Yeast homolog of mammalian mitogen-activated protein kinase, FUS3/DAC2 kinase, is required both for cell fusion and for G₁ arrest of the cell cycle and morphological changes by the CDC37 mutation

The response to extracellular signals is an essential step for the control of growth and differentiation of eukaryotic cells. In yeast, the early steps of conjugation between haploid cells, which resemble the differentiation process in higher eukaryotic systems, involve the mutual exchange of cell type-specific peptide pheromones (reviewed by Cross et al., 1988; and Herskowitz, 1989). Pheromones, termed a- and α-factors, bind to receptors on the target cells of opposite mating type to induce transcriptional activation of a set of genes required for sexual conjugation. The FUS1 and FUS2 genes are required for cell fusion (McCaffrey et al., 1987; Trueheart et al., 1987). The FUS3/DAC2 gene is required both for cell division arrest induced by mating pheromones and for cell fusion (Elion et al., 1990; Fujimura, 1990a).

The FUS3/DAC2 protein kinase belongs to a highly conserved family of kinases called mitogen-activated protein (MAP) kinases or extracellular signal-regulated proteins kinases (ERKs; Boulton et al., 1990, 1991; Pelech and Sanghera, 1992). FUS3/DAC2 inactivates a G₁ cyclin encoded by the CLN3 gene to arrest cell division in the G₁ phase and activates a transcriptional factor STE12. The FAR1 gene is specifically required for pheromone-induced G₁ arrest (Chang and Herskowitz, 1990). Genetic evidence indicates that FAR1 functions mainly by antagonizing CLN2 (G₁ cyclin) activity. Recently FUS3/DAC2 was shown to act on FAR1 protein to mediate the final steps in the mating pheromone signal transduction pathway for G₁ arrest (Peter et al., 1993; Tyers and Futcher, 1993).

Biochemical studies indicate that three protein kinases, encoded by the FUS3/DAC2, STE7 and STE11 genes, are struc-turally related to mammalian protein kinases: FUS3/DAC2 is a MAP kinase; STE7 is a MEK homolog; STE11 is a MEK kinase homolog. These kinases act in the order STE11 (MEK kinase) → STE7 (MEK) → FUS3/DAC2 (MAP kinase; Crews et al., 1992; Errede et al., 1993; Gartner et al., 1992; Lange-Carter et al., 1993; Neiman and Herskowitz, 1994).

In the present study, I constructed dac2 cln3 double mutants and investigated their pheromone responsiveness and mating behavior to clarify the role of FUS3/DAC2 in the mating process. I also constructed cdc dac2 double mutants and inves-tigated their cellular morphology to clarifiy the possible interaction between FUS3/DAC2 and CDC gene products during G₁ progression. Experimental data show that FUS3/DAC2 is required not only for completion of cell fusion in the dac2 cln3 double mutants in which pheromone responsivenes is restored, but also for G₁ arrest conferred by the cdc36, 37 and 39 mutations. The data also provide evidence that STE7 and STE11 are necessary for the G₁ block of cdc37 mutants.

The Saccharomyces cerevisiae strains used in this study are listed in Table 1. The media used were YEPD (rich medium) and SD (minimal medium; Sherman et al., 1983). SD was also supplemented with appropriate bases and/or amino acids (Sherman et al., 1983). Yeast transformation was carried out according to the lithium acetate method (Ito et al., 1983).

Isolation and restriction enzyme digestion of plasmid DNAs, ligation to form new recombinants and transformation of Escherichia coli were carried out by standard procedures (Maniatis et al., 1982).

The dac2::LEU2 mutation was introduced by the one-step gene disruption method (Rothstein, 1983) as follows. Haploid strains were transformed with the NcoI- and NdeI-digested pDAC2-LEU2(X2) plasmid (Fujimura, 1990b) or BamHI- and NdeI-digested pDAC2-Δ5LEU2 (Fujimura, 1992), and Leu⁺ transformants were selected by leucine prototrophy. The cln3::URA3 mutation was introduced as described previously (Cross, 1988). Ura⁺ transformants were selected by uracil prototrophy.

The plate halo assay was carried out as described previously (Fujimura, 1990b) as follows. MATa strains to be tested were grown at 30°C for 3 days in YEPD medium, and about 10⁵ cells were spread onto a YEPD plate containing 20 mM citrate buffer (pH 4.5) and 1 mM Chloroquine (Sigma), which enhanced the sensitivity of MATa cells to α-factor. Cells of the MATα strain XF64-54D, which produces α-factor, were spotted on the plate. Halos were zones of growth inhi-bition that were clearly visible after 2 days of incubation at 30°C.

The mating behavior between cells of opposite mating types in culture was examined under the light microscope. Haploid cells to be tested were inoculated into 400 μl of YEPD medium at an absorbance of about 1 A₆₁₀ unit (4×10⁷ cells/ml). Overnight cultures of the tester strains, XF64-52A (MATa) or XF64-54D (MATα), were then added to the 400 μl of YEPD medium containing cells to be tested. Zygote formation was observed after 7 hours of incubation at 30°C. Quanti-tative mating tests were performed as described previously (Fujimura, 1990a).

Cells harboring a 2 μm-based plasmid that carries a FUS1-lacZ fusion gene (pSB234; Trueheart et al., 1987) were grown in mid-logarithmic phase in SD medium without uracil at 24°C and then incubated for 3 hours at 24°C or 37°C. FUS1-lacZ produced in the cells was assayed for β-galactosidase activity as described previously (Fujimura, 1992).

Since the dac2 mutation reduces the mating ability of wild-type strains, the dac2 cln3 double mutant, which restored pheromone responsiveness, was investigated for mating ability. As shown in Table 2, dac2 disruption mutants showed decreased mating competence, to 1% of that of isogenic wild-type strains. The cln3 mutation itself did not affect the mating competence, implying that full sets of G₁ cyclins (encoded by the CLN1, -2 and -3 genes) are not always required for the mating process as well as for cell growth (Richardson et al., 1989). In contrast to the case of pheromone responsiveness, the mating competence of dac2 cln3 mutants was not elevated (Table 2).

Mating behavior of the mutants was investigated micro-scopically. Yeast strains were mixed with a tester strain of opposite mating type to allow the mating reaction to occur, and observed microscopically after 7 hours of incubation. As shown in Fig. 1, wild-type and cln3 mutant strains formed typical zygotes with the tester strain in which nuclear fusion was complete. On the other hand, dac2 cln3 mutants formed mating pairs, but more than 99% of the mating pairs failed to form zygotes. This result suggests that FUS3/DAC2 is necessary for cell fusion during conjugation.

Since mating pheromones arrest cell division in the G₁ phase through inducing FUS3/DAC2, possible targets of FUS3/DAC2 may be CDC gene products as well as G₁ cyclins. To examine this possibility, cdc dac2 double mutants were con-structed and their mating ability was quantitatively analyzed. As shown in Table 3, the mating defect of dac2 mutants was not suppressed by any cdc mutations tested (cdc28, -36, -37 and -39). This result is in contrast to the suppression of the sterility of the ste2 mutant by the cdc36 and -39 mutations (Lopes et al., 1990; Neiman et al., 1990; Shuster, 1982).

START mutants (cdc28, -36, -37 and -39) arrest at the G₁ phase of the cell cycle and show morphological changes at the restrictive temperature. Since the FUS3/DAC2 kinase controls the activity of G₁ cyclins (Peter et al., 1993; Tyers and Futcher, 1993), there is some possibility that FUS3/DAC2 may interact with the CDC28, -36, -37 or -39 gene products in G₁ progression or in the pheromone response pathway. To clarify this possibility, cdc dac2 double mutants were constructed and their morphologies were investigated at the restrictive temperature.

At the restrictive temperature the cdc28 dac2 mutant formed typical ‘shmoo’-like structures just as the cdc28 mutant does (Fig. 2C and D), indicating that FUS3/DAC2 acts upstream of CDC28. On the other hand, cdc36 dac2 and cdc39 dac2 double mutants formed irregular morphological structures in which cells underwent morphological changes but formed new buds, just like α-factor-treated dac2 mutants (Fig. 2F and J; Elion et al., 1990; Fujimura, 1990a); the dac2 mutation suppressed the G₁ arrest elicited by the cdc36 and cdc39 mutations, but did not suppress the morphological changes induced by the cdc36 and cdc39 mutations. In addition, the double mutants showed a temperature-sensitive growth phenotype.

CDC37 is believed to control the synthesis or activity of CDC28, and the cdc37 mutant arrests in G₁ like the cdc28 mutants at the restrictive temperature (Fig. 2G). However, cdc37 dac2 mutants did not show typical ‘shmoo’-like structures but arrested with a budded phenotype (Fig. 2H), in contrast to the case of the cdc28 dac2 mutant; the dac2 mutation suppressed the G₁ arrest caused by the cdc37 mutation. The same result was observed in a MATα cdc37 dac2 mutant (data not shown). These results indicate that FUS3/DAC2 is necessary for G₁ arrest and the morphological changes in cdc37 mutants, and also suggest the possible involvement of CDC37 in pheromone signal transduction. To clarify this possibility, I investigated the induction of the FUS1 gene in cdc37 mutants. As shown in Table 4, FUS1 was not induced in cdc37 mutants at the restrictive temperature, suggesting that CDC37 may not be involved in pheromone signal transduction. The cdc37 dac2 mutants showed a temperaturesensitive growth phenotype.

The STE7 and STE11 protein kinases are responsible for the activation of FUS3/DAC2 in the pheromone signal transduction pathway (Gartner et al., 1992; Errede et al., 1993). To investigate the possibility that G₁ arrest and morphological changes caused by the cdc37 mutation require STE7 or STE11, cdc37 ste7 and cdc37 ste11 double mutants were constructed and their morphology and G₁ arrest were studied. As shown in Fig. 3, both cdc37 ste7 and cdc37 ste11 arrested with a budded phenotype at the restrictive temperature, suggesting that STE7 and STE11 are required for G₁ arrest and morphological changes by the cdc37 mutation.

In the present study I have obtained evidence that FUS3/DAC2 has a role in cytogamy during conjugation and is not necessary for the morphological changes in cdc28, -36 and -39 mutants at the restrictive temperature.

Cross (1988) reported that the α-factor-resistant daf1/cln3 mutant shows a moderate mating competence. His results indicate that synchronization of the haploid cell cycle by αfactor is not essential for cell fusion. In the present study, I have observed that the dac2 cln3 double mutants that restored pheromone responsiveness fail to undergo cell fusion. Taking these results into consideration, I postulate a dual function for FUS3/DAC2 in mating: it being required for cell fusion as well as for pheromone-induced division arrest in the G₁ phase. FUS3/DAC2 may promote cell fusion by phosphorylating proteins that are required for cell fusion. In connection with this postulation, Rose et al. (1986) suggested that activation of cells by mating pheromones is a prerequisite for cell fusion, because the dependence of cell fusion on mating pheromone treatment cannot be replaced by synchronization in G₁ phase by the cdc28 or cdc35 mutation.

KSS1, a functionally redundant kinase closely related to FUS3/DAC2, is also involved in the pheromone signal transduction pathway (Courchesne et al., 1989; Elion et al., 1991; Gartner et al., 1992). Either FUS3/DAC2 or KSS1 is sufficient to perform a shared function required for signal transduction: activation of the transcriptional factor STE12 by phosphorylation. FUS3/DAC2 and KSS1 differ, however, in that only FUS3/DAC2 functions in G₁ arrest induced by αfactor. Since both the dac2 and dac2 cln3 mutants show a defect in cell fusion, KSS1 kinase may have no shared function with FUS3/DAC2 in cell fusion. This idea is supported by the observation of Gartner et al. (1992) that the fus3 KSS1 strain shows decreased mating ability compared with the FUS3 kss1 strain.

Since the dac2 mutants can express a set of specific genes required for sexual conjugation in response to mating pheromones, as previously described by Elion et al. (1991) and Fujimura (1992), it is unlikely that the defect in dac2 mutants in cell fusion may be due to a loss of pheromone inducibility.

I have also shown that FUS3/DAC2 may interact with CDC gene products. Epistatic analysis between the cdc28 and dac2 mutations indicates that FUS3/DAC2 may act upstream of CDC28 in the pheromone signal transduction pathway. This interpretation is consistent with the models proposed by Peter et al. (1993) and by Tyers and Futcher (1993) in which FUS3/DAC2 regulates CDC28-G₁ cyclin complexes through phosphorylating FAR1 protein. Since the cdc36 and cdc39 mutations constitutively activate the pheromone response pathway by acting on G-protein (Lopes et al., 1990; Neiman et al., 1990) and FUS3/DAC2 is not involved in α-factorinduced morphological changes (Fujimura, 1990b), it is expected that the cdc36 dac2 and cdc39 dac2 mutants undergo morphological changes at the restriction temperature (Fig. 2F and J). The bud formation observed in ‘shmoo’-like cells of cdc36 dac2 and cdc39 dac2 mutants implies that the G₁ arrest conferred by the cdc36 and cdc39 mutations may be caused by FUS3/DAC2.

Finally, I have shown that cdc37 dac2 mutants arrest with a budded phenotype, in contrast to cdc37 mutants, indicating that FUS3/DAC2 is required for G₁ arrest and morphological changes in cdc37 mutants at the restrictive temperature. This effect is not specific to FUS3/DAC2. STE7 and STE11 (kinases responsible for activation of FUS3/DAC2) are also required for the G₁ block of cdc37 mutants. KSS1 has no shared function with FUS3/DAC2 in G₁ arrest and morphological changes in cdc37 mutants. Recently, Collart and Struhl (1994) reported that CDC36 and CDC39 are negative regulators of transcription of several genes. Their hypothesis argues that the G₁ arrest of the cdc36/39 mutants is caused by derepression of pheromone-regulated gene expression (e.g. of STE4). The present study indicates that the G₁ arrest of cdc37 mutants involves a different mechanism. From these results I propose two models. One is that CDC37 and FUS3/DAC2 regulate CDC28 positively and negatively in G₁ progression, respectively (Fig. 4A). The other is that CDC37 negatively regulates FUS3/DAC2, which controls CDC28 for G₁ progression (Fig. 4B). It is clear that STE7 and STE11 are necessary for activation of FUS3/DAC2 in both models. Since the cdc37 dac2 mutants showed temperature-sensitive growth at the restrictive temperature (Fig. 2H), CDC37 may play an essential role in other parts of the cell cycle. Further studies may elucidate the biochemical function of CDC37 in G₁ progression of the cell cycle, as well as CDC28 kinase.

I thank F. Cross for providing plasmids.