Combinatorial Control Required for the Specificity of Yeast MAPK Signaling

See allHide authors and affiliations

Science  28 Feb 1997:
Vol. 275, Issue 5304, pp. 1314-1317
DOI: 10.1126/science.275.5304.1314


In yeast, an overlapping set of mitogen-activated protein kinase (MAPK) signaling components controls mating, haploid invasion, and pseudohyphal development. Paradoxically, a single downstream transcription factor, Ste12, is necessary for the execution of these distinct programs. Developmental specificity was found to require a transcription factor of the TEA/ATTS family, Tec1, which cooperates with Ste12 during filamentous and invasive growth. Purified derivatives of Ste12 and Tec1 bind cooperatively to enhancer elements called filamentation and invasion response elements (FREs), which program transcription that is specifically responsive to the MAPK signaling components required for filamentous growth. An FRE in the TEC1 promoter functions in a positive feedback loop required for pseudohyphal development.

Because common signaling components such as the MAPK cascade respond to a large number of different stimuli, it is not clear how specific signals are produced. In Saccharomyces cerevisiae, elements of the MAPK pathway required for the mating pheromone response are also required for haploid invasive growth and diploid pseudohyphal development. These shared factors include Ste20, Ste11, Ste7, and Ste12 (1, 2). The first three act in sequence and are homologs of the mammalian kinases PAK (p21-activated kinase), MEKK (MAP kinase kinase kinase), and MEK (MAP kinase kinase), respectively (3). The transcription factor Ste12 is a terminal component of these signaling cascades, functioning downstream of the kinases to drive either sexual differentiation or filamentous and invasive growth (3). In mammalian cells, myriad stimuli activate MAPK pathways, yet only a handful of target transcription factors have been identified (4). Therefore, we addressed the question of how a common target of MAPK signaling pathways, Ste12, can direct more than one distinct developmental program.

Ste12 binds cooperatively to pheromone response elements (PREs) of the consensus sequence TGAAACA (5, 6), and two or more of these elements are necessary and sufficient to program pheromone-responsive transcription (7). Because Ste12 can act alone during mating, we thought that there might exist a pathway-specific transcription factor that retargets Ste12 during filamentation and invasion through cooperative DNA binding (combinatorial control). The expression of the reporter gene FG(TyA)::lacZ depends specifically on the MAPK signaling components that promote filamentous and invasive growth (8). Transcription of FG(TyA)::lacZ is driven by a fragment of the retrotransposon Ty1, whose expression requires the TEC1 gene (9). TEC1 is also necessary for pseudohyphal and haploid invasive growth (10), which raises the possibility that Tec1 functions within the filamentous signaling pathway. Tec1 contains the conserved TEA or ATTS DNA binding domain, which is shared by several eukaryotic transcription factors, including human TEF-1 and Aspergillus abaA (11). TEF-1 binds to the sequence CATTCC, whereas abaA binds to the sequences CATTCC and CATTCT (12). We term these conserved elements TEA/ATTS consensus sequences (TCS). The Ty1 fragment in FG(TyA)::lacZ contains a binding site for Ste12 (a PRE) and an adjacent presumptive binding site for Tec1 (a TCS); the two sites are separated by 14 base pairs (bp). We call this composite DNA element (PRE plus TCS) a filamentation and invasion response element (FRE). This region of Ty1 binds to a Ste12-containing complex in crude extracts (6). Thus, Ste12 and Tec1 might cooperate to promote pathway-specific transcription of FG(TyA)::lacZ.

To test this, we placed the 27-bp Ty1 FRE upstream of an enhancerless CYC1::lacZ reporter gene to create FRE(Ty1):: lacZ (13). This construct was expressed constitutively in haploid cells grown in rich medium, which is permissive for invasive growth (Fig. 1A). Mutation of either the PRE (TGAAACG to ACTTACG) or the TCS (CATTCT to CAAACT) reduced expression of the reporter (Fig. 1A). Mutations in elements of the pheromone response pathway that are required for haploid invasion and diploid pseudohyphal development also reduced expression; the ste20 mutant showed a moderate reduction and ste11, ste7, and ste12 mutants exhibited stronger defects (Fig. 1A). Disruption of the TEC1 gene also reduced FRE activity (Fig. 1A). In contrast, mutations in components of the pheromone response pathway that are not required for filamentation and invasion, including Ste2 (the α-pheromone receptor), Ste4 (the β subunit of the receptor-coupled guanine nucleotide binding protein), Ste5 (a protein that tethers components of the MAPK cascade), and Kss1 or Fus3 (the redundant MAPKs required for mating), did not reduce FRE activity.

Fig. 1.

The Ty1 FRE confers filamentous pathway-specific gene expression. (A) Expression of FRE(Ty1)::lacZ in haploid a cells. β-Gal activities (in nanomoles per minute per milligram of protein) were measured on three transformants (13). The relevant genotypes are indicated at the bottom of the graph. Strains used in this study were of the Σ1278b background (23). PRE mutant and TCS mutant indicate reporter mutants assayed in the wild-type (WT) haploid strain. Error bars indicate standard deviations. (B) Expression of FRE(Ty1)::lacZ in a/α diploid cells. Expression was measured in homozygous diploids of the indicated genotypes that carried either a vector plasmid or a plasmid encoding the dominant activated allele STE11-4 (13).

To determine whether the FRE is responsive to signals for pseudohyphal development in diploid cells, we introduced FRE(Ty1)::lacZ into this cell type. Activity of the FRE was one-thirteenth of that in haploid cells (Fig. 1B). Much of this reduction was overcome by activation of the pathway with the hypermorphic allele STE11-4. The activation was abrogated by mutation of the PRE or TCS or by mutation of the downstream signaling components STE7 and STE12. Mutation of TEC1 also blocked the activation of the reporter, which supports a downstream role (relative to Ste11) for Tec1.

We identified elements similar to the Ty1 FRE in several genes required for filamentation and invasion (14). One exists in the TEC1 promoter itself, raising the possibility that TEC1 participates in an autoregulatory loop. In the TEC1 FRE, the orientation of the PRE and TCS is reversed compared to that of the Ty1 FRE, and the spacing between the elements is reduced from 14 to 4 bp.

We placed the 17-bp TEC1 FRE upstream of CYC1::lacZ to yield FRE(TEC1)::lacZ (15). Like FRE(Ty1)::lacZ, this construct was expressed in haploid cells in a PRE- and TCS-dependent manner (Fig. 2A; the PRE mutant is TGAAACA to ACTTACA and the TCS mutant is CATTCC to CAAACC). Mutation of STE20, STE11, STE7, STE12, and TEC1 reduced expression of FRE(TEC1)::lacZ, whereas mutation of components specific to the pheromone-responsive MAPK pathway did not. In the fus3 mutant, expression was increased fourfold, whereas the kss1 mutant resulted in reduced expression (71% of the wild type; Fig. 2A). The kss1 fus3 double mutant exhibited expression that was slightly greater than that of wild-type cells. This pattern mimics the effects of these mutations on haploid invasive growth: fus3 mutants are hyperinvasive, kss1 strains exhibit a weak invasion defect, and the double mutants invade approximately as well as does the wild type (2). Expression of FRE(Ty1)::lacZ was not modulated in this way, so the specific arrangement of the PRE and TCS in the FRE might be important for the appropriate response to Fus3 and Kss1.

Fig. 2.

The TEC1 FRE confers filamentous pathway-specific gene expression. (A) Expression of FRE(TEC1)::lacZ in haploid a cells. (B) Expression of FRE(TEC1)::lacZ in a/α diploid cells.

As with FRE(Ty1)::lacZ, expression of FRE(TEC1)::lacZ in diploid cells was lower than that in haploid cells (Fig. 2A). Again, expression was increased by STE11-4 (eightfold), and this induction was blocked by mutation of the PRE and TCS or by mutation of the downstream signaling components STE7, STE12, or TEC1.

To establish the relevance of the TEC1 FRE to the transcriptional and biological activities of the full-length TEC1 promoter, we mutated the PRE and TCS, either together or individually, in the native promoter (16) and examined the effects of these mutations on the activation of a TEC1::lacZ fusion gene in diploid cells (17). Activation of the MAPK pathway with STE11-4 produced a fourfold increase in expression of TEC1::lacZ (Fig. 3A). This increase was largely blocked by the PRE-TCS double mutant as well as by the PRE and TCS single mutants. We also introduced a high-copy plasmid containing the TEC1 gene into cells harboring the reporter. Overexpression of TEC1 increased expression of TEC1::lacZ approximately twofold. This effect was blocked by mutations in the FRE, which supports the positive feedback model.

Fig. 3.

Requirement of the TEC1 FRE for signal-responsive transcription and pseudohyphal development. (A) Expression of TEC1 promoter mutants in wild-type diploid cells. Expression of TEC1::lacZ promoter fusions carrying point mutations in the FRE was measured on three transformants (17). Solid bars show the average activity in nanomoles per minute per milligram of protein, and error bars show the standard deviation. +STE11-4 indicates that these strains contain a plasmid encoding STE11-4; +2μ TEC1 indicates that these strains carry a high-copy plasmid encoding TEC1. Mutations in the PRE and TCS are indicated by the lowercase designations. (B) Pseudohyphal growth of TEC1 promoter mutants. Growth on SLAD media of strain YM108 (atec1::HIS3/tec1::HIS3 ura3-52/ura3-52 his3::hisG/his3::hisG leu2::hisG/leu2::hisG) containing a wild-type TEC1 allele integrated at URA3 (wild type) was compared to that of cells transformed with the empty vector (null) or tec1 alleles containing mutations in the FRE. The strains contained a plasmid encoding STE11-4 on a LEU2, CEN vector (right column) or a LEU2, CEN control vector that lacked an insert (left column). Plates were incubated for 5 days at 30°C.

To ascertain the importance of the TEC1 FRE to pseudohyphal development, a single copy of the native TEC1 gene with wild-type or mutant promoter sequences was introduced into a homozygous tec1/tec1 mutant diploid by integrative transformation (18). The wild-type promoter allele complemented the pseudohyphal growth defect of the null allele (Fig. 3B). The PRE-TCS double mutant blocked filament production, as did the PRE and TCS single mutants. In the presence of the wild-type allele, STE11-4 enhanced filamentation, whereas the activation of filamentation was blocked by the tec1 null allele. The PRE and TCS mutants also blocked the induction of filamentation by STE11-4.

We examined the ability of purified derivatives of Ste12 and Tec1 to interact with FREs in vitro. Full-length Ste12 and Tec1 were expressed in Escherichia coli as protein fusions with maltose-binding protein (MBP). A FLAG peptide tag was added to the COOH-termini to aid in purification (19). Binding of these proteins to the Ty1 FRE was examined with native gel electrophoresis (Fig. 4B) (20). No protein complex was observed upon incubation of a 32P-labeled fragment containing the 27-bp Ty1 FRE with MBP, MBP-Tec1, or MBP-Ste12 alone. However, incubation of the Ty1 FRE with both MBP-Tec1 and MBP-Ste12 resulted in the appearance of a complex of lower mobility. The requirement of both proteins for complex formation implies cooperative binding. A 100-fold excess of the wild-type Ty1 FRE eliminated the formation of the complex, but the same amount of Ty1 FRE containing mutations in either the PRE or TCS did not. A 100-fold excess of the unlabeled TEC1 FRE also eliminated binding to the Ty1 FRE (Fig. 4A, lane 9), whereas TEC1 FREs containing point mutations in either the PRE or TCS did not compete.

Fig. 4.

Cooperative binding of Ste12 and Tec1 derivatives to FREs. (A) Native gel analysis of protein-FRE complexes. An autoradiogram of a nondenaturing acrylamide gel is shown. B, bound; F, free. 32P-labeled Ty1 FRE was applied to the gel after incubation with the indicated components (20). Approximate protein concentrations were as follows: MBP, 5 × 10−9 M; MBP-Tec1, 6 × 10−10 M; and MBP-Ste12, 6 × 10−10 M. 32P-labeled Ty1 FRE concentration was approximately 3 × 10−10 M. Unlabeled competitor DNAs were added as indicated and are labeled as follows: TY1, wild-type Ty1 FRE; TY1-P, PRE mutant; TY1-T, TCS mutant; TEC1, wild-type TEC1 FRE; TEC1-P, PRE mutant; and TEC1-T, TCS mutant. (B) Deoxyribonuclease I footprint analysis of protein binding to the Ty1 FRE. Shown is an autoradiogram of a denaturing polyacrylamide gel (8%). Reactions contained a 32P-end-labeled probe derived from the FRE(Ty1)::lacZ construct (21). Lanes 1 and 4 correspond to reactions lacking MBP-Tec1 and MBP-Ste12; lanes 2 and 3 correspond to reactions containing both proteins. The positions of the TCS (open box) and of the PRE (solid box) are indicated. Arrows indicate a pair of hypersensitive sites induced between the PRE and TCS. (C) Deoxyribonuclease I footprint analysis of protein binding to the TEC1 FRE. Analysis was performed as in (B), except that the probe was derived from FRE(TEC1)::lacZ.

We also performed deoxyribonuclease I (DNase I) protection experiments with end-labeled fragments from the FRE(Ty1)::lacZ and FRE(TEC1)::lacZ constructs (21). Incubation of the FRE(Ty1)::lacZ or FRE- (TEC1)::lacZ probe with MBP-Tec1 and MBP-Ste12 resulted in the protection of specific nucleotides within the PRE and TCS (Fig. 4, B and C). Binding to FRE- (Ty1)::lacZ resulted in the appearance of a pair of adjacent hypersensitive sites between the TCS and PRE that may be indicative of a distortion in the DNA (Fig. 4B).

Our data demonstrate that the appropriate transcriptional response to overlapping, upstream, MAPK signaling components in yeast requires combinatorial control. Ste12 can act as a homomultimer to promote pheromone-responsive transcription (7). During filamentation and invasion, Ste12 acts with a second transcription factor, Tec1, to drive transcription that is specifically responsive to the MAPK pathway that promotes filamentation and invasion. The TEC1 FRE is necessary for normal pseudohyphal development, which establishes at least one role for Ste12 and Tec1 in the expression of a gene involved in filamentation. Mating pheromone does not activate the MAPK pathway that leads to filamentation and invasion (8), yet it activates Ste12. Thus, the mechanism by which Ste12 is switched on by the mating MAPKs, Fus3 and Kss1, must not operate on Ste12-Tec1 complexes. The inhibitory action of Fus3 on FRE-dependent transcription could also play a role in preventing the activation of Ste12-Tec1 during mating. Because Fus3 and Kss1 together are dispensable for filamentation and invasion (although individually they can modulate haploid invasion), there likely exists a MAPK equivalent that specifically activates Ste12-Tec1 complexes.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
View Abstract

Stay Connected to Science

Navigate This Article