When Dictyostelium cells starve they arrest their growth and induce the expression of genes necessary for development. We have identified and characterized a protein kinase, YakA, that is essential for the proper regulation of both events. Amino acid sequence and functional similarities indicate that YakA is a homolog of Yak1p, a growth-regulating protein kinase in S. cerevisiae. Purified YakA expressed in E. coli is able to phosphorylate myelin basic protein. YakA-null cells are smaller and their cell cycle is accelerated relative to wild-type cells. When starved, YakA-null cells fail to decrease the expression of the growth-stage gene cprD, and do not induce the expression of genes required for the earliest stages of development. YakA mRNA levels increase during exponential growth and reach a maximum at the point of starvation, consistent with a role in mediating starvation responses. YakA mRNA also accumulates when cells are grown in medium conditioned by cells grown to high density, suggesting that yakA expression is under the control of an extracellular signal that accumulates during growth. Expression of yakA from a conditional promoter causes cell-cycle arrest in nutrient-rich medium and promotes developmental events, such as the expression of genes required for cAMP signaling. YakA appears to regulate the transition from growth to development in Dictyostelium.

The life cycle of Dictyostelium discoideum is composed of two phases. During the vegetative phase the cells grow in the soil as solitary amoebae, feeding on bacteria. Nutrient exhaustion triggers the developmental phase in which the cells aggregate, form a multicellular organism and construct a fruiting body made up of spores held on top of a cellular stalk. The result of this starvation response is that most of the cells will survive the unfavorable conditions by differentiating into spores. This survival strategy is common to many sporulating microorganisms and parallels the general pattern of tissue formation during metazoan development: growth cessation followed by terminal differentiation.

Starvation results in an extensive modification of the pattern of gene expression within Dictyostelium cells. Prestarvation Factor (PSF) is an extracellular signal used by cells to measure their food supply relative to the local cell density and is thought to mediate low-level expression of some developmental genes four cell divisions before cells experience overt starvation (Clarke and Gomer, 1995). Thus, PSF may allow cells to change their pattern of gene expression gradually in preparation for a future change in the food supply. Upon starvation, the expression of some vegetative genes is reduced, while genes essential for development become induced. Among the earliest genes to be induced after starvation are the ones required to modulate cell motility and allow aggregation of the cells (Parent and Devreotes, 1996). Aggregation is mediated by chemotaxis to cAMP so many of these genes’ products are required for producing and sensing cAMP, including an adenylyl cyclase (ACA), a cell surface cAMP receptor (cAR1) and the cAMP-dependent protein kinase (PKA). The cAMP produced by some cells in the population stimulates nearby cells to migrate in the direction of the signal until an aggregate is formed. In addition to inducing a chemotactic response, cAMP induces the production of additional cAMP, forming part of a relay circuit that is used to organize the behavior of cells over an area of about 1 cm2 (van Haastert, 1995). Extracellular cAMP signaling is also required for the expression of many of the genes required for early development (Reymond et al., 1995; Verkerke-van Wijk and Schaap, 1997).

PKA is a central regulator of Dictyostelium development that mediates many of the changes in gene expression caused by cAMP signaling. Disruption of the PKA-C gene leads to a complete lack of development (Mann and Firtel, 1991), while overexpression of the catalytic subunit of PKA, PKA-C, or disruption of its regulatory subunit, PKA-R, causes rapid development (Anjard et al., 1992; Mann et al., 1992; Simon et al., 1992). This indicates that PKA dictates the timing of the developmental program. PKA is also required later for stalk and spore cell differentiation (Harwood et al., 1992; Mann et al., 1994). During the first 4 hours of development, PKA activity increases about fivefold over levels found in growing cells, and this activity increase parallels the accumulation of PKA-C and PKA-R mRNAs (Leichtling et al., 1984; Mann and Firtel, 1991; our unpublished observations). PKA-C is essential for the expression of cAR1 and ACA 2-4 hours after the start of development (Schulkes and Schaap, 1995; Mann et al., 1997). Thus, upon starvation a plausible dependent sequence of events can be outlined in which PKA subunit expression increases, PKA-C activity increases, and the expression of genes required for early development ensues (Mann et al., 1997; Shulkes and Schaap, 1995). Although PKA activity is present in growing cells it is not required for growth since PKA-C-null mutants are viable (Mann and Firtel, 1991). Thus, the essential role of PKA is to initiate development when conditions warrant.

The influence of the cell cycle on the cell-type differentiation that occurs 8-12 hours after starvation has been well documented (McDonald and Durston, 1984; Weijer et al., 1984; Ohmori et al., 1987; Gomer and Firtel, 1987; Wood et al., 1996; Gomer and Ammann, 1996). For axenically growing cells, the cell cycle is about 9 hours long and consists of a 15-minute mitosis and a 30-minute S phase followed by a long G2 phase. There is no identifiable G1 phase. Cells that starve near the time of mitosis, S phase or early G2 are more likely to differentiate as prestalk cells, while cells that starve during mid-to late G2 are more likely to differentiate as prespore cells. Therefore, the distribution of cells within the cell cycle at the time of starvation biases their differentation into an initial population of prespore cells or prestalk cells. Recent studies have demonstrated that chromosomal DNA synthesis ceases within 6 hours after starvation (Shaulsky and Loomis, 1995). Thus, it is likely that, upon starvation, cells complete chromosomal replication and undergo development while in the G2 phase of the cell cycle. Based on cell synchronization experiments, Maeda and coworkers have proposed that cells must reach a specific point within G2 in order to begin differentiation (Maeda et al., 1989; Araki and Maeda, 1995). From this viewpoint, specific cell-cycle regulators would be expected to facilitate exit from the cell cycle and mediate the transition between growth and development.

We have characterized a new protein kinase, YakA, that regulates growth, is required for the initiation of development and, thus represents a key link in the regulatory network that governs the transition between these two stages of the Dictyostelium life cycle. YakA has a high degree of similarity to the Yak1p kinase from the yeast Saccharomyces cerevisiae. Yak1p appears to act as a growth attenuator in yeast that acts in opposition to the growth-promoting activity of PKA (Garrett et al., 1991). The PKA pathway is required for the progression of yeast cells through the G0/G1 transition of the cell cycle (reviewed by Murray and Hunt, 1993). Temperature-sensitive (ts) mutations in the genes that encode components of this pathway such as CDC25 (a guanine nucleotide exchange factor), Ras, adenylyl cyclase and the PKA catalytic subunit, are all lethal at the restrictive temperature. Null mutations in YAK1 were isolated as suppressors of the lethality of a Ras(ts) mutation (Garrett and Broach, 1989). These mutations also suppress the lethality of temperature-sensitive mutations in the catalytic subunit of PKA. Overexpression of YAK1 in Ras(ts) mutants causes growth arrest, and overexpression in wild-type cells induces thermotolerance (Garrett et al., 1991; Hartley et al., 1994). Yak1p expression is normally induced upon growth arrest in G1, consistent with a role for Yak1p in maintaining cells in a quiescent state. Thus, although Yak1p is not essential for the growth of yeast cells it appears to be involved in responding to growth arrest in G1 and mediating at least one stress response.

In Dictyostelium, YakA is essential for development and may mediate the response to starvation by inducing a growth arrest that leads to the appropriate changes in gene expression required to initiate development. YakA is necessary for the expression of PKA-C, ACA and cAR1. The yakA mRNA accumulates during growth and peaks in abundance at the time of starvation. Thus, YakA is expressed at the appropriate time to link starvation sensing to the developmental response, triggering a cascade of gene expression that culminates with setting up the cAMP chemotactic response system. The function of YakA in Dictyostelium is similar to the role of Yak1p in yeast, since YAK1 overexpression induces growth arrest of PKA-attenuated yeast strains and a change in cell physiology, namely thermotolerance. The Drosophila and human ‘minibrain’ kinases, which are closely related to YakA, also appear to be involved in cell differentiation during brain development (Tejedor et al., 1995; Smith et al., 1997). Thus, YakA, Yak1p and the minibrain kinases may represent a new family of protein kinases that regulate transitions from growth to differentiation in eukaryotic cells.

Strains, growth and development

Dictyostelium discoideum AX4 strain was grown in axenic medium HL-5 or on SM agar plates in the presence of Klebsiella aerogenes (Sussman, 1987). yakA high-copy cells were grown in the presence of G418-resistant Klebsiella aerogenes (a gift from J. Franke and R. Kessin) on SM plates containing 40 μg/ml G418 (Geneticin, GibcoBRL). For growth curves in HL-5, five clones of each strain were grown to 1×106 cells/ml in HL-5, diluted to 5×104 cells/ml and counted using a hemocytometer. For determination of yakA mutant growth, doubling times were determined in side-by side tests with wild-type sibling transformants from the same transformation that generated the YakA-null clones. For growth in the presence of bacteria, 5×10 cells were plated with 0.2 ml of an overnight culture of K. aerogenes. Cells were collected and washed free of the bacteria by differential centrifugation after 41, 44, 47 and 50 hours, at which time growth ceased and development started, with the formation of aggregates by 66 hours and culminants by 72 hours. Cells were developed on Millipore filters as described (Sussman, 1987) and photographed with an Optronics CCD camera. For cAMP pulsing experiments, cells were washed in 20 mM KPO4, pH 6.4, resuspended to 1×107 cells/ml in the same buffer and shaken at 100 rpm for 2 hours, followed by pulses of cAMP (to 100 nM) at 6-minute intervals. Conditioned medium was prepared from NC4 cells as described (Rathi et al., 1991). Briefly, NC4 cells were cultured on a bacterial suspension to a cell density of 5×106 cells/ml. The culture medium was harvested and used to prepare a fresh bacterial suspension. Cells were inoculated in the conditioned medium at 5×103 cells/ml and collected at a density of 2.5×105 cells/ml for the preparation of RNA samples.

Determination of cell size and mass

Cell size was measured by comparing the light scattering profile of wild-type cells (AX4 and HL328) with that of five independent YakA-null clones. An absorbance curve at 600 nm (A600) was obtained for the control cells at different cell densities, and mutant cell readings were compared to determine a relative cross-sectional area (Mallette, 1969). For cell mass determinations, cell pellets of 2×108 cells were lyophilized and weighed.

Transformation of Dictyostelium

Insertional mutagenesis was carried out as described (Kuspa and Loomis, 1992) using DIV2 (6.5 kb) as the mutagenic plasmid. Flanking genomic DNA was recovered from the AK235 genome by plasmid rescue using BglII to liberate a 12-kb fragment that was cloned as p235Bgl, as described (Kuspa and Loomis, 1994). Using the genomic insert as a hybridization probe on Southern blots, a 5.5-kb BglII fragment in wild-type DNA, and a 12-kb BglII fragment in AK235 DNA, were observed. Disruption of the yakA gene by homologous recombination and selection of uracil auxotrophs was performed as described (Kuspa and Loomis, 1994). p235Bgl was linearized with BglII and used to transform wild-type cells. Southern analysis confirmed that the transformants that integrated the fragment into the original insertion site (IS235) had the same developmental phenotype as AK235, whereas transformants that integrated the plasmid elsewhere were wild type. To generate yakA-800 mutants, pDSL4pyr (described below) was linearized with BamHI and used to transform cells. Strains that overexpressed the yakA cDNA were obtained by calcium phosphate-mediated transformation and selection with Geneticin (Nellen et al., 1987). yakA-overexpressing clones started to appear after 5 days of selection and then stopped growing. At this point the cells were harvested and plated for growth on bacteria, conditions in which the actin 6 promoter is much less active.

cDNA cloning and plasmid construction

Standard DNA and RNA manipulations were carried out as described (Sambrook et al., 1989). A cDNA library was made from polyadenylated RNA purified from growing cells using a Zap cDNA Synthesis kit (Stratagene) and the λACT2 vector (S. Elledge, Baylor College of Medicine). The full-length yakA cDNA was obtained by probing the library with the 500-bp EcoRV-HindIII fragment from p235Bgl, which encodes the N terminus of YakA protein. Five positive clones were obtained from 2.5×105 phage. The phage were converted into plasmids as described by Mulligan and Elledge (1994), and the inserts in these plasmids were cloned into the BamHI site of pGEM3 (Promega). The phage with the longest insert, DSL4, was thus converted into pDSL4. Sequence analysis of all of these clones indicated that each had the same extensive open reading frame that is found in the genomic clone p235Bgl, and four ended at the same stop codon. pDSL4 contained a 5.2-kb insert that was found to contain an open reading frame of 4,374 bp that precisely matched the genomic sequence, after allowing for the four introns in the genomic sequence. The cDNA also contained 436 bp upstream of the predicted start codon, which was found to be identical to the genomic sequence and contained stop codons in all three reading frames. The probable 3′ end of the coding region of the gene was identified in DSL4 by a stop codon and 398 bp of additional sequence that contained stop codons in all three reading frames. Another cDNA found to correspond to the 5′ half of DSL4 confirmed the predicted start codon. The knock-out vector used to generate IS800 insertions in yakA was constructed by inserting a BamHI fragment containing the pyr5-6 gene into a unique BclI site in pDSL4 to generate plasmid pDSL4pyr. The yakA overexpression construct was obtained by cloning the full-length cDNA into the vector pDNeo67 (da Silva and Klein, 1990). To obtain the antisense construct of yakA for use in RNase protection experiments, a 389-bp fragment (bases 232-621) of the pDSL4 cDNA was obtained by the polymerase chain reaction (PCR), using oligonucleotides with added PstI and BamHI flanking sites to the 5′ and 3′, respectively, and cloned into the pGEM3 vector. The yeast high-copy, galactose-inducible, yakA expression plasmid was obtained by cloning the yakA cDNA DSL4 into the pBAD vector (B. Desany and S. Elledge, Baylor College of Medicine).

Assay of YakA protein kinase activity

A hexahistidine-tagged segment of YakA was expressed in E. coli as follows. PCR primers were designed to amplify the first 589 codons of the yakA cDNA DSL4, which encodes the kinase core domain. This fragment was introduced into pCR-Script at the SrfI site, excised with BamHI and HindIII and cloned in-frame into pTrcHisA (Invitrogen, Inc.). The resulting plasmid, pTrcHis(YakA1-589) was introduced into BL21 E. coli cells.

E. coli cells containing pTrcHisA or pTrcHisA(YakA1-589) were grown to an A600 of 1.0, when IPTG was added to 2 mM to induce expression from the Trp/lac promoter. Cultures of BL21 containing pTrcHisA or pTrcHisA(YakA1-589) that were induced for 6 hours were lysed by sonication and loaded onto a 1 ml nickel sepharose affinity column. The column was washed extensively with 10 mM imidizole, 20 mM Tris, pH 7.4, and eluted stepwise with 40 mM imidizole and 100 mM imidazole in Tris buffer, pH 7.4. The samples were then analyzed by western blot analysis and histidine-tagged protein was visualized using nickel sepharose conjugated to alkaline phosphatase. A protein with an apparent molecular mass of 74 kDa was present in the 6-hour-induced cells containing pTrcHisA(YakA1-589) that was not present in the cells containing the empty vector. Kinase activity of the eluted fractions was determined by assays containing [γ-32P]ATP (3,000 Ci/mmol) and myelin basic protein (5 μg). Reaction mixtures were incubated at 30°C for 30 minutes and stopped by adding an equal volume of 20% trichloroacetic acid. Samples were separated on 15% polyacrylamide-SDS gels and labeled proteins were visualized by autoradiography.

DNA and RNA analyses

RNA was extracted using the Trizol reagent as described by the manufacturer (GibcoBRL), subjected to electrophoresis in 1.2% agarose/formaldehyde gels and transferred to nitrocellulose filters as described (Sambrook et al., 1989). The DNA fragments used as probes were as follows: a BclI/HincII fragment of yakA, a BamHI/XhoI fragment containing the full-length cDNA of cprD, an EcoRI fragment containing the full-length cDNA of carA, an EcoRI fragment containing the full-length acaA cDNA and a BamHI/HindIII fragment of PKA-C that excludes the non-conserved repeats at the N terminus of the protein. The antisense probe for yakA was obtained by in vitro transcription using the T7 RNA polymerase and the Riboprobe System (Promega). The RNAse protection assay was performed using the RPA II Ribonuclease Protection Assay Kit (Ambion, Inc.) and analyzed using denaturing conditions according to the manufacturer’s instruction. Control experiments confirmed that all reactions were performed with a tenfold excess of probe RNA.

Yeast strains, growth and transformation

S. cerevisiae cdc25ts (Aronheim et al., 1994) and Y300 (Allen et al., 1994) were cultured and transformed as described (Elledge et al., 1993). To analyze the growth profile, transformed clones were diluted in minimal medium containing 2% galactose and the cell density monitored by reading the absorbance at 600 nm.

cAMP, adenylyl cyclase and PKA assays

Total cAMP production of AX4 and AK800 cells was measured using an radioimmunoassay (RIA) kit (Amersham) according to the manufacturer’s procedure. Cells were starved for 4 hours in phosphate buffer, pulsed for 4 hours with 80 nM cAMP, washed and stimulated with 10 μM 2′-deoxy-cAMP. Samples of cells were collected at different times after the stimulation, lysed in perchloric acid, neutralized and assayed for cAMP. Adenylyl cyclase activity of wild-type and AK800 cells was measured as described (Wang and Kuspa,1997). PKA activity measurements were carried out using the SignaTECT PKA assay system (Promega). Samples were prepared from cells that were harvested from bacterial growth plates and incubated in HL-5 for 6 hours to allow expression of the actin 6 promoter driving yakA in those cells containing the act6::yakA construct. Cell extracts containing 10 μg of protein were prepared according to the manufacturer’s instructions and were used in reactions in the presence or absence of 10 μM of the PKA-specific inhibitor PKI, which inhibits the Dictyostelium enzyme (Mann et al., 1992). PKA activity is defined as the amount (nmol/minute/mg protein) of Kemptide substrate phosphorylated in the absence of PKI minus the amount phosphorylated in the presence of PKI.

DNA and protein sequence analyses

The yakA cDNA sequence was compared to the sequences present in the databanks using the BLAST search program from the National Center of Biotechnology Information. An alignment of the protein sequences was obtained using the multiple alignment search tools in the MacVector computer program (Oxford Molecular Group). The YakA amino acid and nucleotide sequences have been deposited in GenBank under the accession number AF045453.

Characterization of YakA, a protein kinase essential for development

To isolate genes involved in early signaling events we used restriction enzyme-mediated integration (REMI) of plasmid DNA to mutagenize Dictyostelium cells and isolated strains that were unable to initiate development (Kuspa and Loomis, 1992). One mutant, AK235, was selected for further study because it displayed no morphological development (Fig. 1). The genomic DNA flanking the insertion site in AK235 was cloned by plasmid rescue. This plasmid was used to recapitulate the original insertion by homologous recombination in a fresh host, and the original phenotype was obtained (see Materials and Methods). This confirmed that the original integration event caused the developmental deficiency in AK235.

Fig. 1.

Developmental phenotypes of YakA-null cells. (A) Wild-type cells, (B) YakA-null cells and (C) YakA-null cells carrying the yakA cDNA under the transcriptional control of the actin 6 promoter, were plated for development on filters and photographed after 36 hours. Bar, 0.5 mm.

Fig. 1.

Developmental phenotypes of YakA-null cells. (A) Wild-type cells, (B) YakA-null cells and (C) YakA-null cells carrying the yakA cDNA under the transcriptional control of the actin 6 promoter, were plated for development on filters and photographed after 36 hours. Bar, 0.5 mm.

The DNA sequence of the genomic clone revealed one long potential coding region (ORF) with four predicted introns, typical of Dictyostelium genes. This coding region was interrupted by the mutating plasmid at a unique BamHI restriction site (IS235, Fig. 2A), as expected for REMI-based plasmid insertion using BamHI. A full-length cDNA clone, pDSL4, was isolated that matched the genomic sequence after allowing for four introns. The cDNA recognized a 5-kb transcript on northern blots of growth-stage RNA samples that was absent in samples of AK235 RNA (data not shown).

Fig. 2.

Alignment of the predicted YakA and Yak1p proteins. (A) The regions of amino acid sequence similarity between YakA and yeast Yak1p are shown as filled boxes. IS800 and IS235 indicate the positions of insertion mutations. (B) An alignment of the amino acid sequences corresponding to the predicted serine/threonine protein kinase domains of YakA, Yak1p and the human MNB-1. Amino acid 1 corresponds to amino acid 137 in YakA, 302 in Yak1p and 91 in MNB-1. Regions A (residues 2-80) and B (residues 338-391) are underlined. The amino acids that are similar in all three proteins are boxed. The 15 amino acid residues found to be nearly invariant in all serine/threonine protein kinases are marked with an asterisk. The 12 serine/threonine protein kinase domains (Hanks and Quinn, 1991) are indicated by roman numerals above the sequences. The arrowhead indicates the position of IS800.

Fig. 2.

Alignment of the predicted YakA and Yak1p proteins. (A) The regions of amino acid sequence similarity between YakA and yeast Yak1p are shown as filled boxes. IS800 and IS235 indicate the positions of insertion mutations. (B) An alignment of the amino acid sequences corresponding to the predicted serine/threonine protein kinase domains of YakA, Yak1p and the human MNB-1. Amino acid 1 corresponds to amino acid 137 in YakA, 302 in Yak1p and 91 in MNB-1. Regions A (residues 2-80) and B (residues 338-391) are underlined. The amino acids that are similar in all three proteins are boxed. The 15 amino acid residues found to be nearly invariant in all serine/threonine protein kinases are marked with an asterisk. The 12 serine/threonine protein kinase domains (Hanks and Quinn, 1991) are indicated by roman numerals above the sequences. The arrowhead indicates the position of IS800.

The ORF within pDSL4 is predicted to produce a 167 kDa protein of 1,458 amino acids. The predicted protein has a high degree of similarity to the Yak1p protein kinase from Saccharomyces cerevisiae (Garrett and Broach, 1989). Because of this sequence similarity, and the functional similarities with Yak1p described below, we named the Dictyostelium protein YakA and the gene yakA. The two proteins also share stretches of poly-glutamine and poly-asparagine located at the N terminus in Yak1p and in an extended C terminus in YakA (Fig. 2A). An alignment between YakA and Yak1p revealed two additional regions of similarity (labeled A and B in Fig. 2A), one upstream of the kinase core and the other between domains X and XI as defined for serine/threonine protein kinases by Hanks and Quinn (1991). Amino acid sequence comparisons were made between YakA, Yak1p and 10 other protein kinases that were found, through standard database searches, to be most similar to the Yak proteins, such as the minibrain kinases. In every case, the A and B regions were most similar between YakA and Yak1p (data not shown). For example, 47% of the residues are identical in region A between YakA and Yak1p, but there is only 24% and 22% identity when Yak1p or YakA are compared to the human MNB1 protein, respectively (Fig. 2B). These observations suggest that YakA and Yak1p represent a distinct subclass of protein kinases.

We tested the protein kinase activity of YakA by expressing, in E. coli, a hexahistidine-tagged amino-terminal fragment (amino acids 1-589) that contained the predicted kinase core domain. This protein was purified by nickel affinity chromatography and was found to phosphorylate myelin basic protein (MBP; Fig. 3). A mock fraction purified from an E. coli strain that contained an empty expression vector was unable to phosphorylate MBP (data not shown). There was also substantial incorporation of radioactive phosphate into a band on the autoradiogram that corresponded to the position of YakA, suggesting that YakA phosphorylates itself under these conditions (Fig. 3).

Fig. 3.

Protein kinase activity of YakA. Hexahistidine-tagged YakA was purified by nickel affinity chromatography and used for protein phosphorylation assays. YakA was incubated with [γ-32P]ATP and substrate, incubated at 30°C for 30 minutes, and analyzed by SDS-PAGE and autoradiography. The positions of YakA and MBP in the original protein gel are indicated.

Fig. 3.

Protein kinase activity of YakA. Hexahistidine-tagged YakA was purified by nickel affinity chromatography and used for protein phosphorylation assays. YakA was incubated with [γ-32P]ATP and substrate, incubated at 30°C for 30 minutes, and analyzed by SDS-PAGE and autoradiography. The positions of YakA and MBP in the original protein gel are indicated.

The plasmid insertion in the yakA mutant AK235 (IS235) was mapped to codon 873, downstream of the kinase core coding region (Fig. 2A). We constructed a new insertion mutation, IS800, by inserting the selectable pyr5-6 gene into the middle of the kinase coding region at codon 257. When this insertion was introduced into the yakA locus in pyr5-6-null cells, the resulting strain (AK800) was completely deficient in aggregation (data not shown). This confirmed that disruption of the yakA gene is responsible for the developmental defect originally observed in AK235, and suggests that yakA is not essential for growth.

The coding region of the pDSL4 cDNA was cloned into a Dictyostelium expression vector under the control of the actin 6 promoter. The actin 6 promoter is active during the first 12 hours of development and when cells are grown in a defined liquid medium (HL-5), but is much less active in cells that are feeding on bacteria (Knecht and Loomis, 1987; M. Clarke, personal communication). An act6::yakA expression construct was introduced into YakA-null cells and the resulting transformants were found to develop normally when allowed to starve within the colonies formed on bacteria (Fig. 1C). This demonstrates that the ORF defined by the pDSL4 cDNA encodes functional YakA.

YakA expression induces growth arrest in yeast and Dictyostelium

The yeast YAK1 gene was isolated by its ability to suppress the lethality of strains that are impaired in the PKA pathway (Garrett et al., 1989). Thus, temperature-sensitive mutations in CDC25, Ras, adenylyl cyclase and PKA are all lethal at the restrictive temperature in yeast but they can be rescued by mutations in YAK1. Conversely, overexpression of YAK1 in these same mutant backgrounds leads to growth arrest, even at the permissive temperature (Garrett et al., 1991). To directly test the conservation of function between Yak1p and YakA we expressed yakA in wild-type yeast cells and in cdc25 mutant cells. The yakA cDNA was introduced into a cdc25ts strain on a high-copy plasmid under the transcriptional control of a galactose-inducible promoter. When these cells were diluted into minimal medium containing galactose, at the permissive temperature, an arrest of cell growth was observed (Fig. 4A). Microscopic examination of the arrested cells showed that 90% were small and unbudded, indicating that they arrested growth in the G1 stage of the cell cycle. However, this arrest was not induced by yakA expression in the parental yeast strain used to construct the cdc25 mutant, nor when a low-copy plasmid was used (data not shown). These results are analogous to those obtained for YAK1 in other PKA-attenuated strains (Garrett et al., 1991), suggesting that Dictyostelium YakA is functionally homologous to the yeast Yak1p protein.

Fig. 4.

YakA expression induces growth arrest in yeast and in Dictyostelium. (A) Growth profile of S. cerevisiae cdc25ts strains transformed with a high-copy plasmid containing the yakA cDNA under the transcriptional control of the GAL1-10 promoter (open squares) or containing no insert (open diamonds). Yeast clones were transferred into medium containing galactose at time 0, incubated at 22°C, and the A600 was measured at the indicated times. The curves are the averages for five independent clones. Standard error bars were smaller than the symbols. (B) AX4 wild-type cells (open squares) and AX4(act6::yakA) cells (open circles) were scraped from bacterial growth plates during exponential growth, washed free of bacteria and suspended at 106 cells/ml in HL-5 liquid medium. Wild-type cell numbers increased by the first time point (10 hours), while yakA-overexpressing cell numbers did not increase at any time. Five independent AX4(act6::yakA) transformants were tested; similar results were obtained for each, and a representative experiment is shown.

Fig. 4.

YakA expression induces growth arrest in yeast and in Dictyostelium. (A) Growth profile of S. cerevisiae cdc25ts strains transformed with a high-copy plasmid containing the yakA cDNA under the transcriptional control of the GAL1-10 promoter (open squares) or containing no insert (open diamonds). Yeast clones were transferred into medium containing galactose at time 0, incubated at 22°C, and the A600 was measured at the indicated times. The curves are the averages for five independent clones. Standard error bars were smaller than the symbols. (B) AX4 wild-type cells (open squares) and AX4(act6::yakA) cells (open circles) were scraped from bacterial growth plates during exponential growth, washed free of bacteria and suspended at 106 cells/ml in HL-5 liquid medium. Wild-type cell numbers increased by the first time point (10 hours), while yakA-overexpressing cell numbers did not increase at any time. Five independent AX4(act6::yakA) transformants were tested; similar results were obtained for each, and a representative experiment is shown.

Expression of the yakA cDNA under the control of the actin 6 promoter was also found to arrest the growth of wild-type Dictyostelium cells. When AX4(act6::yakA) cells were grown on a bacterial lawn and then suspended in liquid medium to activate the actin 6 promoter, the cells never increased in number, indicating that they arrested growth immediately (Fig. 4B). The majority of the growth-arrested cells retained viability for at least 12 hours under these conditions as judged by their ability to undergo development on filters (see below) and to form colonies on bacteria (data not shown). The DNA content of these cells was examined by flow cytometry before and after arrest. Since the growth-arrested yakA overexpressing cells displayed DNA content profiles that were indistinguishable from growing wild-type cells (data not shown), yakA did not appear to induce a G1 arrest. Rather, these data suggest that overexpression of yakA arrests the growth of Dictyostelium cells in the G2 phase of the cell cycle.

yakA is necessary for proper cell cycle regulation

Microscopic examination of YakA-null cells revealed that they were smaller than wild-type Dictyostelium cells. Light-scattering measurements indicated that YakA-null cells had 78±6% (mean ± s.e.m.) of the cross-sectional area of wild-type cells when cells were grown in liquid medium. Also, YakA-null cells had 57±8% the dry-weight of wild-type cells (mean ± s.e.m.). The mutant cells also doubled faster, and grew to a higher cell density, when compared to parental strains. Wild-type cells doubled every 9-10 hours during early exponential growth in liquid medium, while YakA-null cells doubled every 6-7 hours (Fig. 5A). YakA-null cells also consistently grew to a cell density that was about 50% higher than wild-type cells (Fig. 5B). These growth phenotypes were found in strains carrying either the yakA-235 or the yakA-800 allele. Thus, YakA-null cells have a faster cell cycle and are smaller than wild-type cells, suggesting that the coordination between cell growth and cell-cycle progression is altered in these cells.

Fig. 5.

YakA-null cells divide faster than wild-type cells. Growth curves of YakA-null cells (diamonds) and sibling wild-type control cells (squares). Cells were grown in HL-5 liquid medium and cell densities were determined by direct counting. The period of exponential growth (A) and a complete growth curve (B) are shown. Representative curves from more than five comparisons of different isolates of each strain, carried out in parallel, are shown.

Fig. 5.

YakA-null cells divide faster than wild-type cells. Growth curves of YakA-null cells (diamonds) and sibling wild-type control cells (squares). Cells were grown in HL-5 liquid medium and cell densities were determined by direct counting. The period of exponential growth (A) and a complete growth curve (B) are shown. Representative curves from more than five comparisons of different isolates of each strain, carried out in parallel, are shown.

yakA is required to initiate development

Given that YakA-null cells have a phenotype during exponential growth it is reasonable to assume that yakA is expressed at some level in all cells during growth. In growing cells, yakA mRNA is barely detectable on northern blots (Fig. 6A). To determine whether yakA expression is regulated, RNAse protection assays were performed on RNA from cells that were feeding on bacteria up to the point of starvation and during development. During growth on bacteria, the cell doubling time is 3-4 hours and the time of starvation can be defined as the point at which the bacteria are cleared from the plate. Under these conditions yakA mRNA accumulated during the growth phase and peaked at the point of starvation (Fig. 6B), consistent with a role for YakA in the earliest stages of development. After starvation, yakA mRNA decreased but remained present throughout development.

Fig. 6.

Expression of yakA mRNA. (A) yakA mRNA levels in wild-type and AX4(act6::yakA) cells in HL-5 liquid medium. 40 μg of total RNA were analyzed on northern blots probed with a fragment of the yakA cDNA. (B) RNAse protection assays were used to determine the timing of yakA expression. Wild-type cells were grown in the presence of bacteria for 44, 47 and 50 hours. Under these conditions cells divide every 3-4 hours. The bacteria were consumed by the Dictyostelium cells by 50 hours, as judged by microscopic inspection. Cells were also collected from the same set of plates at the point of aggregation (66 hours) and culmination (72 hours). Total RNA samples were extracted from the cells and subjected to RNAse protection analysis using a [32P]CTP-labeled antisense yakA RNA. (C) yakA mRNA levels in cells stimulated with conditioned medium. RNAse protection assays were performed to determine the yakA mRNA levels of cells grown to a density of 2.5×105/ml in the absence (−) or presence (+) of conditioned medium (see Materials and Methods). The yakA signal is 9.5-fold greater with added conditioned medium as judged by densitometry.

Fig. 6.

Expression of yakA mRNA. (A) yakA mRNA levels in wild-type and AX4(act6::yakA) cells in HL-5 liquid medium. 40 μg of total RNA were analyzed on northern blots probed with a fragment of the yakA cDNA. (B) RNAse protection assays were used to determine the timing of yakA expression. Wild-type cells were grown in the presence of bacteria for 44, 47 and 50 hours. Under these conditions cells divide every 3-4 hours. The bacteria were consumed by the Dictyostelium cells by 50 hours, as judged by microscopic inspection. Cells were also collected from the same set of plates at the point of aggregation (66 hours) and culmination (72 hours). Total RNA samples were extracted from the cells and subjected to RNAse protection analysis using a [32P]CTP-labeled antisense yakA RNA. (C) yakA mRNA levels in cells stimulated with conditioned medium. RNAse protection assays were performed to determine the yakA mRNA levels of cells grown to a density of 2.5×105/ml in the absence (−) or presence (+) of conditioned medium (see Materials and Methods). The yakA signal is 9.5-fold greater with added conditioned medium as judged by densitometry.

By analyzing the expression of well-characterized genes that mark the growth to development transition we determined the point of developmental arrest in YakA-null cells. Wild-type and YakA-null cells were allowed to develop on filters and samples were collected at various times to determine their pattern of gene expression by northern blot analysis. The cprD gene encodes a cysteine proteinase that is expressed during growth. The CprD mRNA and protein levels decrease immediately with the onset of development in wild-type cells (Souza et al., 1995). The acaA gene encodes the adenylyl cyclase ACA, carA encodes the cAMP receptor cAR1 and pkaC encodes the catalytic subunit of PKA, PKA-C. All of these genes in the cAMP response pathway are essential for development (Pitt et al., 1992; Klein et al., 1988; Mann et al., 1992; Anjard et al., 1992). The cprD message remained high in the starving YakA-null cells, while it decreased in wild-type cells by 2 hours of development, as expected (Fig. 7A). Conversely, the expression of acaA, carA and pkaC in YakA-null cells was significantly reduced. The expression of these genes in YakA-null cells remained low for the entire 24 hours required for the wild-type cells to complete development (Fig. 7A).

Fig. 7.

Gene expression in YakA-null cells. (A) Wild-type and YakA-null cells were developed on nitrocellulose filters for the times indicated. Samples of total RNA (20 μg) were analyzed on northern blots using fragments of the indicated genes as hybridization probes. (B) YakA-null and wild-type cells were starved in liquid suspension for 2 hours and then stimulated with 80 nM cAMP every 6 minutes for another 6 hours. Cell samples were collected at the indicated times, total RNA was isolated, and the mRNA levels for carA were determined by northern analysis.

Fig. 7.

Gene expression in YakA-null cells. (A) Wild-type and YakA-null cells were developed on nitrocellulose filters for the times indicated. Samples of total RNA (20 μg) were analyzed on northern blots using fragments of the indicated genes as hybridization probes. (B) YakA-null and wild-type cells were starved in liquid suspension for 2 hours and then stimulated with 80 nM cAMP every 6 minutes for another 6 hours. Cell samples were collected at the indicated times, total RNA was isolated, and the mRNA levels for carA were determined by northern analysis.

Some aggregation mutants can be induced to express cAMP response genes when they are stimulated periodically with exogenous cAMP. Thus, mutants that are completely defective in adenylyl cyclase activation and do not induce developmental genes, such as CRAC-null cells, can be induced to express ACA and cAR1 by applied pulses of cAMP (Insall et al., 1994). We pulsed YakA-null cells with cAMP and collected them at various times to determine cAMP production, adenylyl cyclase activity and carA mRNA levels. YakA-null cells did not accumulate high levels of carA mRNA at any time under these conditions (Fig. 7B). In addition, after 4 hours of pulsing with cAMP, followed by a 2-minute stimulation with 2′-deoxy-cAMP, YakA-null cells produced <0.1±0.06 pmoles of cAMP per 107 cells compared with 30±7 pmoles per 107 wild-type cells. Furthermore, adenylyl cyclase activity was below detection limits in the cAMP-pulsed YakA-null cells but was normal in wild-type cells (data not shown). Thus, the cAMP response system is not appreciably induced either in starving YakA-null cells or when they are periodically stimulated with cAMP. These results imply that YakA is required before 2 hours of development to bring about the changes in gene expression that normally occur upon starvation.

PSF is thought to regulate the starvation response by providing a measure of the food supply several generations before overt starvation (reviewed by Clarke and Gomer, 1995). PSF accumulates in growth medium in direct proportion to cell number and as the ratio of PSF to food bacteria becomes high the expression of early developmental genes such as discoidin-I ensues. We sought to determine the timing of YakA function relative to PSF. First, we tested PSF signaling in YakA-null cells. Wild-type and YakA-null cells were grown in liquid buffer containing bacteria and discoidin-I expression was monitored by immunofluorescence. Discoidin-I was expressed in most YakA-null cells four generations prior to the cessation of growth, indicating that PSF signaling is independent of yakA (data not shown). We also tested whether yakA expression could be induced by extracellular signals that accumulate in the growth medium. Conditioned medium was harvested from wild-type cells that had grown to high cell density. The conditioned medium was found to induce the expression of discoidin-I in wild-type cells at low cell density, indicating that the conditioned medium contained PSF (data not shown). This same conditioned medium induced an accumulation of yakA mRNA in wild-type cells (Fig. 6C), suggesting that yakA expression is controlled by an extracellular factor that accumulates during growth.

yakA expression induces the expression of pkaC

An indication that yakA and PKA function as components of the same pathway came from the developmental phenotype of wild-type cells overexpressing yakA. When the expression of yakA was induced in AX4(act6::yakA) cells by transferring them to liquid medium (the same conditions that induced the growth arrest, as described above), small clumps of cells were observed (Fig. 8A). Wild-type cells normally only become mutually adhesive several hours into development, so the adhesion of the AX4(act6::yakA) cells suggested that early developmental events were occurring in the presence of nutrients.

Fig. 8.

Phenotype of cells overexpressing yakA. (A) Bright field microscopy of wild-type cells (AX4) and AK801 cells (AX4 act6::yakA) that were grown in the presence of bacteria, collected, washed and incubated in HL-5 liquid medium for 6 hours. (B) Cells were treated as described in A, then collected, washed and plated for development on filters.

Fig. 8.

Phenotype of cells overexpressing yakA. (A) Bright field microscopy of wild-type cells (AX4) and AK801 cells (AX4 act6::yakA) that were grown in the presence of bacteria, collected, washed and incubated in HL-5 liquid medium for 6 hours. (B) Cells were treated as described in A, then collected, washed and plated for development on filters.

When allowed to develop on filters, AX4(act6::yakA) cells developed asynchronously and formed thick streams of cells and toroid-shaped multicellular structures by 12 hours (Fig. 8B). Development was accelerated in some of the aggregates early in development, while later development was delayed in the majority of aggregates. Finger structures formed in the AX4(act6::yakA) cells 3-4 hours earlier than was observed for wild-type cells. A small percentage of the AX4(act6::yakA) aggregates formed small fruiting bodies by 16 hours, more than 8 hours before wild-type. However, the majority of the AX4(act6::yakA) fingers observed at 16 hours were delayed at the culmination stage and some of them collapsed and reformed thick streams, slugs and large mounds of cells (Fig. 8B). Over the next 24 hours small fruiting bodies formed from a substantial proportion of these structures (not shown). Note that the developmental phenotype of AX4(act6::yakA) cells is different from the yakA-rescued YakA-null cells described in Fig. 1. This is likely due to the increased expression of yakA from the actin 6 promoter when the AX4(act6::yakA) cells were incubated in liquid growth medium prior to filter development. When AX4(act6::yakA) cells are grown on bacterial plates and allowed to develop in situ (conditions used for the yakA rescue of YakA-null cells) the phenotypes described in Fig. 8 are not apparent.

The expression of pkaC and acaA in AX4(act6::yakA) cells was elevated, relative to wild-type cells, 6 hours after the cells were transferred from bacterial growth plates into liquid medium (Fig. 9). The pkaC mRNA levels remained high in the AX4(act6::yakA) cells throughout their development on filters (Fig. 9A). We also measured the PKA activity in AX4(act6::yakA) cells that were incubated in HL-5. AX4(act6::yakA) cell lysates had 12-fold more PKA activity than wild-type cells when assayed without cAMP present (Table 1). The AX4(act6::yakA) cells appear to have 3.5-fold more PKA-C compared with wild-type cells, as judged by assaying PKA activity in the presence of cAMP (Table 1, values in parentheses). As expected, no PKA activity was induced by yakA expression in PKA-C-null cells (Table 1).

Fig. 9.

Induction of PKA-C and ACA expression by yakA. (A) Wild-type and YakA-overexpressing cells were grown on bacteria, harvested, washed and either incubated in HL-5 liquid medium, or plated on filters to initiate development. Wild-type AX4 and AX4(act6::yakA) cells (AK801) were grown on bacteria and harvested, washed and incubated in axenic medium for 6 hours. 10 μg samples of total RNA from these cells were used for northern analysis with a pkaC probe. (B) The 6-hour HL-5 samples probed with acaA.

Fig. 9.

Induction of PKA-C and ACA expression by yakA. (A) Wild-type and YakA-overexpressing cells were grown on bacteria, harvested, washed and either incubated in HL-5 liquid medium, or plated on filters to initiate development. Wild-type AX4 and AX4(act6::yakA) cells (AK801) were grown on bacteria and harvested, washed and incubated in axenic medium for 6 hours. 10 μg samples of total RNA from these cells were used for northern analysis with a pkaC probe. (B) The 6-hour HL-5 samples probed with acaA.

Table 1.

Dependence of PKA activity on yakA expression

Dependence of PKA activity on yakA expression
Dependence of PKA activity on yakA expression

Expression of yakA leads to an increase in both acaA and pkaC mRNA levels (Fig. 9), and PKA-C is necessary for the induction of acaA expression (Mann et al. 1997). In wild-type cells the largest increase in pkaC expression occurs between 2 and 4 hours of development, at the time that cAMP signaling genes such as acaA and carA are expressed (Fig. 7). Thus it is possible that the induction of pkaC expression at 2-4 hours of development requires cAMP. To test whether the increase in pkaC expression observed in the AX4(act6::yakA) cells is due to cAMP signaling we expressed yakA in an ACA-null mutant (Pitt et al., 1992). The acaA(act6::yakA) cells had a similar amount of PKA activity to that observed in AX4(act6::yakA) cells (Table 1). Thus, cAMP signaling does not appear to be required for the induction of PKA-C activity by yakA.

PKA-C and ACA in YakA function

Null mutations in pkaC, yakA and genes required for cAMP signaling such as acaA all result in aggregation-deficient phenotypes and a failure to express many developmental genes. Thus, it is difficult to order the function of early developmental genes by these phenotypic criteria. For instance, it is possible that PKA-C or cAMP signaling is required for some or all YakA-mediated events. As an initial exploration into staging the requirement for YakA function we analyzed the growth and development of PKA-C-null and ACA-null cells overexpressing yakA. When pkaC(act6::yakA) cells were placed in the liquid medium to activate the actin 6 promoter, their growth was arrested, just as was observed for the AX4(act6::yakA) cells (Table 2). yakA overexpression was also able to arrest the growth of the ACA-null cells under these same conditions (Table 2). Thus, neither adenylyl cyclase nor PKA-C are required for yakA-induced growth arrest. Interestingly, yakA expression was not able to rescue the developmental deficiency of the PKA-C-null cells, but overexpression of pkaC was able to rescue the development of YakA-null cells (Table 2). In addition, we found that yakA overexpression in ACA-null cells partially rescued the developmental deficiency of this strain. The cells were able to aggregate, but very few of these aggregates produced fruiting bodies. This rescue is probably due to PKA-C induction in the acaA(act6::yakA) cells (Table 1), since it has been shown that constitutive production of PKA-C rescues the development of ACA-null cells (Wang and Kuspa, 1997).

Table 2.

Effects of yakA and pkaC expression on growth and development

Effects of yakA and pkaC expression on growth and development
Effects of yakA and pkaC expression on growth and development

We have identified a new protein kinase in Dictyostelium that is essential for the initiation of development and the proper control of cell growth. Its predicted amino acid sequence is most similar to the Yak1p kinase of S. cerevisiae. YakA and Yak1p share two regions that have nearly the same degree of similarity as do their kinase core residues; region A is N-terminal to kinase domain I, and region B is between domains X and XI (as defined by Hanks and Quinn, 1991). Based on the crystal structure of the catalytic subunit of PKA (Knighton et al., 1991), these regions would be expected to be at opposite ends of the kinase core, where they might allow interactions with regulatory subunits. A variable region between domains X and XI exists in other kinases within the CMGC group and, in the case of the cyclin-dependent kinases, has been shown to bind regulatory proteins such as Cks1 (Bourne et al., 1996). Within the core residues themselves, YakA and Yak1p are much more similar to each other than they are to any other protein kinase in the databases, including those designated as Yak-related kinases such as the minibrain kinases (see below). Recently, a human protein kinase called PKY has been described that is more similar to the Yak kinases than are the minibrain kinases (Begley et al., 1997). However, the A and B regions of PKY are no more similar to the Yak kinases than are the A and B regions of the minibrain kinases (our unpublished observations). Thus, YakA and Yak1p may represent a new subfamily of kinases within the CMGC group of serine/threonine protein kinases.

The capacity of Dictyostelium yakA to reproduce the growth arrest in yeast that was observed for overexpression of yeast YAK1 supports the idea that these kinases function similarly in yeast and Dictyostelium. Both YAK1 and yakA arrest growth in yeast strains that are attenuated in the PKA pathway, but have no effect on otherwise wild-type yeast cells (Garrett et al., 1991; Fig. 4). The specificity of the growth arrest induced by both kinases also indicates that they have a conserved function and suggests that the expression of Dictyostelium yakA is not simply poisoning yeast cells. In addition, expression of Dictyostelium yakA in PKA-compromised yeast cells arrests the cells in G1, which is likely to be the stage in the cell cycle that YAK1 functions (Garrett et al., 1991). In yeast, Yak1p appears to operate downstream of PKA activity, or in a pathway that antagonizes the PKA pathway (Garrett et al. 1991; Hartley et al., 1994) and YakA-induced growth arrest in Dictyostelium similarly does not require PKA (Table 2). In yeast, Yak1p levels are low during exponential growth and increases when the cells are induced to arrest early in the cell cycle (Garrett et al., 1991) and Dictyostelium yakA mRNA increases as the cells approach growth arrest and peaks at the point of starvation. Thus, the analogous effects of Yak1p and YakA on the cell cycle of yeast and Dictyostelium, the PKA-independence of their cell cycle regulation, and their increased expression prior to cell cycle arrest, suggests that these Yak kinases operate in similar pathways.

Since wild-type Dictyostelium cells have a narrow range of sizes during growth, it is likely that cell size is tightly controlled. YakA is required to maintain the appropriate cell size, presumably by affecting the regulation that links cell growth to the cell cycle. In yeast, part of this regulation involves the translational control of the synthesis of cyclin Cln3p in G1, in which progression through the start of the cell cycle is regulated by the rate of translational initiation (Polymenis and Schmidt, 1997). Reduced rates of initiation (conditions of slow growth) lead to reduced Cln3p accumulation and result in a slower cell cycle progression. Since most cell growth occurs in G2 in Dictyostelium, size regulation would be expected to involve a coordination between the growth of a cell in G2 and its entry into mitosis. The small size of YakA-null cells, and their shorter doubling time, suggests that this coordination is altered in YakA-null cells. According to this idea the YakA-null cells would consistently undergo mitosis prior to attaining normal size, resulting in a smaller steady-state cell size. Since YakA-null cells have about 60% the volume of wild-type cells yet grow to 150% of the cell density of wild-type cells within the same period of time, macromolecular synthesis must occur at roughly the same rate, per volume of culture, in wild-type and YakA-null cells. This supports the idea that the normal function of YakA is to prevent cell division until a specific cell size is reached rather than to regulate the rate of growth.

The expression of some developmentally regulated genes such as dscA is induced to low levels during growth through the action of PSF (Rathi and Clarke, 1992; Burdine and Clarke, 1995). The increase of yakA expression when cells approach starvation is similar to that observed for genes that respond to PSF. We have found that yakA gene expression is induced in cells at low cell density by conditioned medium harvested from cells at high cell density. This is suggestive evidence that yakA expression is induced by PSF and consistent with the idea that YakA is an effector of PSF, but more direct evidence must await the availability of pure PSF or a PSF gene-disruption. Since one PSF-responsive gene, dscA, is induced in YakA-null cells during growth it appears that PSF signaling is operative in YakA-null cells and that dscA expression is independent of yakA function. Thus, YakA-null cells may produce and respond to PSF normally, but fail to initiate development due to a later requirement for YakA function.

The YakA-null mutants and the yakA-overexpressing cells suggest that YakA couples the sensing of nutrient availability to the initiation of development. Several facts argue against the idea that YakA-induced growth arrest is caused by a toxic effect of overexpressing a protein kinase, and in favor of a role for YakA in arresting growth through a specific signaling pathway. As described above, YakA appears to normally impose negative regulation on cell division during the exponential growth of wild-type cells. YakA-induced growth arrest in nutrient-rich medium does not appear to kill cells, but rather it causes the expression of genes essential for early development. In addition, the phenotype of YakA-null mutants is the opposite of the overexpression phenotype; the same developmental genes induced by YakA overexpression are not appreciably expressed in the absence of YakA. Finally, yakA mRNA normally accumulates throughout the growth phase and is maximal at the point cells stop growing and begin development. All of these facts are consistent with a specific regulatory function of YakA in mediating the transition of growth to development.

One attractive model for YakA function is that a low level of YakA regulates the cell cycle during growth while higher levels of YakA cause cell cycle arrest after its enzyme activity surpasses a threshold, or after its activity is induced in response to a specific starvation signal such as PSF. However, it is not clear that the amount of yakA mRNA that accumulates during the later stages of exponential growth could produce sufficient YakA protein to affect a growth arrest in wild-type cells. It is also possible that YakA is regulated during the cell cycle and, for instance, that YakA accumulates during G2 as the cells approach stationary phase and its activity prevents mitosis.

These issues will be resolved only after an adequate assessment can be made of YakA kinase activity within cells. In addition, the YakA pathway cannot be responsible for all forms of growth arrest since YakA-null cells do stop growing when nutrients are limiting. This yakA-independent growth arrest must be mediated by a nutrient-sensing mechanism that is not sufficient for initiating development. One member of the Ras family of proteins, RasG, has recently been implicated in controlling the growth to development transition (Khosla et al., 1996). A mutant form of RasG predicted to be constitutively active for signaling by analogy to mammalian systems (G12T), prevents development when expressed in wild-type cells. Thus, RasG is a possible target for YakA regulation, but any relationship between these two proteins remains unexplored.

In the absence of YakA, essential components of the cAMP signaling pathway are not expressed and the cells do not aggregate. The lack of a cAMP responsiveness in YakA-null cells was confirmed by their failure to induce cAR1 mRNA upon cAMP stimulation. YakA-null cells are also unable to turn off the expression of at least one vegetative gene, indicating that they are defective in an early event in development. These properties distinguish yakA mutants from other early developmental mutants such as those in Gα3, a G-protein alpha subunit, or the ACA regulator CRAC, two proteins required for early development (Insall et al., 1994; Brandon et al., 1997; Brandon and Podgorski, 1997). Gα3-null and CRAC-null cells both fail to express normal amounts of cAR1 and ACA, but the expression of these proteins can be restored by exogenous cAMP pulsing. Our results suggest that the function of YakA in the synthesis of cAR1 and ACA is required prior to Gα3 and CRAC function.

Much of the developmental function of YakA may be mediated by PKA-C. YakA regulates the induction of PKA-C activity by directly or indirectly regulating PKA-C mRNA levels. When yakA is overexpressed, pkaC mRNA levels and PKA-C activity increase, while in YakA-null cells pkaC mRNA is not induced at 2-4 hours. However, the increase in PKA-C activity in response to yakA expression is independent of ACA, which indicates that YakA function does not require cAMP signaling. Moreover, the observations that increased PKA-C expression can rescue the development of YakA-null cells but YakA expression does not rescue PKA-C-null cells are consistent with an essential role for PKA-C in mediating YakA function during early development. However, it is still possible that YakA regulation is largely independent of PKA-C regulation, given that PKA-C overexpression may simply bypass the requirement for YakA function.

Yeast and Dictyostelium are about as evolutionarily divergent from each other as each one is from the mammals (Loomis and Smith, 1995). Since the Yak kinases appear to function in a similar way in these two highly divergent species, it is likely that Yak proteins function similarly in other eukaryotes. The most likely candidates for metazoan Yak homologs are the minibrain kinases found first in Drosophila and later in mammals. The MNB1 gene was isolated from humans and found to map within a region of chromosome 21 associated with the pathogenesis of Down’s syndrome (Smith et al., 1997). In Drosophila, the minibrain gene mnb appears to be required for the production or maintenance of the appropriate number of neuroblast cells in the developing fly brain (Tejedor et al., 1995). The lower number of neuroblast cells in loss-of-function mnb mutants leads to specific structural abnormalities in the brain and behavioral defects. Superficially, the mnb phenotype is the opposite of the Yak kinase phenotypes. Both Yak kinases appear to negatively regulate cell division, whereas mnb appears to be required in a positive manner for neuroblast maintenance. However, it is equally possible that mnb is required for growth arrest in neuroblasts and promotes their differentiation into neurons, and that a failure to properly regulate this transition leads to cell death. The results with the human MNB1 gene support this hypothesis. A modest increase in the copy number of the human MNB1 gene in mice leads to an increase in the density of cortical neurons (Smith et al., 1997). Thus, with respect to the final differentiated cell type, heat-resistant yeast cells, Dictyostelium spores, and Drosophila or mouse neurons, it appears that the Yak and minibrain kinases all promote differentiation, whereas their absence precludes differentiation. It is therefore possible that the Yak and minibrain kinases share a conserved regulatory function of controlling the cell cycle and promoting the differentiation of eukaryotic cells.

We would like to thank William F. Loomis for advice and support during the course of this work. We thank Brian Desany and Steve Elledge for Lambda phage, yeast strains and vectors. We thank Margaret Clarke for advice concerning the actin 6 promoter and PSF. We also thank Gad Shaulsky and Richard Sucgang for comments that improved the manuscript and Richard Sucgang for assistance in preparing the figures. A. K. is a Kinship Foundation Searle Scholar and an American Cancer Society Junior Faculty Research Fellow. S. L. was supported by a NRSA fellowship from the NIH. This work was supported by grant GM52359 from the NIH, and an ATP grant from the State of Texas.

Allen
,
J. B.
,
Zhou
,
Z.
,
Siede
,
W.
,
Friedberg
,
E. C.
and
Elledge
,
S. J.
(
1994
).
The SAD1/RAD53 protein kinase controls multiple cell cycle checkpoints and DNA damage Induced transcription in yeast
.
Genes Dev
.
8
,
2401
2415
.
Anjard
,
C.
,
Pinaud
,
S.
,
Kay
,
R. R.
and
Reymond
,
C. D.
(
1992
).
Overexpression of Dd PK2 protein kinase causes rapid development and affects the intracellular cAMP pathway of Dictyostelium discoideum
.
Development
115
,
785
790
.
Araki
,
T.
and
Maeda
,
Y.
(
1995
).
Cell-cycle progression during the development of Dictyostelium discoideum and its relation to the subsequent cell-sorting in the multicellular structures
.
Dev. Growth Differ
.
37
,
479
485
.
Aronheim
,
A.
,
Engelberg
,
D.
,
Li
,
N.
,
ai-Alawi
,
N.
and
Karin
,
M.
(
1994
).
Membrane targeting of the nucleotide exchange factor Sos is sufficient for activating the ras signaling pathway
.
Cell
78
,
949
961
.
Begley
,
D. A.
,
Berkenpas
,
M. B.
,
Sampson
,
K. E.
and
Abraham
,
I.
(
1997
).
Identification and sequence of human PKY, a putative kinase with increased expression in multidrug-resistant cells, with homology to yeast protein kinase Yak1
.
Gene
200
,
35
43
.
Bourne
,
Y.
,
Watson
,
M. H.
,
Hickey
,
W.
,
Holmes
,
M. J.
,
Rocque
,
W.
,
Reed
,
S. I.
and
Tainer
,
J. A.
(
1996
).
Crystal structure and mutational analysis of the human CDK2 kinase complex with cell cycle-regulatory protein CksHs1
.
Cell
84
,
863
874
.
Brandon
,
M. A.
,
Voglmaier
,
S.
and
Siddiqi
,
A. A.
(
1998
).
Molecular characterization of a Dictyostelium G-protein α-subunit required for development. Gene (in press
).
Brandon
,
M. A.
and
Podgorski
,
G. J.
(
1997
).
Gα3 regulates the cAMP signaling system in Dictyostelium. Mol. Biol
.
Cell
8
,
1677
1685
.
Burdine
,
V.
and
Clarke
,
M.
(
1995
).
Genetic and physiologic modulation of the pre-starvation response in Dictyostelium discoideum. Mol. Biol
.
Cell
6
,
311
325
.
Clarke
,
M.
and
Gomer
,
R. H.
(
1995
).
PSF and CMF, autocrine factors that regulate gene expression during growth and early development of Dictyostelium
.
Experientia
51
,
1124
1134
.
da Silva
,
A. M.
and
Klein
,
C.
(
1990
).
Cell adhesion in transformed Dictyostelium discoideum cells: expression of gp80 and its biochemical characterization
.
Dev. Biol
.
140
,
139
148
.
Elledge
,
S. J.
,
Bai
,
C.
and
Edwards
,
M. C.
(
1993
).
Cloning mammalian genes using cDNA expression libraries in Saccharomyces cerevisiae. Methods: A Companion to Methods in Enzymology
, pp.
65
81
.
Academic Press
,
New York, NY
.
Garrett
,
S.
and
Broach
,
J.
(
1989
).
Loss of Ras activity in Saccharomyces cerevisiae is suppressed by disruptions of a new kinase gene, YAKI, whose product may act downstream of the cAMP-dependent protein kinase
.
Genes Dev
.
3
,
1336
1348
.
Garrett
,
S.
,
Menold
,
M. M.
and
Broach
,
J. R.
(
1991
).
The Saccharomyces cerevisiae YAK1 gene encodes a protein kinase that is induced by arrest early in the cell cycle
.
Mol. Cell. Biol
.
11
,
4045
4052
.
Gomer
,
R. H.
and
Ammann
,
R. R.
(
1996
).
A cell-cycle phase-associated cell-type choice mechanism monitors the cell cycle rather than using an independent timer
.
Dev. Biol
.
174
,
82
91
.
Gomer
,
R. H.
and
Firtel
,
R. A.
(
1987
).
Cell-autonomous determination of cell-type choice in Dictyostelium development by cell-cycle phase
.
Science
237
,
758
762
.
Hanks
,
S. K.
and
Quinn
,
A. M.
(
1991
).
Protein kinase catalytic domain sequence database: identification of conserved features of primary structure and classification of family members
.
Meth. Enzymol
.
200
,
38
81
.
Hartley
,
A. D.
,
Ward
,
M. P.
and
Garrett
,
S.
(
1994
).
The Yak1 protein kinase of Saccharomyces cerevisiae moderates thermotolerance and inhibits growth by an Sch9 protein kinase-independent mechanism
.
Genetics
136
,
465
474
.
Harwood
,
A. J.
,
Hopper
,
N. A.
,
Simon
,
M.-N.
,
Driscoll
,
D. M.
,
Veron
,
M.
and
Williams
,
J. G.
(
1992
).
Culmination in Dictyostelium is regulated by the cAMP-dependent protein kinase
.
Cell
69
,
615
624
.
Insall
,
R.
,
Kuspa
,
A.
,
Lilly
,
P.
,
Shaulsky
,
G.
,
Loomis
,
W. F.
and
Devreotes
,
P.
(
1994
).
CRAC, a cytosolic protein containing a pleckstrin homology domain, mediates G-protein activation of adenylyl cyclase in Dictyostelium
.
J. Cell Biol
.
126
,
1537
1545
.
Khosla
,
M.
,
Spiegelman
,
G. B.
and
Weeks
,
G.
(
1996
).
Over-expression of an activated rasG gene during growth blocks the initiation of Dictyostelium development
.
Mol. Cell. Biol
.
16
,
4156
4162
.
Klein
,
P. S.
,
Sun
,
T. J.
,
Saxe III
,
C. L.
,
Kimmel
,
A. R.
,
Johnson
,
R. L.
and
Devreotes
,
P. N.
(
1988
).
A chemoattractant receptor controls development in Dictyostelium discoideum
.
Science
241
,
1467
1472
.
Knecht
,
D.
and
Loomis
,
W. F.
(
1987
).
Antisense RNA inactivation of myosin heavy chain gene expression in Dictyostelium discoideum
.
Science
236
,
1081
1086
.
Knighton
,
D. R.
,
Zheng
,
J.
,
Ten Eyck
,
L. F.
,
Ashford
,
V. A.
,
Xuong
,
N.
,
Taylor
,
S. S.
and
Sowadski
,
J. M.
(
1991
).
Crystal structure of the catalytic subunit of cyclic adenosine monophosphate-dependent protein kinase
.
Science
253
,
407
414
.
Kuspa
,
A.
and
Loomis
,
W. F.
(
1992
).
Tagging developmental genes in Dictyostelium by restriction enzyme-mediated integration of plasmid DNA
.
Proc. Nat. Acad. Sci. USA
89
,
8803
8807
.
Kuspa
,
A.
and
Loomis
,
W. F.
(
1994
).
Transformation of Dictyostelium – gene disruptions, insertional mutagenesis, and promoter traps
.
Meth. Mol. Genet
.
3
,
3
21
.
Leichtling
,
B. H.
,
Majerfeld
,
I. H.
,
Spitz
,
E.
,
Schaller
,
K. L.
,
Woffendin
,
C.
,
Kakinuma
,
S.
and
Rickenberg
,
H. V.
(
1984
).
A cytosolic cAMP-dependent protein kinase in Dictyostelium discoideum. II. Developmental regulation
.
J. Biol. Chem
.
259
,
662
668
.
Loomis
,
W. F.
and
Smith
,
D. W.
(
1995
).
Consensus phylogeny of Dictyostelium
.
Experientia
51
,
1110
1115
McDonald
,
S. A.
and
Durston
,
A. J.
(
1984
).
The cell cycle and sorting behaviour in Dictyostelium discoideum
.
J. Cell Sci
.
66
,
195
204
.
Maeda
,
Y.
,
Ohmori
,
T.
,
Abe
,
T.
,
Abe
,
F.
and
Amagai
,
A.
(
1989
).
Transition of starving Dictyostelium cells to differentiation phase at a particular position of the cell cycle
.
Differentiation
41
,
169
175
.
Mallette
,
M. F.
(
1969
).
Evaluation of growth by physical and chemical means
.
Meth. Microbiol
.
1
,
521
566
.
Mann
,
S. K. O.
,
Brown
,
J. M.
,
Briscoe
,
C.
,
Parent
,
C.
,
Pitt
,
G.
,
Devreotes
,
P.
and
Firtel
,
R. A.
(
1997
).
Role of cAMP-dependent protein kinase in controlling aggregation and post-aggregative development in Dictyostelium
.
Dev. Biol
.
183
,
208
221
.
Mann
,
S. K. O.
,
Richardson
,
D. L.
,
Lee
,
S.
,
Kimmel
,
A. R.
and
Firtel
,
R. A.
(
1994
).
Expression of cAMP-dependent protein kinase in prespore cells is sufficient to induce spore cell differentiation in Dictyostelium
.
Proc. Nat. Acad. Sci. USA
91
,
10561
10565
.
Mann
,
S. K. O.
and
Firtel
,
R. A.
(
1991
).
A developmentally regulated, putative serine/threonine protein kinase is essential for development in Dictyostelium
.
Mech. Dev
.
35
,
89
102
.
Mann
,
S. K. O.
,
Yonemoto
,
W. M.
,
Taylor
,
S.
and
Firtel
,
R. A.
(
1992
).
DdPK3, which plays essential roles during Dictyostelium development, encodes the catalytic subunit of cAMP-dependent protein kinase
.
Proc. Nat. Acad. Sci. USA
89
,
10701
10705
.
Mulligan
,
J. T.
and
Elledge
,
S. J.
(
1994
). The construction and use of cDNA libraries for genetic selections. In
Molecular Genetics of Yeast: A Practical Approach
, pp.
65
81
.
Oxford University Press
,
Oxford, UK
.
Murray
,
A.
and
Hunt
,
T.
(
1993
).
The Cell Cycle. W. H
.
Freeman, New York, NY
.
Nellen
,
W.
,
Datta
,
S.
,
Reymond
,
C.
,
Sivertsen
,
A.
,
Mann
,
S.
,
Crowley
,
T.
and
Firtel
,
R. A.
(
1987
).
Molecular biology in Dictyostelium: tools and applications
.
Methods Cell Biol
.
28
,
87
100
.
Ohmori
,
R.
and
Maeda
,
Y.
(
1987
).
The developmental fate of Dictyostelium discoideum cells depends greatly on the cell-cycle position at the onset of starvation
.
Cell Differ
.
22
,
11
18
.
Parent
,
C.
and
Devreotes
,
P. N.
(
1996
).
Molecular genetics of signal transduction in Dictyostelium
.
Annu. Rev. Biochem
.
65
,
411
440
.
Pitt
,
G. S.
,
Milona
,
N.
,
Borleis
,
J.
,
Lin
,
K. C.
,
Reed
,
R. R.
and
Devreotes
,
P. N.
(
1992
).
Structurally distinct and stage-specific adenylyl cyclase genes play different roles in Dictyostelium development
.
Cell
69
,
305
315
.
Polymenis
,
M.
and
Schmidt
,
E. V.
(
1997
).
Coupling of cell division to cell growth by translational control of the G1 cyclin CLN3 in yeast
.
Genes Dev
.
11
,
2522
2531
.
Rathi
,
A.
,
Kayman
,
S. C.
and
Clarke
,
M.
(
1991
).
Induction of gene expression in Dictyostelium by Pre-Starvation Factor, a factor secreted by growing cells
.
Dev. Genet
.
12
,
82
87
.
Rathi
,
A.
and
Clarke
,
M.
(
1992
).
Expression of early developmental genes in Dictyostelium discoideum is initiated during exponential growth by an autocrine-dependent mechanism
.
Mech. Dev
.
36
,
173
182
.
Reymond
,
C. D.
,
Schaap
,
P.
,
Veron
,
M.
and
Williams
,
J. G.
(
1995
).
Dual role of cAMP during Dictyostelium development
.
Experientia
51
,
1166
1174
.
Sambrook
,
J.
,
Fritsch
,
E. F.
and
Maniatis
,
T.
(
1989
).
Molecular cloning: A Laboratory Manual
.
Cold Spring Harbor Laboratory
,
Cold Spring Harbor, NY
.
Schulkes
,
C.
and
Schaap
,
P.
(
1995
).
cAMP-dependent protein kinase activity is essential for pre-aggregative gene expression in Dictyostelium
.
FEBS Lett
.
368
,
381
384
.
Shaulsky
,
G.
and
Loomis
,
W. F.
(
1995
).
Mitochondrial DNA replication but no nuclear DNA replication during development of Dictyostelium
.
Proc. Nat. Acad. Sci. USA
92
,
5660
5663
.
Simon
,
M.-N.
,
Pelegrini
,
O.
,
Veron
,
M.
and
Kay
,
R. R.
(
1992
).
Mutation of protein kinase A causes heterochronic development of Dictyostelium
.
Nature
356
,
171
172
.
Smith
,
D. J.
,
Stevens
,
M. E.
,
Sudanagunta
,
S. P.
,
Bronson
,
R. T.
,
Makhinson
,
M.
,
Watabe
,
A. M.
,
O’Dell
,
T. J.
,
Fung
,
J.
,
Weier
,
H.-U. G.
,
Cheng
,
J.-F.
and
Rubin
,
E. M.
(
1997
).
Functional screening of 2 Mb of human chromosome 21q22.2 in transgenic mice implicates minibrain in learning defects associated with Down syndrome
.
Nature Genetics
16
,
28
36
.
Souza
,
G. M.
,
Hirai
J.
and
Freeze
,
H. H.
(
1995
).
Identification of two novel Dictyostelium discoideum cysteine proteinases that carry N-acetylglucosamine-1-P modification
.
J. Biol. Chem
.
270
,
28938
28945
.
Sussman
,
M.
(
1987
).
Cultivation and synchronous morphogenesis of Dictyostelium under controlled experimental conditions
.
Methods Cell Biol
.
28
,
9
29
.
Tejedor
,
F.
,
Zhu
,
X. R.
,
Kaltenbach
,
E.
,
Ackermann
,
A.
,
Baumann
,
A.
,
Canal
,
I.
,
Heisenberg
,
M.
,
Fischbach
,
K. F.
and
Pongs
,
O.
(
1995
).
Minibrain-a new protein kinase family involved in postembryonic neurogenesis in Drosophila
.
Neuron
14
,
287
301
.
Wang
,
B.
and
Kuspa
,
A.
(
1997
).
Dictyostelium development in the absence of cAMP
.
Science
277
,
251
254
.
Weijer
,
C. J.
,
Duschl
,
G.
and
David
,
C. N.
(
1984
).
Dependence of cell-type proportioning and sorting on cell cycle phase in Dictyostelium discoideum
.
J. Cell Sci
.
70
,
133
145
.
Wood
,
S. A.
,
Ammann
,
R. R.
,
Brock
,
D. A.
,
Li
,
L.
,
Spann
,
T.
and
Gomer
,
R. H.
(
1996
).
RtoA links initial cell type choice to the cell cycle in Dictyostelium
.
Development
122
,
3677
3685
.
van Haastert
,
P. J. M.
(
1995
).
Transduction of the chemotactic cAMP signal across the plasma membrane of Dictyostelium cells
.
Experientia
51
,
1144
1154
.
Verkerke-van Wijk
,
I.
and
Schaap
,
P.
(
1997
). cAMP, a signal for survival. In
Dictyostelium: A Model System For Cell and Developmental Biology
(ed.
K.
Inouye
), pp.
145
162
.
Japan Academic Press
,
Tokyo
.