AIGR Gene Hypothesis Deep Research — Drosophila melanogaster Mst27D (Q8IPI4)

Focus: function_assignment · Hypothesis slug: microtubule-plus-end-capacity-and-polymerization
Gene: Mst27D / CG31907 / FBgn0051907 (DROME, NCBITaxon:7227) · UniProt: Q8IPI4 · Length: 424 aa
Source: genes/DROME/Mst27D/Mst27D-ai-review.yaml

Summary

The seed hypothesis proposes that Mst27D (a) localizes to the microtubule plus end, (b) binds microtubule plus ends, and (c) regulates microtubule polymerization or depolymerization. Evaluated independently against primary experimental evidence, none of these three specific claims is directly supported. Every microtubule-related GO annotation on Q8IPI4 that asserts plus-end localization, plus-end binding, spindle localization, or regulation of polymer dynamics carries the evidence code IBA (Inferred from Biological Ancestor, GO_Central). These are computational propagations from the MAPRE/EB1 (RP/EB) family ancestral node — PANTHER family PTHR10623, InterPro IPR027328 (MAPRE) — and reflect the demonstrated behavior of canonical EB proteins, not of Mst27D itself.

The single primary functional study of the gene, Nuclear elongation during spermiogenesis depends on physical linkage of nuclear pore complexes to bundled microtubules by Drosophila Mst27D (Li, Messina & Lehner 2023, PLoS Genetics, PMID: 37428798), demonstrates a fundamentally different function. Its EB1-like N-terminal calponin-homology (CH) domain binds microtubules generically; it promotes microtubule bundling only under high/ectopic overexpression in cultured cells; its C-terminal region binds the nucleoporin Nup358; and its endogenous, physiological role is to physically link the nuclear-pore-bearing nuclear envelope (NPC-NE) to the "dense complex" microtubule bundle, driving spermatid nuclear elongation during postmeiotic spermiogenesis. This is a structural adaptor/cross-linking function, not plus-end tracking (+TIP behavior) and not regulation of microtubule dynamics.

Verdict: over-annotated / partially supported. What is genuinely supported is generic microtubule binding, a Nup358-binding activity, and a microtubule-bundling / nuclear-elongation role — all distinct from the plus-end and polymerization-dynamics claims of the seed. The plus-end, spindle, and dynamics-regulation terms are best treated as EB-family IBA carry-over (paralog over-annotation). Consistent with the seed's own caution about "inferring loss merely from specialization," we do not claim Mst27D cannot track plus ends; we conclude there is no positive evidence that it does, alongside strong positive evidence for a competing structural function.

Key Findings

Finding 1 — The plus-end / spindle / polymerization GO terms are EB1-family IBA over-annotation

UniProt Q8IPI4 (Mst27D, 424 aa) carries a cluster of microtubule-related GO terms whose evidence code is uniformly IBA (GO_Central), inherited from the MAPRE/EB1 family node. The specific terms include microtubule plus-end (GO:0035371, CC), microtubule plus-end binding (GO:0051010, MF), regulation of microtubule polymerization or depolymerization (GO:0031110, BP), spindle midzone (GO:0051233, CC), spindle assembly (GO:0051225, BP), and protein localization to microtubule (GO:0035372, BP). IBA annotations represent a hypothesis of conserved function propagated across a phylogenetic family, not a measurement of the gene in question. The family node (PANTHER PTHR10623, "MICROTUBULE-ASSOCIATED PROTEIN RP/EB FAMILY MEMBER"; InterPro IPR027328; Pfam PF00307 CH domain, residues ~26–129; C-terminal coiled-coil ~286–344) derives its reference plus-end/dynamics functions from canonical EB proteins such as EB1/MAPRE1–3, which genuinely track plus ends and modulate polymer dynamics.

The primary experimental record for Mst27D (PMID: 37428798) demonstrates only four activities, none corresponding to plus-end tracking or dynamics regulation. First, the N-terminal CH domain binds microtubules — the paper states "The N-terminal CH domain of Mst27D, which is similar to that of EB1 family proteins, binds to microtubules." This is characterized as EB1-like homology and generic lattice binding, not as a measured plus-end preference. Second, bundling occurs at high expression: "At high expression levels, Mst27D promotes bundling of microtubules in cultured cells" — an ectopic-overexpression readout, and a bundling (cross-linking) activity, not a polymerization-dynamics activity. Third, the C-terminus binds Nup358: "The C-terminal region of Mst27D binds to the nuclear pore protein Nup358." Fourth, the integrated function is structural linkage: "We demonstrate that Mst27D establishes physical linkage between NPC-NE and dense complex." Crucially, no assay in the paper demonstrates preferential plus-end tracking, plus-end-specific binding, spindle localization, or regulation of MT polymerization/depolymerization.

Finding 2 — Null-mutant phenotype and endogenous localization define a bundling / nuclear-elongation function

The Li, Messina & Lehner 2023 study (PLoS Genet 19(7):e1010837; PMCID PMC10359004) provides the loss-of-function and localization evidence that positively identifies the role. Endogenous localization is to microtubule bundles and the NPC/nuclear envelope, not to plus ends or the spindle: "Microscopic analyses indicated co-localization of Mst27D with Nup358 and with the microtubule bundles of the dense complex." This directly conflicts with the plus-end and spindle cellular-component claims. The loss-of-function phenotype genetically defines the requirement: "In Mst27D null mutants, this bundling process does not occur and nuclear elongation is abnormal." The requirement is for microtubule bundling and nuclear elongation, not for any dynamics-regulation readout. The authors' mechanistic model integrates these: "Mst27D permits normal nuclear elongation by promoting the attachment of the NPC-NE to the microtubules of the dense complex, as well as the progressive bundling of these microtubules."

Expression is spermatid-specific (postmeiotic), consistent with the "Mst" (male sterile testis) family naming and the spermiogenesis context, and it argues strongly against any mitotic/meiotic spindle role. The seed's instruction to distinguish endogenous postmeiotic expression from ectopic assays is material: the only microtubule-remodeling activity (bundling) was observed under ectopic overexpression in cultured cells, whereas the endogenous, physiologically relevant activity is structural linkage of NPCs to a pre-existing bundle. Importantly, the null-mutant loss of bundling elevates the bundling role from an overexpression-only artifact to a genetically supported in-vivo function.

Finding 3 — Phylogenetic placement: a divergent testis EB/MAPRE paralog lacking the +TIP tail

The Drosophila EB/RP family (PANTHER PTHR10623) contains multiple members: Eb1/CG3265 (canonical +TIP), CG32371, CG2955, CG15306, EB-SUN/CG18190 (a testis ER-associated EB), and Mst27D. A computed Needleman–Wunsch alignment (BLOSUM62) over the N-terminal CH region places Mst27D as a genuine but divergent paralog with no close sister — no family member exceeds ~45% CH-domain identity to Mst27D. Decisively, the C-terminal acidic EEY/F +TIP-recruitment tail is present only in canonical Eb1; Mst27D and the other divergent/testis paralogs lack it. The EEY/F tail is the canonical motif through which EB proteins recruit CAP-Gly +TIP partners (e.g., CLIP-170/CLIP-190, p150Glued); its absence, combined with the C-terminus being repurposed for Nup358 binding, is the structural signature of functional divergence away from the +TIP role.

In keeping with the seed's caution, we treat tail loss as supporting — not proving — the absence of plus-end tracking, since the CH domain can in principle bind plus ends autonomously and several Drosophila EB paralogs also lack the tail. The decisive evidence remains the endogenous localization (bundles/NE, not plus ends) and the null phenotype (bundling/elongation defect). The pattern nonetheless cleanly explains the IBA problem: the family node's reference plus-end/dynamics functions come from canonical Eb1-type proteins, and are being propagated onto specialized testis-branch members that have repurposed both termini.

Mechanistic Model / Interpretation

The evidence converges on a coherent, non-canonical model in which Mst27D is a bifunctional molecular adaptor/scaffold that mechanically couples the nuclear envelope to a cytoskeletal bundle during sperm-head shaping — a role orthogonal to the plus-end-dynamics function implied by the seed hypothesis.

   Spermatid (postmeiotic, testis-specific)
   ┌─────────────────────────────────────────────────────────────┐
   │  NUCLEAR ENVELOPE  (with Nucleoporins / NPCs, incl. Nup358)  │
   │        ▲                                                      │
   │        │  C-terminal region  ──► binds Nup358                │
   │   ┌────┴─────┐                                               │
   │   │  Mst27D  │  (424 aa; divergent EB/MAPRE paralog)         │
   │   └────┬─────┘                                               │
   │        │  N-terminal CH domain ──► binds microtubules       │
   │        ▼                                                      │
   │  ═══════════════  MICROTUBULE "DENSE COMPLEX" (bundle) ═══════│
   │        (progressive bundling; drives NUCLEAR ELONGATION)     │
   └─────────────────────────────────────────────────────────────┘

   Loss of Mst27D  ⇒  no bundling  ⇒  abnormal nuclear elongation

Direct molecular activities (supported): microtubule binding via the EB1-like N-terminal CH domain; nucleoporin (Nup358) binding via the C-terminal region; microtubule bundling (cross-linking) observed at high expression.

Cellular function (supported, in vivo): structural tethering of the NPC-bearing nuclear envelope to a bundled microtubule array, and contribution to progressive bundle formation.

Developmental outcome (supported, downstream): normal spermatid nuclear elongation during spermiogenesis; null mutants show abnormal nuclear elongation. This is the consequence of the linkage/bundling activity, not itself a molecular activity of Mst27D.

NOT supported by any Mst27D assay: plus-end tracking (+TIP behavior), plus-end-specific binding, spindle localization, or regulation of microtubule polymerization/depolymerization dynamics. The seed hypothesis conflates two distinct EB-family capabilities: generic lattice binding (a shared CH-domain property) versus plus-end tracking plus dynamics regulation (which requires additional determinants including the EEY/F tail and plus-end-recognition behavior). Mst27D retains the former and shows no evidence of the latter.

Evidence Matrix

GO Curation Implications (leads — require curator verification)

Note: avoid a bare "protein binding" recommendation — the microtubule partner supports GO:0008017 and the protein partner (Nup358) supports a named-interactor annotation.

Mechanistic Scope

The immediate molecular activities under test are plus-end binding and regulation of microtubule polymerization/depolymerization. The gene product's direct molecular activities are, in fact, (a) generic microtubule lattice binding via an EB1-like CH domain and (b) Nup358 binding via the C-terminus; bundling is a direct cross-linking property but manifests physiologically as a structural outcome (progressive bundle formation), not modulation of tubulin subunit kinetics. Everything about plus-end tracking and dynamics regulation belongs to the inferred/family-ancestral category, not the measured-activity category. The developmental outcome (nuclear elongation) is downstream of the direct linkage/bundling activity and is inferred from loss of function — a biological process the gene contributes to, but distinct from its direct molecular function. We deliberately avoid promoting the loss-of-function phenotype into a molecular-activity claim.

Conflicts and Alternatives

• Paralog over-annotation (primary conflict, and the correct interpretation): the IBA terms derive from canonical EB proteins (EB1/Eb1, MAPRE1–3) that genuinely track plus ends and modulate dynamics. Mst27D shares the CH domain (~38% identity to Dm EB1) but lacks the C-terminal acidic EEY/F tail EB proteins use to recruit CAP-Gly +TIP partners, and its C-terminus is instead repurposed for Nup358 binding. This is functional divergence, not conserved function — a textbook case of family IBA carry-over onto a specialized paralog.
• Ectopic vs endogenous / truncated vs full-length: bundling was observed at high expression in cultured cells using domain-resolved constructs. Endogenous Mst27D operates in postmeiotic spermatids on a largely static bundled array, arguing against a dynamics-regulation role; the null-mutant loss of bundling nonetheless confirms bundling is genuinely Mst27D-dependent in vivo. Because N- and C-terminal activities were mapped separately, the integrated full-length behavior (e.g., any cryptic plus-end preference) is inferred, not exhaustively tested.
• "Loss vs specialization": the seed correctly cautions against inferring loss from specialization. The conservative reading is that plus-end tracking is undemonstrated, not proven absent; the EEY-tail argument is supportive but not decisive. For curation, absence of positive evidence for a specific molecular claim, combined with strong positive evidence for a competing function and a contradictory endogenous localization, is sufficient to remove the unsupported specific terms.

Paralog comparison (computed this run — Drosophila EB/MAPRE family, PANTHER PTHR10623)

Interpretation: Mst27D is a genuine but divergent EB/MAPRE paralog (no family member >45% CH identity). Only canonical Eb1 retains the acidic EEY +TIP-recruitment tail. The IBA plus-end/dynamics annotations propagate from this family node, whose reference functions derive from canonical EB proteins (Eb1-type), not from the specialized testis-branch paralogs to which Mst27D belongs.

Evidence Base (Literature)

• Nuclear elongation during spermiogenesis depends on physical linkage of nuclear pore complexes to bundled microtubules by Drosophila Mst27D. Li, Messina & Lehner, 2023, PLoS Genetics 19(7):e1010837. PMID: 37428798; PMCID PMC10359004. The single primary characterization of Mst27D. Establishes CH-domain microtubule binding, C-terminal Nup358 binding, overexpression bundling, spermatid-specific bundle/NE localization, and a null-mutant nuclear-elongation phenotype. Directly refutes the plus-end/spindle/dynamics claims and defines the true linkage/bundling function.
• Synergistic actions of Nup43 and Myosin VI drive actin cone assembly during Drosophila spermiogenesis. PMID: 42374827. Contextual support that nucleoporins (Nup43) have non-canonical structural roles in Drosophila spermiogenesis, including nuclear elongation and shaping — consistent with the Mst27D–Nup358 linkage paradigm, though not direct Mst27D evidence.
• Database/orientation: UniProt Q8IPI4; PANTHER family PTHR10623; InterPro IPR027328 (MAPRE/EB1); Pfam PF00307 (CH). Used to establish that the microtubule GO terms are IBA family carry-over rather than direct evidence.

Limitations and Knowledge Gaps

1. No direct plus-end assay exists. Checked: full abstract and domain-mapping results of PMID 37428798. Why it matters: GO:0035371/GO:0051010 hinge on plus-end behavior; without a negative live-imaging result, these can be classed "unsupported" but not strictly "refuted." Resolve with: live TIRF comet-tracking of GFP-Mst27D on dynamic microtubules, with EB1 co-tracking.
2. Polymerization/depolymerization effect untested. Checked: no dynamics assay in the paper. Why it matters: GO:0031110 requires a measured effect on polymer dynamics. Resolve with: in-vitro turbidity/TIRF assays on purified tubulin ± Mst27D.
3. Sequence analyses are computational, not functional. Checked: CH-domain identity and EEY/F tail presence via Needleman–Wunsch/BLOSUM62. Why it matters: tail loss correlates with, but does not prove, loss of +TIP recruitment. Resolve with: +TIP-partner co-IP (CLIP-190/CLASP orthologs) with Mst27D.
4. Exact live GO annotation set not re-pulled in this run. Checked: summarized as six MT-related terms, all IBA, with specific IDs above. Why it matters: term IDs and any non-IBA terms must be confirmed against current GOA/QuickGO before edits. Resolve with: live QuickGO/GOA pull for Q8IPI4.
5. High-resolution endogenous CC assignment. Checked: co-localization with Nup358 and MT bundles shown. Why it matters: precise NPC-vs-bundle CC assignment affects term choice. Resolve with: immuno-EM.

Discriminating Tests

• Live +TIP comet assay in S2 cells or spermatids: does GFP-Mst27D form growing-end comets and co-migrate with EB1? Distinguishes plus-end tracking from lattice binding.
• In-vitro reconstitution with purified tubulin ± Mst27D, and an EEY-tail chimera, to test whether restoring the acidic tail confers +TIP behavior and whether Mst27D alters growth/catastrophe kinetics.
• Domain-swap / ΔNup358-binding mutants in vivo to confirm the linkage function is separable from any MT-dynamics effect.
• Phylogenetic/synteny analysis across Drosophilids to date the EB-paralog duplication and confirm tail loss is ancestral to the Mst27D lineage.
• Live GOA/QuickGO pull to enumerate exact GO term IDs and evidence codes for scoping edits.

Curation Leads (require curator verification)

• Reference to cite: PMID 37428798 (Li, Messina & Lehner, 2023). Snippets to verify verbatim:
• "The N-terminal CH domain of Mst27D, which is similar to that of EB1 family proteins, binds to microtubules."
• "At high expression levels, Mst27D promotes bundling of microtubules in cultured cells."
• "The C-terminal region of Mst27D binds to the nuclear pore protein Nup358."
• "Microscopic analyses indicated co-localization of Mst27D with Nup358 and with the microtubule bundles of the dense complex."
• "In Mst27D null mutants, this bundling process does not occur and nuclear elongation is abnormal."
• "We demonstrate that Mst27D establishes physical linkage between NPC-NE and dense complex."
• Action changes: generalize plus-end binding → microtubule binding (GO:0008017); remove the plus-end (GO:0035371), spindle-midzone (GO:0051233), spindle-assembly (GO:0051225), and MT-polymerization-regulation (GO:0031110) IBA terms as unsupported paralog carry-over; add IDA/IPI/IMP leads for microtubule binding, Nup358 binding, microtubule bundle formation (GO:0001578), and spermatid nuclear elongation.
• Suggested question for curator: Should EB-family IBA terms be retained on a spermatid-specialized paralog that lacks the diagnostic EEY/F +TIP tail and has an experimentally different (Nup358-linkage) function?
• Suggested experiments: live +TIP comet-tracking and in-vitro dynamics assays (above).

Proposed Follow-up Experiments / Actions

1. Curator action: pull the live QuickGO/GOA record for Q8IPI4, confirm the six IBA MT term IDs, and stage removals with the verified refuting/qualifying snippets above.
2. Curator action: draft additions for microtubule binding (retain/upgrade), Nup358 binding, microtubule bundle formation, microtubule bundle localization, and spermatid nuclear elongation, each tied to PMID 37428798.
3. Wet-lab (definitive refutation): EB-comet live imaging and in-vitro TIRF dynamics assays to convert "unsupported" into "directly refuted" for the plus-end/dynamics terms.
4. Comparative genomics: formalize the PTHR10623 tree with tail-motif annotation to document, for the record, that only canonical Eb1 retains the EEY/F +TIP tail — provenance for the IBA-over-annotation determination.

Bottom line

Mst27D's plus-end localization, plus-end binding, spindle, and polymerization-regulation GO terms are EB1/MAPRE-family IBA carry-over, not evidence about Mst27D. The gene's demonstrated function is structural: an EB1-like CH domain binds microtubules generically, the C-terminus binds Nup358, and the protein physically links nuclear pore complexes to a bundled spermatid microtubule "dense complex" to drive nuclear elongation. Generalize plus-end binding to microtubule binding (GO:0008017), remove the unsupported plus-end/spindle/dynamics IBA terms, and add the supported Nup358-binding and spermatid-nuclear-elongation functions — all pending curator verification.

Provenance: GO decision table and CH-domain alignment computed this run (Needleman–Wunsch/BLOSUM62; Mst27D CH vs Dm EB1 Q9XZ57 = 37.9% identity; EEY/F tail absent). Paralog table computed over PANTHER PTHR10623 members. Sequences and domains from UniProt Q8IPI4 and family accessions.

Artifacts

• OpenScientist final report
• OpenScientist final report