INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Amoebae of Dictyostelium discoideum aggregate into mounds containing up to 10⁵ cells before forming fruiting bodies in which the spore mass is held up by a tapering stalk (9, 19, 20, 26). Although prespore and prestalk cells diverge and sort out soon after aggregation, they do not form spores and stalk cells until fruiting body formation is initiated. Since spores are unable to move once they have encapsulated, it is essential that their terminal differentiation be coordinated with their position on the elongating stalk; otherwise they would be left at the base. When the mass of prespore cells has been lifted off the substratum, a wave of expression of the spore-specific gene spiA can be seen to start in the prespore cells nearest the prestalk region that passes down through the mass of prespore cells in an hour or two (28). This temporal pattern suggests that prespore cells are responding to a signal secreted from prestalk cells as they undergo terminal differentiation.

The peptide signal SDF-2 is released from prestalk cells in a manner dependent on the ABC transporter TagC and induces rapid encapsulation of prespore cells (6, 31, 33). A candidate receptor for this signal is the putative histidine kinase DhkA, since strains carrying null mutations in dhkA form few spores even when they develop in chimeric mixtures with wild-type cells. When developed as pure populations, dhkA cells proceed normally to culmination but then form fruiting bodies with long thin stalks that tend to topple over (42). Cells that are deficient in DhkA fail to respond to the sporulation-inducing factor SDF-2 (6), further implicating DhkA in the response to SDF-2. The predicted product of dhkA has two hydrophobic domains near the N terminus that could cross the membrane, leaving an extracellular loop of about 310 amino acids on the outside while keeping the putative catalytic domain and the aspartate relay site within the cytoplasm. This putative topology, together with the genetic evidence, suggests that DhkA may be the SDF-2 receptor.

DhkA is a member of the family of two-component signal transduction systems, in which autophosphorylation of a receptor kinase leads to modification of the activity of a response regulator (4, 38). These two-component systems are found in both bacteria and eukaryotes (3, 10, 11, 12, 14, 21, 23, 24, 29). Phosphorelay to an aspartate moiety on the response regulator either can be direct from the histidine phosphate of the sensor kinase or may go through various intermediates (13, 16, 25). DhkA is referred to as a "hybrid kinase" because it has both a catalytic domain and a response regulatory domain (42). Such hybrid kinases autophosphorylate the histidine in the sequence conserved just upstream of the catalytic domain and then pass the phosphate to the aspartate moiety in the response regulatory domain. In the yeast osmoregulatory signal transduction pathway, the phosphate linked to histidine on SLN1 is relayed to an aspartate near the carboxy end of SLN1 before being passed to a histidine carried by the small intermediate protein YPD1 and then on to its final destination on SSK1 (25). Phosphorylation of SSK1 keeps it from activating the mitogen-activated protein kinase kinase kinases SSK2 and SSK22 (22).

Dictyostelium cells that overexpress the catalytic subunit of PKA (40) as a result of carrying multiple copies of the pkaC gene can be induced to form spores rapidly after 24 h in monolayer cultures by addition of SDF-2 (6). While SDF-2 induces up to 50% of dhkA⁺ pkaC::pkaC cells to encapsulate within 30 min, it does not affect encapsulation in dhkA pkaC::pkaC cells (6). These results demonstrate that DhkA plays an essential role in the response to SDF-2 and raises the possibility that this peptide is a ligand that activates DhkA and leads to rapid encapsulation. Since activation of PKA by addition of the membrane-permeable derivative of cyclic AMP (cAMP), 8-Br-cAMP, leads to rapid encapsulation even in dhkA mutant cells (42), it is likely that PKA functions downstream of DhkA.

It has recently been shown that another histidine kinase, DhkB, functions in spores to ensure dormancy by maintaining high PKA activity (45). The germination inhibitor that accumulates during culmination, discadenine, was proposed as a candidate for the ligand that activates DhkB. Moreover, Zinda and Singleton (45) suggested that PKA activity is kept high in spores by inhibiting the cytoplasmic cAMP phosphodiesterase RegA (34, 41). Thus, both of the signal transduction pathways initiated by these histidine kinases may result in elevated levels of cAMP and PKA activity.

In the present study, we show that DhkA is a membrane-spanning histidine kinase and that modification of the extracellular domain by insertion of a MYC epitope reduces sensitivity to SDF-2. Moreover, the presence of the MYC epitope in the extracellular loop of DhkA makes it sensitive to activation by monoclonal antibody to MYC. We have found that inactivating either regA, the gene encoding the cytoplasmic cAMP phosphodiesterase, or pkaR, a gene encoding the regulatory subunit of PKA, partially suppresses the block to sporulation resulting from inactivation of dhkA. Previous studies have shown that activation of PKA in prespore cells leads to rapid encapsulation (15, 27). These findings are consistent with a signal transduction pathway in which the SDF-2 peptide activates DhkA on the surface of prespore cells, leading to inhibition of RegA and to subsequent activation of PKA.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. Proteinase K was purchased from Boehringer-Mannheim, Indianapolis, Ind. Monoclonal antibody 9E10 against c-Myc was purchased from Santa Cruz Biotechnology, Santa Cruz, Calif. Alkaline phosphatase-conjugated goat anti-mouse immunoglobulin G (IgG) and goat anti-rabbit IgG were purchased from Sigma, St. Louis, Mo. Custom-synthesized oligonucleotides were from Operon Technologies Inc., Alameda, Calif.

Plasmids. The knockout vector for disruption of pkaR was a gift from A. Kuspa. The expression vector pkaC::pkaC (A-7 Neo) has been previously described (5). pBluescript-MYC₆ was a gift from M. Yaffe. Knockout vectors for disruption of regA (30) and expression of dhkA under its own promoter (42) were as described previously. Site-directed mutagenesis, insertion of DNA sequences encoding six consecutive c-Myc epitope tags, and in-frame deletion in dhkA were performed by standard molecular biology techniques, and the mutated alleles were cloned into the dhkA expression vector (42). All of the new constructs were verified by sequencing with an ABI Prism 377 DNA sequencer. The dhkA mutant alleles generated were dhkA^H1395Q, which carries a T-to-A substitution at bp 4185 such that histidine 1395 is changed into glutamine; dhkA^D2075N, which carries a G-to-A substitution at bp 5223 such that aspartate 2075 is changed into asparagine; dhkA^900MYC6, in which a 302-bp PstI DNA fragment was generated by PCR from pBluescript-MYC₆ and introduced into the PstI site at bp 2701 in dhkA; and dhkA^2025MYC6, in which a 291-bp HindIII DNA fragment was generated by PCR from pBluescript-MYC₆ and introduced into the HindIII site at bp 6076 in dhkA.

Cells, growth, transformation, and development. Strains used in this work were wild-type AX4 cells (17), dhkA null mutants (42), and regA null mutants (30). All strains were grown in HL5 medium and maintained on SM nutrient agar in association with bacteria (39). Synchronous development of cells on nitrocellulose filters and sporulation assays were performed as described previously (31, 32).

Subcellular fractionation of DhkA. Cells (10⁸) expressing the 900MYC6 epitope-tagged DhkA protein were developed on nitrocellulose filters for 18 h. The cells were collected into buffer (20 mM Tris [pH 8.3], 150 mM NaCl, 1 mM EDTA) and frozen at 80°C. Cells were thawed on ice, the sample was split in two, and 0.5% Nonidet P-40 (NP-40) was added to one of the aliquots. After centrifugation at 1,250 × g for 5 min at 4°C, the supernatant was centrifuged at 40,000 × g for 30 min at 4°C. The resulting supernatant and pellet were saved. One-fifth of each fraction or one-fifth of a sample containing an equal number of unfractionated cells was mixed with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer, and the proteins were resolved by electrophoresis on a 5.5% polyacrylamide-SDS gel. Proteins were subjected to Western blot analysis with the 9E10 monoclonal antibody against c-Myc (Santa Cruz Biotechnology), as described previously (37).

Sensitivity of DhkA to proteinase K treatment of intact cells. Cells expressing the 900MYC6 epitope-tagged DhkA protein or the 2025MYC6 epitope-tagged DhkA protein were developed on nitrocellulose filters for 18 h. Multicellular structures were collected into isotonic buffer (30 mM HEPES [pH 7.5], 10 mM magnesium acetate, 10 mM NaCl, and 10% sucrose) and triturated by passage through a 22-gauge needle. Samples of 2.5 × 10⁷ cells per ml were treated with freshly prepared proteinase K solution at 0°C as indicated. After an hour, 1 mM phenylmethylsulfonyl fluoride (PMSF) was added to inhibit the protease, and the cells were washed three times by centrifugation and resuspended in 1 ml of isotonic buffer containing 1 mM PMSF. The cells were then lysed in SDS-PAGE sample buffer, and one-fifth of each sample was resolved by electrophoresis on a 7.5% polyacrylamide-SDS gel. Proteins were subjected to Western blot analysis with the 9E10 monoclonal antibody against c-Myc (Santa Cruz Biotechnology) or with a polyclonal antibody against the intracellular TipA protein (37), as described previously (31).

Encapsulation assay. Cells were dissociated from early to mid-culminants and deposited at 2 × 10⁴/cm² in 24-well Falcon plastic plates with 0.5 ml of buffer containing 10 mM MES [2-(N-morpholino)ethanesulfonic acid] (pH 6.5), 20 mM NaCl, 20 mM KCl, 1 mM CaCl₂, and 1 mM MgSO₄. Purified SDF-2 or anti-MYC monoclonal antibody 9E10 was added at various concentrations, and spore-like cells were scored by phase-contrast microscopy (7). Each assay was repeated three to six times, and at least 200 cells were scored. One unit of SDF-2 activity is defined as the amount necessary to induce 50% of K-P cells to form spores.

Protein kinase assay. Extracts were prepared from E. coli M15(pREP4) carrying pdhkAcat, pdhkAcat^H1395Q, or the vector pQE-9 (Qiagen) with no insert. When the cultures reached an optical density at 600 nm of 0.7, they were induced with 1 mM IPTG for 1 h and 1 ml was harvested in a solution of 10 mM Tris-HCl (pH 8.0), 50 mM sodium phosphate (pH 8.0), 100 mM NaCl, and 0.1 mM PMSF before being broken by sonication. After centrifugation at 16,000 × g for 5 min, the supernatants were mixed with 50 µl of Talon Metal Affinity Resin (Clontech Laboratories, Inc., Palo Alto, Calif.) for 20 min at 4°C. Talon beads were washed with a solution of 50 mM Tris-HCl (pH 7.5), 50 mM KCl, 5 mM MgCl₂, and 0.1 mM PMSF; collected by centrifugation; and washed a second time with the same buffer supplemented with 2 mM 2-mercaptoethanol.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

DhkA spans the plasma membrane. The primary sequence of dhkA suggested that the encoded protein might be membrane associated, due to the presence of two hydrophobic stretches of about 20 amino acids each between amino acids 770 and 790 and between amino acids 1100 and 1120 (42). To directly determine the membrane topology of DhkA, we generated two independent epitope-tagged alleles of dhkA by introducing a DNA fragment encoding six successive MYC epitopes into different regions in the gene. One of the MYC₆ epitopes was introduced into the predicted extracellular region of DhkA between amino acids 900 and 901 (dhkA^900MYC6), and the other was introduced into the carboxy terminus between amino acids 2025 and 2026 (dhkA^2025MYC6). The tagged genes were transformed into dhkA null mutants, and the localization of the tagged protein was tested.

View larger version (25K): [in this window] [in a new window]   FIG. 1.   Membrane association and orientation of the DhkA protein. (A) Cells expressing the 900MYC6 epitope-tagged DhkA protein were disrupted in the presence or absence of 0.5% NP-40, as indicated. The supernatant (S) and pelleted material (P) following high-speed centrifugation, as well as unfractionated extract (T), were resolved by gel electrophoresis and analyzed by Western blotting with a monoclonal antibody against the MYC epitope. Each sample contains material from 2 × 107 cells. The sizes and positions of protein markers are indicated on the left, in kilodaltons. (B) Intact cells expressing the 900MYC6 epitope or the 2025MYC6 epitope were disaggregated by trituration in isotonic buffer and treated with 0 to 500 µg of proteinase K per ml, as indicated above the respective lanes. Cells were washed free of the protease and resuspended in sample buffer. Samples representing 2 × 107 cells each were resolved by gel electrophoresis and analyzed by Western blotting with a monoclonal antibody against the MYC epitope. Numbers on the left indicate sizes, in kilodaltons. As a control, aliquots from the same samples were analyzed with antibodies against the intracellular protein TipA. (C) Schematic representation of the DhkA protein (2150 amino acids) relative to the plasma membrane. The MYC6 epitopes are represented by shaded areas. H 1393 and D 2075 represent the conserved histidine residue in the H motif and the conserved aspartic acid residue in the D motif, respectively. The drawing is not to scale.

Modification of the extracellular loop reduces sensitivity to SDF-2. Insertion of the MYC₆ epitope tag either into the extracellular loop or near the C terminus was found to affect the function of dhkA, since neither the dhkA^900MYC6 allele nor the dhkA^2025MYC6 allele was able to fully complement the sporulation defect in dhkA null cells (Table 1). If DhkA is a receptor kinase, then the epitope tag in amino acid 900 could interfere with ligand binding. The epitope tag in amino acid 2025 is adjacent to the predicted D motif around the aspartic acid residue at amino acid 2075 and could possibly interfere with phosphorylation of that residue.

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Reduced sensitivity to SDF-2 in dhkA900MYC6 cells. Wild-type cells of strain AX4 () and dhkA cells carrying the DhkA 900MYC6 construct () were developed on filters until early culmination, dissociated, and deposited as a monolayer under buffer. Purified SDF-2 was added at various concentrations, and spores were counted after 6 h. Fold stimulation was calculated relative to the number of spores seen in cultures incubated in the absence of added SDF-2. As a control, we added 50 units of SDF-1 and found that sporulation was stimulated 2.3 ± 0.2-fold for both wild-type dhkA+ cells and dhkA900MYC6 cells.

Site-directed mutations. DhkA is similar to other hybrid kinases in that it carries both an H motif in the predicted catalytic histidine kinase domain and a D motif in the carboxy terminus. The importance of these two domains for the activity of DhkA is demonstrated by the results shown in Table 1. The dhkA gene was modified by site-directed mutagenesis to change either the conserved histidine residue in the H motif into glutamine (dhkA^H1395Q) or to change the conserved aspartic acid residue in the D motif into asparagine (dhkA^D2075N). The mutant alleles were cloned into separate dhkA expression vectors and transformed individually or in combination into dhkA null mutant cells. Sporulation of the resulting transformed cell lines was measured. As shown in Table 1, the wild-type dhkA expression vector was able to fully complement the dhkA mutation, and site-directed mutations in either the H or D motif reduced that ability, demonstrating the importance of each of these residues for the activity of DhkA. When both the H1395Q and D2075N constructs were cotransformed into the same dhkA null host cell line, they complemented each other and were able to fully rescue the sporulation defect of the dhkA null host (Table 1). This result indicates that, as in the yeast osmoregulatory sensor kinase SLN1 (25), phosphorylation of the aspartic acid in the D motif is not necessarily a monomolecular event. In addition, this finding indicates that these single amino acid substitutions did not have a general effect on the structure of the protein but rather a specific effect on the predicted functional domains.

Partial redundancy of DhkA and DhkB. Although sporulation is severely reduced and delayed in dhkA strains, it is not totally blocked (Table 1). Residual sporulation may result from partial overlap of other pathways functioning during culmination. The histidine kinase DhkB has been shown to play an essential role in maintaining dormancy once spores have encapsulated (45). It has been proposed that DhkB responds to the adenine derivative discadenine, which accumulates during culmination and is essential for maintaining dormancy (1). Since it appears that activation of either DhkA or DhkB can result in accumulation of cAMP to levels that activate PKA, DhkB may be responsible for the residual sporulation seen in the absence of DhkA. Therefore, we knocked out dhkB in a dhkA null mutant by using homologous recombination and determined the sporulation efficiency of the double mutant. Cells lacking both histidine kinases form fruiting bodies with long thin stalks resembling those of dhkA mutants and make very few spores (Fig. 3; Table 3). Since inactivation of dhkB alone does not significantly affect encapsulation (45), we would have expected the double mutants to form at least a few percent spores, but less than 1 in 10⁴ of the cells encapsulated. The dhkA dhkB double mutant could be induced to sporulate efficiently by incubating cells dissociated from early culminants in the presence of 20 mM 8-Br-cAMP to activate PKA (data not shown). These results indicate that PKA activation by DhkB may account for the low level of sporulation that is seen in dhkA strains.

View larger version (76K): [in this window] [in a new window]   FIG. 3.   Terminal differentiation of the dhkA dhkB double mutant. Wild-type (AX4) cells (left) and a dhkB derivative of strain AK299 (dhkA) (right) were developed for 36 h on filter supports. The fruiting bodies of the double-mutant (dhkA dhkB) strain had long weak stalks that often toppled over before they could be photographed. Prespore cells of the double mutant ascended the stalks but never encapsulated to form spores.

View this table: [in this window] [in a new window]   TABLE 3.   Sporulation of dhkA dhkB double mutants

dhkA dhkB

Autophosphorylation of the catalytic domain of DhkA. To test whether the putative catalytic domain of DhkA is able to catalyze its own phosphorylation, we expressed the region encoding amino acids 1275 to 1884 of DhkA (42), fused to a His₆ tag, in E. coli. This DhkA derivative lacks the transmembrane domains and intervening loop as well as the carboxy-terminal portion where the relay aspartate is found. We also prepared a variant of this protein in which the essential histidine codon was modified to encode glutamine, as in the dhkA^H1395Q mutation. Both of these proteins were expressed at high levels in E. coli (see Materials and Methods). The His₆ fusion proteins were bound to metal affinity beads and incubated with radiolabelled ATP while still on the beads. The proteins were then eluted, electrophoretically separated on gels, and autoradiographed. Material prepared from bacteria transformed with the construct expressing the wild-type catalytic domain incorporated label into a protein of the expected size (Fig. 4A). A protein of the same size (70 kDa) was recognized by antibodies to the His₆ epitope (data not shown). Extracts from bacteria expressing the construct modified by site-directed mutagenesis did not incorporate label into a protein of this size (Fig. 4B), although a protein with the His₆ epitope was present at the same level (data not shown). To test whether the label in the 70-kDa material had the properties of histidine-phosphate, blots of the gels were exposed before and after being treated with 1 M HCl for 2 h (25). As expected, all of the radioactivity in the 70-kDa band was removed by mild acid hydrolysis (Fig. 4D). Together with the fact that replacement of histidine₁₃₉₅ by glutamine precluded phosphorylation, these results show that DhkA is an autophosphorylating histidine kinase.

View larger version (40K): [in this window] [in a new window]   FIG. 4.   Autophosphorylation of DhkA. Proteins from an equal number of bacteria carrying pdhkAcat (A), pdhkAcatH1395Q (B), or vector alone (C) were bound to metal affinity beads and incubated with [32P]ATP. Proteins were eluted, resolved by gel electrophoresis, and exposed to X-ray film. (D) Proteins were then transferred to a nitrocellulose filter that was treated with 1 M HCl for 2 h, and the portion shown in lane A was exposed for the same period of time to X-ray film. The arrow indicates the position of the 70-kDa protein.

Genetic analysis. Null mutations in dhkA result in reduced sporulation and enhanced prestalk differentiation (42), a phenotype opposite that observed with either regA null mutants or pkaR null mutants where prestalk and stalk differentiation are compromised and prespore differentiation is accelerated (2, 30, 35). We therefore generated double null mutants in which either regA or pkaR were disrupted in a dhkA null background. Both the dhkA regA double null mutants and the dhkA pkaR double null mutants sporulated well in comparison to the parental dhkA strain, as did a dhkA null mutant that was transformed with a vector leading to overexpression of the catalytic subunit of PKA encoded by the pkaC gene (Table 4). Thus, RegA and PKA appear to function downstream of DhkA. Previous results have shown that all of these genes function downstream of the ABC transporter TagC and that the effects of RegA are mediated by PKA (30, 34).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

SDF-2 is released from prestalk cells during culmination and appears to diffuse throughout the sorus (6, 7). Prestalk cells respond by releasing a burst of SDF-2 within a few minutes while prespore cells respond by encapsulating within an hour (6, 7). The responses of both cell types are dependent on the histidine kinase encoded by dhkA (6). Histidine kinases are widespread in two-component signal transduction mechanisms of bacteria and have been found to mediate a variety of responses in plants, fungi, and Dictyostelium (4, 21). They all share a conserved motif surrounding the histidine that is autophosphorylated and show sequence similarity in the catalytic domain. Likewise, there is a telltale sequence surrounding the aspartate to which the phosphate is relayed.

The primary sequence of DhkA shows that, in addition to the conserved motifs, there are two potential transmembrane domains near the N terminus separated by a 310-amino-acid loop (42). When a MYC₆ epitope tag was positioned at amino acid 900 in the loop, it was rapidly degraded when proteinase K was added to the extracellular medium. Protease treatment was carried out at 0°C to minimize internalization of the enzyme by endocytosis. The lack of significant internalization was verified by showing that a cytoplasmic protein, TipA, was not degraded under these conditions. When the MYC₆ epitope was positioned at amino acid 2025 of DhkA near the receptor aspartate, it was protected from protease degradation although DhkA was trimmed. The topology of DhkA was confirmed by the demonstration that addition of antibody to the MYC₆ epitope induced sporulation in cells expressing dhkA^900MYC6. Insertion of the MYC₆ epitope at amino acid 900 of DhkA appears to reduce its sensitivity to SDF-2 by about 100-fold. Thus, the 310-amino-acid loop between the transmembrane domains of DhkA is exposed to the intercellular medium and appears to be critical for ligand binding and activation of DhkA. The carboxy-terminal portion of DhkA that carries the conserved aspartate to which the phosphate is relayed appears to be internal, where it can affect its response regulator.

Using a His₆ derivatized version of the central portion of dhkA, we showed that DhkA is a protein kinase able to autophosphorylate on a histidine residue (Fig. 4). Site-directed mutagenesis of histidine₁₃₉₅ to glutamine in the catalytic domain identified that residue as essential for autophosphorylation (Fig. 4). The physiological significance of the conserved histidine in DhkA was further demonstrated in vivo. We found that a full length dhkA construct carrying this site-directed mutation failed to fully rescue sporulation in dhkA mutants (Table 1). Likewise, the importance of the conserved aspartate near the C terminus of DhkA was demonstrated by site-directed mutagenesis to asparagine to preclude phosphorylation. The dhkA^D2075N mutation also compromised the ability of the gene to complement the dhkA null mutation (Table 1). Whereas each of the alleles was unable to fully complement the dhkA null mutation, cotransformation of dhkA cells with both of the mutant constructs gave strains that sporulated normally, indicating an interaction between the modified proteins.

Phosphotransfer from endogenous DhkB to the aspartate in the H1395Q version of DhkA may account for the increased sporulation seen in dhkA cells expressing dhkA^H1395Q relative to untransformed dhkA cells, while phosphotransfer from the histidine in the D2075N version of DhkA to the aspartate in DhkB may account for the fact that dhkA cells expressing dhkA^D2075N make about a third as many spores as wild-type cells (Table 1). Double mutants lacking both histidine kinases, DhkA and DhkB, make almost no spores even when transformed with constructs expressing either dhkA^H1395Q or dhkA^D2075N (Table 3). Intermolecular cooperation has also been observed in the hybrid kinase SLN1p, which is responsible for osmoregulation in yeast (25). The results presented here show that DhkA is a membrane-spanning histidine kinase and is likely to be a receptor which mediates the cellular response to SDF-2 in the genetic pathway that eventually leads to PKA-dependent differentiation.