New Aspects of Invasive Growth Regulation Identified by Functional Profiling of MAPK Pathway Targets in Saccharomyces cerevisiae

MAPK pathways are universal regulators of morphogenesis and stress responses in eukaryotes. In this study, Vandermeulen and Cullen identified the function of transcriptional targets of a MAPK pathway in yeast....

Abstract

MAPK pathways are drivers of morphogenesis and stress responses in eukaryotes. A major function of MAPK pathways is the transcriptional induction of target genes, which produce proteins that collectively generate a cellular response. One approach to comprehensively understand how MAPK pathways regulate cellular responses is to characterize the individual functions of their transcriptional targets. Here, by examining uncharacterized targets of the MAPK pathway that positively regulates filamentous growth in Saccharomyces cerevisiae (fMAPK pathway), we identified a new role for the pathway in negatively regulating invasive growth. Specifically, four targets were identified that had an inhibitory role in invasive growth: RPI1, RGD2, TIP1, and NFG1/YLR042c. NFG1 was a highly induced unknown open reading frame that negatively regulated the filamentous growth MAPK pathway. We also identified SFG1, which encodes a transcription factor, as a target of the fMAPK pathway. Sfg1p promoted cell adhesion independently from the fMAPK pathway target and major cell adhesion flocculin Flo11p, by repressing genes encoding presumptive cell-wall-degrading enzymes. Sfg1p also contributed to FLO11 expression. Sfg1p and Flo11p regulated different aspects of cell adhesion, and their roles varied based on the environment. Sfg1p also induced an elongated cell morphology, presumably through a cell-cycle delay. Thus, the fMAPK pathway coordinates positive and negative regulatory proteins to fine-tune filamentous growth resulting in a nuanced response. Functional analysis of other pathways’ targets may lead to a more comprehensive understanding of how signaling cascades generate biological responses.

SIGNAL transduction pathways mediate cellular responses, which can include the response to stress, cell differentiation, and morphogenetic changes. One type of signaling pathway that functions in eukaryotes as a driver of development and stress responses are mitogen-activated protein kinase (MAPK) pathways, which regulate transcription factors that modify gene expression to induce a cellular response (Seger and Krebs 1995; Madhani et al. 1999; Chang and Karin 2001; Zeitlinger et al. 2003; Seger 2010; Morrison 2012). Because transcription factors can have many transcriptional targets, the individual functions of all targets must be considered to understand the complete phenotype of a signaling pathway. Thus, characterizing the transcriptional targets of a MAPK pathway may lead to new insights into the regulation of biological responses.

In the budding yeast Saccharomyces cerevisiae, the filamentous growth MAPK (fMAPK) pathway is one of multiple pathways that regulates the cellular response to nutrient limitation known as filamentous growth (Carlson et al. 1981; Gimeno et al. 1992; Lorenz and Heitman 1998; Pan and Heitman 1999, 2000; Cullen and Sprague 2000, 2012; Crespo et al. 2002; Lamb and Mitchell 2003; Borneman et al. 2006; Chavel et al. 2010, 2014; González et al. 2017; Norman et al. 2018; Mutlu et al. 2019; Brito et al. 2020). Filamentous growth occurs in many fungal species, and, in pathogenic fungi, such as the human pathogen Candida albicans, it is critical for virulence, making filamentous growth an important aspect of fungal biology (Lo et al. 1997; Wendland 2001; Nobile et al. 2006; Sohn et al. 2006; Labbaoui et al. 2017; Zhao et al. 2018; Brito et al. 2020). Filamentous growth involves a switch from yeast-form growth (round cell morphology) to filamentous-form growth, where cells produce filament-like structures. The filament-like structures result from three major changes to the cell: an increase in cell length, a reorganization of cell polarity, and increased cell-to-cell adhesion (Roberts and Fink 1994; Cullen and Sprague 2012). Filamentous growth causes cells to invade into substrates, a behavior called invasive growth (Roberts and Fink 1994). Invasive growth is presumed to be a scavenging response for cells to search for nutrients because it is mainly induced by nutrient limitation, such as fermentable carbon source (Cullen and Sprague 2000, 2012) and nitrogen (Gimeno et al. 1992) limitation. It can also be induced by high cell density through quorum sensing molecules (Chen and Fink 2006; González et al. 2017; Lenhart et al. 2019). When cells adhere and invade together in high cell density, they can form a gouge into surfaces, which is called aggregate invasive growth (Chow et al. 2019a).

The fMAPK pathway controls the activity of transcription factors that includes Ste12p and Tec1p [Figure 1A, (Gimeno et al. 1992; Gimeno and Fink 1994; Borneman et al. 2006; Heise et al. 2010; Cullen and Sprague 2012; van der Felden et al. 2014)]. These proteins induce the expression of many target genes (Madhani et al. 1999; Roberts et al. 2000; Heise et al. 2010; Adhikari and Cullen 2014; van der Felden et al. 2014; Chow et al. 2019b; Zhou et al. 2020). Several highly induced targets of the fMAPK pathway positively regulate filamentous growth, such as BUD8, which encodes a protein involved in bud-site-selection at the distal pole [Figure 1A, (Zahner et al. 1996; Taheri et al. 2000; Ni and Snyder 2001; Cullen and Sprague 2002)]; FLO11, which encodes the major cell adhesion mucin-like flocculin [Figure 1A, (Lambrechts et al. 1996; Lo and Dranginis 1996, 1998; Madhani et al. 1999; Rupp et al. 1999; Guo et al. 2000; Cullen and Sprague 2012)]; and CLN1, which encodes a G₁ cyclin (Hadwiger et al. 1989), whose induction leads to a delay in the cell cycle resulting in an elongated cell morphology [Figure 1A, (Loeb et al. 1999; Madhani et al. 1999)]. Many other transcriptional targets remain uncharacterized, raising the possibility that the fMAPK pathway may have unappreciated roles in regulating filamentous growth.

Figure 1.

Phenotypic analysis for invasive growth of transcriptional targets of the fMAPK pathway identified by comparative RNAseq analysis. (A) A model for the MAPK pathway that regulates filamentous growth by inducing target genes (green) that promote cell adhesion (FLO11), cell elongation at G₁ (CLN1), and distal-pole budding (BUD8). Pathway components are highlighted in red (MSB2, KSS1, STE12, TEC1). Not all pathway components are shown. (B) Portion of a volcano plot showing RNAseq data from a previous study (Adhikari and Cullen 2014). x-axis, log₂(FC); y-axis, −log₁₀(P-value). Fold change in gene expression between ste12∆ and wild-type cells grown in YP-Gal (2%) for 5.5 hr. All genes labeled have │log₂(FC)│ > 0.85 and P-value < 2.5 × 10⁻¹¹. Transposable elements and dubious open reading frames not shown in the graph. Green text, genes tested in the study. Yellow, more invasive mutant phenotype. Blue, less invasive mutant phenotype. Black, no invasive mutant phenotype. Gray, not tested.

A longstanding problem surrounding fMAPK pathway targets has been identifying phenotypes. One reason may be that some genes have a phenotype only noticeable under some conditions. Another reason is that targets might only contribute to a phenotype in a small way, if the cumulative effect of many genes is required to produce a phenotype. This means that some targets might have subtle phenotypes that could be overlooked. By examining cells lacking individual fMAPK pathway target genes under a variety of conditions for subtle but reproducible phenotypes, we identified new roles for five fMAPK pathway targets. One unexpected discovery that came from this approach was that the fMAPK pathway, which positively regulates invasive growth, can also negatively regulate aspects of invasive growth under some conditions. The other unexpected finding came from the characterization of a newly identified target, the transcription factor SFG1 (Fujita et al. 2005; White et al. 2009), which enabled the fMAPK pathway to regulate cell adhesion and the cell cycle by multiple mechanisms. Our study suggests that these new functions for the fMAPK pathway provide an additional level of versatility, which presumably allows for more nuanced responses in different environments. Therefore, characterizing the targets of a signaling pathway can lead to new insights about how pathways regulate biological responses.

Materials and Methods

Yeast strains and plasmids

Yeast strains are listed in Table 1. Gene deletions were made through homologous recombination, constructed using auxotrophic or antibiotic resistance markers amplified by polymerase chain reaction (PCR) and introduced into yeast by lithium acetate transformation as described (Gietz 2014). Primers for PCR are listed in Supplemental Material, Table S1. Strains were verified by PCR southern analysis and phenotype, when possible. All strains are isogenic with HYL333 of the ∑1278b background [provided by G. Fink, Whitehead Institute for Biomedical Research, Cambridge, MA, (Liu et al. 1993)]. pRS316 plasmid is a control vector containing URA3 as described in Sikorski and Hieter (1989) for experiments that use ura^– strains. Yeast extract, peptone, dextrose (YPD) medium was used at the concentration of glucose specified (2%, 10%, 16%). For high osmolarity medium, sorbitol (sorb) is added to YPD medium (2% Glu + 8% Sorb). YP-GAL (2%), YPD medium except 2% galactose is used instead of dextrose. Synthetic complete medium (yeast nitrogen base without amino acids, dextrose (2%) or galactose (2%), amino acids) was also used. SD+AA, synthetic media with dextrose and amino acids; SD-URA, synthetic media with dextrose and amino acids minus uracil. SGAL-URA, synthetic medium with galactose and amino acids minus uracil. SLAD, synthetic low ammonium, dextrose (2%) (Gimeno et al. 1992).

Table 1. Yeast strains used in this study.

Strain	Description	Reference
PC538	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52	Cullen et al. (2004)
PC539	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 ste12::URA3	Pitoniak et al. (2009)
PC611	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 ste11::URA3	Cullen and Sprague (2002)
PC1029	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 flo11::KanMX6	Karunanithi et al. (2010)
PC2043	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 FLO11-HA at 1000aa	Karunanithi et al. (2010)
PC2712	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 GAL-FLO11	Karunanithi et al. (2010)
PC3039	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 Msb2-HA at 500aa dig1::KlURA3	Chavel et al. (2010)
PC7144	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 sfg1::KlURA3	This study
PC7145	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 rpi1::KlURA3	This study
PC7146	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 rgd2::KlURA3	This study
PC7147	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 nfg1::KlURA3	This study
PC7164	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 dse1::KlURA3	This study
PC7165	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 dse2::KlURA3	This study
PC7166	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 dse4::KlURA3	This study
PC7167	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 svs1::KlURA3	This study
PC7168	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 ndj1::KlURA3	This study
PC7169	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 prm5::KlURA3	This study
PC7170	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 wsc2::KlURA3	This study
PC7198	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 scw11::KlURA3	This study
PC7200	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 wsc3::NAT	This study
PC7201	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 ynl058c::NAT	This study
PC7202	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 svs1::KlURA3 srl1::NAT	This study
PC7203	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 prm5::KlURA3 ynl058c::NAT	This study
PC7238	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 aad15::KlURA3	This study
PC7239	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 hug1::KlURA3	This study
PC7240	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 pyr2::KlURA3	This study
PC7241	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 mnn5::KlURA3	This study
PC7243	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 wsc2::KlURA3 wsc3::NAT	This study
PC7277	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 tip1::KlURA3	This study
PC7280	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 flo11::Km sfg1::KlURA3	This study
PC7281	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 GAL-FLO11 sfg1::KlURA3	This study
PC7306	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 rho5::NAT	This study
PC7321	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 FLO11-HA at 1000aa sfg1::KlURA3	This study
PC7536	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 nfg1::KlURA3 tip1::NAT	This study
PC7556	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 nfg1::KlURA3 tip1::NAT rgd2::KanMX6	This study
PC7557	MATa SY3089 ste4 FUS1-lacZ FUS1-HIS3 ura3-52 nfg1::KlURA3 tip1::NAT rgd2::KanMX6 rpi1::HYG	This study

Analysis of RNA sequencing data

RNA sequencing (RNAseq) analysis was previously performed in Adhikari and Cullen (2014). Here, the RNAseq data were visualized in a volcano plot generated using the program Instant Clue (http://www.instantclue.uni-koeln.de/). The plot was cropped to show targets induced by fMAPK (genes with a negative-fold change in the ste12∆ mutant).

Microscopy

For DIC (differential interference contrast) imaging, a Zeiss Axioplan 2 microscope (Oberkochen, Germany) was used. The digital images were acquired with an Axiocam MRm camera (Zeiss). For image acquisition and analysis, Axiovision 4.4 software (Zeiss) was used.

Plate-washing assay

The plate-washing assay was used to visualize differences in filamentous growth between the wild-type strain and mutants (Roberts and Fink 1994; Cullen 2015). Briefly, cells were spotted on medium as indicated at 30° for 1–10 days. Cells were spotted equidistant to each other and the edge of the plate to ensure uniform growth. Plates were placed under a stream of water, and colonies were rubbed gently by hand to remove noninvasive cells. Cells that remained in the agar after washing were considered to be part of the invasive scar. Images of the invasive scars were captured by ChemiDoc XRS+ molecular imager (from Bio-Rad Laboratories, Hercules, CA) under immunoblot/chemicoloric setting with no filter or a Nikon D3000 (Nikon, Garden City, NY) digital camera after the plate wash.

To quantify invasive growth, images from the plate-washing assay were imported into ImageJ (National Institutes of Health, Bethesda, MD; https://imagej.nih.gov/ij/). Each image was inverted and treated with identical parameters for adjusting brightness and contrast. For each image, the background was subtracted. Using the set threshold tool with light background set, a threshold was set to convert the invasive scars into pixel images. The threshold was set so that the area around the scar was excluded and areas of invasive growth were highlighted. The pixel area of each invasive scar was measured by the analyze particles tool. This was performed again for two additional higher thresholds (i.e., 10, 30, and 50). The measured values from the three different threshold settings were totaled for a final value. Significance was determined for three replicates, separately for each type of media.

To quantify an invasive growth pattern, images of washed colonies were cropped to 350 × 350 pixels, inverted, and imported into ImageJ. Each image was treated with identical parameters for adjusting brightness and contrast. Images had their background subtracted with a value of 10,000 particles. A box was drawn across the midsection of the image with a pixel height of 40. Using the plot profile tool, which measures the gray value for pixels, a plot profile was generated for each strain of this region of the invasive scar and overlaid onto a graph in excel.

Measuring cell adhesion in liquid and from cells grown on semisolid agar media

To analyze cell adhesion in liquid media, cells were grown for 24 hr in YP-GAL (2%) media at 30°. Images were captured at 5× by microscopy and imported into ImageJ. The background was subtracted by 50 particles. A threshold was applied, set to 170, to convert the image into a binary pixel image. A scale of 1.266 μm per pixels was applied. Using the analyze particles tool, the area of cell clusters was measured and averaged. The averages of three replicates were used to calculate significance. Cells behaved the same if imaged directly in media or after being washed with water.

To analyze cell adhesion on semisolid media, cells were grown for 16 hr in SD+AA at 30°, washed in dH₂O, and cells were spotted onto YP-GAL (2%) medium. Plates were incubated at 30° for 3 days. Cells were harvested from colonies using a metal spatula with care not to excise the agar. Cell biomass was determined by weight. Cells were resuspended in 20 ml dH₂O in 50 ml conical tubes. Tubes were inverted vigorously by hand 10 times. The contents of the tube were poured into a Petri dish, and particles were photographed by ChemiDoc XRS+ molecular imager under immunoblot/chemicoloric setting with no filter. Images were imported into the GIMP2 program and cropped by 970 × 970 pixels circularly. The background was subtracted by 50.0 particles. A threshold was applied, set to 10, to generate a binary pixel image. Images were imported into ImageJ. A scale of 970 pixels = 82.13 mm based on measurements from the ChemiDoc XRS+ molecular imager and GIMP2 program (verified with ruler) was set. Using the analyze particles tool, the total area of cell adhesion was measured. Significance was determined for three replicates.

Colony immunoblots for Flo11p shedding

Colony immunoblots were performed as described (Karunanithi et al. 2010). Cells were grown in 3 ml SD+AA for 16 hr. Cells were pelleted and washed with dH₂O and spotted onto a nitrocellulose membrane directly on top of YP-GAL (2%) or YPD (2% Glu) plates. Plates were incubated at 30° for 3 days. Cells were washed off of the nitrocellulose by plate-washing. The nitrocellulose membrane was examined by immunoblot analysis with anti-HA antibodies and imaged by ChemiDoc XRS+ molecular imager. Signal intensity was measured with the volume tool in Image Lab (https://www.bio-rad.com/en-us/product/image-lab-software?ID=KRE6P5E8Z). Wild-type values were set to 1. Significance was determined for three replicates.

Biofilm/mat assays

Biofilm/mat assays were performed as described (Reynolds and Fink 2001; Karunanithi et al. 2012). Cells were grown in SD+AA for 16 hr and spotted onto semisolid agar (0.3%) medium for 3 days. To analyze plastic adhesion, cells were spotted onto YP-GAL (2%) plates and incubated at 30° for 3 days. Cells were then removed from the agar using a toothpick, resuspended in water, and adjusted to an optical density of A600 = 1.3. Aliquots (100 μl) of cell suspensions were added to polystyrene wells (96-well Falcon Microtest Tissue culture plate) and incubated for 4 hr. An equal volume of 1% crystal violet dye (DIFCO) was added to each well for 20 min. Wells were washed five times and photographed. Quantification was performed with ImageJ. Each well was circularly cropped 250 × 250 pixels from the center of the well. A threshold of 120 was set, then the analyze particle tool measured the total pixel area. Wild-type values were set to 1. Significance was determined for three replicates.

Comparative protein and gene sequence assessments

Comparative assessments for Nfg1p, Rgd2p, Rpi1p, Tip1p, and Sfg1p protein sequences were performed by BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Nonredundant protein sequences (nr) was set for database. The algorithm was blastp (protein–protein BLAST). Sequences used were from the reference strain (S288c) downloaded from the Saccharomyces Genome Database (SGD) (https://www.yeastgenome.org/). Comparative assessment of synteny for SFG1 was performed with the Yeast Gene Order Browser [(http://ygob.ucd.ie/), (Byrne and Wolfe 2005, 2006)].

Quantitative reverse transcription PCR

Quantitative reverse transcription PCR (RT-qPCR) was used to measure the relative expression of FLO11 in wild type with pRS316 and the nfg1∆, rgd2∆, rpi1∆, tip1∆, ste12∆, and dig1∆ mutants. Cells were spotted onto YPD (10% Glu) and incubated at 30° for 2 days. Cells were scraped from the surface of the agar, washed in 1 ml dH₂O, and harvested by centrifugation. RT-qPCR was also used to measure the relative expression of FLO11, DSE1, DSE2, DSE4, and SCW11 in wild type and the sfg1∆ mutant. Cells were grown in 5 ml YPD (2% Glu) cultures grown at 30° for 23 hr. YPD (2% Glu) (1.5 ml) was pelleted and washed with 1 ml dH₂O; a 100 μl aliquot of washed cells was pipetted into 2 ml liquid YP-GAL (2%) cultures and incubated at 30° with shaking for 32 hr. After 32 hr, 2 ml of each sample was washed with 1 ml dH₂O and harvested by centrifugation. RT-qPCR was also used to verify targets of the fMAPK pathway by measuring the relative expression of NFG1, RGD2, RPI1, TIP1, and SFG1 in wild type and the ste12∆ mutant. Cells were grown in 5 ml YPD (2% Glu) cultures grown at 30° for 16 hr; 1.5 ml of 16 hr cultures were pelleted and washed with 1 ml dH₂O. A 100 μl aliquot of washed cells was pipetted into YP-GAL (2%) liquid medium and incubated for 5.5 hr at 30°, washed with 1 ml dH₂O, and harvested by centrifugation. Cells not immediately used in RNA extractions were stored at −80°.

RNA was harvested by hot-acid phenol–chloroform extractions as described (Adhikari and Cullen 2014). Samples were further purified using a QIAGEN RNeasy Mini Kit (catalog number 74104; QIAGEN, Valencia, CA). RNA purity and concentration was measured with NanoDrop (NanoDrop 2000C; Thermo Fisher Scientific, Waltham, MA). RNA stability was determined by agarose gel electrophoresis. cDNA was generated and RT-qPCR was performed as previously described (Chow et al. 2019b). cDNA was generated using iScript Reverse Transcriptase Supermix (catalog number 1708841; Bio-Rad). RT-qPCR was performed using iTaq Universal SYBR Green Supermix (catalog number 1725121; Bio-Rad) on the Bio-Rad CFX384 Real Time System. Primers were obtained from Sigma (Sigma Chemical, St. Louis, MO). Primer sequences can be found in Table S2. ACT1 housekeeping gene primers were based on Chow et al. (2019b). Primer sequences used for FLO11 were based on Chen and Fink (2006). All starting gene concentrations were normalized to the housekeeping gene ACT1 (Chavel et al. 2010; González et al. 2017). Relative gene expression was calculated using the 2−ΔCt formula; Ct was defined as the cycle where fluorescence was statistically significant above background (González et al. 2017); ΔCt is the difference in Ct between a target gene and the housekeeping gene (ACT1; González et al. 2017). RNA was prepared from three biological replicates. Average values are reported.

Measurement of fMAPK pathway activity

To analyze fMAPK pathway activity by the β-galactosidase (lacZ) assay, cells were grown in synthetic medium (SD-URA) for 16 hr. Cells were washed once in dH₂O and resuspended in the medium indicated for 4.5–6.5 hr of growth. Cells were harvested by centrifugation and stored at −80°. The lacZ assays were then performed as described (Jarvis et al. 1988; Cullen et al. 2000) using a FUS1-lacZ reporter as the readout of fMAPK pathway activity. To analyze fMAPK pathway activity by the FUS1-HIS3 transcriptional (growth) reporter, strains were spotted onto SD-HIS+ATA (3-amino-1,2,4-triazole) medium and observed for growth after 3 days.

To analyze fMAPK pathway activity by phosphoblot analysis, cells were grown to saturation in SD-URA medium. Cells were washed and inoculated in 5 ml SD-URA for 5.5 hr at 30°. Cell extracts were prepared for immunoblot analysis according to established procedures (Lee and Dohlman 2008; Adhikari and Cullen 2014). Proteins were precipitated by trichloroacetic acid (TCA). Cells were lysed in TCA buffer (10 mM Tris-HCl pH 8.0; 10% TCA; 25 mM ammonium acetate; 1 mM EDTA) containing glass beads by vortexing for 1 min then placing on ice for 1 min five times. Cells were centrifuged at 15,000 × g for 10 min at 4° and the pellet was mixed in 150 μl of resuspension buffer (0.1 M Tris-HCl pH 11.0; 3% SDS) and boiled for 5 min at 95°. Samples were centrifuged at 15,000 × g for 5 min; 10 μl of each sample was used to measure protein concentration using Pierce BCA Protein Assay Kit (catalog# 23225; Thermo Scientific). An equal volume of 2× sodium dodecyl sulfate (SDS) loading dye (100 mM Tris-HCl pH 6.8; 4% SDS; 0.2% Bromophenol Blue; 20% glycerol; 200 mM β-mercaptoethanol) was added to the supernatant. Protein samples were separated on 10% SDS polyacrylamide gels (SDS-PAGE) and transferred to nitrocellulose membranes (Amersham Protran Premium 0.45 μm NC; GE Healthcare Life Sciences). The membrane was blocked in immunoblot buffer [5% nonfat dry milk (for Pgk1p and Kss1p) or 5% bovine serum albumin (BSA) (for P∼Kss1p), 10 mM Tris-HCl (pH 8), 150 mM NaCl, and 0.05% Tween 20] for 16 hr at 4° rocking. Radiance Plus Chemiluminescent substrate from Azure Biosystems (Dublin, CA) was used for detection. Mouse α-Pgk1p antibodies (#459250; Thermo Fisher Scientific, Rockford, IL) were used to detect Pgk1p as a loading control. Secondary antibodies, goat α-mouse (#170-6516; Bio-Rad Laboratories), were used to detect primary antibodies (Pgk1p) for 1 hr at 20° with rocking. Phosphorylated Kss1p was detected by p42/p44 antibodies (#4370; Cell Signaling Technology, Danvers, MA) and total Kss1p was detected using α-Kss1p antibodies (#6775; Santa Cruz Biotechnology, Santa Cruz, CA). Secondary antibodies, goat anti-rabbit IgG-HRP (#111-035-144; Jackson ImmunoResearch Laboratories, West Grove, PA), were used to detect primary antibodies (Kss1p and P∼Kss1p) and incubated for 1 hr at 20° with rocking. The blot was imaged by ChemiDoc XRS+ molecular imager. Signal intensity was measured by using the volume tool in Image Lab (https://www.bio-rad.com/en-us/product/image-lab-software?ID=KRE6P5E8Z).

Data availability

The authors state that all data necessary for confirming the conclusions presented in the article are represented fully within the article. Strains and plasmids are available upon request. The Gene Expression Omnibus (GEO) accession number for the previously reported expression profiling data are GSE61783 (Adhikari and Cullen 2014). Supplemental material available at figshare: https://doi.org/10.25386/genetics.12609710.

Results

Characterizing transcriptional targets of the fMAPK pathway

Transcriptional targets of the fMAPK pathway have been identified by comparative expression profiling (Madhani et al. 1999; Roberts et al. 2000; Heise et al. 2010; Adhikari and Cullen 2014; van der Felden et al. 2014; Chow et al. 2019b). In Adhikari et al. (2014), wild-type cells and a fMAPK pathway mutant (ste12∆) were compared in liquid YP-GAL (2% galactose) medium. YP-GAL (2%) medium is an fMAPK pathway-inducing condition that triggers the filamentous growth response (Karunanithi and Cullen 2012; Basu et al. 2020). The targets of the pathway identified in Adhikari et al. (2014) are shown here in a volcano plot cropped to display only induced targets (Figure 1B). Some highly induced genes are known targets of the pathway (Figure 1, A and B, blue circles, FLO11 near the center of figure, CLN1 right side of figure, BUD8 right side of figure). Other well-characterized targets included SUC2 (Figure 1B, blue circle near center of figure), which encodes the invertase responsible for hydrolyzing sucrose (Carlson et al. 1981) that contributes to social behaviors (Greig and Travisano 2004; Craig Maclean and Brandon 2008; Koschwanez et al. 2011) like the formation of invasive aggregates (Chow et al. 2019a); the fMAPK pathway components, MSB2 [mucin sensor, (Cullen et al. 2004; Vadaie et al. 2008; Pitoniak et al. 2009)], KSS1 [MAP kinase, (Courchesne et al. 1989; Roberts and Fink 1994; Bardwell et al. 1998a)], STE12 and TEC1 (Laloux et al. 1990; Chou et al. 2006) are induced by the fMAPK pathway to generate positive feedback (Figure 1, A and B, blue circles). PGU1 encodes a pectinase (endopolygalacturonase) that does not affect filamentous growth but breaks down plant tissue and may impact nutrient scavenging in the wild (Madhani et al. 1999) (Figure 1B, black circle in center of figure). Several mating pathway targets were also identified that are under the control of Ste12p (Figure 1, BAR1, STE2, and STE4, black circles); however, the mating pathway is not thought to be required for filamentous growth (Roberts and Fink 1994; Sabbagh et al. 2001; Flatauer et al. 2005; Meem and Cullen 2012).

The fMAPK pathway also regulates the expression of targets whose functions remain uncharacterized. Thirteen new transcriptional targets were investigated (Figure 1B, green text, Figure S1) based on their fold change in expression as Log₂FoldChange > │0.8│. Gene disruptions were constructed in wild-type cells of the filamentous (Σ1278b) background, and deletion mutants were examined for a role in invasive growth. The plate-washing assay was used, where colonies washed off of a surface leave a visible invasive scar (Roberts and Fink 1994). We compared invasive scars of wild-type cells to mutants, looking for an invasive growth phenotype. Because invasive growth occurs in response to limiting carbon (Cullen and Sprague 2000, 2012) and nitrogen (Gimeno et al. 1992) and can be induced by high cell density through alcohols (Chen and Fink 2006; González et al. 2017; Lenhart et al. 2019), the plate-washing assay was performed on different media: YPD (2% Glu), YPD (10% Glu), YPD (16% Glu), YPD high osmolarity medium (2% Glu + 8% Sorbitol), SLAD (low nitrogen), SLAD + 2% ethanol, synthetic dextrose (SD), and YP-GAL (2%).

Most mutants tested did not show an obvious phenotype in invasive growth (Figure S1). Four mutants (ylr042c∆, rgd2∆, rpi1∆, and tip1∆, Figure 1B, yellow circles) did not show a phenotype on YPD (2% Glu), but showed increased invasive growth on YPD (10% Glu) (see below). This indicates unexpectedly that several highly induced targets of the fMAPK pathway function to negatively regulate invasive growth. Based on data shown below, YLR042c was named NFG1 for Negative Regulator of the Filamentous Growth MAPK pathway 1. Among many mutants tested, one showed a clear invasive growth defect (sfg1∆, Figure 1B, blue circle with green text, and Figure S1) and was also characterized in the study. Five target genes have paralogs that might mask their mutant phenotypes due to genetic redundancy or buffering (Wolfe and Shields 1997; Costanzo et al. 2010). Gene disruptions generating prm5∆ ynl058c∆, svs1∆ srl1∆, and wsc2∆ wsc3∆ double mutants did not show an invasive growth phenotype (Figure S1). Gene disruptions for RIB4, SRD1, HPF1, ADA2, AAD3 (paralog to AAD15), and PRY1 (paralog to PRY2) failed to obtain positive isolates. A genome-wide deletion collection in the Σ1278b background did not contain deletion mutants of SRD1, AAD3, RIB4, HPF1, and AAD15 (Ryan et al. 2012), but did for ADA2, PRY1, and PRY2, which did not show an invasive growth phenotype (Ryan et al. 2012; Chavel et al. 2014). We were successful at assigning roles based on phenotype to 5 of 13 (38%) of the target genes tested. However, 62% of the genes failed to produce a phenotype. It is plausible that these genes function in aspects of filamentous growth that are unrelated to agar invasion [for example, Pgu1p (Madhani et al. 1999)]. Genes that showed a phenotype in invasive growth were verified as targets of the fMAPK pathway by examining their expression by RT-qPCR analysis under conditions that promote filamentous growth (Figure S2, YP-GAL medium). NFG1, RGD2, RPI1, TIP1, and SFG1 may be direct targets because the transcription factors Ste12p and Tec1p bind to their promoters based on the repository YEASTRACT [http://www.yeastract.com/index.php, (Zeitlinger et al. 2003; Harbison et al. 2004; Borneman et al. 2007; Lefrançois et al. 2009; Zheng et al. 2010)]. Thus, four negative regulators of invasive growth (NFG1, RGD2, RPI1, and TIP1) and one positive regulator of invasive growth (SFG1) were identified here as fMAPK pathway targets.

The fMAPK pathway induces target genes that negatively regulate invasive growth

A major function of the fMAPK pathway is to positively regulate invasive growth [Figure S1, ste12∆ (Roberts and Fink 1994; Cook et al. 1997; Roberts et al. 2000)]. On YPD (10% Glu) medium, the nfg1∆, rgd2∆, rpi1∆, and tip1∆ mutants showed increased invasive growth compared to wild type (Figure 2A, Washed), which was confirmed by quantification by ImageJ (Figure 2A, Invasion). Thus, Nfg1p, Rgd2p, Rpi1p, and Tip1p have a negative effect on invasive growth. NFG1 is a highly induced ORF by the fMAPK pathway that has been established as a target for some time with no described function in invasive growth [YLR042c, (Caro et al. 1997; Hamada et al. 1999; Madhani et al. 1999; Roberts et al. 2000; Giaever et al. 2002; Hohmann 2002; García et al. 2004; Kim and Levin 2010; Parachin et al. 2010; Adhikari and Cullen 2014; Chow et al. 2019b)]. TIP1 encodes a mannoprotein of the fungal cell wall (Kondo and Inouye 1991; Fujii et al. 1999; Chow et al. 2018) and RGD2 encodes a GTPase-activating protein (RhoGAP for Cdc42p and Rho5p; Roumanie et al. 2001; Tkach et al. 2012), both with no established role in invasive growth. RPI1 encodes a transcription factor that inhibits the Ras/cyclic AMP pathway (Kim and Powers 1991), promotes preparation of cells for the stationary phase in part by fortification of the cell wall (Sobering et al. 2002), and increases stress tolerance during fermentation (Puria et al. 2009). RPI1 was previously shown to promote filamentous growth in some strain backgrounds but not in the ∑1278b strain background (Chin et al. 2012).

Figure 2.

Four targets of the fMAPK pathway negatively regulate invasive growth. (A) Plate-washing assay. Wild-type cells (PC538+pRS316) and the nfg1∆ (PC7147), rgd2∆ (PC7146), rpi1∆ (PC7145), tip1∆ (PC7277), ste12∆ (PC539), and dig1∆ (PC3039) mutants were spotted on YPD (10% Glu) for 3 days. Top row, colonies, middle row, inverted images of plates after wash, Bar, 0.5 cm. Bottom row, close-up of washed plates showing aggregates at 5× magnification, Bar, 400 µm. Invasion, quantification of invasive scars by ImageJ in triplicate, with wild-type values set to 1. Error represents the SEM, which varied <20% across trials. Asterisks, P-value <0.035, by Student’s t-test compared to wild type. FLO11 expression, fold change in FLO11 mRNA levels by RT-qPCR analysis normalized to ACT1. Wild-type values were set to 1. Variance by SD. was <20% across three trials for all strains, except the dig1∆ mutant, which was one trial. Asterisks, P-value ≤0.01, by Student’s t-test compared to wild type. FUS1-lacZ, β-Galactosidase (lacZ) assays. Cells were grown in SD-URA for 16 hr, washed, and resuspended in SGAL-URA for 4.5 hr prior to harvesting cells by centrifugation. (B) Plate-washing assay for wild-type cells (PC538+pRS316) and the nfg1∆ mutant (PC7147) grown on YP-Gal (2%) medium. Top row, colonies, bottom row, inverted images of plates after wash, Bar, 0.5 cm. Low Density, cells spotted with OD₆₀₀ = 1.5 for 3 d. High Density, cells spotted with OD₆₀₀ = 11 for 2 days. (C) Plate-washing assay on YPD (2% Glu), high osmolarity medium [YPD (2% Glu + 8% Sorb)], and YPD (16% Glu) for 3 day. Inverted images of plates after wash for indicated strains, Bar, 0.5 cm. Colonies (not shown) were similar in size and appearance. Invasion, quantification of invasive scars by ImageJ in triplicate, with wild type values set to 1. Error represents the SEM, which varied <30% across trials, except the rpi1∆ mutant on YPD (16% Glu) varied by <75%. Asterisks, P-value < 0.05, by Student’s t-test compared to wild type. (D) Immunoblot analysis of wild type cells (PC538+pRS316) and the nfg1∆ (PC7147) and ste11∆ (PC611) mutants grown in SD-URA for 5.5 hr. Cell extracts were probed with antibodies to detect phosphorylated Kss1p (P∼Kss1p) [α-p42/p44], total Kss1p, and Pgk1p as a control for protein levels. Numbers refer to the ratio of P∼Kss1p to Pgk1p with wild type set to 1. The MAP kinase for the mating pathway, Fus3p, also showed some elevated phosphorylation, as might be expected based on a previous report (Basu et al. 2016).

When compared to the dig1∆ mutant, which lacks a known negative regulator of the fMAPK pathway (Cook et al. 1996; Tedford et al. 1997; Bardwell et al. 1998b; Olson et al. 2000), the nfg1∆, rgd2∆, rpi1∆, and tip1∆ mutants had more subtle phenotypes (Figure 2A). This suggests Nfg1p, Rgd2p, Rpi1p, and Tip1p might not turn off invasive growth like Dig1p, but instead modulate it in a specific context. One way the fMAPK pathway regulates invasive growth is by regulating the expression of FLO11, which encodes the cells’ major adhesion molecule (Lo and Dranginis 1996; Madhani et al. 1999; Rupp et al. 1999; Roberts et al. 2000; Halme et al. 2004; Borneman et al. 2006; Veelders et al. 2010; Adhikari and Cullen 2014; Kraushaar et al. 2015; Barua et al. 2016; Reynolds 2018; Chow et al. 2019b; Brückner et al. 2020). RT-qPCR analysis showed that the expression of FLO11 was elevated in the nfg1∆, rgd2∆, and tip1∆ mutants compared to wild type, indicating these genes have an inhibitory effect on FLO11 expression (Figure 2A, FLO11 expression). The effect was modest (∼0.5-fold), which supports the idea that these genes may be involved in fine tuning invasive growth. As in previous findings (Chin et al. 2012), the rpi1∆ mutant showed no change in the expression of FLO11 compared to wild type (Figure 2A, FLO11 expression).

Closer inspection of the invasive scars showed an increase in aggregate invasive growth (Figure 2A, Close up), which results from the interaction of groups of cells that make gouges in the agar (Chow et al. 2019a). Likewise, the nfg1∆, rgd2∆, rpi1∆, and tip1∆ mutants showed elevated aggregate invasive growth on YP-GAL (2%) medium; however, this occurred only when cells were spotted at high cell density (Figure 2B, the complete data set is in Figure S3), which stimulates aggregate invasive growth due to an increased abundance of quorum-sensing molecules (Chow et al. 2019a). At standard glucose concentrations [YPD (2% Glu) medium], the rgd2∆, rpi1∆, and tip1∆ mutants were not more invasive than wild type, and nfg1∆ was only slightly more invasive at a P-value < 0.062 [Figure 2C, YPD (2% Glu)]. These results indicate Nfg1p, Rgd2p, Rpi1p, and Tip1p inhibit invasive growth more noticeably at higher glucose levels. This observation was puzzling because glucose inhibits invasive growth (Cullen and Sprague 2000). One possibility is that high glucose levels might lead to higher cell density as a result of an elevated carrying capacity (Spor et al. 2008). High carrying capacity may lead to enhanced density-dependent invasion after depletion of glucose. Thus, the Nfg1p, Rgd2p, Rpi1p, and Tip1p proteins negatively regulate aggregate invasive growth.

Nfg1p, Rgd2p, Rpi1p, and Tip1p might act separately or in the same pathway. To address this question, the nfg1∆, rgd2∆, rpi1∆, and tip1∆ mutants were compared by different assays and in different conditions to see if they share the same phenotype. Sharing the same phenotype would suggest that the proteins act in the same pathway. The nfg1∆ and tip1∆ mutants were phenotypically similar, showing increased invasive growth on different types of media: YPD (10% Glu) [Figure 2A, invasion], YP-GAL (2%) (Figure S3B), high osmolarity medium [YPD (2% Glu + 8% Sorb), Figure 2C], and YPD (16% Glu) (Figure 2C). The nfg1∆ and tip1∆ mutants also showed the same pattern of FLO11 expression (Figure 2A, FLO11 expression). These results support the idea that Nfg1p and Tip1p act in the same pathway.

The rgd2∆ and rpi1∆ mutants were phenotypically similar to the nfg1∆ and tip1∆ mutants on some media, showing increased invasive growth on YPD (10% Glu) [Figure 2A, invasion] and YP-GAL (2%) (Figure S3B). However, the rgd2∆ and rpi1∆ mutants were phenotypically different from the nfg1∆ and tip1∆ mutants because they did not show increased invasive growth on high osmolarity medium [YPD (2% Glu + 8% Sorb), Figure 2C]. The rgd2∆ and rpi1∆ mutants were also phenotypically different from each other on high osmolarity medium [YPD (2% Glu + 8% Sorb), Figure 2C] and YPD (16% Glu) (Figure 2C). Furthermore, Rgd2p but not Rpi1p regulated FLO11 expression (Figure 2A, FLO11 expression). Overall, these results suggest Rgd2p and Rpi1p function in different pathways.

Mutant combinations were generated (nfg1∆ tip1∆ double mutant, nfg1∆ tip1∆ rgd2∆ triple mutant, and nfg1∆ tip1∆ rgd2∆ rpi1∆ quadruple mutant) to determine if they had additive phenotypes. Additive phenotypes would suggest the proteins operate in different pathways. The nfg1∆ single mutant, the nfg1∆ tip1∆ double mutant, the nfg1∆ tip1∆ rgd2∆ triple mutant, and the nfg1∆ tip1∆ rgd2∆ rpi1∆ quadruple mutant showed increased invasive growth compared to wild type but did not show strong phenotypic differences from each other by the plate-washing assay (Figure S4, A and B). Collectively, evidence from the plate-washing assay of single mutants and mutant combinations suggests Rgd2p and Rpi1p function separately from each other and from Nfg1p and Tip1p, while Nfg1p and Tip1p may act in the same pathway.

The fMAPK pathway is one of the pathways that regulates FLO11 expression (Madhani et al. 1999; Rupp et al. 1999; Borneman et al. 2006; Chavel et al. 2010, 2014; Cullen and Sprague 2012). Given that Nfg1p, Rgd2p, and Tip1p have a negative effect on FLO11 expression, they might do so by dampening the fMAPK pathway. The nfg1∆ mutant, but not the tip1∆, rgd2∆, or rpi1∆ mutant showed elevated fMAPK pathway activity based on a transcriptional reporter [Figure 2A, FUS1-lacZ]. This indicates that Nfg1p negatively regulates the fMAPK pathway. Double, triple, and quadruple mutant analysis showed that the nfg1∆ tip1∆ double mutant had an additional increase in fMAPK pathway activity compared to the nfg1∆ single mutant (Figure S4A, FUS1-lacZ). Thus, Tip1p might also negatively regulate the fMAPK pathway under some conditions separately from Nfg1p, although we have not explored this possibility. These results indicate Nfg1p and Tip1p act, at least in part, in separate ways. Immunoblot analysis with antibodies that detect phosphorylated (P∼) Kss1p (the MAP Kinase of the fMAPK pathway) showed that P∼Kss1p levels were higher in the nfg1∆ mutant (Figure 2D), compared to wild-type cells and the ste11∆ mutant [Ste11p is the MAP kinase kinase kinase that phosphorylates the MAP kinase kinase, Ste7p, which phosphorylates Kss1p (Liu et al. 1993; Roberts and Fink 1994)]. Thus, Nfg1p, Rgd2p, Rpi1p, and Tip1p have separate functions in the negative regulation of invasive growth, and Nfg1p (and perhaps Tip1p) negatively regulates the fMAPK pathway.

We performed comparative assessments of Nfg1p, Rgd2p, Rpi1p, and Tip1p by BLAST. Nfg1p protein sequence had similarity only within the Saccharomyces clade, with Saccharomyces eubayanus being the most distant relative with a recognizable homolog (Figure S5, A and B); therefore, Nfg1p is not a conserved protein that regulates the fMAPK pathway across all yeasts. The protein sequences of Rgd2p, Rpi1p, and Tip1p had homologs in other yeasts outside the Saccharomyces clade (Figure S5, A and B), including Candida glabrata—a human pathogen that undergoes filamentous growth (Fidel et al. 1999; Csank and Haynes 2000; Rodrigues et al. 2014). Rgd2p also had protein sequence similarity to a homolog in the human pathogen C. albicans (Figure S5, A and B). Thus, Rgd2p, Rpi1p, and Tip1p are conserved in several yeast species and could be regulators of filamentous growth in pathogenic yeasts.

Sfg1p negatively regulates the transcription of cell separation genes

SFG1 was identified as a target of the fMAPK pathway (Figure 1B, blue circle with green text) and positive regulator of invasive growth (Figure S1, Third column). Sfg1p is a transcription factor that induces superficial pseudohyphal growth [a type of growth where cells spread across a surface in filament-like structures (Fujita et al. 2005)] and transcriptionally represses genes that induce cell separation, including DSE1, DSE2, DSE4, and SCW11 (Doolin et al. 2001; Baladrón et al. 2002; Draper et al. 2009; White et al. 2009). The inhibition of cell separation leads to filament formation (King and Butler 1998; Doolin et al. 2001). DSE2, DSE4, and SCW11 have similarity to glucanases and may promote cell separation by degrading the cell wall between mother and daughter cells. To determine whether Sfg1p regulates invasive growth by this mechanism, the transcriptional targets of Sfg1p that induce cell separation were tested for a role in invasive growth. Wild-type cells and the sfg1∆, dse1∆, dse2∆, dse4∆ and scw11∆ mutants were examined for invasive growth by the plate-washing assay (Figure 3A). The dse1∆, dse2∆, dse4∆ and scw11∆ mutants had increased invasive growth compared to wild-type cells, supporting the idea that these genes have an inhibitory effect on invasive growth. DSE1, DSE2, DSE4 and SCW11 were transcriptional targets of Sfg1p by RT-qPCR analysis being upregulated in the sfg1Δ mutant under conditions that promote filamentous growth (Figure 3B, YP-GAL medium). Thus, in support of previous findings, Sfg1p inhibits the transcription of genes that promote cell separation, which results in increased cell attachment and invasive growth.

Figure 3.

Transcriptional targets of Sfg1p that induce cell separation inhibit invasive growth. (A) Plate-washing assay for wild-type cells (PC538+pRS316) and the sfg1∆ (PC7144), dse1∆ (PC7164), dse2∆ (PC7165), dse4∆ (PC7166), and scw11∆ (PC7198) mutants spotted onto YPD (10% Glu) for 3 days. Top row, colonies, bottom row, inverted images of invasive scar after plate wash, Bar, 0.5 cm. Invasion, quantification of invasive scars by ImageJ in triplicate, with wild-type values set to 1. Error represents the SEM, which varied <45% across trials, except the sfg1∆ mutant which varied by <56%. Asterisks, P-value <0.035, by Student’s t-test compared to wild type. (B) Relative gene expression by RT-qPCR of target gene (DSE1, DSE2, DSE4, and SCW11) mRNA levels, normalized to ACT1 expression, between wild-type (PC538) and sfg1∆ (PC7144) cells grown in YP-Gal (2%) liquid medium for 32 hr. Wild-type values set to 1. Error represents SD across three trials. Asterisks, P-value < 0.02, by Student’s t-test compared to wild type.

SFG1 regulates invasive growth independently from FLO11

One requirement for invasive growth is cell adhesion by Flo11p (Lo and Dranginis 1996; Madhani et al. 1999; Rupp et al. 1999; Halme et al. 2004; Borneman et al. 2006; Veelders et al. 2010; Kraushaar et al. 2015; Barua et al. 2016; Reynolds 2018). Flo11p binds in a homotypic manner to other Flo11p molecules to maintain adhesive contacts between cells (Kraushaar et al. 2015; Brückner et al. 2020). The expression of FLO11 is regulated by the fMAPK pathway (Madhani et al. 1999; Rupp et al. 1999; Roberts et al. 2000; Borneman et al. 2006; Adhikari and Cullen 2014; Chow et al. 2019b). Presumably, Sfg1p (by inhibiting cell separation) and Flo11p (by promoting homotypic contacts) function in different ways to control filamentous growth. The fact that SFG1 and FLO11 expression are both regulated by the fMAPK pathway suggests that the pathway may have versatility in regulating cell adhesion through a combination of mechanisms.

To explore how Sfg1p and Flo11p regulate cell adhesion and invasive growth, a double gene deletion mutant was generated (sfg1∆ flo11∆). Wild-type cells were compared to the sfg1∆ and flo11∆ single mutants and the sfg1∆ flo11∆ double mutant by the plate-washing assay. On YP-GAL (2%) medium, the sfg1∆ and flo11∆ single mutants showed a defect in invasive growth [Figure 4A, YP-GAL (2%), the complete data set is in Figure S6A]. The sfg1∆ flo11∆ double mutant had a more severe invasive growth defect than either single mutant [Figure 4A, YP-GAL (2%)]. These data demonstrate that Sfg1p and Flo11p play separate roles in regulating invasive growth. Moreover, the sfg1∆ flo11∆ double mutant retained some invasive growth, which indicates that a third (FLO11- and SFG1-independent) mechanism also regulates invasive growth under this condition. Flo11p is a member of the Flo gene family (Guo et al. 2000; Smukalla et al. 2008; Veelders et al. 2010), and, although other members of this family are typically silenced, another Flo gene may be expressed under this condition. FLO10 is expressed at least to some degree in our strains (Birkaya et al. 2009; Chow et al. 2018) and may contribute to invasive growth in this setting.

Figure 4.

Sfg1p is required for invasive growth and has a different phenotype than Flo11p. (A) Plate-washing assay for wild type (PC538), sfg1∆ (PC7144), flo11∆ (PC1029), and sfg1∆flo11∆ (PC7280) strains spotted on YP-Gal (2%) medium and YPD (2% Glu) medium for 7 days. Left columns, colonies, Washed, inverted images of invasive scar after plate wash, Bar, 0.5 cm. Inset, close up of invasive scars on YP-GAL (2%) marked on washed images by blue box. Orange arrows, ring region of invasive scar. Green arrows, region directly outside invasive ring. Purple arrows, periphery of invasive scar. (B) Plot of invasion across each invasive scar from (A) (Washed, red brackets represent outside edge of region used for measurement). X-axis, distance in pixels from left edge; Y-axis, intensity of invasion measured by gray area intensity of pixels in ImageJ. High values, less invasive growth, low values more invasive growth. Colored arrows mark regions denoted in (A).

The sfg1∆ and flo11∆ single mutants had different invasive patterns. To better visualize the patterns, invasive scars were quantified by ImageJ and represented graphically by a plot profile. ImageJ was used to measure pixel intensity across the invasive scar (Figure 4A, Washed, see red brackets), with higher values (lighter pixels) representing less invasion and lower values (darker pixels) representing more invasion. For wild-type cells, invasive growth occurred in a unique pattern, with the most intense invasive growth occurring as a ring (Figure 4A, Inset, orange arrow), which corresponded to two troughs in the graph [Figure 4B, YP-Gal (2%), blue line]. The sfg1∆ mutant was less invasive than wild-type cells but still produced an invasive ring [Figure 4B, YP-Gal (2%), black line]. The sfg1Δ mutant showed a similar level of invasion as the flo11Δ mutant [Figure 4B, YP-Gal (2%), compare black and red lines] but was more invasive in the ring region (orange arrows) and just outside of the ring region (green arrows) than the flo11∆ mutant. Thus, in these regions Flo11p plays a bigger role in invasive growth than Sfg1p. Along the periphery of the invasive scar, the flo11∆ mutant was more invasive than the sfg1∆ mutant (Figure 4, A and B, YP-Gal (2%), purple arrows), indicating that in this region, Sfg1p plays a bigger role in invasive growth than Flo11p. Surprisingly, the flo11∆ mutant was more invasive than wild type at the periphery (purple arrows). Thus, colonies do not invade in an ‘all or none’ manner. Rather, cells in different parts of the colony show different levels of invasion that are differentially regulated by Sfg1p and Flo11p.

Many adhesion-dependent responses are regulated by the fMAPK pathway (Chow et al. 2019b). For example, cells can form adherent flocs in liquid culture (Verstrepen et al. 2003; Halme et al. 2004; Fidalgo et al. 2006; Barua et al. 2016). To examine the role of Sfg1p and Flo11p in this aspect of cell adhesion, we developed an assay to quantify cell adhesion in liquid cultures. This was done by measuring the average area of a group (cluster) of adherent cells by ImageJ after growth in liquid YP-GAL (2%) medium for 24 hr. The flo11∆ and sfg1∆ single mutants showed a defect in forming clusters (Figure 5A). The sfg1∆ flo11∆ double mutant showed a more severe defect. Thus, Sfg1p and Flo11p contribute equally in regulating cell adhesion in liquid.

Figure 5.

Sfg1p and Flo11p affect cell adhesion differently. (A) Clusters, images of wild-type (PC538) and the sfg1∆ (PC7144), flo11∆ (PC1029), and sfg1∆flo11∆ (PC7280) mutant cells grown in YP-Gal (2%) liquid medium for 24 hr at 5× magnification, Bar, 50 µm. Area (μm²), area quantification for the average size of cell clusters by ImageJ. Error represents SEM which varied < 10% across three trials. Asterisks, P-value < 0.05, by Student’s t-test with wild type. (B) Colony morphology, close up of colony for indicated strains after 6 days on YP-Gal (2%), Bar, 0.5 cm. Particles, adhesive particles from the colony surface. Colonies were grown on YP-Gal (2%) for 3 days, then colony was scraped into dH₂O, mixed, and imaged. Black particles represent groups of cells that remained adherent after mixing. Total, the total area of adherent particles for one colony (mm²) quantified by ImageJ in triplicate. Error represents the SEM and was <10% across three trials. Asterisk, P-value <0.005, by Student’s t-test with wild type. (C) Plastic, adherence estimated by cell adherence to a polystyrene plastic 96-well plate. Indicated strains grown on YP-Gal (2%) medium for 3 days. Cells were scraped from the medium, added to wells, and stained with crystal violet dye. Wells were washed 5× with water. Adhesion, quantification of plastic adhesion by ImageJ in triplicate, with wild-type values set to 1. Error represents the SD, which varied <20% across trials. Asterisks, P-value <0.0001, by Student’s t-test compared to wild type. (D) Biofilm/mat, cells were spotted onto 0.3% agar YP-Gal (2%) for 3 days and imaged. Bar, 1 cm. Inset, close up of biofilm/mat (blue square) to highlight colony pattern.

Environmental impacts on Sfg1p- and Flo11p-mediated adhesion

Other adhesion-dependent responses require Flo11p, such as complex colony morphology, where patterns/ruffles form on the colony surface (Granek and Magwene 2010; Karunanithi et al. 2012; Chow et al. 2019b). The sfg1∆ mutant had an intermediate complex colony morphology phenotype between wild-type cells and the flo11∆ mutant (Figure 5B, Colony Morphology). By this criterion, the sfg1∆ flo11∆ double mutant was indistinguishable from the flo11∆ mutant. To further investigate this cell-adhesion phenotype, we developed an assay to quantify cell adhesion within colonies. Cells were scraped from the surface of colonies grown on semisolid agar medium, resuspended in dH₂O and mixed. Particles made up of adherent cells were imaged and quantified as the area of all particles (total) per colony. Cells derived from wild-type colonies formed particles that were visible to the eye, while cells of the flo11∆ mutant separated and were not visible by eye [Figure 5B, Particles and Total]. The sfg1∆ mutant had an intermediate phenotype (Figure 5B, Particles and Total). The sfg1∆ flo11∆ double mutant was indistinguishable from the flo11∆ mutant by this assay. Thus, Sfg1p plays a minor role compared to Flo11p in this adhesion-dependent phenotype. This is different from the role of Sfg1p in invasive growth and adhesion in liquid where it played the same role as Flo11p.

Some species of yeast, such as C. albicans, are pathogens whose adhesion-related behaviors promote virulence. For example, many species of fungi, including pathogens, form biofilms or mats (Lo et al. 1997; Reynolds and Fink 2001; Kabir et al. 2012; Karunanithi et al. 2012; Silva-Dias et al. 2015). Biofilm/mats occur when cells adhere together in a complex multicellular community (Costerton et al. 1999; Reynolds and Fink 2001; Flemming and Wingender 2010; Kabir et al. 2012; Karunanithi et al. 2012; Azeredo et al. 2017). In this growth mode, cells can adhere to inert surfaces, like plastics, which occurs on medical devices and hospital settings (Kennedy et al. 1989; Reynolds and Fink 2001; Kabir et al. 2012; Karunanithi et al. 2012; Silva-Dias et al. 2015). Biofilm/mat formation and plastic adhesion also occur in S. cerevisiae, and requires Flo11p [Figure 5, C and D, (Reynolds and Fink 2001; Karunanithi et al. 2012)]. Sfg1p was not required for plastic adhesion (Figure 5C) or biofilm/mat expansion and ruffling (Figure 5D). This result indicates Sfg1p is required for a subset of Flo11p-dependent cell-adhesion phenotypes. To summarize, depending on the cell-adhesion phenotype, Sfg1p contributes equally to cell adhesion compared to Flo11p, contributes less, or does not contribute at all.

We also asked whether the environment might impact the way that Sfg1p and Flo11p regulate invasive growth. Compared to YP-GAL (2%), on YPD (2% Glu) medium, the sfg1∆ mutant was only slightly defective for invasive growth, whereas the flo11∆ mutant was more defective [Figure 4A, YPD (2% Glu), the complete data set is in Figure S6B]. The difference in invasive growth was also evident in the invasive patterns. In particular, the flo11Δ mutant was less invasive than the sfg1Δ mutant across the entire plot profile [Figure 4B, YPD (2% Glu), compare red and black lines]. The sfg1∆ flo11∆ double mutant showed no invasive growth on YPD (2% Glu) [Figure 4, A and B, compare the yellow lines in GLU and GAL], indicating that Sfg1p and Flo11p solely control invasive growth under this condition. Furthermore, Flo11p showed different requirements in liquid compared to surface growth. In liquid, the flo11∆ mutant had a ∼1.5-fold decrease in adhesion, compared to ∼183-fold decrease on semisolid agar medium (Figure 5A, Area, and Figure 5B, Total). Sfg1p regulated cell adhesion in liquid and on semisolid agar media similarly, because the sfg1∆ mutant showed ∼twofold decrease under both conditions (Figure 5A, Area, and Figure 5B, Total). Therefore, Sfg1p and Flo11p play different roles in adhesion-dependent responses depending on the environment.

Sfg1p regulates multiple aspects of filamentous growth

Biofilm/mats are embedded in a matrix that is synthesized by the microbial community (Costerton et al. 1999; Flemming and Wingender 2010; Kabir et al. 2012; Azeredo et al. 2017). In S. cerevisiae, Flo11p is shed in biofilms/mats into the extracellular milieu (Karunanithi et al. 2010). Given that Sfg1p impacts the expression of cell wall enzymes, Sfg1p was tested for a role in regulating Flo11p shedding. A wild-type HA-tagged Flo11p strain (FLO11-HA) and a sfg1∆ FLO11-HA mutant were grown on a nitrocellulose membrane laid on YP-Gal (2%) semisolid medium. The membrane was washed and probed by antibodies for the HA epitope. Flo11p-HA shedding was reduced in the sfg1∆ FLO11-HA mutant compared to the wild-type FLO11-HA strain (Figure 6A). Many transcription factors converge on the FLO11 promoter (Borneman et al. 2006); therefore, Sfg1p might impact Flo11p shedding by regulating expression of the FLO11 gene. RT-qPCR analysis showed that FLO11 expression was reduced in the sfg1∆ mutant compared to wild-type cells (Figure 6A, FLO11 expression). Therefore, Sfg1p regulates cell adhesion in part by regulating FLO11 expression. Sfg1p might also impact Flo11p shedding through cell wall remodeling.

Figure 6.

Sfg1p regulates FLO11 expression and cell elongation. (A) Colony immunoblot to detect HA-Flo11p with anti-HA antibodies. Wild type (PC538), FLO11-HA (PC2043), and sfg1∆FLO11-HA (PC7321) strains were grown on nitrocellulose membranes atop YP-Gal (2%) semisolid agar medium for 3 days. Numbers refer to the intensity of anti-HA quantified by image lab. Experiments were performed in triplicate. Error is SEM with <20% variation across trials. Asterisks, P-value <0.02, by Student’s t-test compared to wild type. FLO11 expression, fold change in FLO11 mRNA levels by RT-qPCR analysis normalized to ACT1, wild-type values set to 1. FLO11 expression was measured in wild-type (PC538) and sfg1∆ (PC7144) strains. Error represents SD, which varied <30% across trials. Asterisks, P-value <0.03, by Student’s t-test with wild type. (B) Left and middle columns, Plate-washing assay for wild type (PC538), Gal-FLO11 (PC2712) sfg1∆Gal-FLO11 (PC7281) strains on SGAL+AA after 6 days. Left column, before wash and middle column, after wash, Bar, 0.5 cm. Right column, images of cells grown in YP-Gal (2%) liquid medium for 24 hr imaged at 5× magnification, Bar, 50 µm. Area values represent ImageJ quantification for the average size of cell clusters by area (μm²) for indicated strains. Error represents SEM which varied <15% across three trials. Asterisk, P-value = 0.01, by Student’s t-test with wild type. Wild-type data are from Figure 3. (C) Cell morphology compared after growth in YP-Gal (2%) for 4 hr between wild-type cells (PC538), sfg1∆ (PC7144), and wild type transformed with a plasmid containing an overexpression of SFG1 by a galactose-inducible promoter (pGAL-SFG1). Bar, 10 µm.

To determine the role of Sfg1p in regulating cell adhesion independent of FLO11 expression, a strain where FLO11 is expressed from a galactose-inducible promoter (GAL-FLO11) was compared to the sfg1∆ GAL-FLO11 mutant for invasive growth and cluster formation. Overexpression of FLO11 caused increased invasive growth [Figure 6B, washed, (Chow et al. 2019a)] and the formation of large clusters (Figure 6B, Clusters and Area). Deletion of SFG1 in the GAL-FLO11 strain led to a decrease in invasive growth and a reduction in cluster size (Figure 6B). This data indicates that Sfg1p primarily regulates cell adhesion independent of FLO11 expression. As shown above, Sfg1p had no effect on some Flo11p-dependent responses, like biofilm/mat formation and plastic adhesion. Thus, Sfg1p might not regulate FLO11 expression under all conditions. This idea is supported by the fact that Sfg1p did not regulate Flo11p shedding under all conditions (Figure S7).

In addition to cell adhesion, cells undergoing filamentous growth also regulate cell elongation. Cells elongate by a delay in the cell cycle that leads to extended apical growth (Kron et al. 1994; Edgington et al. 1999). The fMAPK pathway causes a delay in the cell cycle by inducing expression of the CLN1 gene (Loeb et al. 1999; Madhani et al. 1999), which encodes a G₁/S specific cyclin (Hadwiger et al. 1989). However, this is not the only way the fMAPK pathway induces a delay in the cell cycle (Ahn et al. 1999). One additional way may be through regulating the expression of SFG1 because SFG1 regulates the cell cycle (White et al. 2009). A sfg1∆ mutant is delayed in the G₁ phase of the cell cycle, and SFG1 overexpression causes a delay in G₂/M (White et al. 2009). We found by microscopy, the sfg1∆ mutant had a round-cell morphology compared to wild-type cells, and overexpression of SFG1 by a galactose-inducible promoter [pGAL-SFG1 (Gelperin et al. 2005)] induced an elongated morphology (Figure 6C). Thus, the fMAPK pathway may regulate cell elongation by multiple mechanisms, such as by controlling the expression of the CLN1 and SFG1 genes.

Sfg1p was previously shown to be a distantly related member of a family of transcriptional regulators of fungal development in nonpathogenic and pathogenic fungi because it has weak similarity in protein sequence to a family of transcription factors involved in pseudohyphal/hyphal development (Fujita et al. 2005). These include Phd1p and Sok2p in S. cerevisiae (Gimeno and Fink 1994; Ward et al. 1995; Fujita et al. 2005); Efg1p in C. albicans (Stoldt et al. 1997; Fujita et al. 2005); StuA in Aspergillus nidulans (Miller et al. 1992; Fujita et al. 2005); and Asm-1 in Neurospora crassa (Aramayo et al. 1996; Fujita et al. 2005). We found that SFG1 shows synteny (by the Yeast Gene Order browser) and protein sequence similarity (by BLAST) to other fungi species as well (Figure S5, A and B), including an uncharacterized ORF (CAGL0I09856g) in the human pathogen C. glabrata (Fidel et al. 1999; Csank and Haynes 2000; Rodrigues et al. 2014). Presumably, SFG1 is required for filamentous growth in other fungal species besides S. cerevisiae, and may be an important regulator in some pathogenic yeasts.

RHO5 regulates the activity of the fMAPK pathway

Rho5p is a small GTPase of the Rho family (Garcia-Ranea and Valencia 1998; Singh et al. 2008, 2019; Schmitz et al. 2018). RHO5 was not a target of the fMAPK pathway; however, it is regulated by the GTPase-activating protein Rgd2p (Annan et al. 2008), which was a target of the fMAPK pathway (Figure 1B). The plate-washing assay showed that the rho5∆ mutant was defective for invasive growth (Figure 7A, Washed and Invasion) supporting previous observations (Ryan et al. 2012; Foster et al. 2013). To explore if Rho5p regulates invasive growth through the fMAPK pathway, the rho5∆ mutant was tested for fMAPK pathway activity by a transcriptional (growth) reporter [FUS1-HIS3] and a lacZ reporter [FUS1-lacZ]. Both reporters reflect the activity of the fMAPK pathway in cells lacking an intact mating pathway [ste4∆ (Cullen et al. 2004)]. The rho5∆ mutant was defective for fMAPK pathway activity based on growth on SD-HIS media (Figure 7A, SD-HIS). This was not due to a growth defect on synthetic media (Figure 7A, SD+AA). The fMAPK pathway also showed reduced activity in the rho5∆ mutant by the FUS1-lacZ reporter (Figure 7A, FUS1-lacZ). These results indicate that Rho5p may play a subtle role in regulating the fMAPK pathway. We did not find a link between Rgd2p and Rho5p in the regulation of the fMAPK pathway because the rgd2∆ mutant, unlike the rho5Δ mutant, did not show a change in fMAPK pathway activity by the FUS1-HIS3 or the FUS1-lacZ reporters (Figure 7A). By microscopy, the rho5∆ mutant also showed misshaped cell morphology and improper budding (Figure 7B). Overall, the data establishes Rho5p as a positive regulator of the fMAPK pathway.

Figure 7.

Rho5p regulates the fMAPK pathway. (A) Plate-washing assay of wild type (PC538+pRS316) and the rho5∆ (PC7306), rgd2∆ (PC7146), and ste12∆ (PC539) mutants spotted on YPD (10% Glu) and grown for 3 days. Left column, colonies, second column, inverted images of plates after wash, Bar, 0.5 cm. Invasion, quantification of invasive scars by ImageJ in triplicate, with wild type values set to 1. Error represents the SEM, which varied ≤ 50% across three trials. Asterisks, P-value < 0.035, by Student’s t-test compared to wild type. The rgd2∆ mutant invasion value is from Figure 2. SD+AA, strains spotted onto SD+AA and grown for 3 days. SD-HIS, transcriptional (growth) reporter [FUS1-HIS3]. Strains grown on SD-HIS+ATA (3-amino-1,2,4-triazole) medium for 3 days. FUS1-lacZ, β-Galactosidase (lacZ) assays. Cells grown in SD-URA for 16 hr, washed, and resuspended in YPD (10% Glu) medium for 6.5 hr prior to harvesting cells by centrifugation. Error represents SEM, which varied < 10% across three trials. Asterisk, P-value < 0.01. (B) Cell morphology compared after growth in YPD (2% Glu) for 16 hr between indicated strains. Bar, 10 µm.

Discussion

Signaling pathways can regulate the activity of transcription factors that control the expression of many genes that collectively generate cellular responses. To have a full understanding of the cellular responses a pathway generates, one must characterize the functions of its individual targets of the signaling pathway. Here, we characterized targets of the fMAPK pathway in S. cerevisiae. This led to the discovery that, even though the fMAPK pathway overwhelmingly regulates filamentous growth positively, the pathway can also negatively regulate or modulate filamentous growth under some conditions. This also led to the discovery of new positive roles for the pathway in controlling cell adhesion and the cell cycle (Figure 8). In addition, by trying to identify how the target RGD2 regulates invasive growth, we uncovered that RHO5 positively regulates the fMAPK pathway (Figure 8).

Figure 8.

Model of how newly identified targets of the fMAPK pathway impact cell adhesion and cell elongation during filamentous growth. Nfg1p negatively regulates the fMAPK pathway. Other targets also negatively regulate filamentous growth (not shown). The fMAPK pathway induces cell adhesion by regulating FLO11 expression, in part through Sfg1p, and by preventing cell separation through Sfg1p-dependent repression of DSEs and SCW11. The fMAPK pathway induces cell elongation by regulating the G₁-specific cyclin CLN1 and SFG1, which promotes extension of G₂/M. Rho5p regulates the fMAPK pathway. Pathway components are highlighted in red (MSB2, KSS1, STE12, TEC1).

A major role of the fMAPK pathway is to positively regulate invasive growth (Roberts and Fink 1994; Cook et al. 1997; Roberts et al. 2000; Cullen and Sprague 2012). Here, we show that the fMAPK pathway also negatively regulates invasive growth. This occurred under certain conditions by regulating the expression of NFG1, RGD2, RPI1, and TIP1. Moreover, the fMAPK pathway induces the expression of these negative regulators to modulate the formation of invasive aggregates. This adds four new proteins to the large group of proteins that negatively regulate filamentous growth, including Sfl1p (Fujita et al. 1989; Robertson and Fink 1998; Song and Carlson 1998; Pan and Heitman 2002), Nrg1p (Park et al. 1999; Zhou and Winston 2001; Kuchin et al. 2002), Sok2p (Ward et al. 1995; Pan and Heitman 2000, 2002), and Dig1p (Cook et al. 1996; Tedford et al. 1997; Bardwell et al. 1998b; Olson et al. 2000). Rgd2p, Rpi1p, and Tip1p are conserved in several yeast species, including pathogens, and might have similar functions in these species. Nfg1p, however, only has homology within Saccharomyces yeast. Perhaps Nfg1p aids in a specific aspect of Saccharomyces ecology not found in other fungi.

Because the fMAPK pathway is involved in both the negative and positive regulation of filamentous growth, it implies the importance of fine tuning in the regulation of this response. Modulation ensures cells do not “overdo” filamentous growth, which, in some environments, could have negative impacts. For example, when Dig1p is overexpressed, it gives cells a growth advantage in liquid cultures, but reduces growth on semisolid surfaces (Tan et al. 2013). Furthermore, a dig1∆ mutant has decreased biofilm/mat expansion (Karunanithi et al. 2012), which could make it more difficult to scavenge nutrients. Finally, elevated levels of Flo11p, although beneficial for invasive growth, dampens biofilm/mat expansion (Karunanithi et al. 2010).

Nfg1p has been an established, highly induced target of the fMAPK pathway with a function that has remained elusive for some time [YLR042c, (Caro et al. 1997; Hamada et al. 1999; Madhani et al. 1999; Roberts et al. 2000; Giaever et al. 2002; Hohmann 2002; García et al. 2004; Kim and Levin 2010; Parachin et al. 2010; Adhikari and Cullen 2014; Chow et al. 2019b)]. Here, we show Nfg1p regulates invasive growth by dampening the activity of the fMAPK pathway (Figure 8). This fits a common theme among some pathway targets that are induced to dampen pathway activity, resulting in negative feedback (Borneman et al. 2006). Rgd2p, Rpi1p, and Tip1p act at least somewhat separately from Nfg1p and each other to modulate invasive growth. Rgd2p, Rpi1p, and Tip1p may dampen other pathways that regulate filamentous growth (Gimeno et al. 1992; Lorenz and Heitman 1998; Carlson 1999; Pan and Heitman 1999; Cullen and Sprague 2000, 2012; Crespo et al. 2002; Lamb and Mitchell 2003) because extensive cross regulation between pathways occurs in a complex regulatory network (Bharucha et al. 2008; Chavel et al. 2010, et al.2014; Chow et al. 2019b). As currently appreciated, it is not clear how signal amplification is curbed in the network. Here, we provide a possible explanation for this, by pathways making products that presumably dampen the activity of other pathways from the signaling network. For example, the fMAPK pathway may target RPI1 because it dampens the Ras/cyclic AMP pathway (Kim and Powers 1991; Sobering et al. 2002), which also regulates filamentous growth (Mosch et al. 1996; Pan and Heitman 1999; Rupp et al. 1999; Cullen and Sprague 2012).

We also show that the transcriptional repressor SFG1 is a target of the fMAPK pathway. Sfg1p regulates an entire filamentation program—it prevents cell separation by repressing genes encoding daughter-cell-wall degrading enzymes, it triggers cell cycle delay resulting in an elongated cell morphology, and it induces FLO11 expression (Figure 8). Sfg1p also regulates cell adhesion separate from Flo11p. Thus, the regulation of SFG1 expression by the fMAPK pathway identifies a new mechanism by which the fMAPK pathway regulates cell adhesion. Sfg1p and Flo11p do not always contribute equally to cell-adhesion responses, and cell-adhesion regulation by both proteins was affected by the environment. Flo11p regulated cell adhesion more intensely on semisolid than in liquid media, and both Sfg1p and Flo11p regulated invasive growth differently depending on the carbon source present. These new conditional mechanisms indicate that cell adhesion regulation is more complex than currently appreciated and suggests that, in yeast, there is an ‘adhesion code’. For example, we show here that the adhesion code is dependent on the regulation of adhesion molecules, cell-wall-degrading enzymes, and transcription factors, which are controlled differentially depending on the environment. Given the large number of adhesion molecules in C. albicans and other species (Tronchin et al. 1991; Brandhorst et al. 1999; Sheppard et al. 2004; Dranginis et al. 2007; Linder and Gustafsson 2008; Younes et al. 2011; de Groot et al. 2013; Lipke 2018; Takahashi-Nakaguchi et al. 2018), it is likely that the adhesion code in other species is similarly (or more) complex.

Sfg1p also regulated cell elongation, and based on previous work has been shown to induce a delay in G₂/M (White et al. 2009). Overall, it appears the fMAPK pathway integrates separate regulatory modes of filamentous growth into one response: (1) regulating cell adhesion by repressing the expression of genes that encode proteins involved in cell separation and inducing the expression of FLO11 and (2) regulating the cell cycle at G₁ through CLN1 and G₂ through SFG1 to promote cell elongation (Figure 8). Having multiple mechanisms to regulate the same response increases the fine tuning capabilities of the pathway, making slight adjustments for different environments possible. Sfg1p is conserved across some species of yeast, including pathogens like C. glabrata, and could represent an important regulator of filamentous growth that leads to nuanced responses in other species.

In conclusion, by characterizing transcriptional targets of the fMAPK pathway, we have identified novel roles for the pathway in regulating invasive growth, cell adhesion, and the cell cycle. Some of these mechanisms may be conserved in pathogenic yeasts and could assist in understanding fungal infections. Here, we focused on highly induced targets of the fMAPK pathway; however, there are many other targets genes that are induced at lower levels that could impact phenotype. Moreover, there are also many targets whose expression is repressed that may tell us phenotypic information about the fMAPK pathway if explored. Overall, these findings suggest characterizing the genetic targets of other signaling pathways could lead to important advances in understanding signal transduction regulation.