Hypothesis slug: nuclear-localization-and-cell-cycle
Focus type: function_assignment
Source: genes/DICDI/pten/pten-ai-review.yaml (free-text)
Seed hypothesis under evaluation: Dictyostelium discoideum PTEN localizes to the nucleus (GO:0005634) and regulates the cell cycle (GO:0051726) — to be evaluated separately using direct evidence and justified transfer from characterized orthologs, distinguishing cortex/membrane localization from exclusivity, cytokinesis phenotypes from cell-cycle regulation, and mammalian-specific roles from conserved mechanisms.
Executive judgment: REFUTED / OVER-ANNOTATED (both legs of the hypothesis). Neither half of the seed hypothesis survives contact with the direct experimental record for Dictyostelium discoideum PTEN. The proposed nuclear localization (GO:0005634) has no experimental support in this organism: it rests entirely on phylogenetic inference (IBA, Inferred from Biological Ancestor), whereas every direct localization assay (IDA) places the protein at the plasma membrane and cell cortex (posterior/rear, PI(4,5)P₂-dependent). The proposed regulation of cell cycle (GO:0051726) is not curator-annotated at all, and it is mis-scoped: the loss-of-function phenotype is a cytokinesis failure in which the nuclear/DNA-replication cycle continues normally, producing large multinucleate cells. The directly supported process term is mitotic cytokinesis (GO:0000281, IMP), not regulation of cell cycle.
The hypothesis instructs the evaluator to treat localization and process separately, and to test whether ortholog transfer from mammals is justified. On both counts the transfer fails. For localization, the mammalian nuclear-PTEN paradigm depends on a monoubiquitination-driven import switch (the K289 site) and a C-terminal regulatory tail — features that are not conserved in the Dictyostelium protein. In a pairwise alignment, the position aligning to human K289 carries a glutamate (E) in Dictyostelium, i.e. the organism is naturally in the exact import-defective state (K289E) that abolishes nuclear accumulation in Cowden-syndrome patients (PMID: 17218261). The mammalian nuclear roles (heterochromatin maintenance, mitotic-checkpoint-complex regulation, Plk1 crosstalk, CBX8 interaction) form a metazoan-specific program built on machinery Dictyostelium lacks.
The consolidated Dictyostelium model is membrane-centric and cytokinesis-linked: PTEN is a PIP3 3-phosphatase recruited to the posterior cortex, where it establishes back-of-cell polarity that is required for both chemotaxis and cleavage-furrow function. The curation recommendation is therefore to not add GO:0005634 or GO:0051726 as experimentally supported terms, to flag the existing IBA nucleus annotation as phylogenetic carry-over rather than direct evidence, and to anchor the process description on GO:0000281 (mitotic cytokinesis) and the membrane/cortex cellular-component terms already supported by IDA evidence.
Every direct localization study of D. discoideum PTEN reports plasma-membrane and cortical (posterior/rear) enrichment; none reports a nuclear signal. In the foundational study, Iijima & Devreotes (PMID: 12062103) showed that "Exogenously expressed PTEN-GFP localized to the surface membrane at the rear of the cell. Membrane localization required a putative PI(4,5)P2 binding motif and was required for chemotaxis." The dependence of localization on a PI(4,5)P₂-binding motif ties the protein mechanistically to the inner leaflet of the plasma membrane, not to nuclear-import machinery.
This picture is reinforced by Matsuoka & Ueda (PMID: 30367048), who described PTEN and PIP3 as "enriched mutually exclusively on the anterior and posterior membranes of eukaryotic motile cells," establishing a bistable membrane-partitioning mechanism in which PTEN occupies the posterior cortex. Janetopoulos et al. (PMID: 15809030) placed PTEN function at cell poles and the cleavage furrow, and live-imaging studies of chemotactic signaling (PMID: 25300796; PMID: 27505897) consistently track PTEN as a membrane/cortex marker. Crucially, no primary study of Dictyostelium PTEN reports GO:0005634 (nucleus) occupancy. This negative is strong because localization is precisely the assay these studies were designed to measure — live-cell GFP imaging that would readily reveal nuclear accumulation had it existed.
The seed hypothesis explicitly asks the curator to distinguish "cytokinesis phenotypes from cell cycle regulation," and the Dictyostelium data land firmly on the cytokinesis side. Janetopoulos et al. (PMID: 15809030) report that cells lacking PTEN (with intact PI3K) "are defective in cytokinesis, and cannot divide in suspension. The cells continue to grow and duplicate their nuclei, generating large multinucleate cells." This is the diagnostic signature of a cytokinesis defect: DNA replication and mitotic nuclear division proceed on schedule, but the physical division of the cell body fails, so nuclei accumulate. A genuine regulation of cell cycle (GO:0051726) role would be expected to alter S-phase entry, checkpoint control, or nuclear-division rate — none of which is reported.
Downstream and mechanistic studies are consistent. Tang et al. (PMID: 21169559) show PTEN maintains polarity "required for cytokinesis and chemotaxis" by limiting PIP3, with the key effector being phosphorylation of PKB (Akt) substrates. Consalvo et al. (PMID: 38940195) implicate PTEN in AprA-induced inhibition of proliferation — but this is a population-level chemorepellent response mediated through Ras/PIP3 signaling, a downstream physiological effect rather than direct engagement of the cell-cycle machinery. None of these supports GO:0051726 as a direct molecular function of the gene product.
To test whether nuclear localization could be justified by transfer from the well-characterized human ortholog, I performed a pairwise global alignment (Needleman–Wunsch) of human PTEN (P60484, 403 aa) against D. discoideum PTEN (Q8T9S7, 533 aa). Global identity was 45.0% (181/402 aligned positions), and the core phosphatase + C2 domain region (human residues 1–350) aligned at 44.4% identity. The catalytic P-loop signature motif HCKAGKGR is perfectly conserved at the equivalent position (~123) in both proteins, confirming that the lipid-phosphatase catalytic function is conserved — this is the bona fide transferable feature.
However, the residues that specifically govern nuclear import in mammals are not conserved. Trotman et al. (PMID: 17218261) demonstrated that "A lysine mutant of PTEN, K289E associated with Cowden syndrome, retains catalytic activity but fails to accumulate in nuclei of patient tissue due to an import defect." K289 is the major monoubiquitination site that drives PTEN into the nucleus. In my alignment, the Dictyostelium residue aligning to human K289 is a glutamate (E) — i.e., the protein is naturally in the import-defective state that causes disease when engineered into human PTEN. The minor import-associated lysine K13 is conserved (K), but it is insufficient on its own. In addition, the human C-terminal regulatory tail (the phospho-cluster and PDZ-binding motif) is absent in Dictyostelium, which instead carries a long low-complexity Asn/Ser/Thr C-terminal extension unrelated to nuclear trafficking. Ortholog-based transfer of GO:0005634 is therefore mechanistically unjustified.
Caveat: the aligned K289 falls in a divergent C2-domain loop (CBR3/Cα2 region, local identity ~21%), so the exact single-residue call is low-confidence; the robust conclusion is that there is no conserved lysine at the aligned position and the entire mammalian import-signal region is poorly conserved.
| Feature | Human PTEN (P60484) | Dictyostelium PTEN (Q8T9S7) | Transferable? |
|---|---|---|---|
| Length | 403 aa | 533 aa | — |
| Global identity | — | 45.0% (181/402) | — |
| Catalytic P-loop (HCKAGKGR) | Present (~pos 123) | Present (~pos 123) | Yes (catalysis) |
| K289 nuclear-import / monoubiquitination site | K (import-competent) | E (import-defective state) | No |
| Minor import lysine K13 | K | K | Partial (insufficient alone) |
| C-terminal regulatory tail / PDZ motif | Present | Absent (Asn/Ser/Thr low-complexity extension) | No |
An audit of the curated GO cross-references for UniProt Q8T9S7 clinches the evidence-code argument. The cellular-component term GO:0005634 (nucleus) is supported only by IBA:GO_Central — phylogenetic Inferred from Biological Ancestor — with no experimental backing. Every experimental (IDA:dictyBase) localization is non-nuclear:
Consistently, the UniProt subcellular-location comments list only Cell membrane, Cytoplasm, and Cytoplasm/cell cortex — no nucleus. On the process side, GO:0051726 (regulation of cell cycle) is NOT annotated for this gene. The directly supported biological-process term is GO:0000281 mitotic cytokinesis [IMP:dictyBase], together with GO:1903665 negative regulation of asexual reproduction [IMP]. This evidence-code stratification is the single most decisive line for a curator: the seed hypothesis's nucleus term is carry-over inference, and its cell-cycle term does not exist in the curated record.
Targeted PubMed searches ("Dictyostelium PTEN nucleus development proliferation growth"; "PTEN nuclear localization amoeba lower eukaryote evolution conservation") returned zero papers describing nuclear PTEN in Dictyostelium. Across the full set of 21 papers reviewed, nuclear PTEN function is documented only in mammalian systems — heterochromatin maintenance (PMID: 25946202), mitotic-checkpoint-complex regulation (PMID: 28670501), CBX8/PRC1 interaction (PMID: 34592789), Plk1 crosstalk (PMID: 40117175), and Ndfip1/Nedd4-dependent import (PMID: 17218261; PMID: 22213801). The nuclear-PTEN literature is a mammalian-specific body of work built on an import mechanism that Dictyostelium PTEN lacks (Finding 3).
The consolidated evidence supports a single coherent, membrane-centric model for Dictyostelium PTEN with no nuclear or direct cell-cycle role:
ANTERIOR (front) POSTERIOR (rear/cortex)
┌───────────────────┐ ┌────────────────────────┐
│ PI3K active │ mutual exclusion │ PTEN active │
│ PIP2 → PIP3 ↑ │◄───────────────────►│ PIP3 → PIP2 (posterior)│
│ Actin, pseudopods │ (bistable switch) │ Myosin II, uropod │
└───────────────────┘ └────────────────────────┘
│
▼
Maintains cortical polarity
│
┌───────────────────────────┴───────────────┐
▼ ▼
Chemotaxis Cytokinesis (furrow)
│ │
(directed migration) pten⁻ → furrow fails,
nuclei keep duplicating
→ multinucleate cells
(GO:0000281 defect)
PTEN's immediate molecular activity is PIP3 3-phosphatase at the plasma membrane/cortex, recruited via a PI(4,5)P₂-binding motif. This activity partitions the membrane into a PIP3-high anterior and a PTEN/PIP2 posterior (mutual inhibition → bistability; PMID: 30367048). The resulting polarity is required both for directional chemotaxis and for assembling a functional cleavage furrow. Loss of PTEN elevates PIP3 and PKB-substrate phosphorylation (PMID: 21169559), disrupting furrow completion — while the nuclear-division cycle proceeds unabated, giving the classic multinucleate phenotype. Every arrow in this model is anchored in direct membrane/cortex data; none requires or predicts nuclear PTEN.
The mammalian nuclear-PTEN program (heterochromatin, MCC, Plk1, CBX8) is a later evolutionary elaboration built on a C-terminal tail and K289 monoubiquitination import switch that arose in the metazoan lineage. Because Dictyostelium PTEN lacks both, the nuclear paradigm is organism-specific and non-transferable.
| Citation (PMID) | Evidence type | Supports/Refutes/Qualifies | Claim tested | Key finding | Context | Confidence & limitations |
|---|---|---|---|---|---|---|
| 12062103 | Localization (direct, GFP) | Refutes nuclear; supports membrane | Where does Dicty PTEN localize? | PTEN-GFP at rear surface membrane; needs PI(4,5)P₂ motif | D. discoideum, live cells | High; overexpressed GFP fusion |
| 30367048 | Localization + single-molecule/modeling | Refutes nuclear; supports cortex | Membrane partitioning of PTEN | PTEN/PIP3 mutually exclusive on posterior/anterior membranes | D. discoideum motile cells | High |
| 15809030 | Mutant phenotype | Refutes cell-cycle; supports cytokinesis | Cell-cycle or cytokinesis defect? | pten⁻ fails cytokinesis; nuclei keep duplicating → multinucleate | D. discoideum | High; core discriminating evidence |
| 21169559 | Mutant/genetic epistasis | Qualifies; supports cytokinesis | Mechanism of pten⁻ defect | PKB-substrate phosphorylation is key downstream regulator; polarity for cytokinesis/chemotaxis | D. discoideum | High |
| 38940195 | Mutant phenotype | Qualifies (downstream) | Does PTEN affect proliferation? | PTEN needed for AprA-induced proliferation inhibition (population effect) | D. discoideum chemorepulsion | Medium; downstream/pathway, not cell-cycle machinery |
| 24292679 | Structure–function (heterologous) | Qualifies membrane mechanism | Membrane localization determinants | Membrane-binding regulatory interface; recruitment required for function | Human PTEN in Dictyostelium | High for membrane; uses human protein |
| 17218261 | Direct assay (mammalian) | Refutes transfer | Is K289 required for nuclear import? | K289E retains catalysis but fails nuclear accumulation (import defect) | Human, patient tissue | High; defines non-conserved determinant |
| 22213801 | In vivo (mouse) | Qualifies transfer | Nuclear-import machinery | Ndfip1/Nedd4-driven ubiquitination imports Pten | Mouse neurons | Medium; mammal-specific machinery |
| 25946202, 28670501, 34592789, 40117175 | Direct (mammalian) | Competing (organism-specific) | Nuclear PTEN roles | Heterochromatin, MCC, CBX8, Plk1 — all mammalian, tail/K289-dependent | Human/mouse | High but not transferable |
| This report — sequence analysis | Computational (evolutionary) | Refutes transfer | Is nuclear-import machinery conserved? | 45% identity; catalytic motif conserved; K289→E, C-tail absent in Dicty | Q8T9S7 vs P60484 | Low–medium at K289 (divergent loop); needs MSA/structure |
| UniProt Q8T9S7 / dictyBase | Review/database (evidence codes) | Refutes nuclear as experimental | What is the annotation basis? | Nucleus = IBA only; all IDA localization non-nuclear; GO:0051726 absent; GO:0000281 IMP present | Curated record | High for provenance |
GO:0005634 (nucleus, CC) — do NOT retain as experimentally supported; flag as IBA-only carry-over. The term has no IDA/IMP backing in Dictyostelium; it is Inferred from Biological Ancestor. Because the specific import determinant (K289) and C-terminal tail that justify nuclear localization in the ancestor-defining mammalian orthologs are not conserved, the IBA propagation is a false-transfer candidate. Curator action: consider a NOT qualifier or removal, and flag the GO_Central ortholog set for review.
GO:0051726 (regulation of cell cycle, BP) — do NOT add. Not curator-annotated; not supported by any Dictyostelium primary study. The phenotype is cytokinesis, not cell-cycle regulation.
GO:0000281 (mitotic cytokinesis, BP) — retain as the correct, IMP-supported process term (with GO:1903665 negative regulation of asexual reproduction). This accurately represents the pten⁻ phenotype.
Retain the IDA cellular-component terms already in place: GO:0005886 (plasma membrane), GO:0051285 (cell cortex of cell tip), GO:0032154 (cleavage furrow), GO:1990753 (equatorial cell cortex), GO:0031254/0031257 (trailing edge), GO:0005829 (cytosol), GO:0001931 (uropod).
Molecular-function anchor: the primary MF is phosphatidylinositol-3,4,5-trisphosphate 3-phosphatase activity — the conserved, transferable feature — not "protein binding."
The immediate molecular function tested is membrane-localized PIP3 3-phosphatase activity — dephosphorylation of PI(3,4,5)P₃ → PI(4,5)P₂ at the posterior plasma membrane/cortex, dependent on a PI(4,5)P₂-binding motif. Downstream/pathway consequences that must not be conflated with the seed terms: reduced PKB-substrate phosphorylation, maintained cortical polarity, functional cleavage-furrow assembly, directed chemotaxis, and — at the population level — AprA-mediated proliferation inhibition and chemorepulsion. The multinucleate phenotype is an inferred-from-loss-of-function cellular outcome of failed cytokinesis, not evidence of a direct cell-cycle-regulatory activity. The mammalian nuclear functions (heterochromatin, MCC/Plk1, CBX8) are context- and organism-specific roles that depend on trafficking machinery absent in Dictyostelium.
The seed hypothesis is refuted / over-annotated for Dictyostelium discoideum PTEN. Nuclear localization (GO:0005634) is unsupported by any direct evidence and rests on non-transferable phylogenetic inference; regulation of cell cycle (GO:0051726) is neither annotated nor phenotypically supported, the true defect being mitotic cytokinesis (GO:0000281). The protein's evidenced identity is a membrane/cortex-localized PIP3 3-phosphatase that maintains polarity for cytokinesis and chemotaxis.