STAT signaling in Dictyostelium development

Authors


Author to whom all correspondence should be addressed.
Email: tkawata@bio.sci.toho-u.ac.jp

Abstract

Signal transducers and activators of transcription (STAT) proteins are one of the important mediators of phosphotyrosine-regulated signaling in metazoan cells. These proteins are components of JAK/STAT signal transduction pathways, which regulate immune responses, cell fate, proliferation, cell migration, and programmed cell death in multicellular organisms. The cellular slime mould, Dictyostelium discoideum, is the simplest multicellular organism using molecules homologous to STATs, Dd-STATa–d. The Dd-STATa null mutant displays delayed aggregation, no phototaxis and fails culmination. Here, the functions of Dictyostelium STATs during development and their associated signaling molecules are discussed.

Introduction

Dictyostelium discoideum is a simple protozoan widely used as a model organism in developmental biology and cell biology. The non-sexual life cycle includes cell division, cell differentiation, and cell sorting. Cell aggregation is initiated in response to pulsatile cAMP after several hours of starvation. At the early loose mound stage (6–9 h), up to approximately 100 000 aggregated amoeboid cells form a weakly connected multicellular mound-shaped structure. As development proceeds, the cells form stronger connections that are accompanied by the secretion of extracellular matrices. The resulting tight mound forms a tip at its apex (approximately 13 h), which then elongates upward to form the first finger structure (approximately 15 h). In most cases, the finger collapses onto the substratum to form a slug-shaped structure (approximately 17 h) that is capable of migration towards light and a region of optimum temperature. Cells in the slug structure differentiate into two basic cell-types: anterior prestalk (pst) and posterior prespore (psp) cells. Promoter analyses of ecmA and ecmB genes, which encode extracellular matrix proteins, using β-galactosidase fusions have shown the existence of multiple prestalk-cell subtypes that differ in their patterns of gene expression (Williams et al. 1989). These subtypes include prestalk AB (pstAB), prestalk A (pstA), prestalk O (pstO), and prestalk B (pstB) cells. In addition, another prestalk cell type, tip-organiser cells exist in the slug structure (Williams 2006; see also the review article by Fukuzawa 2011 in this issue). As these cells express the ecmA gene, they were formerly included in the pstA zone. However, as demonstrated by monitoring cudA, a marker gene encoding a transcription factor that regulates slug/culmination choice (Fukuzawa et al. 1997), the anterior region also shows a unique expression pattern in a sub-set of pstA cells (Williams 2006, 2010).

Following formation of the slug structure, a drop in humidity or reduction of ammonia concentration triggers culmination (19 h). At this time, the slug stops horizontal movement to form a Mexican hat-like structure, and the stalk initiates elongation from the top part of the Mexican hat structure to penetrate the spore mass (sorus). During culmination, the basal part of the Mexican hat structure forms the basal disc, and the upper cup lifts the sorus to form the fruiting body (24–26 h) (Loomis 1975; Sternfeld 1998; Kessin 2001). The stalk is composed of highly vacuolated dead cells. The formation and extension of the stalk entails caspase-independent programmed cell death (Olie et al. 1998; Roisin-Bouffay et al. 2004). PstAB cells eventually form the inner parts of the basal disc (Jermyn et al. 1996; Dormann et al. 1996; Shimada et al. 2005), while spores continue to mature in the sorus during culmination. As spores are surrounded by a coat consisting of glycoproteins and polysaccharides, they are highly resistant to environmental stresses (Srinivasan et al. 2000; West 2003). In an appropriate environment of high humidity and temperature, amoeboid cells germinate from spores and then begin a new life cycle (Xu et al. 2004).

A transcription factor necessary for commitment to culmination

In D. discoideum, culmination is the commitment step leading to terminal determination. It is triggered by a drop of humidity or ammonia concentration, high osmotic pressure, low pH and overhead light exposure (Raper 1940; Slifkin & Bonner 1952; Schindler & Sussman 1977; Bonner et al. 1985). Weak acids also induce culmination, whereas weak bases, such as ammonia, inhibit the process (Gee et al. 1994). Mechanisms exist for repressing precocious stalk and spore differentiation before and during culmination, which is initiated when prestalk cells enter the “primordial” stalk tube and begin the conversion into stalk cells. Signals repressing the transition from prestalk into stalk cells regulate the choice between slug migration and culmination (Harwood et al. 1992, 1993). Two mutually redundant repressor elements, which contain an imperfect TTGA inverted repeat sequence, prevent precocious expression of the ecmB gene in pstA cells during both slug migration and culmination (Harwood et al. 1993). A putative repressor protein that binds to the TTGA element was proposed to prevent the expression of ecmB in prestalk cells. During culmination, this repressor would be removed from the promoter, resulting in the activation of ecmB expression and commitment of pstA cells to stalk cells (Harwood et al. 1993). Although the precise signaling pathway is unknown, cAMP-dependent protein kinase (PKA) activation is necessary for lifting the repression of ecmB (Harwood et al. 1992).

In addition to the TTGA inverted repeat sequence, a 53-base element containing a TTGA direct repeat present in the ecmA promoter region was identified, which was initially proposed to be the ecmA activator region (Kawata et al. 1996). In an attempt to identify the TTGA-binding protein, a biochemical approach was undertaken to purify the protein using DNA affinity chromatography with the 53-base element and a large number of slug cells (approximately 1 × 1012 cells). Subsequent cloning and sequencing analysis of the purified protein’s cognate cDNA clones revealed the encoded protein was a Dictyostelium STAT. This study represented the first identification of a non-metazoan STAT (signal transducer and activator of transcription) protein and the first indication of the existence of SH2 signaling pathways in non-metazoan organisms (Kawata et al. 1997). To date, four STATs have been identified in the Dictyostelium genome. The first identified STAT is now termed Dd-STATa (STATa hereafter in this review); Dd-STATb (STATb) and Dd-STATc (STATc) were isolated by low stringency hybridization (Fukuzawa et al. 2001; Zhukovskaya et al. 2004); and Dd-STATd was discovered in a computer search of the Dictyostelium genome (Gao et al. 2004; Abe & Williams, unpubl. data).

JAK-STAT signaling and its evolution

In mammals, seven STATs, STAT1–4, STAT5a, STAT5b and STAT6, have been identified. This family of proteins is activated by cytokines, such as interferon and interleukin (IL), or by growth factors, such as epidermal growth factors (EGF) and platelet-derived growth factor (PDGF) (Akira 1999). In response to these extracellular signals, a signaling cascade termed the JAK-STAT pathway is initiated (Fig. 1) (Levy & Darnell 2002). In this pathway, latent cytosolic STAT monomers first bind to specific receptors, which then induce receptor-associated JAK kinases to phosphorylate tyrosine residues near the STAT C-termini. Following phosphorylation, STAT molecules form dimers that then translocate into nucleus to bind target DNA sequences either directly or indirectly to regulate transcription of target genes. STAT3 is activated through gp130 as a receptor component by IL-6 family cytokines, including IL-6, oncostatin M (OSM), IL-11, and leukemia inhibitory factor (LIF) (Hirano et al. 2000; Ernst & Jenkins 2004). STAT3 is a positive regulator of cell growth and is an oncogene that is constitutively active in cancer cells (Bromberg et al. 1999; Bowman et al. 2000; Hirano et al. 2000; Yu et al. 2007, 2009). In contrast, the activities of STAT1, which is a negative regulator of cancer and an activator of p53 and p21, which induce caspase to initiate apoptosis of cancer cells (Agrawal et al. 2002; Bhanoori et al. 2003).

Figure 1.

 Schematic diagram of the JAK-STAT pathway and homologues in various model organisms. The upper panel shows the mammalian JSK-STAT pathway. P, phosphorylated tyrosine residue. Lower panel indicates the homologous components of the JAK-STAT pathway in Dictyostelium, Drosophila, and mouse. cAR1, cAMP receptor 1.

Although yeasts have no known STAT genes, the fact that the simple eukaryote Dictyostelium possesses STAT homologues indicates this family of proteins must have arisen early in evolution, even though Dictyostelium is not believed to be a direct ancestor of metazoans (Williams et al. 2005; Williams 2010). In Drosophila, all JAK-STAT components have been identified, from Unpaired to Marelle (or Stat92E) (Fig. 1) (Arbouzova & Zeidler 2006). As mentioned, four STATs are present in Dictyostelium and a single STAT exists in Caenorhabditis elegans (Wang & Levy 2006); however, no canonical JAK kinase has been found in either organism (Plowman et al. 1999) Since STATs are phosphorylated at a tyrosine residue and are functional in Dictyostelium and C. elegans, there must be tyrosine kinases in both organisms that phosphorylate their STATs. Alternatively, STATs are phosphorylated by kinases totally unrecognizable as tyrosine-specific in these two organisms.

Mammalian STATs contain several structurally and functionally conserved domains (Fig. 2) (Soler-Lopez et al. 2004). For example, the amino-terminal region of human STAT1 is known to enhance cooperative DNA binding and regulate nuclear localization (Vinkemeier et al. 1998). A second conserved region is a coiled-coil domain containing a four-helix bundle. This domain associates with a number of regulatory proteins and is also implicated in the antiparallel non-phosphorylated dimmer formation of STAT1 (Zhong et al. 2005). The central region contains an immunoglobulin-like (Ig) domain that serves as a DNA-binding domain recognizing a γ-interferon activating sequence (GAS) consensus sequence, TTCN2–4GAA. The Ig domain is followed by a “connector” or “linker” domain consisting of an EF-hand fold (EF-domain). In addition, an SH2 domain is located near the C-terminus followed by a phosphorylated tail segment. Notably, the SH2 domain is the most highly conserved in the STAT family and mediates recruitment to specific receptors and the dimerization of STAT molecules. Finally, a transcriptional activation domain (TAD) is located at the C-terminus, which is conserved between homologues in different organisms, but varies between STAT family members (Schindler 2002).

Figure 2.

 Schematic structure of human STAT1 and Dictyostelium STATs. The illustration shows the comparison of domains between human STAT1 and the Dictyostelium STATs, Dd-STATa–Dd-STATd. The boundaries of each domain are based on the crystal structure described by Soler-Lopez et al. (2004) and each domain is indicated by a different colour. Activator domain, transcriptional activator domain; EF-hand, a connector or linker domain with EF-hand hold; Ig-like, immunoglobulin-like domain; N-domain, N-terminal conserved domain; pY, phosphorylated tyrosine residue; SH2, Src homology 2 domain. A triangle represents the position of inserted amino acids in the SH2 domain of STATb (see text).

Recently, non-canonical JAK-STAT signaling was found in Drosophila. In this pathway, a portion of non-phosphorylated STAT is localized in the nucleus on heterochoromatin in association with heterochromatin protein 1 (HP1). Unphosphorylated STAT is essential for HP1 localization and heterochromatin stability (Shi et al. 2007; Brown & Zeidler 2008; Li 2008). Similarly, nuclear localization of unphosphorylated STAT3 and STAT5 has also been reported in mammals, and both proteins have been shown to influence transcription (Reich & Liu 2006; Yang & Stark 2008).

Dd-STATa

The STATa protein is composed of 707 amino acids and has a molecular mass of 79.964 kDa, as determined by mass spectrometric analysis (Kawata et al. 1997). STATa is phosphorylated at the tyrosine residue at position 702 near the C-terminus, and this modification is necessary for reciprocal interaction with the SH2 domain of the dimerising partner STATa molecule (Fig. 2) (Kawata et al. 1997). After phosphorylation of the tyrosine residue by extracellular cAMP, STATa translocates to the nucleus (Araki et al. 1998). The tyrosine kinase that specifically phosphorylates STATa has not yet been identified (Fig. 1). The domain organization of STATa is fundamentally similar to that of STAT1, with the exception of an N-terminal domain unrelated to mammalian STATs, consisting of poly(Asn) and poly(Gln) residues (Fig. 2) (Soler-Lopez et al. 2004). STATa also lacks the transcriptional activator domain (TAD) at the C-terminus, which is consistent with evidence that STATa acts as a repressor of the ecmB gene in pstA cells. However, STATa is also a confirmed activator of the cudA gene in tip-organiser cells (Fukuzawa & Williams 2000). Additional accessory protein(s) interacting with STATa, or unidentified modification of STATa, may be involved in transcriptional activation.

Despite substantial differences in the primary sequences between STATa and mammalian STATs, nearly all secondary structure elements are well conserved. These elements include the coiled-coil (residues 242–356), Ig (residues 357–484), EF-hand (residues 485–575), and SH2 (residues 576–690) domains, followed by the phosphorylated tail segment (residues 691–707) (Fig. 2). Crystal structure analysis has revealed that the DNA-unbound form of STATa adopts a stable, fully extended conformation. Upon DNA-binding, the conformation of this protein changes drastically. The SH2 domain of STATa belongs to the prototypical Src SH2 domain family, whereas those of mammalian STATs do not fit this consensus nearly as well. Therefore, STATa appears to be less diverged from the primordial STAT than mammalian STATs (Soler-Lopez et al. 2004).

STATa is a multi-functional protein, as indicated by the defects that a STATa null strain of Dictyostelium displays in both early and late development (Mohanty et al. 1999). In this null strain, the aggregation of cells is delayed due to inefficient chemotaxis to cAMP sources. Although the mutant cells form slugs, they do not exhibit phototaxis and have an aberrant pattern of gene expression. Particularly, the ecmB gene is ectopically expressed in pstA cells, suggesting that STATa functions as the repressor of ectopic ecmB gene expression. The STATa null strain remains in the slug stage for several days, and will occasionally form aberrant terminal structures with a small spore mass supported by an undifferentiated column. Although STATa null cells show almost no stalk differentiation in the slug stage, the mutant cells are hypersensitive to differentiation-inducing factor (DIF)-1 and form stalk cells easily in a monolayer assay (Mohanty et al. 1999).

Dd-STATb

STATb possesses a highly aberrant SH2 domain, as a 15-amino-acid insertion is present in the middle of the domain, while the conserved SH2 domain arginine residue, which is essential for interaction with the phosphotyrosine of the dimerising partner in all other STATs, is replaced by leucine (Fig. 2) (Zhukovskaya et al. 2004). Interestingly, substitution of the phosphotyrosine residue of STATb with phenyalanine does not block homodimer formation, and immunoprecipitation analysis indicates that STATb does not hetero-dimerise with either STATa or STATc (Zhukovskaya et al. 2004). A STATb null strain develops normally, and cells grow at an apparently normal speed; however, when STATb null and STATb+ cells are mixed and cocultured, STATb null cells are gradually lost from the population. In addition, microarray analysis has revealed altered gene expression in STATb null cells, including lowered expression of the HGPRT gene and elevated expression of the smlA, dicoidin 1, and DdCAD-1 genes (Zhukovskaya et al. 2004).

Dd-STATc

STATc resembles STATa with respect to primary amino acid sequence, although STATc has a longer N-terminal domain and is 124 amino acids larger than STATa (Fig. 2). The carboxyl halves of STATa and STATc are highly conserved, but STATc responds to DIF-1 instead of cAMP (Fukuzawa et al. 2001). Nuclear accumulation occurs by the repression of STATc nuclear export (Fukuzawa et al. 2003). The level of tyrosine phosphorylation rises to reach an approximate plateau at the slug stage, and subsequently drops at culmination. Similar to STATa, the tyrosine kinase that specifically phosphorylates STATc has not been identified.

STATc is selectively enriched in the nuclei of pstO cells. In a proportion of slugs, cells located at the rear also show STATc nuclear enrichment (Fukuzawa et al. 2001). When grown on a bacterial lawn, a STATc null mutant forms smaller plaques than those of the parental strain. The mutant plaques also have rough edges and are surrounded by satellite fruiting bodies. Until the slug stage, the null strain develops 1–2 h quicker than the parent strain. The rapid development correlates with accelerated gene expression of genes such as csA, which encodes a cell adhesion molecule, and pspA, which encodes a cell-surface protein of prespore cells. The DIF-1 inducible ecmA gene is overexpressed in the null mutant (Fukuzawa et al. 2001). Although STATc null slugs exhibit a strong tendency to continue migration under conditions where parental slugs culminate, they form normal-looking fruiting bodies. The ecmA promoter sequences necessary for activation in pstA cells are also active in the pstO cells of STATc null slugs. Therefore, STATc is a DIF-1-dependent repressor protein that prevents the binding of a pstA activator protein to the cis-acting element of the ecmA gene in pstO cells (Fukuzawa et al. 2001).

In addition to DIF-1, STATc is also activated by hyperosmotic stress, heat shock, and oxidative stress (Araki et al. 2003; Na et al. 2007). Similar responses were also found in mammalian STAT1, which is activated by interferon and also by hyperosmotic stress (Gatsios et al. 1998). Hyperosmotic stress is known to elevate intracellular cGMP and cAMP levels (Kuwayama et al. 1996; Ott et al. 2000). However, STATc is activated in the null mutants of the known cGMP- and cAMP-mediated intracellular response pathways. Microarray analyses showed that approximately 20% of the hyperosmotic stress regulated genes use the STATc pathway (Na et al. 2007). Two hyperosmotic stress-induced genes, gapA and rtoA, were identified in a different microarray study and characterized (Fig. 3; Araki et al. 2003). Although osmotic stress induces these two genes in a STATc-dependent manner, neither gene is inducible by DIF-1. Thus, STATc functions as a transcriptional activator in a genetically separable stress-response pathway from the DIF-1 pathway (Araki et al. 2003).

Figure 3.

 STAT signaling networks in Dictyostelium and STATa binding sites. Upper panel represents the signaling networks showing linking molecules to STATa, STATb, and STATc. cAR1, cAMP receptor1; cAR3, cAMP receptor 3; ACA, adenylnyl cyclase; AmtA, ammonium transporter A; AmtC, ammonium transporter C; ecm, extracellular matrix; GskA, glycogen synthase kinase A; PKA, cAMP-dependent kinase; DhkC, Dictyostelium histidine kinase C; cudA, culmination defective A; aslA, acetyl-CoA synthetase like A; PIAS, protein inhibitor of activated STAT; DdCAD-1, calcium-dependent cell adhesion molecule-1; HGPRT, hypoxanthine phosphoribosyltransferase; gapA, RasGTPase-activating protein A; rtoA, RatioA; PTP3, protein tyrosine phosphatase 3; CblA, Casitas B-lineage lymphoma A. Lower panel shows the sequences of STATa binding sites of the cudA and ecmB promoters. Nucleotides essential for STATa binding are shown in red letters. Arrows indicate the dyad structures.

Recently, protein tyrosine phosphatase PTP3 was shown to regulate the phosphorylation level of STATc (Araki et al. 2008). In the absence of DIF-1 or hyperosmotic stress, PTP3 is active to dephosphorylate STATc. However, upon DIF-1 treatment or hyperosmotic stress, PTP3 is serine-phosphorylated, which reduces PTP3 activity and increases the level of tyrosine phosphorylation of STATc (Araki et al. 2008). The phosphorylation of two serine residues of PTP3, S448 and S747, is independent of cGMP signaling that is necessary for STATc activation. Instead, the elevation of intracellular Ca2+ levels, which is also a potent STATc activating signal, activates PTP3 serine phosphorylation. Therefore, two parallel pathways are possible for the activation of STATc by osmotic stress (Araki et al. 2010). Another Dictyostelium SH2 domain-containing protein, CblA, also positively regulates STATc tyrosine phosphorylation (Langenick et al. 2008). CblA is a homologue of Casitas B-lineage lymphoma (Cbl) protein, which serves as a regulator of receptor tyrosine kinases (RTKs) and downregulates PTP3 activity to increase STATc phosphorylation.

cAMP signaling linked to the Dd-STATa pathway

STATa is activated by extracellular cAMP acting through the serpentine receptor cAR1 in a Gβ protein-independent manner (Araki et al. 1998). Consistent with this role, STATa is activated in slug’s pstA cells, which contain the source of extracellular cAMP signaling (Araki et al. 1998). In addition, when cAMP is injected into the rear prespore region of slugs, GFP:STATa fusion proteins accumulate in the surrounding nuclei (Dormann et al. 2001). Similarly, when the ACA adenylyl cyclase is overexpressed from a semi-constitutive actin15 promoter, cudA is ectopically expressed throughout the slug (Verkerke-van Wijk et al. 2001). Expression of the endogenous ACA (acaA) gene in the slug stage is restricted to cells, known as tip-organizer cells, located at the front of the slug. These observations suggest that the expression of ACA leads to localized cAMP accumulation, which then activates STATa to direct the transcription of the cudA gene in tip-organizer cells (Fig. 3). Similarly, transcription of the ecmF and aslA genes appears to be activated by STATa, although the location of cells expressing these genes is distinct from tip-organizer cells (Shimada et al. 2004a, 2005). The accumulation of intracellular cAMP is also necessary for the activation of PKA, which is required for culmination.

In Dictyostelium, intracellular cAMP levels are downregulated by the cAMP phosphodiesterase RegA (Fig. 3). The histidine kinase DhkC is thought to regulate slug/culmination choice by a phosphorelay mechanism through the relay protein RdeA (Fig. 3) (Kirsten et al. 2005). When the phosphorelay is active, RegA is activated to lower the intracellular cAMP level and hence, slugs continue to migrate. Since ammonium transporter C (AmtC) inhibits the activity of DhkC, it effectively lowers the activity of RegA and activates PKA to promote culmination (Fig. 3) (Kirsten et al. 2005). In an amtC null strain, there is little or no activated nuclear localization of STATa, and no observable expression of the cudA gene in tip-organizer cells during the slug stage (Kirsten et al. 2005). Interestingly, another ammonium transporter, AmtA, functions antagonistically through DhkC to activate the phosphorelay system (Fig. 3) (Singleton et al. 2006).

Extracellular cAMP also activates the tyrosine kinase Zak1 through the cell-surface cAMP receptor, cAR3 (Kim et al. 1999; Plyte et al. 1999). Zak1 phosphorylates and activates GskA, a Dictyostelium homologue of GSK-3 implicated in cell fate decision, which then phosphorylates nuclear STATa on serine residues within the Asn/Gln-rich N-terminal domain to promote nuclear export (Fig. 3) (Ginger et al. 2000). Conversely, extracellular cAMP inhibits GskA activity by increasing tyrosine phosphatase (PTPase) activity through the cAR4 receptor (Kim et al. 2002). Protein tyrosine kinase PTP1 negatively regulates STATa, as indicated by the increased nuclear enrichment of STATa in a ptp1 (ptpA) null strain (Early et al. 2001).

Function of tip-organizer cells and Dd-STATa

The slug tip behaves like an organizer region in vertebrate embryogenesis, as demonstrated by the generation of a secondary slug following the grafting of another tip onto the rear prespore zone (Raper 1940). The distinctive characteristics of the STATa null mutant are its “slugger” phenotype, failure to culminate, and lack of phototaxis (Mohanty et al. 1999). Interestingly, the genes thought to be involved in the identical signaling cascade display similar slugger phenotype when disrupted. These include cudA, dhkC, and amtC genes. Null mutants of other potential STATa target genes, including ecmF, alsA, ahhA, and expL7, show little or no obvious phenotypic defects (Shimada et al. 2004a, 2005; Aoshima et al. 2006; Ogasawara et al. 2009). Notably, CudA directs the expression of expL7. The CudA is known to regulate the slug/culmination switch and localizes in the prespore zone and in the tip-organizer cells at the extreme anterior of the slug (Fukuzawa et al. 1997; Fukuzawa & Williams 2000; Wang & Williams 2010a). Importantly, cudA expression is lacking only in the latter population of cells in the STATa null strain. CudA was determined to be a transcriptional factor because it is located in the nucleus, expression of the prespore-specific cotC gene is greatly reduced in a cudA null strain (Fukuzawa et al. 1997), and it directly binds to the promoter elements of both cotC and expL7 (Yamada et al. 2008; Wang & Williams 2010a,b). Moreover, CudA is suggested to be derived from an ancestral STAT-like DNA binding protein (Yamada et al. 2008). The cudA promoter contains several regions that are responsible for directing its expression in prespore and tip-organizer cells (Fukuzawa & Williams 2000). For example, a dyad sequence in the tip-organizer-specific region that loosely resembles the repressor sequences in ecmB promoter is a binding site for STATa in vitro (Fig. 3). Taken together, these findings suggest that STATa is the direct activator of CudA expression.

Other signals linking to Dd-STATa

Database searches of the Dictyostelium genome indicate the presence of a homologue of the protein inhibitor of activated STAT (PIAS) protein, Dd-PIAS. Although Dd-PIAS does not appear to interact with STATa directly, Dd-PIAS has been demonstrated to affect the expression of several putative STATa target genes (Kawata et al. unpubl. data). Recently, a number of cDNA clones that functioned as overexpression suppressors of a “partially active” hypomorphic STATa strain were isolated. One of the isolated cDNA clones contains a segment of the dutA gene that encodes an mRNA-type non-coding RNA (Shimada & Kawata 2007). Interestingly, the phosphorylation level of STATa elevated significantly in the suppressor strain, suggesting regulation of tyrosine kinase activity by the non-coding RNA. Interaction of dutA RNA with a yet unidentified cytosolic protein was also detected (Kawata & Watanabe, unpubl. data). Another interesting suppressor gene is sunB, which encodes a protein with a Sad1/UNC-84 (SUN) homology domain located at its centre. It was noted that sunB mutant failed to culminate, and the observed decrease in SunB protein in prestalk cells may be regulated by STATa (Shimada et al. 2010). Genes whose products are homologous to mammalian proteins involved in programmed cell death, and zinc transporters have also been identified in the Dictyostelium genome. The expression of several of these genes, in addition to a number of cellulase genes, is downregulated in the STATa null strain (Sunaga et al. 2008; Eiguchi et al. unpubl. data; Kawata et al. unpubl. data).

Dictyostelium discoideum is the simplest eukaryote known to use STAT signaling, with STATs serving several vital roles in Dictyostelium development. Despite its relative simplicity, Dictyostelium uses four STATs, although only one STAT has been identified in both C. elegans and Drosophila. Although it is possible that the latter organisms lost a few STATs during evolution, it would be interesting to determine whether eukaryotic organisms less complex than Dictyostelium also use several STAT molecules. Indeed, the phylogenetically lower cellular slime mould Acytostelium subglobosum also possesses four STAT genes (Urushihara et al., pers. comm.), and Acrasis rosea contains at least a STATa homologue (Kawata, unpubl. data).

One of the major unresolved issues is the identification of the kinases that phosphorylate and activate the Dictyostelium STATs. Analysis of the Dictyostelium kinome indicates there is no canonical receptor-tyrosine kinase group, but numerous tyrosine-kinase-like (TKL) genes exist in this organism (Goldberg et al. 2006). The three-part (writer, reader, and eraser) toolkit model for phosphotyrosine signaling has been reported recently (Lim & Pawson 2010). In this proposed model, tyrosine kinase represents the newest invented toolkit since Dictyostelium has only reader (SH2 domains) and eraser (PTPs) tools. Interestingly, at least one TKL kinase is likely to phosphorylate STATa in a cAMP-induction assay, but is not required for STATa phosphorylation during normal development (Sasaki & Kawata, unpubl. data). If this is true, the Dictyostelium STAT signaling pathway is not as simple as the metazoan JAK/STAT pathway, and likely involves several TKLs.

Among pstAB cells of Dictyostelium, some heterogeneity and sub-types exist (Yamada et al. 2005). The expressions of nearly all these cell-type marker genes are missing in the STATa null mutant (Shimada et al. 2004b, 2005; Sunaga et al. 2008). Although the most important function of STATa appears to be the production of tip-organizer cells, this has yet to be proven. However, if this holds to be true, it raises several questions. For example, which cell-type of pstAB cells is the real organizer in this organism? Does the organizer correspond to the “primordial” stalk? If one of these pstAB type cells functions as an organizer, then when is its differentiation determined and the final commitment of culmination triggered? The elucidation of these remaining questions would provide novel insight into the roles of STAT proteins and linked signaling pathways.

Acknowledgments

This work was supported by a Grant-in-Aid for Scientific Research C from the Japanese Society for the Promotion of Science (JSTS) (no. 2150230) to T. Kawata. I give thanks to Mr Norimitsu Sasaki, Toho University, Japan, for providing his illustration.

Ancillary