|Home | About | Journals | Submit | Contact Us | Français|
The Snf1 protein kinase of Saccharomyces cerevisiae has been shown to have a role in regulating haploid invasive growth in response to glucose depletion. Cells contain three forms of the Snf1 kinase, each with a different β-subunit isoform, either Gal83, Sip1, or Sip2. We present evidence that different Snf1 kinases play distinct roles in two aspects of invasive growth, namely, adherence to the agar substrate and filamentation. The Snf1-Gal83 form of the kinase is required for adherence, whereas either Snf1-Gal83 or Snf1-Sip2 is sufficient for filamentation. Genetic evidence indicates that Snf1-Gal83 affects adherence by antagonizing Nrg1- and Nrg2-mediated repression of the FLO11 flocculin and adhesin gene. In contrast, the mechanism(s) by which Snf1-Gal83 and Snf1-Sip2 affect filamentation is independent of FLO11. Thus, the Snf1 kinase regulates invasive growth by at least two distinct mechanisms.
In fungi such as Candida albicans, Cryptococcus neoformans, and Ustilago maydis, the ability of cells to undergo a dimorphic transition between yeast-like growth and filamentous or hyphal growth is an important determinant of pathogenicity. The budding yeast Saccharomyces cerevisiae also exhibits a dimorphic transition in response to nutrient limitation and provides a convenient genetic system for studying this process. The nature and regulation of the dimorphic transition is determined by ploidy (for reviews, see references 10, 22, and 28). Haploid S. cerevisiae cells initiate filamentous invasive growth upon glucose depletion (7, 37), whereas diploid cells make a transition to pseudohyphal growth in response to nitrogen starvation (12), although filamentous growth can also occur in response to carbon source depletion (8, 21). The distinct form of pseudohyphal growth that occurs in mutants lacking the forkhead transcription factors, which control cell cycle-regulated genes, is nutrient independent (15, 51).
Haploid invasive growth depends on FLO11, a gene encoding a cell surface glycoprotein that functions as a flocculin or adhesin (13, 21, 24, 25, 36). FLO11 has a large and complex promoter that is regulated by the cyclic AMP-dependent protein kinase and mitogen-activated protein kinase pathways (24, 27, 32, 38, 40). The Snf1 kinase, which is required for haploid invasive growth (7), also regulates FLO11 expression in response to glucose depletion by antagonizing Nrg1- and Nrg2-mediated repression of FLO11 (20, 47). One aspect of invasive growth is adherence to the support (36), and Snf1 and the Nrg repressors correspondingly affect the Flo11-dependent adherence of cells to a plastic surface (20). Another aspect of invasive growth is filamentation, which entails cell elongation and a switch from axial to unipolar budding; these morphological changes also require Snf1 (7).
The Snf1 kinase is highly conserved in fungi, plants, and animals (called AMP-activated kinase in mammals), and this family of kinases has broad roles in transcriptional and metabolic regulation in response to stress (for reviews, see references 14 and 19). In the pathogenic yeast C. albicans, Snf1 function is essential for viability (9, 34). In S. cerevisiae, Snf1 is required for many aspects of transcriptional and metabolic adaptation to glucose limitation (5, 11) and has been implicated in other stress responses (1, 43). Besides haploid invasive growth, Snf1 also affects developmental processes such as diploid pseudohyphal growth (20), aging (3, 23), and meiosis and sporulation.
The participation of Snf1 in diverse regulatory responses in S. cerevisiae is facilitated by the existence of multiple forms of the kinase, as is also the case in mammals. S. cerevisiae cells contain three forms, each comprising the catalytic subunit Snf1, the activating subunit Snf4, and one of three β-subunit isoforms, Gal83, Sip1, or Sip2 (17, 49). We will refer to these forms by designating the β subunit, for example, Snf1-Gal83. Although the β subunits exhibit significant functional redundancy, they also have important roles in regulating the specificity of the kinase (3, 42, 45, 49); for example, Gal83 mediates the physical interaction of the kinase with Sip4, a Snf1-dependent transcriptional activator of gluconeogenic genes (45). The β subunits also regulate the subcellular localization of the kinase and presumably its access to different substrates. All of the β subunits are cytoplasmic when cells are grown in abundant glucose; upon glucose depletion, Gal83 directs Snf1 to the nucleus, Sip1 is relocalized apparently to membranes and then to the vacuole, and Sip2 remains cytoplasmic (46). Finally, at least one kinase form is subject to multiple regulatory inputs: Snf1-Gal83 is regulated both by the glucose signaling pathway that inhibits its catalytic activity and by a distinct pathway that controls its localization in response to fermentable carbon sources (46). Thus, the β subunits both confer specificity and provide versatility in the control of different functions of Snf1.
To explore the functions of the Snf1 kinase in regulating invasive growth, we have examined the roles of different Snf1 kinases in adherence to the support and filamentation. We present evidence that Snf1 affects adherence by a pathway involving the Snf1-Gal83 form, the Nrg repressors, and FLO11. In contrast, both Snf1-Gal83 and Snf1-Sip2 affect filamentation by a FLO11-independent pathway(s). Thus, these studies reveal two distinct mechanisms for regulation of invasive growth by the Snf1 kinase.
S. cerevisiae strains used in this work are listed in Table Table1.1. All strains were in the Σ1278b genetic background and were derived by standard genetic methods (39) from the isogenic strains MY1401, MY1402, and MY1411 of the Sigma2000 series (Microbia, Cambridge, Mass.). gal83Δ::KlURA3 is a complete disruption of the GAL83 coding sequence created by homologous recombination with the direct repeat-flanked Kluyveromyces lactis URA3 marker, using the method of adaptamer-directed gene disruption (35). Popout recombinants of gal83Δ::KlURA3 were selected on 5-fluoroorotic acid to generate gal83Δ. sip1Δ::KanMX6 (46) and snf1Δ10 (6) were introduced as described previously. sip2::KanMX6 is a complete deletion of the SIP2 coding sequence created by homologous recombination with the KanMX6 marker by using the method of PCR synthesis of marker cassettes (48). The FLO11 upstream region was replaced with the Schizosaccharomyces pombe adh+ promoter as described previously (31), using promoter sequence amplified from plasmid spADH-CLB2 (2), to generate the SpADHp-FLO11 allele. All mutations were first introduced into MY1401, and combinations of these alleles were then obtained by genetic crossing. Genotypes were established by mutant phenotypes and by analysis of genomic DNA using PCR. Complete synthetic medium (CSM) was purchased commercially (BIO 101 Systems; Q-BIOgene, Carlsbad, Calif.), and rich medium was yeast extract peptone (YEP).
The plate washing assay described by Roberts and Fink (37) was modified as follows. Fresh colonies from YEP-plus-2% glucose plates were resuspended in TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.5), or fresh overnight cultures were grown in YEP plus 2% glucose and cells were collected, washed, and resuspended in TE buffer; both methods gave similar results in invasive growth assays. Cells were spotted onto CSM containing 0.1% glucose and 2% agar, incubated at 26°C for 1 to 4 days, and photographed. Plates were washed under a stream of water by rubbing with a bent glass rod and then photographed again. Invasion was also assayed on YEP containing 2% glucose and 2.5% agar.
Assays for adherence to the wells of a polystyrene 96-well microtiter plate (Falcon Microtest flat bottom plate, 35-1172; Becton-Dickinson Labware) were carried out as described elsewhere (36). Cells were grown in CSM plus 2% glucose to an optical density at 600 nm (OD600) of 0.5 to 1.5, collected, washed, and resuspended to an OD600 of 1 in CSM with 2% or 0.1% glucose. Cells (0.1 ml) were transferred to the wells of a microtiter plate and incubated at 30°C for 1 to 7.5 h, as indicated. The cells were then stained with crystal violet, and the wells were washed repeatedly with water.
Cells were grown to stationary phase in YEP plus 2% glucose liquid medium, washed twice with TE buffer, diluted, and spread at low concentration onto CSM with 0.1% or no glucose; 0.1% glucose was as repressive for filamentation as 2% glucose, but overall growth was more closely matched to that occurring on medium with no added glucose. Plates were incubated at 26°C for 16 to 24 h, and microcolonies were viewed using a Nikon Eclipse E800 fluorescence microscope. Images were taken with an Orca100 (Hamamatsu) camera, using Open Lab (Improvision) software, and processed in Adobe Photoshop 5.5.
To test the requirement for the different β subunits in invasive growth, we examined isogenic sip1Δ, sip2Δ, and gal83Δ mutants in the Σ1278b genetic background. Cell suspensions were spotted on solid CSM containing a low concentration of glucose (0.1%); previously, Cullen and Sprague showed that invasion occurs in response to glucose limitation (7). After incubation for 2 days, the plates were washed. Cells that remained on the plate, having adhered to or penetrated the agar, are described here as exhibiting invasive growth (Fig. (Fig.1A).1A). The gal83 mutant was markedly defective, as were the snf1 and gal83 sip1 sip2 mutants. The sip1 mutant resembled the wild type, as reported previously (7), as did the sip2 mutant. Although the snf1 and gal83 sip1 sip2 strains grew poorly on 0.1% glucose, the gal83 mutant grew as well as the wild type; hence, there was no correlation between the extent of growth and invasion.
Because the β subunits exhibit significant functional redundancy, we also constructed double mutants (Fig. (Fig.1A).1A). The sip1 sip2 double mutant exhibited invasive growth, indicating that the Gal83 subunit is both necessary and sufficient. The gal83 sip1 mutant was noticeably more proficient in invasive growth than the gal83 single mutant, and the gal83 sip2 sip1 strain was more proficient than the gal83 sip2 strain; these findings suggest that Sip1 inhibits this process, either by a direct mechanism or perhaps indirectly through effects on the metabolic state of the cell. The sip2 mutation slightly reduced invasiveness of both gal83 and gal83 sip1 strains, suggesting that Sip2 has an auxiliary positive role in invasive growth. A positive effect of Sip2 was similarly apparent in a reg1 gal83 mutant background (see below) (Fig. (Fig.1B).1B). This evidence that the Sip2 subunit contributes in some manner to invasive growth was confirmed by subsequent experiments revealing a role in filamentation (see below). Together, these findings indicate that the Snf1-Gal83 kinase has the primary role in invasive growth, but Sip1 and Sip2 also affect this process.
We also assayed invasive growth on YEP plus 2% glucose, which is commonly used for such assays; however, the snf1 and gal83 sip1 sip2 mutants showed a less reproducible defect than on CSM with 0.1% glucose or YEP with 0.1% glucose (data not shown). Invasive growth of the wild type occurs more rapidly on 0.1% glucose than on 2% glucose (Fig. (Fig.1B).1B). We suggest that on 0.1% glucose, the Snf1 kinase is critical for an immediate invasive response, whereas during growth on 2% glucose, wild-type and mutant cells differ in their capacity for metabolic adaptation to glucose depletion, which in turn affects the time course of invasion. In previous studies, a snf1::URA3 mutant exhibited a pronounced defect in invasiveness on YEP plus 2% glucose relative to the SNF1 ura3 parent (7), but the magnitude of the defect may have been enhanced by the auxotrophic difference because, in our strains, the ura3 mutation improves invasive growth (data not shown).
To confirm that Snf1-Gal83 is required for invasive growth on YEP plus 2% glucose, we took advantage of our previous finding that a reg1 mutant is much more invasive than the wild type on this medium (20). Reg1 directs protein phosphatase type 1 to inhibit the Snf1 kinase (16, 26, 29, 41, 44), and mutation of REG1 relieves glucose repression of many genes, including FLO11 (20). The effect of reg1 on invasive growth requires the Snf1 kinase (20) and presumably reflects the relief of glucose inhibition of the kinase, because the reg1 mutant was more invasive than wild type on 2% glucose but not on 0.1% glucose (Fig. (Fig.1B).1B). The gal83 mutation caused a defect in invasive growth of the reg1 mutant on both YEP plus 2% glucose and CSM with 0.1% glucose; moreover, sip2 further reduced invasion of the reg1 gal83 mutant on both media. These findings exclude the possibility that the effects of gal83 and sip2 are specific to a particular growth medium.
The ability of yeast cells to adhere to a plastic surface is related to their capacity for invasive growth in that both processes require the Flo11 flocculin and adhesin, which is essential for adherence to the support (13, 36), and both are dependent on Snf1 kinase activity (20). We therefore tested the role of Gal83 in adherence to plastic. Wild-type and gal83 mutant strains were grown to exponential phase in 2% glucose, resuspended in 0.1% glucose, and transferred to the wells of a microtiter plate. The mutant cells adhered to the plastic less well than the wild-type cells, although the difference was not striking (Fig. (Fig.2A).2A). We also tested strains carrying the reg1 mutation, which enhances plastic adherence (dependent on the Snf1 kinase and FLO11 ) and therefore improves the sensitivity of the assay. The reg1 gal83 mutant was clearly defective in adherence relative to the reg1 strain and resembled the reg1 snf1 and reg1 sip1 sip2 gal83 strains (Fig. (Fig.2B).2B). Thus, Snf1-Gal83 has a role in adherence.
Previous studies showed that one of the roles of the Snf1 kinase in regulating invasive growth is antagonism of the zinc-finger proteins Nrg1 and Nrg2, which repress FLO11 (20). Both Nrg repressors interact physically with Snf1 (47), and mutation of NRG1 and NRG2 alleviates snf1 mutant defects in invasive growth, adherence to plastic, and expression of the STA2 (glucoamylase) promoter (20), which is nearly identical to that of FLO11 for 3.5 kb. To determine whether the Snf1-Gal83 form of the kinase is responsible for antagonism of Nrg-mediated repression, we examined the genetic relationships of nrg1 and nrg2 to gal83 with respect to these phenotypes. First, we assayed invasive growth; in a gal83 mutant, each single nrg mutation had some effect and both nrg mutations together strongly restored invasive growth (Fig. (Fig.3).3). The nrg mutations also alleviated the gal83 mutant defect in adherence to plastic (Fig. (Fig.2A).2A). Finally, we assayed expression of lacZ fused to the STA2 promoter in β subunit mutants during growth in glucose and after a shift to low (0.05%) glucose. Derepression of STA2-lacZ was greatly reduced by gal83 but was not affected by mutation of the other β subunits (Fig. (Fig.4A),4A), and expression was restored in a gal83 nrg1 nrg2 triple mutant (Fig. (Fig.4B).4B). These data implicate Snf1-Gal83 in relief of glucose repression by the Nrg proteins. Taken together, these genetic findings support the view that the Gal83 subunit mediates the functional interaction of the Snf1 kinase with Nrg1 and Nrg2.
We next tested the role of different Snf1 kinases in filamentation, which entails cell elongation and a switch from axial to unipolar budding. Filamentation occurs in response to low glucose, and a snf1 mutant is defective in both elongation and unipolar budding (7). We used the assay devised by Cullen and Sprague (7) to observe morphological changes. Cells were spread onto CSM lacking glucose and allowed to form microcolonies. Microscopic examination of the β subunit mutants showed that each single mutant was fully competent for filamentation, as were the sip1 sip2 and sip1 gal83 double mutants (Fig. (Fig.5A).5A). The sip2 gal83 double mutant, however, was defective and resembled the snf1 and sip1 sip2 gal83 mutants (Fig. (Fig.5A5A).
We further examined the effects of a reg1 mutation. The reg1 mutant exhibited filamentous growth on medium containing no, 0.1, or 2% glucose (Fig. (Fig.5B5B and data not shown), consistent with its other glucose-insensitive phenotypes. However, reg1 snf1 cells did not filament, indicating that the enhanced filamentation depends on Snf1 activity. This finding allowed us to test the phenotypes caused by β subunit mutations on medium containing glucose and thereby rule out the possibility that the impaired filamentation of sip2 gal83 cells results from their poor growth in the absence of glucose. The combinations gal83 sip2 and gal83 sip1 sip2 impaired filamentation in the reg1 mutant background, whereas other combinations had no noticeable effect (Fig. (Fig.5B).5B). These findings indicate that either Gal83 or Sip2 is sufficient for filamentation.
The role of Sip2 in filamentation most likely accounts for the observation that mutation of SIP2 in a gal83 strain further diminishes invasive growth (Fig. (Fig.1).1). The following evidence also suggests that Snf1-dependent filamentation contributes to invasive growth. Replacement of the endogenous FLO11 promoter with the S. pombe adh+ promoter was shown to give constitutive, low-level FLO11 expression and to abolish invasion of wild-type cells (31); in our strain background, this replacement similarly abolished invasive growth (Fig. (Fig.6B).6B). However, reg1 mutant cells carrying the SpADHp-FLO11 allele filamented in the presence of glucose, unlike wild-type cells (Fig. (Fig.6A),6A), and exhibited detectable invasive growth on agar (Fig. (Fig.6B).6B). Although it is possible that reg1 upregulates other genes besides FLO11 to affect adherence to agar, these findings are consistent with the idea that Snf1-dependent filamentation contributes to invasiveness.
If Snf1-Gal83 and Snf1-Sip2 affect filamentation through different, redundant pathways, then Snf1-Gal83 could conceivably promote filamentation through the same pathway by which it affects adherence. To address this possibility, we examined the roles of the Nrg repressors and FLO11 in filamentation. The nrg1 nrg2 double mutation did not significantly affect filamentation in an otherwise wild-type genetic background or suppress the filamentation defect of a snf1 mutant on medium containing 0.1% or no glucose (data not shown). Previous studies showed that diploid filamentation requires FLO11 (24, 40); in contrast, we found that in haploids, FLO11 is dispensable, as both flo11Δ and reg1 flo11Δ mutant cells exhibited filamentous morphology similar to that of the corresponding FLO11 strains in either 0.1% or no glucose (Fig. (Fig.6C).6C). These results are consistent with recent evidence that flo11 does not affect haploid filamentous morphology in another strain background (8). Thus, the filamentation pathway involving Snf1-Gal83 is distinct from the pathway by which Snf1-Gal83 affects adherence to agar and plastic.
We have investigated the roles of the Snf1 kinase in regulating haploid invasive growth in response to glucose depletion. We show that different forms of the kinase are required for two different aspects of invasive growth, namely, adherence to the agar substrate and filamentation. The Snf1-Gal83 form of the kinase affects adherence by a mechanism involving the Nrg repressors and FLO11. The Snf1-Gal83 and Snf1-Sip2 kinases together affect cell morphology by a mechanism(s) that is independent of Nrg proteins and FLO11. Thus, the Snf1 kinase has multiple roles in the complex process of invasive growth (Fig. (Fig.7).7). These findings also provide further evidence that the different β subunits confer specificity to the Snf1 kinase in cellular regulatory responses.
We first showed that the effects of Snf1 on invasive growth are mediated primarily by Snf1-Gal83. Gal83 is both necessary and sufficient for invasive growth, as a gal83 mutant was defective and a sip1 sip2 strain was proficient. However, the sip2 mutation caused a noticeable decrease in invasive growth in a gal83 mutant background, which we attribute to the defect in filamentation. In contrast, mutation of SIP1 improved invasive growth in a gal83 mutant, suggesting an inhibitory role for Sip1. Snf1-Sip1 may directly inhibit some function that promotes invasive growth or may act indirectly through effects on the metabolic state of the cell. We think it unlikely that loss of Sip1 simply affects the distribution of the Snf1 catalytic subunit among the different isoforms because Sip1 is not an abundant β subunit and because loss of Sip1 has an effect in a gal83 sip2 mutant.
Genetic evidence indicates that the effects of Snf1-Gal83 on adherence occur, at least in part, by inhibition of the Nrg1 and Nrg2 repressors. These repressors may in fact be the major targets, because adherence to both agar and plastic substrates was effectively restored in a gal83 mutant by the double nrg1 nrg2 mutation. Snf1 interacts physically with Nrg1 and Nrg2 (47), and one possibility is that Gal83 interacts directly with these repressors, as it does with the activator Sip4 (45). We were unable to detect two-hybrid interaction of Gal83 with Nrg1 or Nrg2 in several different assays (V. K. Vyas and C. D. Berkey, unpublished results), but these negative results do not exclude the possibility that Gal83 contributes to the Snf1-Nrg interaction. Alternatively, Gal83 could simply facilitate the interaction between Snf1 and these repressors by virtue of its role in the nuclear localization and nuclear function of the kinase (46). Gal83 is the β subunit that is primarily responsible for the nuclear localization of Snf1 upon glucose depletion (46), and Nrg1 is a nuclear protein (47).
We present evidence that one of the functional targets of the Snf1-Gal83/Nrg pathway is FLO11, which is essential for invasive growth and adherence to plastic (21, 24, 25, 36). This pathway may also regulate other genes needed for invasive growth, as only a few Nrg1- and Nrg2-repressed genes besides FLO11 have been identified (20, 33, 47, 50). It is noteworthy that an ortholog of the Nrg proteins in the fungal pathogen C. albicans has been shown to repress filamentous growth and expression of the hypha-specific adhesin genes HWP1, ALS3, and ALS8 (4, 30).
We also examined the role of the Snf1 kinase in promoting filamentation and found that either the Snf1-Gal83 or Snf1-Sip2 form is sufficient. The simple model is that the Snf1 kinase affects filamentation by a pathway in which both Snf1-Gal83 and Snf1-Sip2 can participate, such that the presence of either one alone suffices for function of the pathway (Fig. (Fig.7A).7A). Although Gal83 becomes enriched in the nucleus in response to glucose depletion while Sip2 remains predominantly cytosolic (46), their different localizations do not preclude performance of an identical function. Alternatively, the Snf1 kinase could promote filamentation by two redundant pathways, one involving Snf1-Gal83 and the other Snf1-Sip2; however, Snf1-Gal83 must act through a pathway different from that by which it regulates adherence (Fig. (Fig.7B).7B). The function(s) of the Snf1 kinase in filamentation could potentially involve transcriptional control or regulation of metabolic enzymes or other cellular components. Very recent work indicates that one function of Snf1 is to promote the disappearance of the axial budding determinant Axl1 in response to glucose limitation (8).
This study thus identifies at least two distinct signaling mechanisms by which the Snf1 kinase contributes to control of two aspects of invasive growth. It is likely that Snf1 participates in yet other regulatory interactions besides those represented in Fig. Fig.7;7; for example, Snf1-Gal83 may relieve Nrg-mediated repression of other genes besides FLO11, as mentioned above. The involvement of both Snf1-Gal83 and Snf1-Sip2 also raises the possibility that adherence and filamentation are differentially regulated by nutrient signals. A glucose signaling pathway inhibits the Snf1 catalytic activity, but a distinct pathway regulates the nucleocytoplasmic distribution of Snf1-Gal83 in response to fermentable carbon sources, whereas Snf1-Sip2 is constitutively cytoplasmic (46). Further studies are needed to resolve the various roles of Snf1 in the very complex regulatory network that controls invasive growth.
We thank A. Amon, A. Wach, R. Rothstein, and R. Reid for plasmids.
This work was supported by NIH grant GM34095 to M.C.