MYMX (myomixer / Minion / myomerger) review notes

UniProt A0A1B0GTQ4 (MYMX_HUMAN), 84 aa, HGNC:52391, chr6. Previously annotated as a lncRNA
(LOC101929726 / RP1-302G2.5). Member of the MICROPROTEINS project.

2026-09-30 session (claude-code)

Identity and discovery

• Three simultaneous 2017 discoveries of the same mouse gene (Gm7325):
• Bi et al. Science 2017 PMID:28386024 (CRISPR screen in C2C12).
• Quinn et al. Nat Commun 2017 PMID:28569755
• Zhang et al. Nat Commun 2017 (Minion) PMID:28569745. Note: the task brief said "Zhang et al. 2017 Genes Dev"; PubMed shows the Minion paper is in Nat Commun (PMID:28569745), not Genes Dev.
• Human ORF first identified by Zhang et al. PMID:28569745, and shown to be protein-coding and functional PMID:28569745; frameshifts abolish rescue, so the activity is from the protein, not the RNA.

Topology / location

• Single-pass membrane protein; N-terminal hydrophobic helix anchors in plasma membrane, C-terminal region extracellular (ectodomain) PMID:30197239.
• Bi 2017: PMID:28386024.
• ER/Golgi: the only primary evidence is a crude fractionation PMID:28569745 — this fraction does not discriminate among the three compartments. It is the likely source of the mouse ER/Golgi IDA that seeds the IBA and Ensembl-Compara IEA. ER/Golgi presence would be biosynthetic transit at most; the functional site is the cell surface. -> over-annotation.

Mechanism (the MF question)

• Myomaker (MYMK, 7-TM) drives hemifusion; MYMX drives pore formation PMID:30197239.
• MYMX works independently of MYMK in a heterologous HA-driven hemifusion system PMID:30197239.
• Recombinant ectodomain (aa 26-84) rescues fusion of Mymx-null myoblasts in trans PMID:30197239.
• Biophysics: ectodomain induces positive spontaneous curvature in the outer leaflet PMID:33479215; two ectodomain helices regulated by externalized PS PMID:36095199.
• Required on only one of the two fusing cells PMID:35642635.
• Co-IP with Myomaker PMID:28386024 but interaction not required for function PMID:30197239. So "protein binding"/MYMK binding is not the core MF.
• MF term: GO:0140522 fusogenic activity ("The activity of joining two lipid bilayers to form a single membrane.") is the closest existing term. None of its children fit (GTPase-dependent, refolding-mediated, helical-bundle zippering). GO:0140912 membrane destabilizing activity is defined around permeabilisation/lysis — not a good fit.
• Participation test: MYMX itself does the work of the pore-opening step (its ectodomain stresses the outer leaflet); it is not a substrate. Passes.
• Comparator check (QuickGO 2026-09-30): GO:0140522 has very few experimental annotations (e.g. Chlamydomonas HAP2 IDA). Mouse Mymx (Q2Q5T5) and Mymk carry no MF other than protein binding (IPI) — i.e. no MF at all has been assigned; the absence reflects the recency of the fusogen MF branch rather than a deliberate convention excluding MYMX. Proposed as NEW with caution; flagged as a question for experts whether a partial (pore-completing) fusogen should carry the parent term.

Processes

• Mouse KO: no multinucleated myofibers, perinatal lethality PMID:28386024.
• Zebrafish KO abolishes fusion PMID:29078404.
• Satellite cell cKO abolishes regeneration PMID:29581287 — this is the mouse IMP source for GO:0014905/GO:0043403 propagated to human.
• Human disease: CFZS2 (MIM:619941), R46* truncation removes ectodomain incl. AxLyCxL motif PMID:35642635. Patient iPSC myoblasts fail to fuse; knock-in mice die perinatally.
• PMID:35642635 IMP was annotated to GO:0014905 (myoblast fusion involved in skeletal muscle regeneration). Full text checked: no injury/regeneration experiments; "regenerat" appears only in introduction citing earlier work. Experiments are iPSC-derived myoblast fusion, developmental muscle phenotype of knock-in mice, and heterologous fusion. -> MODIFY to GO:0007520.
• Evolution: de novo vertebrate gene PMID:36054355; poorly conserved sequence, fish orthologs ~62-75 aa but functional.

Microprotein-project observations

• Classic smORF misannotated as lncRNA; rescue with frameshift controls is the gold standard coding proof.
• A microprotein can be a bona fide MF-bearing effector (membrane-remodelling ectodomain), not only a regulator of a larger partner — contrasts with SERCA-regulating micropeptides (SLN, PLN, MRLN, DWORF).
• GO currently gives MYMX no MF; fusogenic activity is the gap.

Action summary

• PM (IBA, IEA, ISS): ACCEPT
• Golgi/ER membrane (IBA x2, IEA x2): MARK_AS_OVER_ANNOTATED
• myoblast fusion (IBA, IEA, IGI, ISS): ACCEPT
• GO:0014905 IEA, ISS: ACCEPT; IMP PMID:35642635: MODIFY -> GO:0007520
• GO:0043403, GO:0060538 (all): KEEP_AS_NON_CORE
• NEW: GO:0140522 fusogenic activity