Overexpression of Dd PK2 protein kinase causes rapid development and affects the intracellular cAMP pathway of Dictyostelium discoideum

cAMP has two distinct roles during Dictyostelium dis - coideum development. Its extracellular role as the chemoattractant during aggregation is well understood: cAMP is released by cells, detected by surface receptors linked to G proteins and then evokes three principal responses from the cells: chemotaxis towards cAMP, release of more cAMP and induction of the expression of a number of developmental genes (Sonneborn et al., 1963; Town et al., 1976; Sampson, 1977; Kessin, 1977).

An intracellular role for cAMP in development as a second messenger has also been suspected, principally because the cAMP-dependent protein kinase (cAPK) is developmentally regulated (Sampson, 1977; Rutherford et al., 1982; de Gunzburg et al., 1986). In Dictyostelium, this enzyme consists of single catalytic (C) and regulatory subunits (R). The R-subunit has been cloned and shown to be highly homologous to that from higher organisms (Mutzel et al., 1987), whereas the C-subunit is only known as a biochemical activity. Overexpression of the R-subunit, or of a mutant R-subunit unable to bind cAMP (effectively a dominant negative mutant of cAPK) blocks aggregation, indicating a role for cAPK in early development (Simon et al., 1989; Firtel and Chapman, 1990). A role for intracellular cAMP in spore maturation has been suggested by several lines of evidence. Addition of 8-bromo-cAMP (a cell-permeable analogue of cAMP) to submerged amoebae causes them to differentiate into spores, whereas cAMP itself only causes them to become prespores (Kay, 1989). A number of mutants able to form spores without 8-bromo-cAMP have been isolated (Sonneborn et al., 1963; Town et al., 1976; Kessin, 1977; Abe and Yanagisawa, 1983). These sporogenous and rapidly developing (rde) mutants develop rapidly and all form spores prematurely in development, so that the final “fruit” varies from a mound of stalk and spore cells (rdeC mutants) to a stump with spores at the base (rdeA mutants). Recently, the rdeC locus has been shown to encode the R-subunit of cAPK (Simon et al., 1992). Since elimination of the R-subunit in the rdeC mutants is expected to yield a constitutively active C-subunit, this result provides further strong evidence that activation of cAPK induces spore maturation. Finally, expression of an R-subunit of cAPK with mutated cAMP-binding sites driven by a prespore-specific promoter blocks the maturation of prespores into spores (Hopper, N., Harwood, A., Veron, M., Bouzid, S. and Williams, J.G., personal communication). These results indicate a central role for cAPK during cell differentiation in Dictyostelium.

We previously isolated a gene from Dictyostelium dis - coideum (Dd PK2) encoding a putative protein kinase of 648 amino acids whose C-terminal half is highly homologous (54% sequence identity) to the C-subunit of mammalian and yeast cAPKs (Bürki et al., 1991). Despite this homology, Dd PK2 seems unlikely to be the C-subunit of cAPK because the putative protein is too large and has an N-terminal domain with no homology to the C-subunit of cAPK. The Dd PK2 mRNA increases during development (Bürki et al., 1991) and an essential role for the protein in development is indicated by the lack of aggregate formation of a null mutant (Ddpk3⁻ in Mann and Firtel, 1991). To try to understand the function of Dd PK2, we overexpressed the complete protein or a truncated version with half the catalytic domain in Dic - tyostelium. Our results indicate a link between Dd PK2, rapid development and spore formation.

An XbaI cassette containing the actin 15 promotor/Tn903/actin 15 terminal part of pDNeo2 (Witke et al., 1987) was ligated into XbaI linearized pGEM3 DNA (Promega) containing the 6 kb genomic fragment coding for Dd PK2 (Bürki et al., 1991). The resulting construct named K-Neo was linearized with KpnI, blunt ended with T4 polymerase (Boehringer-Mannheim) in the presence of the four dNTPs (Sambrook et al., 1989) and recircularized with T4 DNA ligase (Boehringer-Mannheim). The resulting plasmid was named Kdel-Neo (Fig. 1).

K-Neo and Kdel-Neo plasmids were electroporated (0.67 kV, 3 mF in 0.2 cm cuvette) into AX2 in sterile water. Cells were plated on tissue culture plates (Falcon 25020) with 10 ml of HL5 broth (Ashworth and Watts, 1970). After 24 hours, selection was applied at 10 μg/ml of G418 (Gibco BRL) for 48 hours, followed by repetitive media changes (every 2 to 3 days) keeping a constant level of 5 μg/ml of G418. After 10-14 days, Dictyostelium clones were removed from plates with cloning rings and the rest of the cells harvested as a population. Once the cells were growing under shaking conditions (160 revs/minute) at 22°C, the G418 concentration was progressively raised to 50 μg/ml.

For filter development, cells were grown up to 3–6 ×10⁶/ml and harvested by centrifugation at 1000 g. After resuspension in the same volume of PDF buffer (20 mM KCl, 1.2 mM MgSO₄, 6.7 mM K₂HPO₄, 13.3 mM KH₂PO₄), the cells were centrifuged again and resupended at 10⁸/ml in PDF. 100 μl of cell suspension was deposited on 25 mm nitrocellulose filters (Schleicher and Schuell AG) on a pad (Whatman no. 1003125) saturated with PDF containing 500 μg/ml streptomycin and incubated at 22°C in a humid chamber for the indicated times.

The sporogenous phenotype was detected as described previously (Kay, 1989). Briefly, cells were incubated at 5 × 10⁵/ml in a 5 cm tissue culture dish (Sterilin) with 2 ml of 10 mM MES, 20 mM NaCl, 20 mM KCl, 1 mM CaCl₂, 1 mM MgSO_4, 200 μg/ml streptomycin, 15 μg/ml tetracyclin, 5 μg/ml discadenine, with 5 mM cAMP, with or without 20 mM 8-Br-cAMP. Differentiation was scored by phase microscopy after 24 hours of incubation, and each condition was done in duplicate.

Filters supporting developing cells were lysed in Eppendorf tubes by vortexing with 50 mM Tris pH 8.5, 0.1% SDS, 1 mM ribonucleotide vanadyl complex (Sigma). After removal of the filter, the cell lysate was extracted six times with phenol/chloroform (1:1 with 8-hydroxy-quinoline) and the RNA was ethanol-precipitated twice. 20 μg of RNA was loaded per lane on 1.2% agarose gels in 20 mM MES, pH 6.5 (Fluka) as described (Reymond et al., 1984), transferred under vacuum (Moretronic) to Genescreen membrane (NEN) and UV-crosslinked for 5 minutes. Blots were hybridised in Rapid Hybridisation Buffer (Amersham) with either random labelled (Boehringer-Mannheim kit) HindIII-KpnI fragment of Dd PK2 or probes derived from CP2, 2H3, discoidin I and actin 6.

For cAMP measurements, developing cells were harvested as cAPK activities were determined using the Protein Kinase A Assay System from Gibco BRL. Essentially, 2×10⁷ vegetative cells of the different strains were centrifuged at 1000 g for 5 minutes and resuspended in 200 μl of extraction buffer containing protease inhibitors (100-fold concentrated mix contains: 0.5 M benzamidine, 10⁴ units/ml aprotinin, 5 mg/ml trypsin inhibitor, 1 mg/ml leupeptin, 2 mg/ml antipain, 100 mM TLCK). The cells were lysed by freezing in dry ice and rapid thawing at 20°C (water bath). After centrifugation at 10000 g for 2 minutes at 4°C, cAPK activity was measured at 20°C in 10 μl of each Dictyostelium extract as phosphorylation of Kemptide substrate (in the presence of 0.1 mM cAMP) which can be inhibited by PKI inhibitor, according to the instructions of the Kit. The linearity of the reaction was verified by using twofold dilutions of the same extracts.

In order to obtain overexpression of Dd PK2, we ligated a neo^R cassette (Knecht et al., 1986) into the Dd PK2 genomic clone A7 (Bürki et al., 1991) in pGEM3 (Fig. 1). The genomic DNA fragment of about 6 kb, present in clone A7, contains the complete Dd PK2 coding region in its center and thus probably contains promoter and termination sequences as well. The unique K-Neo clone obtained showed an internal deletion of about 800 bp located 170 nucleotides upstream of the putative AUG start codon. This deletion within the promoter region does not seem, however, to prevent overexpression of the RNA (see below). We deleted the KpnI internal site of Dd PK2 (Fig. 1). The resulting construct, Kdel-Neo, has a frame-shift mutation resulting in premature termination in the middle of the putative catalytic C-terminal domain (loss of the last 151 amino acids). The encoded protein contains the complete N-terminal half of Dd PK2 as well as the phosphate anchor site (Knighton et al., 1991), but lacks peptide-binding sites and catalytic loop. The presence of the introduced mutation was verified by the disappearance of the KpnI site (data not shown).

K-Neo and Kdel-Neo DNAs were electroporated into Dic - tyostelium discoideum (Howard et al., 1988). We isolated a few individual clones (e.g. K-2 and Kdel-41) from the 40 to 50 obtained and kept the rest of the cells together as populations (K-P, Kdel-P). The pDNeo 2 (Witke et al., 1987) vector containing only the neo^R cassette was electroporated in parallel into Dictyostelium as control and called P-3.

The presence of 20 to 100 copies of the constructs in Dic - tyostelium cells was shown by hybridization of genomic blots with a Dd PK2 probe (data not shown). RNA was isolated from cells starved for 6 hours, since at that time Dd PK2 is fully expressed (Bürki et al., 1991) and analysed by hybridization with a HindIII-KpnI probe from Dd PK2 (Fig. 2). Scanning the autoradiogram indicated a 3-to 18-fold overexpression of Dd PK2 in K-Neo- and Kdel-Neo-containing cells (K- and Kdel-cells respectively), as compared to P-3 control cells. An actin probe was included in the hybridizations to allow standardization of RNA levels (Fig. 2).

We then observed development of K- and Kdel-cells. Both formed fruits after 16–18 hours on filters, whereas either P-3 or the parental strain AX2 required about 24 hours. The timing of aggregation was unchanged but development was accelerated thereafter. In addition K-cells had broader stalks resulting in the formation of shorter fruits on filters (Fig. 3). When plated at high density on agar, K-cells had even shorter fruits with enlarged stalks and some prostrate forms. Kdel-cells produced relatively short fruiting bodies but otherwise resembled AX2 wild-type (Fig. 3). No difference was observed between population of cells (K-P, Kdel-P) and the single clones (K-2, Kdel-41), indicating that the observed phenotype was due to the reintroduced DNA and not to a peculiarity of clonal isolates (data not shown).

In a further construct, we removed most of the Dd PK2 coding region by deleting the internal NdeI fragment (Fig. 1). After electroporation into AX2 cells, the integrity of vector sequences was verified by genomic Southern blot hybridization (data not shown). Such Dictyostelium cells showed wild-type development, strongly suggesting that rapid development results from overexpression of either complete or truncated Dd PK2 protein and not from any other sequence present on the vector (data not shown).

In order to analyse the pattern of gene expression during development, cells were starved on filters and RNA was extracted at time intervals (Fig. 4). Dd PK2 RNA is overexpressed throughout development with a relatively high level between 6 and 10 hours both in K- and K-del cells. We do not observe the shift in RNA size between 3 and 6 hours as in wild-type AX2 (Bürki et al., 1991), probably because of the internal 800 bp deletion within the upstream region in both K- and Kdel-Neo constructs. The presence of early and prespore genes was detected by hybridization with discoidin I and 2H3 probes, respectively. Even though CP2 is expressed both in prestalk and prespore cells (Pears et al., 1985), its timing of expression coincides with the appearance of prestalk cells. The P-3 control pattern corresponds to wild-type expression, with discoidin I expressed early, CP2 after about 14 hours (slug stage) and 2H3 after about 22 hours (culmination). We observed discoidin I expression in vegetative cells since they were grown axenically (Vauti et al., 1990). In K- and Kdel-cells, the expression of discoidin I was brought forward by about 3 hours only. In Kdel-cells the expression of both CP2 and 2H3 was brought forward by about 8 hours, but they retained their wild-type order of appearance, as in rdeA or C mutant development (Saxe and Firtel, 1986). K-cells also express CP2 earlier than wild type. However, most striking is the expression of 2H3 in K-cells which, unexpectedly, was advanced by about 12 hours and was expressed simultaneously with CP2.

The rapid development of K- and Kdel-strains resembled previously described rapid developing (rde) mutants which are affected in the intracellular cAMP signal transdution pathway (Abe and Yanagisawa, 1983). We therefore measured cAMP levels in the transformed strains (Fig. 5). The level of intracellular cAMP in P-3 cells during late development (up to about 6.5 pmol cAMP/mg of protein) is in good agreement with results from others (Abe and Yanagisawa, 1983) on wild-type cells. Both K- and Kdel-cells showed a lower level of intracellular cAMP after aggregation (10 hours) and for the rest of development compared to P-3. The levels measured at 0 and 5 hours are close to the limit of detection and do not differ significantly between the strains. The low level of intracellular cAMP observed in K- and Kdel-cells during late development resembles rdeC rather than rdeA mutants.

rdeC mutants are sporogenous, that is they are able to form spore cells when the amoebae are incubated in submerged culture in buffered salts containing cAMP (Kay, 1989). Under the same conditions, wild-type cells form only prespores. K-cells are sporogenous by this criterion whereas their parents, P-3 and Kdel-P are not (Table 1). However, spores can be induced in both strains P-3 and Kdel-P by 8-Br-cAMP, an analogue that is able to penetrate cells and activate cAMP-dependent protein kinase (cAPK). This shows that these two strains have kept the potential to form spores and that the phenotypic effects of overexpression of the Dd PK2 gene mimics a treatment activating cAPK.

cAPK levels were measured directly in K-P, Kdel-P and P-3 cells. The basal level of cAPK in crude extracts varied from sample to sample, even from cells of the same strain, possibly due to activation of the enzyme by varying levels of contaminating cAMP (data not shown). More reliable results were obtained after addition of saturating amounts of cAMP to the extracts (Table 2). We therefore measured cAPK activity as kemptide phosphorylation that was inhibitable by PKI, a specific inhibitor of cAPK. P-3 control and Kdel-P cAPK activities could not be statistically distinguished, whereas K-cells showed about 4-fold more activity. These results indicate that Dictyostelium cells, overexpressing the whole Dd PK2, possess increased cAPK activity.

Overexpression of the Dd PK2 mRNA in Dictyostelium cells (K-cells) modifies their phenotype. Development after aggregation is speeded up and aberrant fruiting bodies formed. On nitrocellulose filters, squat fruiting bodies with enlarged stalks are seen and on agar, at high cell density, many prostrate lumps also form. K-cells are also sporogenous, that is they can form spores when developed as sub-merged monolayers in the presence of cAMP. In these respects K-cells resemble members of the rapid developing/sporogenous class of mutants (rde) (Kessin, 1977; Abe and Yanagisawa, 1983). However, K-cells differ from rde mutants (Saxe and Firtel, 1986) in that a marker of prespore differentiation (2H3) is expressed at the same stage of development as the CP2 mRNA. Furthermore the fruit morphology of K-cells differs slightly from rdeA mutants which show pyramid-shaped fruits and from rdeC which form hemispheric mounds. The sporogenous phenotype and the premature expression of 2H3 both indicate that differentiation of spores is facilitated in K-cells.

The sporogenous phenotype has been shown previously to be associated with activation of the intracellular cAMP signal transduction pathway (Abe et al., 1981; Kay, 1989). This was confirmed recently by showing that the sporogenous rdeC mutants have a defective regulatory subunit of cAPK (Simon et al., 1992). Overexpression of the Dd PK2 gene also activates the intracellular cAMP signal transduction pathway, as evidenced by the high level of cAPK enzymatic activity found in K-cell lysates. The low level of intracellular cAMP measured in K-cells could be due to a negative feedback mechanism between cAPK and adenylyl cyclase as proposed earlier (Simon et al., 1992).

The phenotype of the Kdel-cells indicates that rapid development can be dissociated from spore formation in submerged monolayers. The presence of the N-terminal domain together with the ATP binding region (phosphate anchor, Knighton et al., 1991) is sufficient to confer rapid development without the sporogenous phenotype. The respective roles of the N-terminal part and ATP-binding site in the acceleration of development remain to be determined.

Although the Dd PK2 gene strongly resembles the catalytic subunit of cAPK (Knighton et al., 1991) in its C-terminal domain, the N-terminal half of the protein differs from any known sequence (Bürki et al., 1991). It therefore seems unlikely that this gene does encode the Dictyostelium catalytic subunit of cAPK, unless the protein is processed to remove the N-terminal domain. Furthermore, the absence of cAMP-binding site(s) and pseudosubstrate sequence within the N-terminal domain renders unlikely a regulation of Dd PK2 catalytic activity analogous to protein kinase C. The Dd PK2 gene, thus, resembles the SCH9 protein kinase, which has been found to substitute functionally for cAPK activity in yeast (Toda et al., 1988). Biochemical characterization of the Dd PK2 should clarify its relationship to cAPK.

We would like to thank Dr R.A. Firtel for the gift of 2H3, CP2 and discoidin I probes, M. Knetsch for helpful discussions, P. Dubied and B. Allegrini for photographic work. This work was supported by grant FN 31-28589.90 to C.D.R.