. The KanMX6 cassette was amplified by PCR with flanking sequences to LEU2 locus on the YCplac181-CDC12-GFP plasmid. Complementary primers were designed: 5′-ATGTCTGCCCCTAAGAAGATCGTCGTTTTGCCAGGTGACCACGTTGCGGATCCCCGGGTTAATTA-3′ and 5′-TTAAGCAAGGATTTTCTTAACTTCTTCGGCGACAGCATCACCGACTGAATTCGAGCTCGTTTAAAC-3′ (Millipore Sigma, St Louis, MO, USA). The PCR product was co-transformed with YCplac181-CDC12-GFP digested with AflII to target integration at the LEU2 locus. Plasmids were rescued from isolates that were KanMX6 positive. YCplac181-CDC12-GFP::NAT was made by the same strategy.

pRS316-GFP-CDC42::KanMX6 was made from pRS316-GFP-CDC42 (PC6454) by targeting integration at the URA3 locus using complementary primers 5′-ATGTCGAAAGCTACATATAAGGAACGTGCTGCTACTCATCCTAGTCGGATCCCCGGGTTAATTA-3′ and 5′-TTAGTTTTGCTGGCCGCATCTTCTCAAATATGCTTCCCAGCCTGGAATTCGAGCTCGTTTAAAC-3′ (Millipore Sigma). pRS316-pACT1-EGFP-CRIBCLA4-CDC42^ΔCAAX-mCherry-CAAX::KanMX6 was made from pRS316-pACT1-EGFP-CRIBCLA4-CDC42^ΔCAAX-mCherry-CAAX (PC7137) (Smith et al., 2013) and pSHO1^D16H-GFP::KanMX6 was made from pSHO1^D16H-GFP (PC1964) by targeting integration at the URA3 locus using the same complementary primers as for pRS316-GFP-CDC42::KanMX6. The PCR product was co-transformed with plasmids digested with BsmI to target integration at the URA3 locus. Plasmids were rescued from isolates that were KanMX6 positive. pRS316-GFP-CDC42::NAT and pSHO1^D16H-GFP::NAT were made by the same strategy.

Plasmids were recovered from cells with Promega Wizard Plus SV Miniprep DNA Purification system (A1460, Promega, Madison, WI, USA) used as specified by the manufacturer with the following modifications: 100 μl of glass beads were added at the lysis step, and cells were vortexed for 5 min before the addition of neutralization buffer. Rescued plasmids were transformed into Escherichia coli, and positive clones were confirmed by DNA sequencing, by GENEWIZ (South Plainfield, NJ, USA) and by fluorescence microscopy after transformation in yeast.

Yeast and bacterial strains were manipulated by standard methods (Rose et al., 1990; Sambrook et al., 1989). The single-cell invasive growth assay (Cullen and Sprague, 2000) and the plate-washing assay (Roberts and Fink, 1994) were performed as described. Actin staining by rhodamine phalloidin was performed as described (Yuzyuk and Amberg, 2003). In cells lacking an intact mating pathway (ste4Δ), the FUS1-HIS3 reporter (McCaffrey et al., 1987) was used to evaluate fMAPK pathway activity (Cullen et al., 2004). FUS1-HIS3 activity was measured by spotting equal concentrations of cells onto SD-HIS medium and SD-HIS medium containing 3-amino-1,2,4-triazole (3-ATA).

Halo assays were performed as described (Jenness et al., 1987). A saturated culture of cells (A₆₀₀=0.1) was spread on YEPD medium and allowed to dry. α-Factor (3 μl and 10 μl, at 1 mg/ml) was spotted on the plates. Plates were incubated at 30°C and photographed at 24 and 48 h. Experiments were performed in triplicate. Halo diameter (in centimeters) was measured by ImageJ analysis and plotted as a function of α-factor concentration. For Lat-A sensitivity, saturated culture of cells (A₆₀₀=0.1) was spread onto SD-URA-LEU medium containing 0.5 M sorbitol. Lat-A (10 μl; L5163, Sigma) at 0, 0.1, 0.2 and 0.5 mM (in 10 mM DMSO) were spotted on plates. Halo diameter (in centimeters) was measured by ImageJ analysis and plotted as a function of Lat-A concentration.

Immunoblot analysis was used to detect phosphorylated MAP kinases as described (Basu et al., 2016; Lee and Dohlman, 2008; Sabbagh et al., 2001). In cdc24-4 mutant combinations, cells were grown to mid-log in 10 ml SD-URA at 30°C and then incubated at 37°C for 4 h. Cells were harvested by centrifugation. Pellets were washed once with water and flash-frozen in liquid nitrogen. In experiments that did not involve cdc24-4, cells were grown in YEPD or YEP-GAL media for the times indicated. Proteins were precipitated by trichloroacetic acid (TCA) and analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (10% acrylamide). Proteins were transferred to nitrocellulose membranes (Amersham Protran Premium 0.45 μm NC; 10600003, GE Healthcare Life Sciences).

Cell synchronization experiments were performed as described (Breeden, 1997). The strain harboring Clb2p-HA (PC2744) was transformed with a plasmid containing the STE4 gene (pSTE4; Stevenson et al., 1992). Cells were grown to an optical density (OD) A₆₀₀ of 0.2 in SD-URA medium. Cells were washed and resuspended in equal volume of YEPD and incubated for 90 min at 30°C. α-Factor was added to a final concentration of 5 μg/ml and the culture was incubated for 90 min to arrest cells in the G₁ phase of the cell cycle. Arrested cells were washed twice with water and resuspended in fresh YEPD or YEP-GAL medium to release cells into the cell cycle. Aliquots (10 ml) were harvested every 10 min, flash-frozen in liquid nitrogen and stored at −80°C.

ERK-type MAP kinases (P∼Kss1p and P∼Fus3p) were detected using p44/42 antibodies (no. 4370; Cell Signaling Technology, Danvers, MA, USA) at a 1:5000 dilution. Kss1p was detected using α-Kss1p antibodies (no. 6775; Santa Cruz Biotechnology, Santa Cruz, CA, USA) at a 1:5000 dilution. Clb2p-HA was detected using the α-HA antibody at a 1:5000 dilution (no. 12CA5; Roche Diagnostics). For secondary antibodies, goat anti-rabbit IgG-HRP antibodies were used at a 1:10,000 dilution (no. 111-035-144; Jackson ImmunoResearch Laboratories, West Grove, PA, USA). Mouse α-Pgk1p antibodies were used at a 1:5000 dilution as a control for total protein levels (no. 459250; Thermo Fisher Scientific, Waltham, MA, USA). Goat α-mouse secondary antibodies were used at a 1:5000 dilution to detect primary antibodies (no. 170-6516; Bio-Rad Laboratories, Hercules, CA, USA). Blots were visualized by chemiluminescence using a Bio-Rad ChemiDoc XRS+ system (1708265, Bio-Rad). Quantitation of band intensities for immunoblot analysis was performed with Image Lab Software (Bio-Rad). For blots to evaluate phosphorylated MAP kinase proteins, membranes were incubated in 1×TBST (10 mM TRIS–HCl, pH 8, 150 mM NaCl, 0.05% Tween-20) with 5% bovine serum albumin. For other immunoblots, membranes were incubated in 1×TBST with 5% non-fat dried milk. All primary incubations were carried out for 16 h at 4°C. Secondary incubations were carried out at 25°C for 1 h.

Wild-type and temperature-sensitive mutants containing desired plasmids were grown to saturation in SD-URA media. For each strain, 0.1 OD₆₀₀ of cells were serially diluted four times in distilled water and spotted onto SD-URA-LEU+ 0.5 M sorbitol plates. The plates were incubated at 30 and 37°C. For cdc12-6 suppression, cells were spotted onto SD-URA plates and incubated at 24 and 30°C as the cdc12-6 mutant had a severe growth defect at 30°C and failed to grow completely at 37°C. For all suppression assays, plates were photographed every day for 4 days using an Evolution MP Color Camera (Media Cybernetics) and Q Capture software. Images were imported into ImageJ software. Cells of interest were selected and the measure tool was used to generate values for the integrated density, area and the mean signal intensity. Growth of a colony was quantified by measuring the signal intensity of the colony against the background using the formula: corrected signal intensity=integrated density−(area of selected cell×mean signal intensity of the background cell). Growth at 37°C was compared with growth at 30°C for each strain and then normalized to wild-type. Quantitation was reported for the day where the temperature-sensitive mutant containing STE11-4 plasmid showed around 60% of growth compared with the wild-type, which was the maximum growth detected across various mutant combinations and independent trials. For most analyses, only the first or the second dilution of cells was used for the quantitation. All growth assays were performed in triplicate. Error bars show standard error of mean (s.e.m.) among the three trials.

Cells for qPCR were concentrated (OD A₆₀₀=20) and spotted in 10 µl aliquots onto YEP-GAL (2% agar) for 24 h. Cells were spotted in six colonies per plate equidistant to each other and the plate center. All six colonies were harvested for each trial, and two separate trials were compared for each strain. The entire colony surface was scraped into 500 µl of distilled water, harvested by centrifugation, washed and stored at −80°C. RNA was harvested by hot acid phenol chloroform extraction as described previously (Adhikari and Cullen, 2014). Samples were further purified using Qiagen RNeasy Mini Kit (catalog no. 74104). RNA concentration and purity were measured using NanoDrop (NanoDrop 2000C). RNA stability was determined by 1% agarose Tris-borate-EDTA (TBE, 89 mM Tris base, 89 mM boric acid, 2 mM EDTA) gel electrophoresis.

cDNA libraries from RNA samples were generated using iScript Reverse Transcriptase Supermix (no. 1708840; Bio-Rad). qPCR was performed using iTaq Universal SYBR Green Supermix (no. 1725120; Bio-Rad) on a Bio-Rad CFX384 Real-Time System. Fold changes in expression were determined by calculating ΔΔCt using ACT1 mRNA as the housekeeping gene for each sample. Experiments were performed with biological replicates, and the average of multiple independent experiments was recorded. Primers for pPCR for GIC2 were: forward 5′-GCGCCAACAAGACAAATCACAAA-3′ and reverse 5′-GCAATTGCTCATCTTGGAATCC-3′. Primers for GIC1 were: forward 5′-GCCGAACAAGAACAACATCAA-3′ and reverse 5′-GTTTTGGCAGACCCATGTCTC-3′. Primers for ACT1 were: forward 5′-GGCTTCTTTGACTACCTTCCAACA-3′ and reverse 5′-GATGGACCACTTTCGTCGTATTC-3′ as published (Chavel et al., 2010).

Differential interference contrast (DIC) and fluorescence microscopy using FITC, Cyan Fluorescent Protein (CFP), Yellow Fluorescent Protein (YFP), rhodamine and DAPI filter sets were performed using an Axioplan 2 fluorescent microscope (Zeiss) with a Plan-Apochromat 100×/1.4 (oil) objective (NA 1.4) (cover slip 0.17) (Zeiss). Digital images were obtained with the Axiocam MRm camera (Zeiss) and Axiovision 4.4 software (Zeiss). Adjustments to brightness and contrast were made in Adobe Photoshop. Some images were obtained using structural deconvolution with a Zeiss Apotome filter. Multiple polarization events were assigned by examining cells over multiple focal planes by DIC and fluorescence microscopy. Time-lapse microscopy was performed on a Zeiss 710 confocal microscope equipped with a Plan-Apochromat 40×/1.4 oil DIC M27 objective on a heated stage at 37°C. GFP was imaged using 488 nm laser excitation and an emission window from 496 to 572 nm. Time-lapse Z-stacks were captured at 10 and 20 min intervals to monitor bud emergence and Cdc42p clustering.

Cells were grown for 16 h on SD+AA medium. Using a toothpick, cells were removed from the plate, washed twice with water and resuspended in 1 ml water. Cells (50 μl) were spread onto SD+AA or S-GAL+AA media and incubated at 30°C. Cells were visualized at 0, 5, 10 and 18 h intervals at 20×. Budding rate, ln(n)/ln(2), was determined as described (Hall et al., 2014), where n is the number of daughters produced. Budding rate was adjusted for the time interval (5, 5 and 8 h). More than 30 cells were examined for each interval.

For time-lapse microscopy, cells were placed onto semi-solid agarose pads that were prepared as described (Skinner et al., 2013) with the following modifications. Approximately 700 μl of SD-URA medium, prepared with agarose (1%), was placed inside 12 mm Nunc glass base dishes (no. 150680; Thermo Fisher Scientific, Waltham, MA, USA) and allowed to set at 25°C for 5 min.

For GFP-Cdc42p clustering in cdc24-4 mutant combinations, cells were taken from colonies grown at 30°C for 16 h on SD-URA+G418 semi-sold agar media [2% agarose, 0.67% yeast nitrogen base (YNB) without ammonium sulfate, 0.1% monosodium glutamate (MSG), 2% dextrose, 1× amino acid stock without uracil, 0.36 mg/ml G418], resuspended in 7 μl of synthetic broth (0.67% YNB without ammonium sulfate, 0.1% MSG, 2% dextrose), and placed under the agarose pads by gently lifting the pad with a scalpel. Water (100 μl) was placed in the dish adjacent to the agarose pad to prevent moisture loss, and the Petri dishes were incubated at 37°C for 4 h. For co-localization of Sho1p-GFP and Gic2p-PBD-tdTomato, cells were prepared as above except that they were imaged at 30°C. For GFP-Cdc42p clustering in cells carrying MSB2^Δ100-818, 500 μl of saturated culture was washed twice with water and resuspended in 500 μl water. An aliquot of 5 μl was placed under the agarose pad, and the Petri dish was incubated at 30°C for 10 h before imaging. For strains expressing GFP-Cdc42p and Cdc3p-mCherry, cells were prepared from colonies grown at 30°C for 16 h on SD-URA.

Plasmids containing the Cdc42p biosensor have been described (Smith et al., 2013) and were a generous gift from Dr Rong Li (Johns Hopkins University, Baltimore, MD, USA). Cells containing pRS316-pACT1-eGFP-CRIBCLA4-Cdc42p^ΔCAAX–mCherry-CAAX (PC7137) and Q61L (PC7138) or D57Y (PC7139) derivatives were grown on SD-URA medium to maintain selection for the plasmids. Semi-solid agarose pads were prepared as described above. Cells were taken from colonies grown for 16 h on semi-solid agar medium (SD-URA), resuspended in 7 μl of SD broth, and placed under the agarose pads.

Confocal microscopy and FLIM was performed at the SUNY-Buffalo North Campus Confocal Imaging Facility. FLIM-FRET has been described (Bassard and Halkier, 2018; Gratton et al., 2003; Osterlund et al., 2015; Padilla-Parra et al., 2015; Periasamy and Clegg, 2009; Sun and Periasamy, 2015). FLIM-FRET images were acquired with a Simple Tau TCSPC 150 and HPM-100-40 GaAsP detector (Becker & Hickl, Berlin, Germany) employing the direct coupled port of the LSM 710 and Zeiss Intune Laser excitation at 490 nm. At least 1000 photons per pixel were acquired for the highest count region for each cell.

FLIM-FRET data analysis was performed using SPCImage 7.3 (Becker & Hickl). Data were fitted by a two-component exponential decay model and automatically generated instrument response function (IRF). Bin factor, shift and offset were adjusted to obtain good fit (low χ²). For measuring the average lifetime, the decay curve was pooled for the region of interest. The mean fluorescence lifetime for the two-component decay model was calculated according to the equation: τ_m=a₁τ₁+a₂τ₂, where a_i and τ_i are the amplitude and lifetime of the ith component and τ_m is the average lifetime of the donor fluorescence.

Scanning electron microscopy was based on established methods (Piccirillo and Honigberg, 2011) and performed as described (Basu et al., 2016). For some experiments, cells were grown for 16 h in liquid medium at 30°C. Cells were washed in 0.1 M sodium phosphate buffer at pH 7.4 and diluted to about 10⁶ cells, which were passed over 0.2 μm Whatman nucleopore polycarbonate filter paper (catalog no. 889-78084GE Whatman) with a 10 ml syringe (no. 309604 BD Syringe). Cells were rinsed with one round of buffer by syringe, fixed with 1% glutaraldehyde for 15 min, and rinsed again. Cells were treated by a graded series of ethanol washes (30, 50, 70, 85 and 100%) by syringe to dehydrate the samples. The filter paper was removed from the holder, placed in a Petri dish and treated with hydroxymethyldiazane (HMDS). Samples were placed at 4°C for 16 h and imaged the following day. All solutions were filter sterilized before use and stored in clean containers free of corrosion products. Samples were carbon coated and imaged on a field emission scanning electron microscope (Hitachi SU70).

Gic2p-PBD-tdTomato clustering was quantified as described in Okada et al. (2017) with the following modifications. Raw fluorescence and DIC images were imported in ImageJ. DIC images were used to draw the cell boundary with the polygon tool. The same region of interest (ROI) was applied to the fluorescence image. Signal intensities for all pixels inside the ROI were exported as a CSV file using save XY coordinates option under the Analyze tool menu. A custom MATLAB (MATLAB R2016b, The MathWorks, Natick, MA, USA) code (available on request) was designed to analyze the CSV files. In the analysis, mean and standard deviation (s.d.) of all the pixels of a cell were calculated. The pixel intensity that was greater than the mean+2 s.d. was selected. A region of the image where no cell was present was chosen as the background and the average pixel intensity of the background region was subtracted from the selected pixels of the cell. The signal intensity for each of these background subtracted pixels was normalized to the peak value which was set to 1. The sum of these normalized pixels was used to represent each cell. Over 60 cells were measured for each sample.

Time-lapse microscopy images were processed in ImageJ. Grayscale fluorescence images were converted into maximum intensity projection and inverted. Quantitation of timing of Cdc42p clustering was performed as described (Lai et al., 2018; McClure et al., 2016) with following modifications. Threshold was applied to the 16-bit grayscale image to highlight cells and converted to binary image. Any merged cells were separated using the watershed plugin after applying a Gaussian Blur filter with sigma value of 2. Using Analyze Particles tab, the binary image was used to analyze the grayscale image to calculate mean intensity and s.d. of each cell. CV was then calculated from the mean and s.d., and plotted as a function of time.

Kymographs were made on inverted maximum intensity projection images. Brightness and contrast were adjusted uniformly for each time frame and each strain. Line tool with pixel width of 1 was used to define the ROI and Reslice tool was used to generate slices for each time frame.

GO term analysis (Ashburner et al., 2000) was performed using the Gene Ontology enRIchment analysis and visuaLizAtion algorithm (GOrilla) (Eden et al., 2007, 2009) using the two unranked lists mode. Genes encoding proteins associated with cell polarity were identified by GO term analysis (Ashburner et al., 2000; GO:0000282, cellular bud site selection; GO:0030010, establishment of cell polarity; GO:0031106, septin ring organization; GO:0006887, exocytosis; GO:0007120, axial cellular bud site selection; GO0007121, bipolar cellular bud site selection; GO:0030427, site of polarized growth). GO terms and descriptions come from SGD (http://www.yeastgenome.org). Statistical significance for pairwise comparisons was performed using Student's t-test in Microsoft Excel. For multiple comparisons, one-way ANOVA with Tukey’s test was performed in Minitab (www.minitab.com).

We thank Matthias Peter (ETH Zürich), D. Lew (Duke University), John Pringle (Stanford University), Rong Li (Johns Hopkins University), Charlie Boone (University of Toronto), Alan Davidson (University of Toronto), Scott Emr (Cornell University), Peter Novick (Yale University), Erfei Bi (University of Pennsylvania) and Mike Yu (SUNY-Buffalo) for providing strains and/or plasmids. We also thank Steve Free (SUNY-Buffalo) for comments. Karnesh Jain (University of North Dakota) wrote the MATLAB code for quantifying Gic2p-PBD-tdTomato signal. Nadia Vadiae, Heather Dionne, Marysa Notaro, Pu Zheng, Minakshi Mukherjee, Matt Vandermeulen and Shally Lin helped with experiments.

The peer review history is available online at https://jcs.biologists.org/lookup/doi/10.1242/jcs.241513.reviewer-comments.pdf