The Hybrid Type Polyketide Synthase SteelyA Is Required for cAMP Signalling in Early Dictyostelium Development

Background In our previous study we found that the expression of stlA showed peaks both in the early and last stages of development and that a product of SteelyA, 4-methyl-5-pentylbenzene-1,3-diol (MPBD), controlled Dictyostelium spore maturation during the latter. In this study we focused on the role of SteelyA in early stage development. Principal Findings Our stlA null mutant showed aggregation delay and abnormally small aggregation territories. Chemotaxis analysis revealed defective cAMP chemotaxis in the stlA null mutant. cAMP chemotaxis was restored by MPBD addition during early stage development. Assay for cAMP relay response revealed that the stlA null mutant had lower cAMP accumulation during aggregation, suggesting lower ACA activity than the wild type strain. Exogenous cAMP pulses rescued the aggregation defect of the stlA null strain in the absence of MPBD. Expression analysis of cAMP signalling genes revealed lower expression levels in the stlA null mutant during aggregation. Conclusion Our data indicate a regulatory function by SteelyA on cAMP signalling during aggregation and show that SteelyA is indispensable for full activation of ACA.


Introduction
Dictyostelium cells live in the soil as unicellular amoebae and consume bacteria and yeast as a food source. When their food source is depleted, Dictyostelium cells aggregate to form a multicellular organism. Cells in the aggregate start to differentiate into prestalk and prespore cells in a heterogeneous mixture which resembles salt and pepper [1]. Aggregates then transform into migrating slugs in which the differentiated cells sort to produce a prestalk-prespore pattern. After the slug stage, cells complete differentiation into their terminal types to form the stalk and spore cells of the mature fruiting body [1,2]. Several small molecules are reported to control Dictyostelium development, amongst which the second messenger cAMP.
In Dictyostelium, cAMP acts not only as an intracellular second messenger but also as a secreted extracellular signal. During aggregation, secreted cAMP acts as a chemoattractant, which causes Dictyostelium cells to move together by chemotaxis [3]. In addition to inducing chemotaxis, the activation of cAMP receptor cAR1 also initiates a signal cascade that results in cAMP synthesis. Synthesized cAMP is then secreted to attract neighboring cells [4,5]. cAMP also has an intracellular role in development, controlling many aspects of the developmental programme through the activation of PKA [6][7][8][9][10].
cAMP is produced by three distinct adenylyl cyclases during Dictyostelium development. These are ACA, ACB, and ACG. ACA produces the cAMP required for cell aggregation in early development [5,11], ACB is involved in the maturation of stalk and spore cells [12,13], and ACG controls prespore differentiation in the slug stage and also acts as an osmosensor that controls spore germination in the fruiting body [10,13,14].
The genome sequence of Dictyostelium was published in 2005 [15]. Dictyostelium, it turned out, has a massive potential for the production of polyketides, with 45 polyketide synthase (PKS) genes found in its genome [15,16]. Two are novel types of hybrid PKSs called ''Steely'' [17], which to our best knowledge, are only found in the cellular slime molds. These novel enzymes are a fusion of iterative type I and type III PKSs. Iterative type I PKSs are large single polypeptides including multifunctional domains encoded by only a single gene and iteratively utilize some of their functional domains for a polyketide biosynthesis, whereas type III PKSs are single enzymes comprising of only keto-synthase domain [18,19]. One of the Steely enzymes, SteelyB, was reported to produce Differentiation Inducing Factor-1 (DIF-1), an inducer of prestalk cell differentiation [17]. We showed earlier that a mutant that lacked the stlB gene made morphologically normal but structurally weak stalk and did not produce a basal disc or lower cup. These defects were recovered by the addition of DIF-1 in the medium, indicating that DIF-1 induces basal disc and lower cup formation among prestalk cells [17,20]. DIF-1 has an additional role in developmental regulation [21], and both DIF-1 and the related molecule DIF-2 function as modulators of cAMP-induced chemotaxis [22]. This suggests that DIFs govern multiple aspects of the developmental programme.
Another Steely enzyme, SteelyA, was reported to make 4methyl-5-pentylbenzene-1,3-diol (MPBD) in vivo and we showed that a stlA knockout mutant has defective spore maturation, which was recovered by the addition of MPBD to the mutant cells [23]. Previously, two distinct patterns of stlA gene expression were reported: one of them was that the gene was expressed during the early stage development [17], whereas another one was that the gene was expressed during only late stage [24]. We resolved these conflicting reports concerning the gene expression pattern, by reexamining the expression pattern with by RT-PCR using two different primer sets. It appeared that the expression of stlA peaked in the pre-aggregation stage, followed by smaller peak during the late culmination stage [23]. We also confirmed that the function of SteelyA during this second peak was the control of spore maturation [23,25]. In subsequent research we specified the timing of MPBD-regulated spore maturation and found that MPBD was critical in the later stage of development [25].
In this study, we performed a detailed analysis of the function of SteelyA during early stage development and found that MPBD is Table 1. Summary of the primer sets used in qRT-PCR.

Cell culture and development
The Dictyostelium discoideum Ax2 strain was grown in an axenic medium (HL-5) at 22uC [26]. The stlA null strain (Strain ID: DBS0236953) was grown in HL-5 medium supplemented with 10 mg/ml blasticidin S [17]. To initiate multicellular development, the axenically grown cells were harvested at a density of approximately 5 6 10 6 cells/ml, washed twice in the phosphate buffer (PB) (2.7 mM Na 2 HPO 4 /10.7 mM K 2 HPO 4 , pH 6.2), and plated on PB agar plates (1.5% agar in PB) or nitrocellulose filters (0.45 mm pore size, Black gridded, 47 mm diameter, Millipore) at a density of 1610 6 cells/cm 2 . For gene expression analysis, developing cells were harvested every 3 hours until 12 h and used for RNA purification. For submerged streaming, the axenically grown cells were harvested and washed twice, and then submerged in 1 ml of PB at 5610 5 cells/cm 2 in a culture dish and the depth of PB was about 3 mm.

Chemotaxis assay and cAMP pulsing
Chemotaxis assay towards cAMP was performed as described [22] with a few modifications. Axenically grown cells in shaking culture were harvested in the log phase, washed in PB, and were starved at a density of 2610 7 cells/ml in PB in the presence or absence of MPBD for 6.5 h. Then, starved cells were washed twice in PB and resuspended in PB to a final density of 3610 5 cells/ml. Small portion of agar (,10 g) was washed with water (500 ml each time) repeatedly and extensively. Then the agar was dried completely. The agar powder was dissolved in PB and boiled to prepare the hydrophobic agar plate. Three droplets of starved cells, 2 ml (3 mm diameter) each, were placed on 1% hydrophobic agar plate in the petri dish. Chemotaxis towards cAMP was tested by placing a 2 ml droplet of various concentrations of cAMP at 3 mm distance from the cell-filled droplet. The distribution of cells within their droplet was observed after 1 h, and droplets were scored 'positive' if at least twice as many cells had migrated to the side closer to the cAMP droplet than to the other side. The chemotaxis rate was calculated as the percentage of the number of 'positive' droplets to that of total droplets.
To find out the chemotactic response to folate, axenically grown cells in shaking culture were harvested in the log phase, washed twice in PB, and were resuspended in PB to a final density of 2610 7 cells/ml. Droplets of cells, 1 ml of the cell suspension, were placed on a hydrophobic agar plate at 3 mm distance from a 1 ml droplet of different concentrations of folate and were observed after 4 h from placing on the plate.
To examine the effect of cAMP pulsing on the stlA null strain, Ax2 or stlA null cells were harvested from the exponential growth phase and washed twice in PB. Washed cells were resuspended at a density of 2610 7 cells/ml in PB and then stimulated with 20 nM cAMP pulses every 6 min for 6 h. The pulsed cells were washed and then used for the chemotaxis assay or developed on nitrocellulose filters.

Assay for cAMP relay response in intact cells
Cells were harvested in the log phase, washed twice, and distributed on non-nutrient agar plates (1.5% agar in 5 mM Na 2 HPO 4 N2H 2 O/5 mM KH 2 PO 4 buffer, pH 6.5) at 1610 8 cells/ plate. Cells were incubated for 16 h at 4uC and some hours at 22uC until aggregation territories started to form (2 h for the case of wild type cells and 3-4 h for the mutant cells) and then were harvested in the phosphate buffer (5 mM Na 2 HPO 4 N2H 2 O/5 mM KH 2 PO 4 , pH 6.5) and resuspended at a density of 1610 8 cells/ ml. Cells were stimulated with 5 mM 29-deoxycAMP in the presence of 5 mM DTT. Stimulation was terminated by adding an equal volume of 3.5% perchloric acid, and lysates were neutralized with 50% saturated KHCO 3 . cAMP levels were analyzed by isotope dilution assay, with [ 3 H]cAMP as the competitor for binding to the bovine PKA regulatory subunit [27,28].

Gene expression analysis by qRT-PCR
Specific primer sets for qRT-PCR were designed to amplify a unique sequence located near the end of 39 region of each gene (Table.1). BLAST similarity searches were used to confirm that each primer sequence amplified the unique sequence. Before using each primer set for qRT-PCR, genomic DNA was used as a template to evaluate the quality of primer set. Total RNA was extracted from cells at the indicating time point using the RNeasy mini kit (Qiagen) according to the manufacturer's instructions and then the contaminating genomic DNA was eliminated using Recombinant DNase I (TaKaRa). DNase treatment was performed as follows: 50 mg total RNA was mixed with 10 units Recombinant DNase I and DNase I buffer (2 mM Tris-HCl; pH 7.5, 5 mM NaCl, 0.01 mM CaCl 2 , 5% glycerol) and incubated for 30 min at 37uC. The reaction was stopped by incubating with 25 mM EDTA at 80uC for 2 min. cDNA was synthesized by reverse transcription of 1 mg total RNA with oligo dT 15 using ImProm-II Reverse Transcription System (Promega). qPCR was carried out with ABI PRISM 7000 Sequence Detection System (Applied Biosystems) and the amplifications were performed using iTaq Universal SYBR Green Supermix (Bio-Rad).
The qPCR program was performed as follows: One cycle of 30 sec at 95uC followed by 40 cycles of 15 sec at 95uC, 1 min at 60uC for annealing, extension, and plate reading. Results of qRT-PCR were analyzed using the comparative C T method [29] with the amplification of ig7 (mitochondrial large rRNA) as a control.

Results and Discussion
In previous studies we found that the expression profile of stlA peaks in the early stage of development and shows a weak peak in the last stage of development [23]. In this study we chose to understand SteelyA's function in the early stage of development by analyzing the stlA null mutant.

StlA null strain shows small aggregation territories
We first studied cell aggregation in the stlA null mutant and found that this process was delayed by about 3 h when compared with its parental strain Ax2. In some cases, the stlA null strain showed developmental delay by more than 7 h (Fig. 1A).
In addition to the aggregation delay, stlA null cells formed smaller aggregates than those of Ax2 cells. When cells were developed on a nitrocellulose filter, the stlA null cells made smaller and more numerous aggregates compared to Ax2 cells (Fig. 1B). As the result of forming small aggregates, fruiting bodies of the stlA null strain were smaller in size than those of the Ax2 strain (Fig. 1C).
To analyze the streaming behavior of stlA null cells in more detail, we observed aggregation under submerged conditions (Fig. 2). Under these conditions, stlA null cells made shorter aggregation streams and showed smaller aggregation territories than Ax2 cells. Since these aggregates were homogeneously distributed on the substratum the smaller aggregate sizes were not due to the uneven distribution of cells. To investigate whether these defects were due to the lack of MPBD, the major product of SteelyA enzyme, we added 200 nM MPBD directly to the mutant cells because 200 nM MPBD was needed to rescue the spore maturation defect in the mutant [23]. We found that the aggregation of stlA null cells was rescued in the presence of MPBD during starvation (Fig. 2). This observation suggests that MPBD is involved in cAMP-mediated aggregation.

StlA null cells are impaired in chemotaxis towards cAMP
Small aggregation territories and delayed aggregation could result from defects in chemotaxis, cell adhesion or cAMP synthesis [30][31][32]. The observation that the stlA null mutant showed aggregation defect suggested that the mutant had a defect in chemotactic response to cAMP or cAMP production. To find out whether the aggregation defect was due to either chemotactic response to cAMP or cAMP production, we first investigated chemotaxis towards cAMP using the two droplet chemotaxis assay. As shown in Figure 3A, stlA null cells showed a chemotaxis defect towards 250 nM cAMP. In contrast, Ax2 cells showed clear cAMP chemotaxis under the same condition. This cAMP chemotaxis defect of the stlA null strain was recovered by addition of 200 nM MPBD in the buffer during 6.5 h starvation. This result indicates that the presence of MPBD during starvation induces competence for chemotaxis towards cAMP. We next examined the effect of various MPBD concentrations on chemotaxis. As shown in Figure 3B, MPBD improved cAMP chemotaxis in a dose dependent manner in the stlA null mutant. The minimal concentration for MPBD recovery of cAMP chemotaxis is about  100 nM and this concentration is similar to the recovery concentration of spore encapsulation in the stlA null mutant [23]. The defect in cAMP-induced chemotaxis was not due to the defect in cell motility because the stlA null strain showed the same chemotactic response to folate as the parent Ax2 (Fig. 4). Figure 4 shows a representative result of chemotactic response to 500 mM folate of each cell. We also examined the response to 250 mM or 100 mM folate and the same results were obtained (data not shown).

cAMP-induced cAMP production is impaired in stlA null mutant
Next, we focused on cAMP-induced cAMP production in the stlA null mutant. When cells were stimulated by extracellular cAMP during the aggregation stage, they soon started to produce extracellular cAMP to transmit the aggregation signal to adjacent cells. Figure 5 shows the results of cAMP relay response assay. When cells were stimulated with cAMP analogue, 29-deoxy cAMP, cAMP receptor was stimulated and then ACA was activated. We measured the cAMP accumulation in the 29-deoxy cAMP stimulated cells that ACA was activated. Figure 5 shows the result with 29-deoxy cAMP stimulation. Without 29-deoxy cAMP stimulation, stlA null cells accumulated less than half of cAMP in Ax2 cells (Fig. 5A). On the other hand, with 29-deoxy cAMP stimulation, stlA null cells eventually showed almost the same cAMP accumulation, but had slower production kinetics than Ax2 cells (Fig. 5B).
These results strongly suggested that the stlA null mutant had a defect in the cAMP signal relay, especially related to the production of cAMP in the early aggregation stage. To confirm this, the cells were starved and stimulated with periodical addition of cAMP. Figure 6A shows the rescue of cAMP chemotaxis in stlA null cells by cAMP pulsing. When stlA null cells were stimulated with periodical cAMP addition during starvation, they showed the normal cAMP chemotaxis. This recovery of cAMP chemotaxis by cAMP pulsing indicates that stlA null cells have a defect in periodical cAMP production. Figure 6B indicates the dose dependency of cAMP chemotaxis in cAMP pulsed cells examined by the two droplet chemotaxis assay. Also, cAMP pulsed stlA null cells showed normal development without aggregation delay, that is, the aggregation defect of the mutant was recovered by cAMP pulsing in the absence of MPBD (Fig. 7). These results suggested that MPBD was involved in cAMP signal relay and regulates the synthesis and/or release of cAMP in the early stage of development.
cAMP signalling genes are down regulated in stlA null mutant The observation that the stlA null mutant responds chemotactically to folate but not to cAMP suggests a defect in cAMP signal transduction. cAMP-induced chemotaxis is mediated by complex parallel signal transduction cascade. The observation that pretreatment with MPBD restores cAMP-induced chemotaxis in the stlA null mutant suggests that MPBD may induce the expression of genes required for this response.
To determine which step(s) caused the defective chemotaxis in the stlA null mutant we examined the expression pattern of cAMP signalling genes in stlA null and Ax2 cells. Figure 8 shows the gene expression patterns of key genes in cAMP chemotaxis by quantitative RT-PCR. We analyzed the genes coding for the cAMP receptor (carA), adenylyl cyclase (acaA), secreted cAMP phosphodiesterase (pdsA) [33], intracellular cAMP phosphodiesterase (regA) [34], positive regulators of adenylyl cyclase (piaA and dagA) [35,36] and an a subunit of G-protein coupled with a cAMP receptor (gpaB) [37]. The mRNA levels of these genes were lower in the mutant cells than in Ax2 cells, with the sole exception of the piaA gene. The reason for this difference in piaA gene expression from other cAMP signalling genes will be discussed later. This decreased expression pattern was especially clear at 3 h when aggregation genes started to express. Although the maximal mRNA levels of these genes were initially lower in the stlA null mutant than in the Ax2, their levels at 9 h were almost the same as those of the Ax2 cells (9 h in Figure 8). At this point (9 h), the Ax2 cells had reached the late aggregation stage, while stlA null cells were early aggregation. We also examined the expression levels of a gene coding for cell-cell adhesion protein (csA) [38]. This gene is strictly developmentally regulated and cAMP inducible [39]. Although the expression of csA was lowered and delayed in the stlA null mutant compared to Ax2, its mRNA level in the stlA null mutant was the same as Ax2 by 9 h. These results indicate that at least at 9 h after starvation (t9), the expression levels of cAMP signalling genes in the stlA null mutant are enough to induce the expression of cAMP inducible genes, and as a result, csA is expressed normally. We also confirmed that MPBD rescued carA and acaA expression defects in the stlA null mutant (data not shown).
As mentioned above, the expression of piaA, a positive regulator of adenylyl cyclase, was not down regulated as other genes in the stlA null mutant but rather up regulated. This regulation of piaA may be due to the difference of pathways activating ACA between CRAC (Cytosolic Regulator of Adenylyl Cyclase) and Pia (coded by dagA and piaA, respectively). Pia is a component of the target of rapamycin complex 2 (TORC2) [40]. TORC2 plays roles in activation of ACA and PKB [40][41][42][43]. By reconstitution experiments using cells lacking both CRAC and Pia, it has shown that both CRAC and Pia are absolutely required for ACA activity [35]. However, it has also reported that CRAC is not required for TORC2 formation [40]. In our stlA null mutant, dagA was down regulated but expressed partially. This means that the CRAC pathway in the stlA null mutant is partially active but has only a small effect on ACA. We propose that the stlA null mutant should make use of the TORC2 pathway to compensate for reduced ACA activity caused by down regulated dagA. As a result, the expression of piaA may be up regulated in the stlA null mutant.
From above data, we concluded that the function of SteelyA during early stage is to make starving cells aggregation competent by regulating the cAMP signalling gene expression. This means both the chemotactic response to cAMP and adenylyl cyclase activity required for aggregation are controlled by MPBD, a major product of SteelyA.
In this study, we focused on the role of SteelyA during the early stage of development and showed that the polyketide MPBD produced by SteelyA is required for the expression of cAMP signalling genes during early development. A stlA null mutant had aggregation defects because their expression in the mutant was down regulated.
The stlA null mutant showed similar phenotype to the padA 2 mutant [30,44]. In both mutant, chemotactic cAMP response was impaired and cAMP signalling genes were down regulated. However, the aggregation defect in the padA 2 mutant is not rescued by cAMP pulses because cAMP pulses are not able to fully induce cAMP relay genes in the mutant. In addition, accumulated extracellular cAMP levels in the padA 2 mutant are almost the same as that in wild type unlike the stlA null mutant [30]. In contrast, the aggregation defect in the stlA null mutant was restored by cAMP pulses and the mutant showed lower accumulated cAMP levels. MPBD signalling cascade for induction of cAMP signalling genes, therefore, is probably different pathway from PadA signalling cascade. MPBD may act on the same pathway as cAMP pulses for stimulation of cAMP signalling gene expression because cAMP pulsed stlA null cells show normal aggregation in the absence of MPBD. MPBD acts as a spore maturation factor during the late stage of development [23,25]. Anjard et al. reported that MPBD might bind to CrlA, the GPCR coupled to Ga1, and inhibit GskA to induce spore differentiation by SDF-1 [45]. CrlA, suggested as an MPBD receptor, is also expressed during early stage [46,47]. However, the crlA knockout mutant was reported as forming larger aggregates than wild type cells and exhibited normal chemotaxis to cAMP [46]. Although there is the difference of parental strains between crlA and stlA mutant (KAx3 and Ax2, respectively), the phenotype of crlA mutant differs from that of stlA null mutant. Therefore, MPBD-CrlA pathway might be not used to modulate the expression of cAMP signalling genes during the early stage of development. MPBD might regulate those expressions by acting through some receptor other than CrlA. Figure 8. Expression patterns of cAMP signalling genes in stlA null mutant. Expression levels of each gene were analyzed by qRT-PCR. The levels were normalized to that of the endogenous control ig7 (mitochondrial large rRNA). In addition, the expression value of Ax2 at some time point (carA, acaA, pdsA, gpaB and csA: 6 h, regA: 3 h, piaA and dagA: 12 h) in each gene was used as an internal standard and values were normalized it. Values indicated in graphs are the means and SD (bar) of 3 independent experiments, which were performed by using independent total RNA. doi:10.1371/journal.pone.0106634.g008