|Home | About | Journals | Submit | Contact Us | Français|
Haploid yeast cells are capable of invading agar when grown on rich media. Cells of the Σ1278b genetic background manifest this property, whereas other laboratory strains are incapable of invasive growth. We show that disruption of the RAS2 gene in the Σ1278b background significantly reduces invasive growth but that expression of a constitutively active Ras2p (Ras2Val19p) in this strain has a minimal effect on its invasiveness. On the other hand, expression of Ras2Val19p in another laboratory strain, SP1, rendered it invasive. These results suggest that a hyperactive Ras2 pathway induces invasive growth and that this pathway might be overactive in the Σ1278b genetic background. Indeed, cells of the Σ1278b are defective in the induction of stress-responsive genes, while their Gcn4 target genes are constitutively transcribed. This pattern of gene expression was previously shown to be associated with an active Ras/cyclic AMP (cAMP) pathway. We show that suppression of stress-related genes in Σ1278b cells is a result of their inability to activate transcription through the stress response element (STRE). Disruption of RAS2, which abolished invasiveness, induced an increase in STRE activity. Further, in the SP1 genetic background, disruption of either the MSN2/4 genes (encoding activators of STRE) or the yAP-1 gene was sufficient to restore invasive growth in ras2Δ cells. We conclude that Ras2-mediated suppression of the stress response is sufficient to induce invasiveness. Accordingly, the fact that the stress response is suppressed in Σ1278b background explains its invasiveness. It seems that invasiveness is a phenotype related to unregulated growth and is therefore manifested by cells harboring an overactive Ras/cAMP cascade. In this respect, invasiveness in yeast is reminiscent of the property of ras-transformed fibroblasts to invade soft agar.
RAS genes encode GTP binding proteins, which mediate the signaling of many extracellular ligands into intracellular effectors (28, 34, 39, 48). In mammalian cells, Ras-activating stimuli include not only biological compounds such as growth factors and cytokines but also stress signals such as reactive oxygen species and UV radiation (12, 13, 30, 45). In response to any of those stimulants, Ras activates an array of intracellular pathways through which it induces the appropriate biological response. This response is cell specific (e.g., proliferation in fibroblasts and differentiation in neuronal cells) and signal specific (mitogenesis in response to growth factors, growth arrest, and/or apoptosis in response to stress). In many cell types, constitutive activation of Ras leads to oncogenic transformation, which is manifested by unregulated proliferation, sensitivity to stresses, and loss of contact inhibition of growth (3, 34). As a result, these cells form foci when grown on plastic plates and are capable of invading soft agar (11).
Similar to the case of mammalian cells, in the yeast Saccharomyces cerevisiae Ras is activated by both growth signals (e.g., glucose [8, 56]) and stress signals (e.g., UV radiation and starvation ). Ras activation in yeast may result in different phenotypes, depending on cell ploidity and growth conditions. In haploid cells, Ras induces mitogenesis through the cyclic AMP (cAMP) cascade (8, 19, 56). In diploid cells, Ras also activates the cAMP-protein kinase A (PKA) pathway and controls cell proliferation, but it also affects meiosis and sporulation (8, 19). Constitutive activation of the yeast Ras2/cAMP cascade leads to unregulated proliferation, sensitivity to various stresses (e.g., heat shock and nitrogen starvation), and extremely low rates of sporulation. All these phenotypes are associated with high intracellular cAMP levels and high PKA activity, which accelerate the transition from the G1 to the S phase of the cell cycle (8). The sensitivity to stress manifested by these mutants is directly related to their inability to arrest in G1 in response to stress and is also correlated with their inability to induce transcription of stress-related genes (4, 16, 37, 49). The unregulated Ras/cAMP pathway suppresses the activity of the stress-responsive transcription factors Msn2 and Msn4, which are responsible for activation of a large number of stress-related genes (21, 36, 37, 49, 53). Msn2 and Msn4 bind a cis element known as the stress response element (STRE), which is found in the promoters of a large number of stress-responsive genes (7, 40, 42). Another transcriptional activator, yAP-1, is also involved in activation of STRE (22). Unlike stress-related genes, which are suppressed in strains harboring an activated Ras/cAMP pathway, Gcn4 target genes (e.g., HIS3 and HIS4) are constitutively expressed in these strains (14). Normally these genes are expressed only in response to amino acid starvation (25).
Diploid cells grown under nitrogen starvation show yet another phenotype that requires an active Ras2p. These cells undergo a developmental switch from a yeast form to filamentous growth (also known as pseudohyphal growth [20, 41]). The filamentous-growth phenomenon of diploids is related to another phenomenon observed in haploid cells, invasive growth (46). This type of growth is specific to haploid cells of the Σ1278b genetic background grown on rich media. Under these conditions, cells penetrate the agar and therefore are not washed away under a water current (46). The molecular mechanisms that control invasive growth seem to be very similar to those governing the pseudohyphal growth in diploids. Both phenotypes require an intact cAMP-PKA pathway as well as the Cdc42/Ste20 signaling system, which activates the Ste11/Ste7/Kss1 cascade (known as the mating mitogen-activated protein (MAP) kinase pathway [35, 41, 50]). These two Ras-dependent pathways activate transcription factors whose activity is essential for invasive growth and filamentous growth. The Ras/cAMP cascade activates (probably through the Tpk2 subunit of PKA [44, 47]) the Flo8 transcription factor, while the Ras/mating-pathway cascade activates the Tec1-Ste12 transcriptional complex. Flo8 and Tec1-Ste12 seem to coactivate the transcription of genes, such as FLO11, which are essential for invasive growth and pseudohypha formation (50). It is not known whether suppression of expression of some genes is also essential for induction of these phenotypes.
Pseudohyphal growth and invasive growth are observed in a very few laboratory strains, in particular those of the Σ1278b genetic background (20). One explanation for the uniqueness of Σ1278b strain is that most laboratory strains carry a mutation that hampers their ability to form filaments or to show invasive growth. Strain S288C, for example, was shown to contain a mutation in its FLO8 gene, an essential gene for pseudohypha formation and invasive growth (31). Other laboratory strains (W303, for example) may harbor yet another mutation (31). In this paper we suggest another explanation for the unique phenotypes of the Σ1278b strain. We show that invasive growth is a phenotype associated with suppression of the cellular stress response by a constitutively active Ras2/cAMP pathway. This pathway is hyperactive in strains of the Σ1278b background.
Although the mating pathway and the cAMP cascades seem to cooperate for induction of filamentous growth, cells harboring a constitutively active Ras2/cAMP cascade (RAS2Val19; ira1Δ) do not require the mating pathway for pseudohyphal growth (33, 44, 50). Also, although expression of the Ras2Val19 protein enhances filamentous growth, disruption of the RAS2 gene does not eliminate completely the ability of the cell to form filaments. To eliminate this ability, disruption of both RAS2, and another gene, GPA2, is required (29). It seems, therefore, that the role of the Ras cascades in invasiveness and pseudohyphal growth is complex and involves interaction with other signaling pathways. The exact role of Ras in filamentous growth was not fully revealed, because many of its downstream targets, essential for pseudohyphal growth, are not known. Here we show that suppression of stress-responsive transcription factors, which were previously shown to be suppressed by the Ras/cAMP pathway, is essential for invasive growth. Hence, invasiveness requires gene suppression in addition to activation of some genes.
We show that cells lacking the RAS2 gene are not capable of invasive growth but that expression of a constitutively active Ras2p (RAS2Val19) in Σ1278b cells had no effect on their invasive growth capability. However, introduction of RAS2Val19 into another strain, SP1, dramatically enhanced its invasive-growth capability. These observations show that Ras2p is an essential component of invasive growth and suggest that the Ras2 cascade may be overactive in Σ1278b strains. We further prove the latter conclusion and show that the pattern of gene expression in Σ1278b strains is similar to that previously reported for mutants with an active Ras/cAMP cascade; namely, stress responsive genes are suppressed, while Gcn4 target genes are constitutively transcribed (4, 14, 16, 37, 49). In a search for Ras-dependent downstream components, essential for invasive growth, we tested the possible involvement of STE7 (a member of the mating pathway), GCN4, and the stress-responsive transcription factors Msn2, Msn4, and yAP-1 (activators of STRE [22, 40]). We report that disruption of either STE7 or GCN4 in a RAS2val19 strain did not affect its efficient invasive growth. On the other hand, disruption of MSN2/4 or yAP-1 in ras2Δ restored invasive growth in this strain. The effect of these disruptions on invasive growth is correlated with their effects on the transcription of stress-related genes. Transcription of these genes is spontaneously elevated in ras2Δ cells and is abolished in SP1ras2Δmsn2Δmsn4Δ and SP1ras2Δyap1Δ cells. In this respect, SP1ras2Δ msn2Δmsn4Δ and SP1ras2Δyap1Δ strains are similar to the SP1RAS2Val19 and Σ1278b strains, in which the stress response is suppressed. We conclude that invasive growth is a phenotype shown by cells in which the general stress response is suppressed by an overactive Ras cascade.
Cultures were grown on YPD medium (2% glucose, 1% yeast extract, 2% Bacto Peptone) or on synthetic dropout medium YNB [0.17% yeast nitrogen base without amino acids, 0.5% NH4(SO4)2, 2% glucose].
The strains used in this study are listed in Table Table1.1. RAS1 and RAS2 were disrupted by using plasmid pRa545 and pRa530, respectively (55). YAP1 was disrupted by using plasmid SM25 (43). MSN2 and MSN4 were disrupted by using plasmids pΔBX and pZfh45-1, respectively (17). STE7 was disrupted by using plasmid pNC149 (10). All disruptions were verified by PCR. Strain ΣL5527LH was made by consecutive disruptions of LEU2 and HIS3 in strain ΣL5227, using the hisG system. The strain SP1ras1Δras2Δ-/p21 was constructed in three steps: (i) disruption of RAS2, (ii) introduction of plasmid YGA-Ras (9) expressing the human Ha-ras gene, and (iii) disruption of RAS1.
pCTT1-lacZ (TRP) is a derivative of pLG-669Z (23). Using XhoI-SmaI digestion, a 1.5-kb fragment of the CYC1 promoter was deleted, creating plasmid pΔss-178. A StuI fragment within the URA3 gene of pΔss-178 was removed and replaced with a TRP1 fragment to create pΔss-178(TRP1). An oligonucleotide containing the STREs from the CTT1 promoter (identical to CTT-20 [reference 37 and see below] but with XhoI-compatible ends) was inserted into the XhoI site (position −178 of the CYC1 promoter) of pΔss-178(TRP1). The plasmid obtained was termed pSTRE-lacZ(TRP1). The simian virus 40 (SV40)-lacZ(TRP1) construct was created by inserting an oligonucleotide containing AP-1 recognition elements from the SV40 enhancer (24) into the XhoI site of pΔss-178(TRP1). The oligonucleotides used were SV40 (5′ TCGACATCTCAATTAGTCAGCAAG 3′ 5′ TCGACTTGCTGACTAATTGAGATG 3′) and CTT1-20 (5′ TCGATTCAAGGGGATCACCGGTAAGGGGCCAAG 3′ 5′ TCGACTTGGCCCCTTACCGGTGATCCCCTTGAA 3′).
RNAs were prepared and primer extension analysis was performed as described previously (16), with the following changes: 5 U of avian myeloblastosis virus reverse transcriptase (Pharmacia) was added per reaction, and the reaction mixture was incubated at 42°C. Gene-specific primers were ACTIN (5′-GTTATCAATAACCAAAGCAGCAAC-3′), CTT1 (5′-GTACGGAAAACCGTTTTGTAGAGA-3′), GPD1 (5′-GCATTCAAGTGGCCGGAAGT-3′), GPP2 (5′-CGTTAACTTTCAAAGATAGA-3′), HSP104 (5′-GATGTTGATGATCCGAAGCC-3′), TRX2 (5′-CCCAGATGCTAAAGCACTGTC-3′), HIS4 (5′-ACTATTGCATGAGGCCAGATCATC-3′), HSP26 (5′-GGTGCGTAGCCTCTTAAGCCG-3′), HSC82 (5′-TATCTAAAGCATCGGAGGCG-3′), and HAL3 (5′-CCCCTTCGCATCCTCATGGT-3′).
β-Galactosidase reactions were performed as previously described (1). In each case, the reaction was performed at least three times in duplicate. Results are shown as means and standard deviations.
Cells were streaked on YPD plates and grown for 5 days at 30°C and a further 2 days at room temperature. Then the plates were washed with water as described previously (46). Briefly, the plates were photographed (with a 35-mm camera), washed under a gentle stream of deionized water (applied through a 1-ml pipette), and immediately photographed again.
Cells were grown on YPD medium to early log phase (0.6 × 107 to 1.0 × 107 cells/ml). Then, 1 ml was removed, washed with water, resuspended in 1 ml of water and mixed with 300 μl of 20% perchloric acid. The mixture was vortexed vigorously for 1 min and incubated at 4°C for 120 min. Following neutralization of the cell extracts with KCHO3, cAMP levels were determined by using the [125I]cAMP scintillation proximity assay system (acetylated procedure) (Amersham Pharmacia Biotech). cAMP assays were performed twice in duplicate for each strain. The results shown are the mean of the two experiments.
Since constitutively active Ras2p enhances pseudohyphal growth in diploid cells (20, 41), we studied the possible involvement of Ras2p in a related phenomenon in haploid cells, invasive growth. We have constructed a strain expressing the Ras2Val19 protein (from an integrated gene) and a strain disrupted for its RAS2 gene. Both strains were of the Σ1278b genetic background. Figure Figure1A1A shows that the invasive-growth capability of the Σras2Δ strain is very weak whereas the ΣRAS2Val19 strain invades as efficiently as the wild type. Thus, although the invasive-growth phenotype of haploid cells is dependent on RAS2, it is not enhanced by the RAS2Val19 mutation in the Σ1278b genetic background. This observation may imply that the Ras2p-dependent cascade is constitutively active in Σ1278b cells. To test this notion, we carried out two types of experiments (i) to test whether expression of Ras2Val19 in another strain would convert it similarly to Σ1278b strains with respect to invasive growth and (ii) to test whether Σ1278b strains possess properties previously reported to be associated with an active Ras/cAMP pathway. To test the effect of the RAS2Val19 mutation in other strains, we searched for a laboratory strain that shows a weaker (but not completely defective) invasive growth than Σ1278b strains. We tested three strains with genetic backgrounds commonly used in research laboratories (S288C, W303, and SP1); (Table I). All three strains exhibited a weaker invasive activity than Σ1278b strains. The YPH102 strain (a S288C derivative) was completely defective in invasive growth (data not shown), while W303 and SP1 showed some invasive activity (data not shown, but see Fig. Fig.11 for strain SP1). We decided to use the SP1 strain and tested the SP1, SP1ras2Δ, and SP1RAS2Val19 strains. As shown in Fig. Fig.1A,1A, the weak invasive activity of the SP1 strain was completely eliminated in SP1ras2Δ. On the other hand, expression of the Ras2Val19 protein in the SP1 genetic background dramatically enhanced its invasive-growth capability (Fig. (Fig.1A).1A). Disruption of the BCY1 gene in the SP1 strain also enhanced its invasiveness (data not shown). We have also tested a strain of the SP1 genetic background in which the only functional Ras is the human Ha-ras protein. This protein, which is constitutively active in yeast (2, 19), also enhanced invasive growth but not as strongly as the yeast Ras2Val19 protein did (Fig. (Fig.1B).1B). We conclude that an active Ras2p is an essential component for invasive growth in both Σ1278b and SP1 genetic backgrounds.
Next, we tested some properties of Σ1278b strains to see whether they are similar to those of mutants harboring a constitutively active Ras/cAMP pathway. In these mutants (e.g., RAS2Val19, bcy1Δ, and ira1Δ), transcription of many stress-responsive genes is suppressed (4, 16, 37, 49) since the activity of Msn2 and Msn4 is suppressed (21, 37, 40, 49). As a result, these mutants are sensitive to stresses such as heat shock and nitrogen starvation (8, 16, 36, 49, 52). Another characteristic of these mutants is the constitutive expression of Gcn4 target genes (e.g., HIS3 and HIS4 ). We monitored the expression of stress-related genes as well as expression of the Gcn4 target gene HIS4 in the SP1, SP1RAS2Val19, and Σ1278b strains. Primer extension analysis (Fig. (Fig.2A)2A) revealed that transcription of stress-related genes was suppressed in Σ1278b. In fact, induction of heat shock genes in response to mild heat shock in this strain was significantly weaker than that in SP1RAS2Val19. On the other hand, transcription of HIS4 (a Gcn4 target gene) was high in Σ1278b cells grown on rich YPD medium and was similar to that of SP1RAS2Val19 (Fig. (Fig.2A);2A); i.e., in Σ1278b strains the stress response was suppressed and Gcn4 was constitutively active. Since many stress-related genes (including HSP26 and HSP104) contain STREs in their promoters, we tested whether expression of stress genes might be suppressed in Σ1278b cells, because transcription from STREs cannot be induced. We measured the activation of a STRE-lacZ reporter gene (Fig. (Fig.2B).2B). In SP1 cells, the activity of this reporter was about threefold higher than that measured in Σ1278b cells following exposure to H2O2. Another stress-responsive reporter gene, SV40-lacZ, carrying AP-1 sites which are activated by the stress-responsive activator yAP-1 (22, 24), was normally active in Σ1278b cells (Fig. (Fig.2B).2B). Thus, the Σ1278b strain seems to be specifically defective in activation of STRE. Finally, we measured intracellular cAMP concentrations in the various strains. The cAMP levels measured in Σ1278b strains (5.9 pmol/107 cells) were significantly higher than those of SP1 (4.0 pmol/107 cells). In fact, they were similar to cAMP levels measured in SP1RAS2Val19 (6.5 pmol/107 cells). The cAMP concentration in ΣRAS2Val19 (8.2 pmol/107 cells) was about 1.4-fold higher than that in the wild-type Σ strain, showing that the Ras/adenylyl cyclase system is not maximally active in the Σ1278b genetic background. We also measured the intracellular cAMP level of the W303 strain, which, similarly to SP1, invades the agar poorly. In this strain, the cAMP concentration was 2.6 pmol/107 cells, lower than the level in SP1.
Collectively, the above results support the notion that the Ras2p cascade is hyperactive in Σ1278b strains.
Having shown that Ras2p is an essential component of invasive growth, we wished to identify Ras2-dependent downstream elements through which it induces invasive growth. Since Ras2p is known to activate the mating pathway in diploid cells (41), to increase the transcriptional activity of Gcn4 (14), and to suppress the transcription of many stress-responsive genes (5, 16, 49, 52), we systematically tested whether these Ras2p-dependent activities are required for invasive growth. To test the possible involvement of the mating pathway, we constructed an SP1RAS2Val19ste7Δ strain. This strain exhibited invasive growth capability similar to that of the SP1RAS2Val19 strain (Fig. (Fig.3).3). To check if Gcn4 activation is required, we tested the invasiveness of an SP1RAS2Val19gcn4Δ strain (14). As shown in Fig. Fig.3,3, this strain invaded the agar as efficiently as the RAS2Val19 strain did. Therefore, Ras2-dependent activation of either the mating pathway or the Gcn4 pathway is not required for invasive growth.
Unlike the mating pathway and Gcn4, which are activated by Ras (14, 41), expression of stress genes is suppressed by Ras. Therefore, a possible role for stress genes in invasive growth should be addressed by using a different strategy.
Some stress genes, in particular heat shock genes, are spontaneously induced in strains harboring mutations in components of the Ras cascade (e.g., cyr1 [6, 16], cdc25ts [5, 22], and tpk1w tpk2 tpk3 ). We therefore determined the expression of stress genes in our strains. The mRNA levels of five genes involved in the oxidative stress response (TRX2 and CTT1), osmotic shock response (GPD1 and GPP2), and heat shock response (HSP104) were monitored (Fig. (Fig.4A).4A). In the wild type strains, expression of these genes was low, in some cases undetectable (Fig. (Fig.4A).4A). However, in both SP1ras2Δ and Σras2Δ strains, the mRNA levels of these genes were significantly elevated in the absence of any stress stimulus (Fig. (Fig.4A).4A). By contrast, in the SP1RAS2Val19 strain the mRNA levels of these genes were lower than in SP1. The situation in strains derived from the Σ1278b background is different. No dramatic differences in the levels of expression of stress-related genes were observed when wild-type and ΣRAS2Val19 strains were compared. Some of the genes tested (CTT1, HSP104, and GPP2) contain STRE or STRE-like elements in their promoters (26, 42, 51). We therefore tested the activity of a STRE reporter in the different strains. Similar to the case for the mRNA levels (Fig. (Fig.4A),4A), expression of the reporter was significantly elevated in SP1ras2Δ and Σras2Δ, the strains that are defective in invasive growth (Fig. (Fig.4B).4B). Notably, the reporter activity in Σras2Δ was less than 50% of that measured in SP1ras2Δ, providing another indication for the weak STRE-dependent transcription in Σ1278b strains. It is clear, however, that in both genetic backgrounds, elevated expression of stress-related genes is correlated with the loss of the invasive-growth capability. These results also suggest that in the Σ1278b genetic background, the hyperactive component of the Ras2/cAMP pathway, which is responsible for suppression of STRE-mediated transcription, functions upstream of Ras2p, since disruption of RAS2 results in relatively efficient activation of the entire downstream stress cascade.
The question remains, however, whether the spontaneous expression of stress genes in ras2Δ strains not only is correlated with but actually is the cause of their inability to undergo invasive growth. To test whether suppression of stress genes is directly responsible for invasive growth, we disrupted MSN2/4 or yAP-1 in the background of SP1ras2Δ. Disruption of these genes should result in the suppression of stress genes, whose levels are spontaneously elevated in ras2Δ strains. We could then use these strains to assay whether the artificial elimination of the STRE-mediated stress response, would be sufficient to restore invasive growth to ras2Δ cells. We constructed SP1ras2Δmsn2Δmsn4Δ and SP1ras2Δyap1Δ strains and tested the expression of stress genes in these strains compared to that in SP1ras2Δ (Fig. (Fig.5A).5A). Expression of all genes tested was significantly lower in the SP1ras2Δmsn2Δmsn4Δ strain than in the SP1ras2Δ strain. Disruption of either MSN2 or MSN4 alone in the ras2Δ background did not dramatically affect the expression of the stress-responsive genes (Fig. (Fig.5A).5A). We further tested the expression of the STRE-lacZ reporter gene in these strains and found that it paralleled the mRNA levels of stress-responsive genes (Fig. (Fig.5B).5B). Disruption of either MSN2 or MSN4 significantly lowered STRE-lacZ activity, but disruption of both was required to reduce this activity to a wild-type level. Disruption of yAP-1 in the ras2Δ background had a similar effect. Expression of all genes tested was lower in ras2Δyap1Δ than in ras2Δ but was not as low as in ras2Δmsn2Δmsn4Δ (Fig. (Fig.5A).5A). Disruption of yAP-1 also caused a significant decrease in STRE-lacZ activity (Fig. (Fig.5B),5B), in accordance with the observation of Gounalaki and Thireos (22). These results show that the effect of Ras2 on many stress-responsive genes is mediated through Msn2/4 and yAP-1 transcription factors.
Having verified that spontaneous expression of stress-responsive genes is abolished, we tested the invasive-growth capabilities of the SP1ras2Δmsn2Δmsn4Δ and SP1ras2Δyap1Δ strains. Figure Figure5C5C shows that these strains display an invasive growth capability similar to (or even better than) that of the parental wild-type strain, SP1. Thus, in the complete absence of RAS2, invasive growth is restored by eliminating MSN2/4 or yAP-1, suggesting that Ras2 controls invasive growth via these factors. Since all known target genes of Msn2/4 and yAP-1 are stress-related genes, it seems that Ras2 induces invasive growth merely by suppression of the stress response. The fact that in the RAS2Val19 and Σ1278b strains (which are highly invasive [Fig. 1A]) the stress response is suppressed (16, 49, 52) (Fig. (Fig.2A)2A) further supports this notion. However, it could be that general suppression of all STRE dependent genes is not required but, rather, suppression of a few or even one particular stress-responsive gene is sufficient to induce invasive growth. In any case, in the SP1 genetic background, these gene(s) are targets of yAP-1 and Msn2/4.
A reporter gene, containing the filamentous response elements (FRE::lacZ [29, 35, 41]), has been used as a marker for invasive growth. We have tested this reporter in our strains (Fig. (Fig.6).6). FRE::lacZ activity measured in strains of the Σ1278b background was correlated with their invasive-growth capability. However, reporter gene activity was not correlated with the invasive-growth ability of mutants of the SP1 background (Fig. (Fig.6).6). This result is in accord with a recent report that FRE::lacZ activity does not always correlate with pseudohyphal growth (32). Also, in some strains of the SP1 background, expression of the FRE::lacZ reporter gene was dramatically affected by the growth conditions. In a given mutant (e.g., SP1RAS2Val19), it was high when cells were grown in liquid culture and very low when they were grown on agar plates (data not shown). The activity in the SP1 wild-type strains was less strongly affected. In contrast to the situation for FRE::lacZ, expression of STRE::lacZ in a given strain did not vary significantly.
We have also constructed a Σras2Δyap1Δ strain and measured its invasive-growth capability and its ability to express stress-related genes. Unlike the case for the SP1ras2Δ strain (Fig. (Fig.5),5), disruption of the yAP-1 gene in Σras2Δ had no effect on expression of stress genes (Fig. (Fig.7A).7A). The mRNA levels of all genes tested were high and similar in Σras2Δ and Σras2Δyap1Δ. Also, expression of the STRE-lacZ reporter was high in both strains (Fig. (Fig.7B).7B). These results suggest that in the Σ1278b background, Ras2 controls stress genes via other transcription factors. The Σras2Δyap1Δ strain, in which the stress response is spontaneously active but yAP-1 is absent, allowed us to test whether invasive growth is dependent on the suppression of stress genes or simply on yAP-1 activity (which may control other groups of genes). We examined the invasive-growth activity of Σras2Δyap1Δ (Fig. (Fig.7C)7C) and found it to be indistinguishable from that of Σras2Δ. Both strains were very inefficient in invading the agar (Fig. (Fig.7C).7C). These results suggest that in Σ1278b strains too, the expression level of stress genes determines the potential of invasive growth but that in this genetic background it is independent of yAP-1 activity.
This study showed that an active Ras2p is essential for invasive-growth activity in yeast. It further showed that a constitutively active Ras2/cAMP pathway is required to induce this phenotype, since invasiveness is a consequence of suppression of stress-responsive transcription factors, rendering them unable to induce transcription. Thus, invasive growth could be considered a pathogenic phenotype, manifested only by strains that harbor an unregulated Ras/cAMP cascade. These include strains with mutations along the pathway (e.g., RAS2Val19, bcy1Δ, and ira1Δ), or laboratory strains such as Σ1278b, in which the reason for the hyperactivation of the pathway is unknown. There seem to be some differences, however, between Σ1278b strains and the SP1RAS2Val19 mutant. An important difference is the role of the mating pathway. In Σ1278b, the mating pathway is essential for invasive growth (35, 50), whereas in SP1RAS2Val19 it is dispensable (Fig. (Fig.3).3). However, in Σ1278b strains, an activating mutation in components of the Ras/cAMP cascade or the GPA2 cascade (RAS2Val19, ira1, and GPA2Val132) makes the mating pathway dispensable for invasiveness (33, 50). These results, together with the high cAMP level measured in ΣRAS2Val19, show that the Ras/cAMP cascade, although overactive in Σ1278b strains (Fig. (Fig.2),2), could be further enhanced by activating mutations and that this enhancement makes the mating pathway dispensable for invasive growth. Another difference between Σ1278b and RAS2Val19 mutants of other genetic backgrounds is the fact that Σ1278b cells are not sensitive to stresses (data not shown). Thus, although their general stress response is defective, they can efficiently cope with heat shock, osmotic shock, and starvation, conditions under which RAS2Val19 cells die. Therefore, although some important similarities between SP1RAS2Val19 and Σ1278b (wild-type) strains have been measured at the level of gene expression (Fig. (Fig.2),2), the differences between the strains suggest that the Σ1278b genetic background should be regarded as strains in which the stress response is suppressed as a result of an overactive but not fully active Ras/cAMP pathway. What could be the physiological role of the mating pathway, which makes it essential for invasive growth in the Σ1278b genetic background (35, 46)? Obviously, some genes essential for invasiveness are induced by both the mating pathway and the Ras/cAMP pathway (35, 50). However, since suppression of the stress response is sufficient to induce invasiveness, it could be that on rich media the mating pathway also assists in suppressing the expression of stress genes. Thus, the Ras/cAMP pathway and the mating pathway, which cooperate in activation of invasiveness related genes (50), may also cooperate to achieve full suppression of gene expression. Another possibility is that an activated Ras/cAMP pathway solely suppresses the expression of stress genes and induces invasiveness without the assistance of the mating pathway. In such a model, the mating MAP kinase pathway could be a regulator of the Ras/cAMP pathway (for example, through inhibition of phosphodiesterases [PDEs]) and not a direct effector of Ras. Interestingly, in mammalian cells, the ERK2 MAP kinase is an inhibitor of PDE4B (27). It should be noted that biochemical experiments had clearly demonstrated a physical interaction between yeast Ras and adenylyl cyclase and RasGTP-dependent activation of the cyclase (18, 58). On the other hand, evidence for an interaction between Ras2p and the Cdc42/STE20/STE7 pathway is mostly genetic and could be explained by several models including the inhibition of PDEs.
The biological function of invasive growth of haploid cells is not known. Given that it is associated with an active Ras/cAMP cascade, it could be a mere consequence of the inability of cells to cease growth when present at high density or when in contact with a hard surface. Under such conditions, cells of wild-type strains would activate their stress response and arrest in the G1 phase of the cell cycle. However, if the stress response is inhibited and proliferation continues, cells near the surface would be forced into the agar by the nondividing dense layer of cells above them. Similar physical force is probably responsible for formation of stalk-like colonies in S. cerevisiae (15). In this respect, invasive growth in yeast is reminiscent of the property of oncogenically transformed mammalian cells, which grow in soft agar (3, 11). In this case, too, contact inhibition of growth is lost and unregulated proliferation continues even in dense colonies (forming foci), and even when in contact with hard material (agar). Interestingly, feral strains, isolated from the wild, are capable of forming filaments (31). It would be interesting to test the status of the Ras/cAMP pathway and the stress response in these strains.
The molecular mechanism that renders Σ1278b strains nonresponsive to stress is not known. It seems, however, that the defect resides along the Ras/STRE cascade, since other stress responsive cascades are intact (Fig. (Fig.2).2). The observation that disruption of RAS2 results in activation of STRE in the Σ1278b background (Fig. (Fig.4)4) suggests that the stress response in this genetic background is defective upstream of Ras2, maybe in sensing the stress. A recent report which showed that disruption (in the Σ1278b background) of MEP2, encoding an ammonium permease which functions upstream of Ras2 and Gpa2, eliminates pseudohyphal growth (32) may also suggest that the putative suppressor of the STRE cascade in Σ1278b strains functions upstream of Ras2 and Gpa2.
Another puzzle regarding Σ1278b strains is the effect of yAP-1 on stress genes. This transcription factor is involved in transcriptional activation from both AP-1 elements (36, 49) and STREs (22) (Fig. (Fig.4B).4B). However, in the Σras2Δ strain, yAP-1 is not involved in STRE activation (Fig. (Fig.7B).7B). Thus, in the Σ1278b genetic background, the Ras pathway controls the expression of stress-related genes via a mechanism not involving yAP-1 (Fig. (Fig.7).7). By contrast, the situation in SP1, in which Msn2/4 and yAP-1 are major downstream targets of Ras2, is in agreement with previously published data (22). It seems that the machinery responsible for activation of stress-related genes in the Σ1278b genetic background is different in some aspects from that in other laboratory strains.
We thank M. Wigler, G. R. Fink, and P. Estruch for yeast strains and plasmids.
This study was supported by the Israel Cancer Association and the U.S.-Israel Binational Science Foundation (grant 96-386).