Adenylate cyclase A acting on PKA mediates induction of stalk formation by cyclic diguanylate at the Dictyostelium organizer

Significance

Organizers are small groups of cells in developing embryos that secrete signals to control behaviors such as cell differentiation or cell movement of larger groups. In Dictyostelia, the apical tip is the site where differentiation of the fruiting body stalk initiates. The cause of tip-specific stalk formation has been unclear, but we show here that the more widely produced stalk-inducing signal cyclic diguanylate activates cAMP synthesis by adenylate cyclase A, which is specifically expressed at the apical tip. cAMP next activates cAMP-dependent protein kinase, which then triggers stalk differentiation.

Abstract

Coordination of cell movement with cell differentiation is a major feat of embryonic development. The Dictyostelium stalk always forms at the organizing tip, by a mechanism that is not understood. We previously reported that cyclic diguanylate (c-di-GMP), synthesized by diguanylate cyclase A (DgcA), induces stalk formation. Here we used transcriptional profiling of dgca− structures to identify target genes for c-di-GMP, and used these genes to investigate the c-di-GMP signal transduction pathway. We found that knockdown of cAMP-dependent protein kinase (PKA) activity in prestalk cells reduced stalk gene induction by c-di-GMP, whereas PKA activation bypassed the c-di-GMP requirement for stalk gene expression. c-di-GMP caused a persistent increase in cAMP, which still occurred in mutants lacking the adenylate cyclases ACG or ACR, or the cAMP phosphodiesterase RegA. However, both inhibition of adenylate cyclase A (ACA) with SQ22536 and incubation of a temperature-sensitive ACA mutant at the restrictive temperature prevented c-di-GMP–induced cAMP synthesis as well as c-di-GMP–induced stalk gene transcription. ACA produces the cAMP pulses that coordinate Dictyostelium morphogenetic cell movement and is highly expressed at the organizing tip. The stalk-less dgca− mutant regained its stalk by expression of a light-activated adenylate cyclase from the ACA promoter and exposure to light, indicating that cAMP is also the intermediate for c-di-GMP in vivo. Our data show that the more widely expressed DgcA activates tip-expressed ACA, which then acts on PKA to induce stalk genes. These results explain why stalk formation in Dictyostelia always initiates at the site of the morphogenetic organizer.

Aggregative multicellularity resulting in fruiting body formation is the most common evolutionary transition from a unicellular to a multicellular lifestyle. Whereas in most aggregating organisms the fruiting bodies consist entirely of either spores or cysts or have stalks consisting of secreted matrix, the Dictyostelia additionally evolved somatic cells. Stalk cells are the ancestral somatic cells of Dictyostelia, and their differentiation starts at the tip of emerging fruiting structures, with prespore cells moving up along the stalk to form the spore head. The model Dictyostelium discoideum additionally differentiates into three more somatic cell types, which form disk and cup-shaped structures to support the stalk and spore head (1).

Similar to animals but unlike plants and fungi, D. discoideum development consists of an integrated program of coordinated cell movement and cell differentiation. This program is initiated by starvation, which causes cells to collect into aggregates, using secreted cAMP pulses, produced by adenylate cyclase A (ACA), as a chemoattractant (2). Secreted cAMP produced by the adenylate cyclases ACG and ACR additionally induces differentiation of prespore cells (3). The prespore cells in turn synthesize the polyketide Differentiation-Inducing Factor 1 (DIF-1), which causes differentiation into precursors of some somatic cell types (4). In cell monolayers, DIF-1 induces the differentiation of vacuolated cells, which are phenotypically identical to stalk and basal disk cells (5). However, in normal development, DIF-1 is only required for lower cup and basal disk differentiation (6).

D. discoideum uses the diguanylate cyclase DgcA to synthesize cyclic diguanylate (c-di-GMP) (7), a well-known second messenger in prokaryotes with a major role in triggering biofilm formation (8). Dictyostelium dgcA is expressed in prestalk cells, and dgca null mutants form normal migrating slugs but do not initiate fruiting body formation. This defect is due to the lack of stalk cell differentiation, and is restored by externally applied c-di-GMP. c-di-GMP also induces stalk cell differentiation in cell monolayers, indicating that c-di-GMP is a secreted signal that triggers stalk cell differentiation (7). The mode of action of c-di-GMP is unknown, as are the c-di-GMP–regulated genes that cause stalk cell differentiation.

We previously used a reporter gene fused to a region of the ecmB promoter, which directs expression in stalk cells, as a marker for c-di-GMP–induced stalk gene expression (7). ecmB (extracellular matrix B) is a commonly used stalk marker, but is also expressed in the basal disk and upper and lower cup from other promoter regions (1). Because absolute expression levels from cells transformed with reporter constructs depend on plasmid copy number, such markers are less suited for comparing gene expression levels between mutants. They are also unsuitable for use in mutants generated by overexpression of genes under the same selectable marker. To identify stalk genes that are directly regulated by c-di-GMP, we performed high-throughput RNA sequencing of wild-type and dgca− multicellular structures. We validated candidate stalk genes by examining their expression pattern and up-regulation by c-di-GMP, and then used the genes as markers to investigate the signal transduction pathway of c-di-GMP. Our results point to crucial roles for ACA and PKA as intermediates for c-di-GMP–induced stalk gene expression.

Results

Identification of Target Genes for c-di-GMP.

To identify genes that are regulated by c-di-GMP, we used high-throughput RNA sequencing to compare all genes transcribed in dgca− mutants and control random integrants at the time point when the control cells just start to form fruiting bodies (Fig. S1A). Due to their stalk defect, the dgca− mutants continue slug migration for days until they die (7). RNAs were isolated from two dgca− knockout clones and two random integrants in two separate experiments, yielding four replicates each for dgca− and control. Cluster analysis of the gene expression profiles showed that the four knockout (KO) and four random integrant (RI) samples formed two distinct clades (Fig. S1B), indicating reproducible differences between the expression profiles of dgca− and control cells. Statistical analysis showed that out of the 8,939 detected mRNAs, 2,565 were significantly up-regulated in the dgca− mutant, 2,281 were down-regulated, and 4,093 genes remained unchanged (Fig. S1C). Gene ontology analysis showed that the genes down-regulated in dgca− were enriched in catalytic activity, oxidoreductase activity, RNA binding, and carbohydrate binding (Fig. S1D). The latter is expected, because cellulose cell wall synthesis is a major aspect of the stalk (and spore) phenotype.

Fig. S1.

Gene expression profiling of dgca− mutants. (A) Experimental setup. Two dgca− knockout (KO) clones and two random integrants (RIs) were developed for 22 h in two separate experiments, “A” and “B.” At this stage, the RI clones were in early-to-mid fruiting body formation. RNAs were isolated, enriched for mRNA, and reverse-transcribed to yield cDNA libraries, which were sequenced using the Illumina HiSeq 2000 platform. (B) Sample clustering. Gene expression across each of the samples was clustered via their per-sample pairwise Spearman correlation coefficients. (C) Differential gene expression in KO vs. RI. Differentially expressed genes (FDR <0.05) are highlighted in red. Blue lines indicate a log₂ fold change of 1, up or down; cyan lines delineate genes that are fivefold up- or down-regulated. Green diamonds mark 25 highly expressed prestalk genes that are potential targets for c-di-GMP. An orange diamond marks the position of the stalk marker ecmB. (D) Gene ontology analysis. A gene ontology analysis was performed using GOseq (v1.12.0) (34) with the D. discoideum gene ontology annotations (dictybase.org/downloads/) to test for molecular function enrichment among all genes that were differentially expressed in the dgca− mutant and in the sets that were respectively up- or down-regulated. Molecular functions are shown, which were significantly enriched (P < 0.05) after correction for false discovery rates using the Benjamini–Hochberg procedure. Gene functions that are enriched in more than one set have identically color-coded descriptions.

The 2,281 genes that are down-regulated in the dgca− mutant are potential c-di-GMP target genes. However, only 390 genes are down-regulated more than fivefold, suggesting that the majority of genes responded pleiotropically to the altered environment of the early fruiting body, rather than being direct targets for c-di-GMP. Additionally, because spore differentiation depends on the stalk being formed first, this set also contains spore genes that are indirectly up-regulated by c-di-GMP. Comparison with earlier expression profiling of prestalk and prespore cells in migrating slugs (9) shows that of the 2,281 genes down-regulated in dgca−, 440 and 484 genes are more than threefold up-regulated in prestalk and prespore cells, respectively (Dataset S1, sheet 4). To isolate experimentally useful stalk marker genes, we first selected genes that were at least 10-fold down-regulated in dgca− cells and had a read count in control cells of >50 (Dataset S1, sheet 6). From this set, we isolated genes that were at least 10-fold overexpressed in prestalk over prespore cells and were strongly up-regulated at the onset of fruiting body formation, as determined earlier (9). We also selected some highly expressed down-regulated genes that had stalk-enriched orthologs in another Dictyostelium species, Polysphondylium pallidum. This left us with 25 well-expressed potential target genes for c-di-GMP (Fig. S1C and Dataset S1, sheet 7). This set also contained ecmB but not other DIF-1–regulated genes, such as ecmA, staA(pDd26), and staB (10, 11).

c-di-GMP Regulation and Expression Patterns of Putative c-di-GMP Target Genes.

The 5′ intergenic regions of eight putative c-di-GMP target genes were fused to lacZ and transformed into wild-type cells. Cells were developed on nitrocellulose filters and, at progressive stages of development, structures were stained with X-gal to visualize lacZ expression. Four genes, DDB_G0271196, abcG18 (ABC transporter G18), DDB_G0277757, and gluA (beta-glucosidase A), were exclusively or predominantly expressed in the stalk, with DDB_G0271196, further called staC, also being expressed late in the upper and lower cup of the mature spore head (Fig. 1). We used quantitative real-time PCR to establish whether the four genes were up-regulated by c-di-GMP in wild-type cells. Recent studies showed that c-di-GMP–induced stalk cell differentiation was strongly reduced in the stlb− and dmta− mutants that cannot synthesize DIF-1 (6, 12, 13), and we therefore also included these mutants in our analysis.

Fig. 1.

Expression patterns and c-di-GMP regulation of candidate c-di-GMP target genes. Expression patterns. Wild-type cells, transformed with lacZ fused to the promoters of eight candidate c-di-GMP target genes, were incubated at 22 °C and stained with X-gal at different developmental stages. ec, mc, and lc, early, mid, and late culmination; f, fruiting body; m, mound; s, slug. Arrows indicate amoeboid cup cells. (Scale bars, 100 µm.) c-di-GMP regulation. Wild-type, dmta−, and stlb− slugs were dissociated, resuspended in stalk salts, and incubated with 0, 1, 3, or 10 µM c-di-GMP for 6 h. Cells were harvested and total RNA was isolated and resuspended to 1 µg/µL. Quantitative (q)RT-PCR reactions were performed using primer sets that hybridized to the coding regions of the candidate genes. Normalized transcript levels are expressed as the percentage of fold-change up-regulation by 10 over 0 µM c-di-GMP (open symbols), with fold change at 10 µM shown as closed symbols. Means and SEM of three experiments performed in triplicate are presented. Gene identifiers are shown without DDB_ prefixes.

In wild-type cells, 1 to 3 µM c-di-GMP optimally induced stalk gene transcription, with transcription decreasing at 10 µM. However, DDB_G0277757 and gluA were only up-regulated five- and twofold, respectively. In the DIF-less mutants, up-regulation of staC and abcG18 only occurred effectively at 10 µM c-di-GMP, suggesting that DIF-1 induces responsiveness of cells to c-di-GMP. The marginally up-regulated staD and gluA genes were even less up-regulated by c-di-GMP in the absence of DIF. There was, however, no marked difference in fold up-regulation of stalk transcripts by 10 µM c-di-GMP between wild-type and DIF-less mutants.

A second set of four genes was expressed very late in fruiting body formation, when spore formation was already completed (Fig. 1). These genes were expressed in the upper and lower cup of the spore head but, unlike the elliptical spores, the cells expressing these genes retained the irregular amoeboid shape (Fig. 1, arrows). The genes all encode proteins of unknown function and were named cupA to D. cupC (DDB_G0282455) is also expressed very late throughout the stalk, whereas cupD (DDB_G0293854) showed scattered expression earlier in development and relatively high expression at the rear of migrating slugs. None of the genes were expressed in the basal disk. The cup genes were also inducible by c-di-GMP but required higher concentrations than the two stalk genes, although this was less pronounced for cupB (DDB_G0278537). The dmta− and stlb− mutants required even higher c-di-GMP concentrations for cup gene induction, and also tended to show lower fold up-regulation by c-di-GMP. Additional experiments using promoter–lacZ constructs showed that cup genes needed at least 100 µM c-di-GMP for optimal expression (Fig. S2). Because DgcA is not expressed in the cup or spore population (7), the requirement for such high c-di-GMP concentrations for cup gene induction in vitro could be artifactual, and not reflect regulation of cup genes in normal development.

Fig. S2.

Extended c-di-GMP dose–response curve. Wild-type AX2 cells transformed with either the abcG18::lacZ or cupA::lacZ construct were developed to the slug stage. Slugs were dissociated into single cells and incubated in stalk salts at 10⁶ cells per mL for 8 h with the indicated concentrations of c-di-GMP. Cells were lysed by three freeze–thaw cycles and incubated at 22 °C with CPRG and Z buffer until the samples with highest activity per transformed strain had reached an OD₅₇₄ of about 1.0. Data are expressed as the percentage of activity obtained at 100 µM c-di-GMP, and represent means and SD of two experiments performed in triplicate.

Components of the c-di-GMP Signal Transduction Pathway.

To identify components of the c-di-GMP signal transduction pathway, we first investigated mutants with a similar phenotype as the dgca− mutant. A mutant that expresses a dominant-negative inhibitor of PKA (PkaRm) from the ecmA prestalk promoter shows, like dgca−, prolonged slug migration, but eventually forms gnarled erect structures without stalk cells (14). We first tested whether c-di-GMP could restore proper fruiting body formation of the ecmA::PkaRm slugs. However, unlike dgca− slugs, ecmA::PkaRm slugs did not form fruiting bodies when exposed to c-di-GMP (Fig. 2A), suggesting that PKA either acts downstream of c-di-GMP or in parallel to the c-di-GMP pathway.

Fig. 2.

Effects of PKA inhibition and stimulation on c-di-GMP–induced gene expression. (A) EcmA::PkaRm rescue. Wild-type AX2 cells, transformed with the prestalk-specific PKA inhibitor ecmA::PkaRm, and dgca− cells were developed into migrating slugs, overlaid with either 10 µL of water (control) or 1 mM c-di-GMP, and photographed after 10 h. (Scale bars, 100 µm.) (B) PkaRm effects on gene induction. AX2 cells transformed with ecmA::PkaRm or an inactive construct, ecmA::PkaRc, were developed into migrating slugs, which were dissociated and incubated with the indicated c-di-GMP concentrations. After 8 h, RNA was isolated and the levels of two stalk- and three cup-specific transcripts were determined by qRT-PCR. Data are expressed as the percentage of transcript levels in the PkaRc controls treated with 10 µM c-di-GMP. Means and SEM of two experiments performed in triplicate are presented. Significant differences between gene induction in Rm and Rc cells, as determined by a rank-sum test, are indicated by *P < 0.05 and **P < 0.005. (C) 8Br-cAMP effects on gene induction. Dissociated dgca− slugs were incubated without additives or with 3 µM c-di-GMP or 10 mM 8Br-cAMP. After 6 h, RNA was isolated and levels of stalk- and cup-specific transcripts were determined by qRT-PCR. Data are expressed as fold change relative to untreated controls and represent means and SEM of two experiments performed in triplicate.

To investigate whether PKA acts downstream of c-di-GMP, we compared c-di-GMP induction of target gene expression in AX2 cells, transformed with either ecmA::PkaRm (which cannot bind to cAMP) or with a control ecmA::PkaRc construct, which can additionally not bind to PkaC (14). Fig. 2B shows that c-di-GMP induction of both the stalk and cup genes is 50 to 70% reduced in PkaRm cells, with repression being stronger for cup genes than stalk genes. The incomplete inhibition by ecmA::PkaRm is likely due to the fact that not all prestalk cells express ecmA, because we found that even at threefold higher G418 selection than the 100 µg/mL used previously (14), 30% of terminal structures still showed a very thin stalk (Fig. S3).

Fig. S3.

Inhibition of stalk formation by ecmA::PkaRm. Phenotype of ecmA::PkaRm culminants. Wild-type AX2 and AX2, transformed with ecmA::PkaRm and selected at 300 µg/mL G418, were developed until erect fruiting structures had formed. Structures were transferred to a slide glass, squashed with a coverslip, and photographed. About a third of PkaRm structures contained a thin stalk as shown in the image.

To validate PKA involvement, we tested whether the membrane-permeant PKA activator 8-bromo-cAMP (8Br-cAMP) could bypass the c-di-GMP requirement for target gene induction. Fig. 2C shows that the stalk genes were induced equally effectively by 3 µM c-di-GMP and 10 mM 8Br-cAMP in dgca− cells. Combinatorial stimulation with c-di-GMP and 8Br-cAMP only slightly improved stalk gene induction (Fig. S4), indicating 8Br-cAMP mimics rather than facilitates the effects of c-di-GMP.

Fig. S4.

Combinatorial effects of c-di-GMP and 8Br-cAMP on stalk gene induction. Wild-type AX2 cells, transformed with abcG18::lacZ, were developed to the slug stage. Slugs were dissociated into single cells and incubated in stalk salts at 10⁶ cells per mL for 6 h with 0 or 3 µM c-di-GMP, 10 or 30 mM 8Br-cAMP, or 3 µM c-di-GMP combined with 10 or 30 mM 8Br-cAMP. Cells were lysed and assayed for β-galactosidase activity. Data are expressed as the percentage of β-galactosidase activity induced by 3 µM c-di-GMP. Means and SEM of three experiments performed in triplicate are presented.

For cup genes, the induction by 8Br-cAMP was at least 10-fold more effective than induction by c-di-GMP. Combined with the inhibitory effects of PkaRm, the stimulatory effects of 8Br-cAMP indicate that PKA is likely to act downstream of c-di-GMP. However, the much greater effect of 8Br-cAMP than c-di-GMP on cup genes suggests that c-di-GMP is unlikely to be the primary signal for induction of cup genes.

Involvement of Adenylate Cyclases and RegA.

To investigate how c-di-GMP increases PKA activity, we first tested the effect of c-di-GMP on cellular cAMP levels. Cells from dissociated slugs were incubated with the cAMP phosphodiesterase (PDE) inhibitors DTT and 3-isobutyl-1-methylxanthine (IBMX) (15) in the presence and absence of 2 µM c-di-GMP. In slugs, cAMP is produced by three adenylate cyclases, ACA, ACG, and ACR (3, 16, 17). Addition of PDE inhibitors allows cAMP to accumulate to about 20 pmol per 10⁷ cells, whereas stimulation with c-di-GMP causes a further twofold increase (Fig. 3A), which persisted for as long as c-di-GMP was present (Fig. S5).

Fig. 3.

Effect of c-di-GMP on cAMP levels. (A) c-di-GMP effects on cAMP. Dissociated wild-type slugs were incubated at 10⁸ cells per mL with the PdsA/RegA inhibitors DTT and IBMX, 2 µM c-di-GMP, and 2 mM ACA inhibitor SQ22536 as indicated. At 0, 3, 10, 30, and 60 min, reactions were terminated and cAMP was assayed by isotope dilution assay. (B) rega−, acra−, and acga− mutants. Null mutants in the intracellular PDE RegA and the adenylate cyclases ACR and ACG were developed to migrating slugs, dissociated, and incubated with DTT/IBMX in the presence and absence of 2 µM c-di-GMP, followed by cAMP assay. (C and D) tsaca2 activation by c-di-GMP. aca−/tsaca2 (C) or wild-type (D) cells were developed to slugs at the permissive temperature (22 °C), followed by 40 min of incubation at either 22 or 28 °C. Structures were then dissociated and incubated at 22 or 28 °C with DTT/IBMX in the presence and absence of 2 µM c-di-GMP, and assayed for cAMP at the indicated time intervals.

Fig. S5.

Prolonged c-di-GMP effects and cycloheximide effects on cAMP levels. (A) Prolonged c-di-GMP effect. Dissociated wild-type slugs were incubated for 5 h at 10⁸ cells per mL with the PdsA/RegA inhibitors DTT and IBMX only or with additional 1 µM c-di-GMP, which was added either once at the start of the experiment or at 1-h intervals during the experiment. At the indicated time points, reactions were terminated and cAMP was assayed by isotope dilution assay. Means and SEM of two experiments performed in triplicate are presented. (B) Cycloheximide effects. Dissociated wild-type slugs were preincubated for 30 min in the presence and absence of 500 µg/mL cycloheximide (CHX), and each sample was subsequently incubated in the presence and absence of 1 µM c-di-GMP (with cycloheximide remaining present). At the indicated time points, reactions were terminated and cAMP was assayed by isotope dilution assay. Means and SEM of three experiments performed in triplicate are presented. The pretreatment with and presence of cycloheximide reduced overall adenylate cyclase activity but not its stimulation by c-di-GMP. This, and its rapid effect on cAMP levels, indicates that c-di-GMP activates ACA activity and not ACA protein synthesis.

Many signals that stimulate cell differentiation by activating PKA in Dictyostelium do so by indirectly inhibiting the intracellular cAMP phosphodiesterase RegA (18). Null mutants in either regA, acgA, or acrA develop to the slug stage, but acaA null cells cannot aggregate and therefore do not acquire competence for c-di-GMP. To investigate whether the c-di-GMP–induced cAMP increase was due to either inhibition of RegA or stimulation of ACG or ACR, we measured the effect of c-di-GMP on cAMP levels in rega−, acga−, or acra− mutants. Fig. 3B shows that all three mutants showed c-di-GMP–induced cAMP accumulation similar to wild-type cells, indicating that neither RegA, ACG, nor ACR mediates the effect of c-di-GMP. To test involvement of ACA, we used an adenylate cyclase inhibitor, SQ22536, which inhibits ACA in vivo, but not ACG or ACR (15). Fig. 3A shows that SQ22536 almost completely inhibits c-di-GMP–induced cAMP accumulation, suggesting that c-di-GMP activates ACA. SQ22536 also inhibits c-di-GMP–induced stalk gene expression, as measured in abcG18::lacZ and staC::lacZ transformed cells (Fig. 4A), supporting evidence that ACA and PKA mediate c-di-GMP induction of stalk cell differentiation.

Fig. 4.

Effect of ACA inhibition on stalk gene induction. (A) SQ22536 effects on stalk genes. dgca− cells transformed with either staC::lacZ or abcG18::lacZ were developed to slugs, dissociated, and incubated with 3 µM c-di-GMP in the presence and absence of 2 mM SQ22536. β-Galactosidase activity was measured after 6 h. Data are expressed as the percentage of activity obtained with only 3 µM c-di-GMP. (B and C) Gene expression in aca−/tsaca2 and wild-type. aca−/tsaca2 (B) or wild-type AX2 (C) slugs, developed at 22 °C, were dissociated and incubated in the presence and absence of 3 µM c-di-GMP at either 22 or 28 °C for 6 h. RNAs were isolated and abcG18 and staC transcript levels were measured by qRT-PCR. Means and SEM of three (A and B) and two (C) experiments performed in triplicate are presented.

Because pleiotropic effects of SQ22536 cannot be excluded, we used an aca− mutant transformed with a temperature-sensitive ACA variant (tsaca2) to validate involvement of ACA. This mutant develops normally at 22 °C but stops development when transferred to 28 °C, a temperature where the wild type still forms normal fruiting bodies (19). After development of aca−/tsaca2 cells into slugs at 22 °C, the cells showed normal c-di-GMP–induced cAMP accumulation at 22 °C. However, at the nonpermissive temperature of 28 °C, c-di-GMP–induced cAMP accumulation was much reduced in the aca−/tsaca2 mutant (Fig. 3C) but not in wild-type cells (Fig. 3D). c-di-GMP induction of stalk gene expression was also strongly inhibited at 28 °C in aca−/tsaca2 (Fig. 4B). This was not due to direct inhibition of gene expression by the higher temperature, because in wild-type, c-di-GMP–induced stalk gene expression was 1.5- to 2-fold higher at 28 °C than at 22 °C (Fig. 4C). Neither DIF-1 nor higher c-di-GMP concentrations could restore stalk gene expression in the aca−/tsaca2 mutant at 28 °C (Fig. S6). When combined, the results obtained with the ACA inhibitor and the aca−/tsaca2 mutant show that ACA mediates c-di-GMP–induced stalk gene expression.

Fig. S6.

Effects of DIF and 10 µM c-di-GMP on stalk gene expression in aca−/tsaca2 cells. aca−/tsaca2 slugs, developed at 22 °C, were dissociated and incubated for 6 h at 22 or 28 °C with 0, 3, or 10 µM c-di-GMP or a combination of 3 µM c-di-GMP and 100 nM DIF, as indicated. RNAs were isolated and abcG18 and staC transcript levels were measured by qRT-PCR. Data represent means and SEM of two experiments performed in triplicate.

Increasing cAMP or activating PKA in prestalk cells blocks slug and fruiting body formation due to interference with chemotactic signaling (20–22). To prove nevertheless that c-di-GMP effects in intact slugs are solely mediated by cAMP, we fused the cyanobacterial adenylate cyclase mPAC, which is activated by light also when expressed in Dictyostelium (23), to the ACA promoter that directs tip-specific expression (24) (Fig. 5C). Fig. 5A shows that in darkness the dgca−/ACAt::mPAC slugs continued to migrate but that light triggered fruiting body formation and stalk cell differentiation (Fig. 5B) within 5 h (Movie S1). dgca− did not form fruiting bodies in darkness or in light (Fig. 5A), and an A15::mPAC construct (23) was much less effective (Fig. S7). When combined, our experiments conclusively show that ACA mediates c-di-GMP–induced stalk formation.

Fig. 5.

Rescue of dgca− culmination by a light-activated adenylate cyclase. (A) Two plates each of dgca− and dgca−/ACAt::mPAC cells were developed for 16 h in darkness until slugs had formed. Subsequently, one plate of each was incubated for 20 h in darkness and the other under ambient room lighting. The experiment was repeated twice with the same result. (Scale bar, 200 µm.) (B) Segment of the stalk of a light-exposed dgca−/ACAt::mPAC structure stained with the cellulose dye calcofluor. (Scale bar, 10 µm.) (C) dgca− cells, transformed with an ACAt::lacZ fusion construct, were developed to slugs and stained with X-gal to confirm the tip specificity of expression from the ACAt promoter. (Scale bar, 100 μm.) (D) Model for induction of stalk formation. DgcA is expressed throughout the anterior quarter of the slug (7) but its target ACA is only expressed at the tip. c-di-GMP therefore only induces stalk formation at the tip by activating ACA and thereby PKA.

Fig. S7.

Effect of transformation of A15::mPAC on dgca− development. The dgca− mutant was transformed with a fusion construct of the constitutive Dictyostelium actin 15 promoter and the light-activated adenylate cyclase mPAC from Microcoleus chthonoplastes. Two plates each of dgca− and dgca−/A15::mPAC cells were developed for 16 h in darkness until slugs had formed. Subsequently, one plate of each was incubated for up to 48 h in darkness and the other under ambient room lighting. Developing structures were photographed after a total incubation period of 40 or 64 h. A few fruiting bodies that were formed in light-treated dgca−/A15::mPAC cells at 64 h are indicated by arrows. (Scale bar, 1 mm.)

Discussion

Transcription Profiling Identified Novel Stalk Genes and Cup Genes.

Comparison of the transcription profiles of wild-type and dgca− mutants yielded 2,281 genes that were down-regulated in dgca− but only 390 genes that were down-regulated more than fivefold. Their stalk defect prevents dgca− cells from entering culmination, and the question arises as to whether the dgca−/wild-type comparison merely identifies culmination-specific genes. However, there is only a weakly supported positive correlation between genes up-regulated in wild type over dgca− on one hand and genes up-regulated in culmination over the slug stage on the other (Fig. S8), indicating that DgcA affects a specific gene set. Comparison of cells stimulated in vitro with and without c-di-GMP is another approach to identify c-di-GMP target genes. However, other signals, such as DIF-1, induce genes such as ecmA and pdsA, and even stalk cell differentiation in vitro (5, 25, 26), without being required for these processes in vivo (6, 26). Our approach to identify DgcA-dependent genes in vivo and then test their expression pattern and c-di-GMP induction in vitro is more likely to yield developmentally relevant DgcA targets. DIF-1 is required for formation of the basal disk (6, 13) and, because basal disk and stalk cells are phenotypically identical, the cells induced by DIF-1 in vitro are most likely basal disk cells.

Fig. S8.

Correlation between dgcA-dependent genes and early culmination genes. For all genes that were significantly up-regulated in control random integrants compared with dgca knockouts, up-regulation was expressed as the sum of normalized reads in the four RI replicates divided by the sum of normalized reads in the four RI and four KO replicates (the result must be >0.5 for up-regulated genes; Dataset S1, sheet 8). For the same genes, normalized reads at 16 and 20 h were retrieved from a duplicate developmental time course of wild-type Dictyostelium cells, which were at those time points at the migrating slug and early-to-midculmination stages, respectively (10). Up-regulation in culminants over slugs was similarly calculated, and plotted against up-regulation in random integrants. Linear regression indicates a weakly supported positive correlation between temporal and dgcA-induced up-regulation. However, many genes that are highly up-regulated in RI over KO cells show poor up-regulation or even down-regulation in culminants over slugs and vice versa. This indicates that selection of putative c-di-GMP target genes on the basis of up-regulation in culmination would yield many false positives. The stalk (red) and cup genes (green) that were used in this study and selected using a combination of criteria (main text) are highly up-regulated under both experimental conditions.

Recent studies found strongly reduced c-di-GMP induction of the vacuolated stalk/basal disk cell type in stlb− and dmta− mutants in vitro (12). We show that stalk gene transcription in stlb− and dmta− mutants in vitro required 10-fold higher c-di-GMP concentrations than in wild-type cells (Fig. 1), suggesting that DIF-1 promotes responsiveness to c-di-GMP. In normal development, stalks are somewhat weaker in stlb− and dmta− mutants (6), suggesting that sufficiently high c-di-GMP concentrations are produced to counteract most of the reduced responsiveness.

In addition to stalk genes, the set of prestalk-enriched genes that was down-regulated in dgca− also contained genes that were expressed very late in the upper and lower cup of the spore head. The cells expressing these genes remained amoeboid, and likely serve to anchor the spore mass to the stalk. The upper and lower cup cells, identified here, differ from those identified by the expression patterns of the prestalk genes ecmA and ecmB, which appear much earlier when the stalk is just being formed, and for ecmA already in slugs (10). The promoter region that directs ecmB lower cup expression also directs expression in the basal disk, whereas the upper cup population largely becomes part of the stalk (27). The cup genes identified in our study are not expressed in the basal disk, and the amoebas expressing these genes move to the upper cup position too late to be incorporated in the stalk.

ACA Acting on PKA Mediates c-di-GMP–Induced Stalk Gene Expression.

c-di-GMP induction of stalk gene expression is reduced in mutants in which PKA activity is down-regulated, whereas PKA activation with 8Br-cAMP is equally effective as c-di-GMP in inducing stalk gene expression (Fig. 2 and Fig. S4). This suggests that the effects of c-di-GMP are mediated by PKA, which was further confirmed by experiments showing that c-di-GMP induces a prolonged elevation of cAMP levels in slug cells (Fig. 3). The c-di-GMP–induced cAMP increase occurred normally in acra−, acga−, and rega− mutants but was prevented by the ACA inhibitor SQ22536 and by down-regulation of ACA in a temperature-sensitive ACA mutant (Fig. 3). SQ22536 and ACA down-regulation also prevented c-di-GMP–induced stalk gene expression (Fig. 4). These data show that c-di-GMP induces stalk gene expression in cell suspension by triggering cAMP synthesis by ACA, which in turn activates PKA.

Stalk formation of the stalk-less dgca− mutant was restored by heterologous expression of a light-activated adenylate cyclase from a tip-specific promoter in slugs following exposure to light. This indicates that also in vivo cAMP synthesis, normally activated by c-di-GMP, triggers stalk formation. dgcA has a fairly broad expression domain throughout the anterior 30% of the slug (7), whereas ACA is highly expressed at just the tip 5%, where stalk formation initiates (17, 24). This suggests a simple model (Fig. 5D) for induction of stalk formation, in which c-di-GMP, although present at the entire prestalk region, can only activate PKA and thereby stalk cell differentiation at the slug tip, where ACA is expressed. The interaction of c-di-GMP with ACA therefore explains why stalk formation always initiates at the tip.

Methods

Cell Culture.

D. discoideum cells were grown in HL5 medium (Formedium), supplemented with 10 µg/mL blasticidin for knockout mutants, 20 µg/mL G418 for aca−/tsaca2 mutants (19) and cells transformed with ACAt::mPAC-YFP or lacZ constructs, and 300 µg/mL G418 for cells transformed with ecmA::PkaRm or ecmA::PkaRc (14).

High-Throughput RNA Sequencing.

RNA isolation, cDNA library construction, and high-throughput sequencing.

Two dgca knockout clones and two clones harboring random integrations of the dgcA knockout construct were plated at 10⁶ cells per cm² on nonnutrient agar and incubated at 22 °C until the random integrant clones had reached the early culminant stage. Total RNA was isolated from two separate experiments and enriched for mRNA using poly-T–linked magnetic beads. Bar-coded cDNA libraries were constructed using the Illumina TruSeq RNA v2 protocol and checked for quality using the Agilent TapeStation DNA D1K Kit. The eight bar-coded libraries were normalized to 10 nM and combined into one pool, which was sequenced as paired-end 100-bp reads using the Illumina HiSeq 2000 platform. Across the pool, 38 to 63 million reads were generated per library (Dataset S1, sheet 1). The raw data are available from the European Nucleotide Archive (project code PRJEB10384, study ID code E-MTAB-3829).

Differential gene expression analysis.

Sequencing data quality was checked with FastQC (bioinformatics.babraham.ac.uk/projects/fastqc/) and reads were aligned to the D. discoideum genome (28) using TopHat v2.0.9, with parameters set at an average Dictyostelium intron length of 146 bp (29). Reads had an alignment rate of 90%, and read counts were generated with HTSeq-count v0.5.4p5 (30). Poorly expressed genes (expression <1 count per million in >4 samples) were filtered out, and differentially expressed genes were determined with edgeR v3.2.0 (31) using the exact test algorithm following TMM (trimmed mean of M values) normalization. All genes with a false discovery rate (FDR) of <0.05 are listed as differentially expressed genes in Dataset S1, sheet 3.

DNA Constructs.

To study expression patterns of putative c-di-GMP target genes, complete 5′ intergenic regions were amplified by PCR from D. discoideum genomic DNA using the primers listed in Table S1, which include XbaI and BglII sites in the forward and reverse primers, respectively. Amplicons were digested with XbaI and BglII and cloned into the XbaI/BglII-digested plasmid pDdGal17 (32). A fragment of 1,258 bp corresponding to acaA promoter 3 that directs tip-specific expression (33) was amplified using the primer pair ACAtF_L and ACAtR_L with XbaI and BglII restriction sites for cloning into pDdGal17 and primer pair ACAtF_mP and ACAtR_mP with SalI and HindIII sites for replacing the actin15 promoter in act15::mPAC-YFP (23). The amplicons were ligated into the plasmids after digestion with the appropriate enzymes. Constructs were verified by DNA sequencing and transformed into AX2 wild-type or dgca− cells.

Table S1.

Oligonucleotide primers used for cloning

Name	DNA sequence, 5′–3′
G0276063prF	AAATCTAGATTGTAAACCATCATCAAAGGA
G0276063prR	AATAGATCTGAAACTTACTTAAAATTGACA
G0282455prF	AAATCTAGAATGTTGTAGCAAACTACATTA
G0282455prR	AATAGATCTAAGTAATGAGAATATCACTA
G0293854prF	ATTTCTAGATGATCGCAAGTTCGATCCCTGTT
G0293854prR	ATTAGATCTTTTTTTGAAACTTGTATATTTTGATATATG
G0278537prF	AAATCTAGAACATCTTGTTCTTTTTCAAGCT
G0278537prR	ATTAGATCTTCTTGAATATTGGCTGATGTCAATG
G0275687prF	ATTTCTAGAAATTAAAAAATCTTTTTTATTGTGTGT
G0275687prR	ATTAGATCTAATATTCTCTTTTTTCTTCATTTTCCAT
G0292810prF	AAATCTAGAATATCTAAAATCAGATGGTCCTTA
G0292810prR	ATTAGATCTCTTTTTATATATATATATAAAATTGAATTTAATTAAATATTC
G0277757prF	ATATCTAGAGTCACAACCATAATAATATTTATATAATTATCA
G0277757prR	ATTAGATCTTTTTTTTTTTTTTTTTTATTATTATTTAAAAAATAATTATTGTCT
G0271196prF	AAATCTAGAATTCAGCGACTCTATTATGGTTTT
G0271196prR	AAAAGATCTCAATTGATGATATTCTTCAGCCATTC
ACAtF_L	GGGGTCGACCCTCACTTCATAAATATATCTTTG
ACAtR_L	CCCAAGCTTCTTTTTTTTTTTGTGATTATTATTATTAC
ACAtF_mP	GCTCTAGACCTCACTTCATAAATATATCTTTG
ACAtR_mP	GAAGATCTCTTTTTTTTTTTGTGATTATTATTATTAC

F, forward; R, reverse.

β-Galactosidase Assays.

Histochemical assay.

Wild-type cells transformed with promoter–lacZ constructs were plated at 4 × 10⁶ cells per cm² on nitrocellulose filters, supported by nonnutrient agar, and incubated at 22 °C until the desired developmental stages had been reached. Structures were fixed with 0.25% glutaraldehyde and stained with X-gal (34). Different developmental stages from cells transformed with the same constructs were stained for equally long periods, but for different constructs this could vary from 15 min to 24 h.

Spectrophotometric assay.

Cells transformed with promoter–lacZ constructs were developed into slugs, harvested, and resuspended to 2 × 10⁶ cells per mL and incubated as 90-µL aliquots on microtiter plates. Plates were shaken at 170 rpm at 22 °C for 6 to 8 h. Cells were lysed by freeze–thawing, and incubated at 22 °C with 30 μL 2.5× Z buffer (5 mM KCl, 1.5 mM MgCl₂, and 0.2 mercaptoethanol in 50 mM Na-phosphate, pH 7.0) and 10 μL 40 mM CPRG (chlorophenol red-β-d-galactopyranoside). The OD₅₇₄ was measured at regular time intervals using a microtiter plate reader.

Quantitative Transcript Analysis by Quantitative RT-PCR.

RNA was isolated from 10⁷ cells using the RNeasy Mini Kit (Qiagen), and DNA contamination was removed using the Turbo DNA-free Kit (Ambion). RNA was transcribed into cDNA using the SensiFAST cDNA Synthesis Kit (Bioline), and quantitative real-time PCR was performed using iQ SYBR Green Supermix (Bio-Rad) on a RealPlex2 thermal cycler (Eppendorf) and the oligonucleotide primers listed in Table S2. Gene expression levels were normalized to the expression level of the constitutively expressed genes DDB_G0282429 and/or DDB_G0280765 (9) in the same sample.

Table S2.

Oligonucleotide primers used for qRT-PCR

Name	DNA sequence, 5′–3′
G0282429F	GATTATGGTTGGCCTGAAGGT
G0282429R	CAACACCAGTTACCCAAGGA
G0280765F	GTTGGTTCCGCCAAACAAGATC
G0280765R	ACAACTATGTCTACAAGGTAC
G0275687F	GCAACATCATTACCAGTTGAAG
G0275687R	CAATGGATAAGAGATGAAATCC
G0278537F	CGTGATAACTCTGATACATCTCT
G0278537R	GGAGGTTTGGAACAAATGCACC
G0276063F	GAAATTGCCTCAAGTTCATTAC
G0276063R	ACACCACTAGCAACACCAACT
G0282455F	CTGTGTCTCCAAGTGTGCTTT
G0282455R	ATGATCTGGACATCCCTTTAC
G0293854F	TCGAACCTTATCAGTACTACG
G0293854R	CTTTTTCGTAACCGCAATCCT
G0277757F	TGGTTACTATGGTCTTCCAGA
G0277757R	TTGGTTAGACGATCTTTGTCA
G0292810F	ATCTTTGGGCTCAAGTTACTC
G0292810R	TATCGACACCAATAAAGGAGA
G0271196F	ATGGCTGAAGAATATCATCAATTG
G0271196R	TGGGCCTTCAATATCGTAATG

cAMP Measurements.

Slugs were dissociated by repeated pipetting and resuspended to 10⁸ cells per mL. Twenty-five-microliter aliquots of cells were mixed with 5 µL of variables (25 mM DTT, 5 mM IBMX, 12 µM c-di-GMP, or 12 mM SQ22536, added either separately or in various combinations). The mixtures were shaken for 0 to 60 min at 22 °C and 170 rpm. Reactions were terminated with 30 µL 3.5% (vol/vol) perchloric acid. Samples were neutralized with 15 µL 50% (wt/vol) saturated KHCO₃ and 30 µL cAMP assay buffer (4 mM EDTA in 150 mM K-phosphate, pH 7.5). Lysates were centrifuged for 15 min at 1,000 × g, and cAMP was measured in 50 µL of supernatant by isotope dilution assay (15).