JAK-independent STAT signaling in Dictyostelium discoideum

The social amoeba Dictyostelium discoideum has four STAT-family transcription factors (STATa to STATd) but no Janus kinase and no kinases of the tyrosine-kinase group. Its STATs are instead tyrosine phosphorylated by tyrosine kinase-like (TKL) kinases, and they dimerize, accumulate in the nucleus and regulate transcription during multicellular development and stress. The two best-studied paralogs are wired to different inputs. STATc is activated by the prestalk morphogen DIF-1 and by hyperosmotic and other stresses. Its activating Tyr922 is phosphorylated by the TKL kinases Pyk2 and Pyk3, which bind the STATc SH2 domain through autophosphorylated tyrosines. The kinase activity is largely constitutive, so the signal is set mainly by the tyrosine phosphatase PTP3: DIF-1 and stress induce serine phosphorylation of PTP3 (in part by the TKL kinase Phg2), which inhibits it and shifts STATc toward the phosphorylated state. Nuclear STATc activates stress-response genes and restricts graded expression of the prestalk gene ecmA. STATa is tyrosine phosphorylated on Tyr702 within seconds of extracellular cAMP binding the serpentine receptor cAR1, apparently without the G-beta subunit. The receptor-like TKL kinase DrkA phosphorylates STATa in vitro and is a candidate STATa kinase. STATa stays nuclear in the prestalk cells of the slug tip, where it represses the stalk gene ecmB and directly activates cudA, a transcription factor needed for culmination.

Boundary. This is a concrete D. discoideum module. Its theme is STAT transcription factors activated without JAK kinases. It has two paralog-specific pathway variants that coexist in the same organism (selection ONE_OR_MORE) but differ in input, cell type and kinase. Each variant has an upstream-input part and a STAT transcriptional-control part. The STATc variant is the best evidenced. Its kinases (Pyk2, Pyk3) are established by null mutants, a double mutant, in vitro phosphorylation of Tyr922 and co-immunoprecipitation, and its phosphatase arm (PTP3, Phg2) by substrate trapping, overexpression and phosphosite mapping. The STATa variant has a well-established receptor input (cAR1) and transcriptional output (ecmB repression, cudA activation). Its kinase is less certain: DrkA rests on overexpression and in vitro phosphorylation, with no loss-of-function test. STATb and STATd are not modeled as variants. STATb is constitutively nuclear, has an SH2 domain lacking the phosphotyrosine-binding arginine, keeps its behaviour when the predicted phosphotyrosine is mutated, and has no known input (PMID:14701681), so it has no input part to model. STATd (dstD, UniProtKB:Q86I20, unreviewed) has no functional characterization in the cached literature. Relation to MODULE:jak_stat_signaling. That module is a taxon-neutral receptor-JAK-STAT trunk: cytokine-receptor engagement, JAK trans-activation, receptor phosphotyrosine docking sites, STAT recruitment and tyrosine phosphorylation by JAK, dimerization and nuclear import, then GAS/ISRE transcription, with SOCS, phosphatase and PIAS feedback. Here the STAT end of that trunk is conserved: tyrosine phosphorylation of a C-terminal tyrosine, SH2-phosphotyrosine dimerization (shown in the STATa crystal structure), nuclear accumulation and binding to promoter elements (STATa binds a mammalian ISRE). The upstream end is not. There is no JAK and no receptor-docked kinase. For STATc the receptor for DIF-1 is unknown and stress is sensed through F-actin remodelling. The kinase is a TKL that the STAT SH2 domain binds directly through the kinase's own autophosphorylated tyrosines, not through a phosphorylated receptor tail. The kinase is largely constitutive, and the signal is gated by inhibiting a tyrosine phosphatase (PTP3) by serine phosphorylation. In the canonical module phosphatases only terminate signaling. Pyk3 does carry a pseudokinase domain N-terminal to its kinase domain that inhibits the kinase domain, as JAK JH2 does, which is convergent architecture, not JAK homology. For STATa the receptor is a G-protein-coupled cAMP receptor, and its signal to STATa does not need the G-beta subunit. This module therefore does not conform to the jak_stat_signaling template. It is a lineage-specific wiring around a conserved STAT core, as in the C. elegans module (MODULE:c_elegans_jak_independent_stat_signaling), but the two lineages differ: Dictyostelium STATs keep phosphotyrosine-SH2 activation and are activated by TKL kinases, whereas STA-2 in the worm has lost the phosphotyrosine motif. GO terms. The pathway term is GO:0097696 cell surface receptor signaling pathway via STAT, whose definition names D. discoideum and TKL kinases. GO:0007259 (via JAK-STAT) is deliberately absent; the statA and statC reviews modify the inherited IBA to GO:0097696. GO:0007260 tyrosine phosphorylation of STAT protein describes the Pyk2/Pyk3 and DrkA step exactly, but it is currently part_of GO:0007259. Asserting it here would imply JAK-STAT signaling by inference. The kinase annotons therefore carry GO:0097696 and give GO:0007260 only in text. GO has since added an only_in_taxon Metazoa constraint on GO:0007259 (TPV-003 in projects/TAXON_PATHWAY_VARIANCE.md; go-ontology#32689) without first re-parenting GO:0007260 under GO:0097696, so GO:0007260 now inherits the constraint by part_of and asserting it for D. discoideum would be a constraint violation, not just an implication. Pyk2 phosphorylating STATc on Tyr922 without a JAK (PMID:22699506) is the counterexample. GO:0007260 is listed for obsoletion in go-ontology#27064. The case is written up for GO and PAINT curators in the public cross-gene briefing JAK-STAT Without JAK (https://claude.ai/artifact/4MLfbbbWWnLchDvspwdPuM). Relation to other Dictyostelium modules. MODULE:dicty_dif1_response_prestalk_patterning models the DIF-1 response through the bZIP factors DimA and DimB and their outputs ecmA and ecmB. STATc is a second, parallel DIF-1 effector. It is not included there and is modeled here only. The two modules share an output: STATc restricts graded ecmA expression among prestalk cells, while DimA/DimB induce it. DIF-1 production is in MODULE:dicty_dif1_biosynthesis. MODULE:dicty_extracellular_camp_relay models the cAR1-Galpha2-CRAC-ACA aggregation oscillator. cAR1 (UniProtKB:P13773) is reused here as the input to STATa, but in a different context: at the mound and slug stage, and independently of G-beta, whereas the relay depends on heterotrimeric G proteins. ACA is not modeled here. Its tip-restricted expression directs the tip-specific nuclear accumulation of STATa (PMID:11356026), so the relay supplies the cAMP that the STATa variant reads. The dif1_response module explicitly leaves the STATa-CudA culmination circuit out; that circuit is the STATa variant here. Other modulators kept as context. CblA, a Cbl homologue, promotes DIF-1-induced STATc tyrosine phosphorylation by downregulating PTP3 (PMID:18840649). Cytoskeletal perturbation (latrunculin A, cytochalasin A) induces PTP3 S747 phosphorylation and STATc activation (PMID:22365144). cGMP (via GbpC) and Ca2+ are proposed parallel second messengers acting on the kinase and phosphatase arms, respectively (PMID:20159963). The PIAS homologue Dd-PIAS has a genetic and expression-pattern link to STATa (PMID:21933174); it is raised as a knowledge gap, not modeled. GskA in other Dictyostelium modules. GskA has larger roles outside this module: cAMP-receptor-driven prespore versus prestalk-B fate choice and chemotaxis via TORC2/RacE (see its gene review). No existing Dictyostelium module models those processes, so GskA appears here only for STATa nuclear export. Family context. All four STATs are PANTHER PTHR11801 members. STATa, STATb and STATc (dictyBase DDB_G0281381, DDB_G0268638, DDB_G0293532) are seeds of the family-root PAINT node PTN000210448 (nucleus, cytoplasm) in interpro/panther/PTHR11801/PTHR11801-paint.tsv. Concrete GENE_PRODUCT selectors are used instead of family descriptors because the module is species-scoped. PTP3 is in PTHR19134:SF549 and GskA in PTHR24057:SF0. The four TKL kinases are in the large Ser/Thr kinase family PTHR23257: Pyk2/splB in SF975, Pyk3 in SF969, DrkA in SF921 and Phg2 in SF776. Their UniProt cross-references (InterPro 110.0 HMM best hits on the same PANTHER 19.0 release) instead give PTHR44329:SF298 for Pyk2, Pyk3 and DrkA and PTHR46716:SF1 (TAK1) for Phg2. The disagreement was resolved in favour of PANTHER's own placement. PANTHER 19.0's classification files and tree put each kinase in a Dictyostelium clade of PTHR23257, and all four take their IBAs from the PTHR23257 node PTN001148170, not from any PTHR44329 or PTHR46716 node. The placement is pinned in interpro/panther/panther-members-overrides.tsv, and the PTHR44329 and PTHR46716 family reviews no longer list these kinases as representatives. Pyk2, Pyk3 and DrkA tyrosine-kinase activity therefore remains a gene-level (and lineage-specific) property and is not a subfamily-level PTHR44329 inference.