A Conserved Signalling Pathway for Amoebozoan Encystation that was Co-Opted for Multicellular Development

The evolution of multicellularity required novel mechanisms for intercellular communication, but their origin is unclear. Dictyostelium cells exchange signals to position specialized cell types in multicellular spore-bearing structures. These signals activate complex pathways that converge on activation of cAMP-dependent protein kinase (PKA). Genes controlling PKA were detected in the Dictyostelid unicellular ancestors, which like most protists form dormant cysts when experiencing environmental stress. We deleted PKA and the adenylate cyclases AcrA and AcgA, which synthesize cAMP for PKA activation, in the intermediate species Polysphondylium, which can develop into either cysts or into multicellular structures. Loss of PKA prevented multicellular development, but also completely blocked encystation. Loss of AcrA and AcgA, both essential for sporulation in Dictyostelium, did not affect Polysphondylium sporulation, but prevented encystation. We conclude that multicellular cAMP signalling was co-opted from PKA regulation of protist encystation with progressive refunctionalization of pathway components.

Polysphondylium pallidum, is like D.discoideum amenable to both forward and reverse genetic approaches. In addition to the D. discoideum genome 9 , the genomes of P.pallidum, residing in group 2, D. fasciculatum, residing in group 1 and D. lacteum, residing in group 3 were recently sequenced by ourselves and colleagues 10 (Schaap, P. and Gloeckner, G. in preparation). Genome sequence of the strictly unicellular amoebozoan Acanthamoeba castellani is also available 11 . This information allows us to retrace conservation and change in known developmental signalling genes. We have developed procedures for successive disruption of multiple genes in Polysphondylium pallidum, which allows us to assess gene function in both unicellular and multicellular development 12 . In this work we test the hypothesis that multicellular sporulation is evolutionary derived from unicellular encystation by investigating the roles of the catalytic subunit of PKA (PkaC) and the adenylate cyclases ACG and ACR in the uniand multicellular life cycles of P. pallidum.

Results
Conservation of PkaC, ACG and ACR in Amoebozoa. Social amoebas can be subdivided into four major groups, which together are members of Amoebozoa, a kingdom of mainly unicellular amoebas 8 . Genomes, representative of the four groups and three genomes of unicellular Amoebozoa have recently become available [9][10][11]13 (http://sacgb.fli-leibniz.de/; http://www.physarum-blast.ovgu.de/). We investigated the presence of genes encoding ACR, ACG and PkaC, the catalytic subunit of PKA in these genomes. BLAST queries detected homologs of PkaC and ACR in all Amoebozoan genomes, except that of the obligatory parasite Entamoeba histolytica, while ACG was only conserved in the Dictyostelid genomes ( Figure 1). Because the PkaC sequence is similar to that of many other protein kinases, we also included the closest homologs of PkaC (PkgB and PkgD) in our query. Phylogenetic inference showed that our initally selected PkaC orthologs grouped together with D. discoideum PkaC, while hits to the PKA homologs grouped with either PkgB or PkgD (Supplementary Figure 1).
All ACG proteins consist of a CHASE domain flanked by one or two transmembrane domains and a cytosolic adenylate cyclase domain ( Figure 1). The ACR homologs also display a similar functional domain architecture as D. discoideum ACR with 6-7 trans-membrane domains, a HATPase-c domain, two receiver domains and a cyclase catalytic domain. The HisKA autophosphorylation/ dimerization domain that is usually located at the N-terminus of the HATPase-c domain is missing in the Amoebozoan ACRs. The closest homologs to ACG or ACR are prokaryote ACs, while a fungal PkaC is closest to the amoebozoan PkaCs.

Disruption of P.pallidum PkaC by homologous recombination.
To assess the function of PkaC in uni-and multicellular development, we disrupted P. pallidum PkaC by replacing a PkaC internal segment with a floxed G418 cassette 14 (Supplementary Figure 2). The pkacamoebas and a control random integrant (RI) proliferated normally, but completely lost the ability to aggregate and form fruiting bodies. After 48 h of starvation, when RI cells had formed fruiting bodies, the lawn of starving pkaccells had somewhat contracted, but aggregates were never formed ( Figure 2A). Strikingly, unlike RI controls, the pkacamoebas also could not form cysts when starved at high osmolarity ( Figure 2B), a condition that triggers encystation of unicellular amoebas 15 .
After removal of the G418 selection cassette by transformation with cre-recombinase, the pkaccells were transformed with the PkaC coding sequence fused to its own promoter. The resulting pkac-/PkaC cells fully regained both development into fruiting bodies and encapsulation of single amoebas into cysts ( Figs. 2A,B), thus demonstrating that PkaC is essential for both multicellular development and encystation.
Disruption of AcgA and AcrA in P.pallidum. ACG has an overlapping role with ACR in prespore differentiation in D. discoideum and mediates inhibition of spore germination by high ambient osmolarity in the spore head 2,4,16 . To investigate whether ACG has essential roles in P. pallidum multicellular development and encystation, we disrupted the ACG gene AcgA in P. pallidum (Supplementary figure 3). The acgamutant formed normal aggregates and fruiting bodies ( Figure 3A). Surprisingly, the fruiting bodies contained viable detergent resistant spores, that showed normal inhibition of spore germination by high osmolarity ( Figure 4D). The acgamutant also showed normal encystation ( Figure 3C).   ACR is essential for spore maturation in D. discoideum 3 and to assess its role in P. pallidum, we next ablated the P. pallidum AcrA gene (Supplementary Figure 4). Surprisingly, unlike D. discoideum acra-, the P. pallidum acramutant displayed normal spore formation in fruiting bodes ( Figure 3A,B). The acramutants also encysted when submerged at high osmolarity, but encystation was less efficient than in wild-type P. pallidum (Figs. 3B, 4F). To assess whether ACG and ACR are functionally redundant, we created a double acraacgamutant. P. pallidum acracells were transformed with crerecombinase to excise the neomycin selection cassette and G418 sensitive clones were selected, which were then transformed with the AcgA knock-out construct. Strikingly, the acra-acgamutant was also not defective in fruiting body morphology or sporulation (Fig. 4A,B). However, the ability to encyst was completely lost (Fig. 4C). To validate that this was due to loss of cAMP synthesis, the neomycin cassette was excised from the acra-acgamutant once more, and cells were transformed with a plasmid that contained a P. pallidum genomic fragment, encompassing the full length AcrA coding and 59intergenic region. The resulting acra-acga-/AcrA mutant regained the ability to encyst (Fig. 4C), confirming that lack of encystation was due to the combined loss of AcgA and AcrA.
Roles for ACG and ACR in regulation of encystation and spore germination by osmolytes. When comparing P. pallidum to D. discoideum, a striking difference is the absence of an obvious role for ACR or ACG in P. pallidum spore differentiation. It is also remarkable that ACR can complement ACG in osmolyte-induced encystation, a role more suited for ACG with its intramolecular osmosensor 17 . In D. discoideum, high osmolarity prevents spores from germinating prematurely, while still in the sorus. We first investigated whether ACG or ACR still regulated spore germination in P. pallidum. Mature spores from wild-type and mutant fruit-ing bodies were incubated with detergent to lyse unencapsulated cells, and plated clonally on E.coli lawns in the presence and absence of 0.4 M of the osmolyte sorbitol. In the absence of osmolyte, wild-type, acra-, acga-, acra-acgaand acra-acga-/AcrA spores germinated with equal efficiency ( Figure 4D). In the presence of osmolyte, germination of wild-type, acgaand acra-acga-/AcrA spores was about 85% reduced and of acraspores about 30%. Only acra-acgaspores still germinated at full efficiency. These data show that ACG and ACR both mediate inhibition of spore germination by high osmolarity, with ACR playing the more dominant role.
Cyst germination on E.coli lawns is only 30% reduced by 0.4 M sorbitol in wild-type P. pallidum and by about 50% in a random integrant of the AcrA KO construct (Fig. 4E). Inhibition of cyst germination by osmolyte was completely lost in the acramutant, but not in the acgamutant. Remarkably, cyst germination was completely inhibited when acracells were complemented with AcrA expressed from its own promoter. This could be a consequence of AcrA overexpression, caused by integration of multiple copies of the expression vector in the P. pallidum genome.
A time course of the progression of encystation in wild type P. pallidum and all adenylate cyclase mutants showed that 60% of wildtype and RI control amoebas have encysted after 72 h of incubation with osmolyte (Fig. 4F). The acra-acgacells have completely lost encystation, while acracells encyst more slowly. Unexpectedly, both the acgacells and the acra-acga-/AcrA cells reached 70-80% encystation within 24 hours. Because acra-acga-/AcrA also showed more efficient inhibition of cyst germination by high osmolarity, this may simply imply that the increased ACR levels facilitated encystation. Why ACG should be required for encystation in an acrabackground, but simultaneously reduce encystation in a wild-type background is less clear. It suggests a scenario in which ACR is the main player, with its expression down-regulated by ACG. Negative regulation of AcrA expression by ACG was previously observed in D. discoideum slugs 2 .  Discussion ACR and ACG have overlapping roles in inducing prespore differentiation and spore maturation in D.discoideum, but appeared to be dispensible for the same processes in P.pallidum spore differentiation. However, the P. pallidum genome contains three copies of the gene encoding the adenylate cyclase ACA 10 , which is mainly involved in chemotactic signalling in D.discoideum 7 . One of the Ppal copies, ACA2, is expressed in prespore cells (Kawabe, Y., in preparation) and likely provides cAMP for spore differentiation.
The most striking outcome of this work is that deletion of PkaC and combined deletion of AcrA and AcgA completely blocks the process of encystation in P. pallidum. PkaC and ACR, but not ACG, are conserved in other Amoebozoa, such as Acanthamoeba castellani and Physarum polycephalum (Figure 1). Other compo-nents of the cAMP signalling pathway such as PkaR and the cAMP phosphodiesterase RegA are also conserved in these amoebozoan genomes and, remarkably, also in the amoeboflagellate Naegleria gruberi, which resides in the kingdom Excavata 11,18,19 .
RegA plays a central role in regulating cAMP levels during D. discoideum multicellular development, where its activity is controlled by intercellular signalling. The signals bind to sensor histidine kinases/phosphatases (SHKPs), which regulate the phosphorylation state of the intrinsic response regulator of RegA and thereby the attached cAMP phosphodiesterase activity 7 . The genomes of Naegleria and all sequenced Amoebozoa contain a large number of SHKPs. In non-dictyostelid amoebas, the roles of SHKPs have not yet been explored, but in D. discoideum their main target is RegA. About 7 out of the 15 Dictyostelium SHKPs mediate effects of secreted Loss of RegA causes precocious encystation in P. pallidum and A. castellani 19 , as expected for a protein that inhibits PKA function. Combined with results obtained in this work, this allows us to propose a universal mechanism for control of amoebozoan and excavate encystation ( Figure 5). In this scheme, cAMP activation of PKA induces encystation and prevents excystation. cAMP is synthesized by ACR in Amoebozoa, while in both Amoebozoa and Excavata, cAMP is hydrolysed by RegA. RegA activity is under both positive and negative regulation from external stimuli that activate a sensor histidine kinase or a sensor histidine phosphatase, respectively. For the former, such stimuli could be stress factors that cause encystation and for the latter, the stimuli could signal the presence of food to cause cyst germination.
Researchers have been intrigued for many years by the fact that PkaC regulates so many aspects of D. discoideum development, starting with the transition from growth to development 20 , the regulation of chemotactic signalling 21 , the differentiation of prespore cells 22 , the maturation of spores and stalk cells [23][24][25] and the germination of spores 4 . We show here that these roles are likely to have emerged from an original role of PKA in mediating stress-induced encystation and stress-inhibited excystation. The trigger for multicellular development is also nutrient stress and its end-point is the formation of two encapsulated cell types, the stalk cells and spores. It therefore appears that PKA retained its ancestral role in encapsulation, when Dictyostelia started to adopt a multicellular survival strategy, but additionally acquired novel roles to assist the organism to proceed through its multicellular differentiation programme in a wellregulated manner. These novel roles required novel mechanisms to regulate cellular cAMP levels. The differences in ACG, ACR and ACA function that we noted in this work between the group 2 species P.pallidum and the group 4 species D.discoideum may reflect different evolutionary trajectories towards effective cell signalling between these groups.
In conclusion, we propose that the range and complexity of the cAMP-mediated signalling cascades that control almost all stages of multicellular development in modern Dictyostelia have been coopted from an ancestrol role for cAMP in stress-induced dormancy. The emergence of complex cell-cell communication from basic environmental sensing in Dictyostelia may prove to be a paradigm for the evolution of multicellularity in other eukaryote lineages.

Methods
Cell growth, development, encystation and sporulation. Growth and development. Polysphondylium pallidum strain PN500, was routinely grown in association with Escherichia coli on lactose peptone (LP) agar. For multicellular development, P. pallidum cells were harvested in 20 mM KH 2 PO 4 /K 2 HPO 4 , pH 6.5 (KK2), washed free from bacteria and incubated on non-nutrient (NN) agar at 10 6 cells/cm 2 and 22uC.
Encystation. For quantification of encystation, P. pallidum cells were grown in a suspension of autoclaved Klebsiella aerogenes in KK2 , until cell density reached 5 3 10 6 -10 7 cells. Cells were washed, resuspend in KK2 with 0.4 M sorbitol at 10 7 cells/ml and shaken at 180 rpm and 21uC. Aliquots of 0.1 ml were sampled at regular intervals and supplemented with 1 ml 0.1% Calcofluor (which reacts to cellulose in the cyst wall). Total amoeba and cyst numbers were determined by counting cells in a haemocytometer under phase contrast and UV illumination, respectively. 100-500 cells were counted for each time point.
Cyst and spore germination. P. pallidum spores were harvested from 5-day old fruiting bodies. For cysts, P. pallidum cells were grown in K.aerogenes suspension as described above, washed and incubated with 0.4 M sorbitol for 3 days to allow mature cysts to form. Cysts or spores were treated for 10 min with 0.1% Triton-X100 to lyse amoeboid cells, washed with KK2, counted and clonally plated on LP with E. coli. Emerging P. pallidum plaques were counted after 4-6 days.
Gene disruption and complementation. Disruption of P. pallidum PkaC. For disruption of PkaC, two fragments of P. pallidum gDNA were amplified using primer pairs PKACI59/PKACI39 and PKACII59/PKACII39 (Supplementary table 1). The fragments were digested with XbaI/BamHI or XhoI/KpnI, using restriction sites incorporated in primer design, and sequentially inserted into inserted into XbaI/ BamHI and XhoI/KpnI digested plasmid pLox-NeoIII 14 to flank the LoxP-neo selection cassette (Supplementary figure 2A). P. pallidum cells were transformed by electroporation with the XbaI/KpnI vector insert, followed by selection of transformants on autoclaved K.aerogenes at 300 mg/ml G418 26 . Genomic DNA was isolated from G418 resistant clones and screened by 2 PCR reactions (Supplementary figure 2B) and Southern blot (Supplementary figure 2C) to diagnose PkaC disruption by homologous recombination. Two knock-out (KO) and two random integrant (RI) clones were identified.
Complementation of pkac-with PkaC. To remove the A6neo cassette, KO (pkac-) cells were transformed with pA15NLS.Cre 27 for transient expression of Cre recombinase. Transformed clones were replica-plated onto autoclaved K.aerogenes on LP agar with and without 200 mg/ml G418 for negative selection. To generate a PkaC expression construct, the 0.8 kb PkaC 59intergenic region was amplified from gDNA using primers P-PKAC p3 and P-PKAC P4r (Supplementary table 1), which contain NheI and BglII restriction sites, respectively, and ligated into NheI/BglII digested vector pDdCGFP 28 . Next, the PkaC coding sequence was amplified using primers palPKAC p9 and palPKAC p11r (Supplementary table 1), which harbour BglII and XbaI sites, respectively, and ligated into the BamHI and XbaI sites of the newly created vector. This fuses PkaC N-terminally to its own 59intergenic region and C-terminally to green fluorescent protein (GFP)(see Supplementary figure 2A). The construct was transformed into pkaccells and G418 resistant cells were selected as described above.
Disruption of P. pallidum AcgA. To disrupt P. pallidum AcgA, two AcgA fragments of 2.8 and 1.1 kb, respectively, were amplified from P. pallidum gDNA using primer pairs Pp-ACG-P5/Pp-ACG-P3 and Pp-ACG-53H/Pp-ACG-53K (Supplementary  table 1). The first fragment was reduced to 2.4 kb by XbaI/BglII digest and inserted into XbaI/BamHII digested vector pLox-NeoIII 14 . The second fragment was digested with HindIII/KpnI, using restriction sites introduced in the primers, and inserted into the HindIII/KpnI sites of the newly generated vector pAcgA-KO (Supplementary figure 3). P. pallidum cells were transformed by electroporation with the linearized vector pAcgA-KO, followed by selection of transformants and diagnosis of gene knock-out as outlined in Supplementary figure 3.
Disruption of P. pallidum AcrA. For P. pallidum AcrA disruption, two DNA fragments of ,2.0 kb were amplified by PCR from P. pallidum gDNA, using primer pairs AcrA59-fw/AcrA59-rev and AcrA39-fw/AcrA39-rev (Supplementary table 1), which harbour KpnI/BamHI and HindIII restriction sites, respectively. The two fragments were sequentially cloned into the KpnI/BamHI and HindIII restriction sites of plasmid pLoxP-NeoI 29 , yielding vector pAcrA-KO. Correct orientation of the HindIII fragment was verified by digest with SalI and XbaI and DNA sequencing. The insert containing the two AcrA KO fragments flanking the floxed neomycin resistance cassette (Supplementary figure 4A) was introduced into P. pallidum cells, and gene knock-outs were identified by PCR and Southern blot (Supplementary figure 4B,C).
Creation of a double acra-acga-cell line. The A6neo cassette was removed from AcrA KO cells (acra-) as described above. A G418 sensitive acraclone was transformed with linearized pAcgA-KO to delete AcgA. The AcgA gene disruption was confirmed by PCR and Southern blot (Supplementary figure 3C).
Complementation of acra-acga-with AcrA. To express AcrA from its own promoter, a fosmid, used for P. pallidum genome mapping 10 and containing AcrA, was digested whith SpeI. This released an 8.65 kb fragment, which contained the entire AcrA Figure 5 | Pathway for encystation in Amoebozoa and Excavata. The choice between feeding trophozoite and dormant cyst stage is controlled by cAMP binding to PkaR causing dissociation from and activation of PkaC. cAMP is synthesized by adenylate cyclases, which may be activated directly by stress. cAMP is hydrolyzed by the cAMP phosphodiesterase RegA, which is respectively activated by sensor histidine kinases that sense food, or inhibited by sensor histidine phosphatases that sense stress. Violet: components that are conserved in Amoebozoa and Excavata, blue: conserved in Amoebozoa, Teal or pink, conserved in only dictyostelia or Excavata, respectively. Arrow: stimulates; crossbar: inhibits.