MYMX (myomixer, also called myomerger or Minion) is an 84-amino-acid, muscle-specific microprotein encoded by a small open reading frame that was originally mis-annotated as a long non-coding RNA. It is a single-pass plasma membrane protein: an N-terminal hydrophobic helix anchors it in the bilayer while two short helices of its C-terminal ectodomain face the outside of the cell. MYMX is expressed transiently in differentiating myoblasts and in activated satellite cells, and together with the seven-pass membrane protein myomaker (MYMK) it drives the fusion of mononucleated muscle cells into multinucleated myofibers. The two fusogens act at different steps: MYMK brings the two bilayers to hemifusion, and the MYMX ectodomain then stresses the outer leaflet of the apposed membranes - shifting its spontaneous curvature to positive values in a phosphatidylserine-dependent manner - so that the hemifusion diaphragm resolves into an expanding fusion pore. MYMX activity is required on only one of the two fusing cells and can operate without a physical MYMK interaction. Loss of the protein leaves myoblasts able to differentiate and assemble sarcomeres but unable to fuse, so mouse and zebrafish nulls lack multinucleated myofibers, and conditional deletion in adult satellite cells prevents skeletal muscle regeneration after injury. In humans, homozygosity for a truncating variant that removes the ectodomain causes an autosomal recessive congenital myopathy resembling Carey-Fineman-Ziter syndrome. MYMX arose de novo in early vertebrates and, despite weak sequence conservation, retains fusogenic function across fish, reptiles and mammals.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0000139 Golgi membrane | IBA GO_REF:0000033 | MARK AS OVER ANNOTATED | Summary: Golgi membrane as a site of MYMX activity rests ultimately on one crude fractionation experiment in mouse, in which the microprotein was recovered in a membrane pellet that pooled plasma membrane, ER and Golgi. That fraction cannot discriminate among the three compartments, and every functional experiment places the working protein at the cell surface with its ectodomain outside. Reason: The PAINT node inherits the mouse Myomixer/Myomerger localisation annotations, and among them the ER/Golgi calls derive from a mixed membrane fraction (PMID:28569745) rather than from imaging or marker co-localisation. Because MYMX acts on the outer leaflet of the plasma membrane from the extracellular side (PMID:30197239), a Golgi pool would at most represent biosynthetic transit and is not a site where the protein is active. The is_active_in qualifier therefore overstates the evidence. Not removed outright, since a single-pass membrane protein does traverse the secretory pathway. Propagation Review Root cause: SOURCE WEAK OR INFERRED Failure modes: COMPARTMENT OR COMPLEX MISMATCH SOURCE EVIDENCE WEAK Sources checked: MGI:MGI:3649059 · Mymx (mouse) SOURCE WEAK OR INFERRED The mouse localisation evidence for this compartment is a membrane pellet that pooled plasma membrane, ER and Golgi (PMID:28569745); no imaging or marker co-localisation resolves the compartments. PANTHER:PTN002222898 · MYMX family node SUPPORTS SOURCE BUT NOT TARGET The node placement is sound - human MYMX lies inside the vertebrate MYMX clade - but the compartment claim propagated from it is far weaker than the plasma membrane claim carried by the same node. Supporting Evidence: PMID:28569745 Subcellular fractionation did however confirm significant enrichment within the membrane-associated fraction containing plasma membrane, ER and Golgi, suggesting insertion into or association with a membrane compartment PMID:30197239 Immunoblotting of the isolated proteins with Myomerger antibodies showed that Myomerger is normally expressed on the cell surface and suggested that the non-transmembrane region of Myomerger is extracellular |
| GO:0000139 Golgi membrane | IEA GO_REF:0000107 | MARK AS OVER ANNOTATED | Summary: Ensembl-Compara transfer of the same mouse Golgi membrane annotation. The orthology call is correct, but the underlying mouse evidence is a mixed plasma membrane/ER/Golgi fraction, so the compartment assignment is not supported at this resolution. Reason: Orthology transfer faithfully reproduces the mouse annotation; the weakness is in the source evidence rather than the projection. located_in is a weaker claim than is_active_in and is defensible as secretory-pathway transit, but it should not be read as a functional location for a protein that works from the cell surface (PMID:30197239). Supporting Evidence: PMID:28569745 Subcellular fractionation did however confirm significant enrichment within the membrane-associated fraction containing plasma membrane, ER and Golgi, suggesting insertion into or association with a membrane compartment |
| GO:0005789 endoplasmic reticulum membrane | IBA GO_REF:0000033 | MARK AS OVER ANNOTATED | Summary: As for the Golgi call, the ER membrane annotation traces to a membrane fraction that contained plasma membrane, ER and Golgi together. MYMX is not known to act in the ER, and unlike the SERCA-regulating muscle micropeptides it has no reported sarcoplasmic/ER target. Reason: No experiment localises MYMX to the ER specifically; the protein resists high-salt and alkaline extraction and is solubilised only by detergent, consistent with integral membrane insertion, but the compartment assayed functionally is the plasma membrane (PMID:30197239). is_active_in ER membrane asserts a site of action for which there is no evidence. Propagation Review Root cause: SOURCE WEAK OR INFERRED Failure modes: COMPARTMENT OR COMPLEX MISMATCH SOURCE EVIDENCE WEAK Sources checked: MGI:MGI:3649059 · Mymx (mouse) SOURCE WEAK OR INFERRED The mouse localisation evidence for this compartment is a membrane pellet that pooled plasma membrane, ER and Golgi (PMID:28569745); no imaging or marker co-localisation resolves the compartments. PANTHER:PTN002222898 · MYMX family node SUPPORTS SOURCE BUT NOT TARGET The node placement is sound - human MYMX lies inside the vertebrate MYMX clade - but the compartment claim propagated from it is far weaker than the plasma membrane claim carried by the same node. Supporting Evidence: PMID:28569745 Subcellular fractionation did however confirm significant enrichment within the membrane-associated fraction containing plasma membrane, ER and Golgi, suggesting insertion into or association with a membrane compartment PMID:30197239 Myomerger was only solubilized with detergent indicating that this protein is tightly associated with the membrane. |
| GO:0005789 endoplasmic reticulum membrane | IEA GO_REF:0000107 | MARK AS OVER ANNOTATED | Summary: Ensembl-Compara projection of the mouse ER membrane annotation, which itself comes from a mixed membrane fraction. Reason: Same limitation as the mouse source annotation: the fraction pooled plasma membrane, ER and Golgi (PMID:28569745). Retained as a weak located_in claim compatible with secretory transit, flagged so that it is not treated as a functional compartment. Supporting Evidence: PMID:28569745 Subcellular fractionation did however confirm significant enrichment within the membrane-associated fraction containing plasma membrane, ER and Golgi, suggesting insertion into or association with a membrane compartment |
| GO:0005886 plasma membrane | IBA GO_REF:0000033 | ACCEPT | Summary: Correct, and the most informative location annotation on this gene. MYMX is a single-pass plasma membrane protein whose ectodomain faces the extracellular space, and its fusogenic activity is exerted on the outer leaflet of the plasma membrane, so is_active_in is exactly the right qualifier. Reason: The PAINT node placement is well supported: mouse Myomixer is detected at the cell surface by immunostaining of intact myoblasts and by cell-impermeable surface biotinylation, and recombinant ectodomain added outside the cell rescues fusion of null myoblasts, which only makes sense for a surface-acting protein. MYMX orthologs from fish and reptiles are fusogenic in mammalian cells, so the function - and with it the surface localisation - is inherited across the clade the IBD covers. Supporting Evidence: PMID:28386024 immunostaining of intact C2C12 myoblasts expressing Myomixer with FLAG tag at the C terminus revealed the presence of the protein on the cell surface PMID:30197239 myoblast fusion depends on the interactions between the ectodomain of Myomerger and the plasma membrane of the myoblasts |
| GO:0005886 plasma membrane | IEA GO_REF:0000120 | ACCEPT | Summary: Combined automatic annotation reaching the correct compartment. Redundant with the IBA and ISS plasma membrane calls but not wrong. Reason: The UniProt subcellular-location assignment (single-pass cell membrane protein) matches the experimental topology work in mouse, and the human protein is the direct ortholog of the characterised mouse protein with the same length and the same N-terminal hydrophobic segment. Supporting Evidence: PMID:35642635 MYMX is a single-pass transmembrane micropeptide that is required on only 1 of 2 fusing cells |
| GO:0005886 plasma membrane | ISS GO_REF:0000024 | ACCEPT | Summary: Curator-judged sequence-similarity transfer from mouse Myomixer (Q2Q5T5). The human and mouse proteins are both 84 residues with a conserved N-terminal membrane anchor, and human MYMX is fusogenic in cells, which requires surface presentation. Reason: The transfer is between close orthologs whose functional equivalence has been demonstrated directly: human MINION/MYMX ORFs rescue fusion in Minion-knockdown mouse myoblasts, and the human protein cooperates with MYMK in heterologous fusion assays. Supporting Evidence: PMID:28569745 both untagged and C-terminally epitope-tagged human MINION ORFs strongly reconstituted cell fusion in MinionKD cells PMID:35642635 In the presence of MYMK, MYMX WT strongly enhanced heterologous fusion, whereas the MYMX R46* variant had minimal fusogenicity |
| GO:0007520 myoblast fusion | IBA GO_REF:0000033 | ACCEPT | Summary: The central process annotation for this gene, at the right level of specificity. MYMX is required for myoblast fusion in mouse, zebrafish and human cells, and its appearance in its own WITH/FROM list reflects the fact that human MYMX itself carries experimental evidence that the PAINT curator used in placing the IBD. Reason: The function is ancestral to the whole vertebrate MYMX family: deletion abolishes fusion in mouse and in zebrafish, and orthologs from fish and reptiles substitute for the mammalian protein. The human gene sits squarely inside that clade and its loss-of-function variant causes a human fusion-deficiency myopathy, so there is no case for lineage-specific loss. Supporting Evidence: PMID:28386024 Disruption of the Myomixer gene prevented fusion of C2C12 cells and primary mouse myoblasts but did not affect expression of myosin heavy chain, a marker for differentiation PMID:29078404 abolishes myoblast fusion in vivo |
| GO:0007520 myoblast fusion | IEA GO_REF:0000120 | ACCEPT | Summary: Automatic annotation via the MYMX-family InterPro signature and orthology to mouse. Correct; the family is small, muscle-specific and functionally homogeneous. Reason: IPR039014 corresponds to the MYMX family itself rather than to a promiscuous domain, so the InterPro2GO mapping to myoblast fusion is safe, and it agrees with the experimental human evidence. Supporting Evidence: PMID:35642635 Myoblasts generated from patient-derived induced pluripotent stem cells displayed defective fusion |
| GO:0007520 myoblast fusion | IGI PMID:28569745 The microprotein Minion controls cell fusion and muscle form... | ACCEPT | Summary: Genetic-interaction annotation from the Minion paper, in which the human ORF complemented the fusion defect of Minion-knockdown mouse myoblasts and frameshift versions of the same cDNA did not. This is direct functional evidence that the human microprotein, as a protein rather than as an RNA, participates in myoblast fusion. Reason: The cross-species complementation design is why the annotation is IGI with mouse Myomixer as the interacting entity. The frameshift controls exclude a non-coding-RNA explanation, which matters here because the locus was originally annotated as a lncRNA. Supporting Evidence: PMID:28569745 both untagged and C-terminally epitope-tagged human MINION ORFs strongly reconstituted cell fusion in MinionKD cells PMID:28569745 These frameshift point mutants failed to complement the fusion defect |
| GO:0007520 myoblast fusion | ISS GO_REF:0000024 | ACCEPT | Summary: Sequence-similarity transfer of myoblast fusion from mouse Myomixer. Redundant with the IGI and IBA annotations but independently sound. Reason: Human and mouse proteins are orthologous 84-residue micropeptides, and the human protein has been shown experimentally to be fusogenic, so the inference is confirmed rather than merely plausible. Supporting Evidence: PMID:35642635 In the presence of MYMK, MYMX WT strongly enhanced heterologous fusion, whereas the MYMX R46* variant had minimal fusogenicity |
| GO:0014905 myoblast fusion involved in skeletal muscle regeneration | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Orthology transfer of the mouse regeneration annotation, which rests on tamoxifen-inducible deletion of Mymx in Pax7-positive satellite cells: satellite cell fusion and muscle regeneration after injury are abolished. Human MYMX is expressed in regenerative myogenesis and the orthology is unambiguous, so the projection is reasonable. Reason: This is the same molecular activity as GO:0007520 applied in the adult injury context; it is a physiological setting for the core function rather than a second function. Retaining it is useful because it records the satellite-cell requirement, but the core annotation is plain myoblast fusion. Direct human evidence for post-injury regeneration is absent, so it is kept as non-core rather than accepted as core. Supporting Evidence: PMID:29581287 abolishes satellite cell fusion and prevents muscle regeneration PMID:35642635 Similarly, deletion of Mymx or Mymk in adult satellite cells prevents skeletal muscle regeneration in response to injury (13, 14). |
| GO:0014905 myoblast fusion involved in skeletal muscle regeneration | IMP PMID:35642635 Impaired activity of the fusogenic micropeptide Myomixer cau... | MODIFY | Summary: The cited paper is the human CFZS2 study, and it is the right paper for a human MYMX loss-of-function process annotation - but its experiments are developmental, not regenerative. Patient-derived iPSC myoblasts fail to fuse, knock-in mice carrying the human R46* allele die perinatally from absent multinucleated myofibers, and the truncated protein is stable but poorly fusogenic. Regeneration appears only in the introduction, citing the earlier mouse satellite-cell work. Reason: The experimental basis supports the parent term GO:0007520 myoblast fusion, in a developmental setting, and not the regeneration-restricted child. No injury model, satellite cell assay or adult regeneration experiment is reported here. This is a term-choice correction, not a challenge to the curator's reading of the gene: the same paper amply supports involvement in myoblast fusion, and the regeneration claim is separately carried by the mouse-derived IEA and ISS rows. Proposed replacements: myoblast fusion Supporting Evidence: PMID:35642635 To assess muscle abnormalities in vitro, skeletal muscle cells were differentiated from iPSCs derived from gingival fibroblasts from the male patient PMID:35642635 the myogenic regulator MYOG, the differentiation marker DES, and the muscle fusogens MYMK and MYMX were expressed at equivalent levels in MYMX R46* myotubes, despite their inability to fuse PMID:35642635 However, 40% of Mymx R46*/R46* mice died shortly after birth, and none survived beyond 14 days of age |
| GO:0014905 myoblast fusion involved in skeletal muscle regeneration | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: Curator sequence-similarity transfer of the mouse satellite-cell regeneration phenotype. Biologically expected for the human ortholog, which is expressed during regenerative myogenesis, but it is the same activity in an adult context rather than a distinct function. Reason: Kept for completeness of the regeneration record while the core annotation remains GO:0007520 myoblast fusion. The mouse evidence is strong and conditional (satellite-cell specific), so there is no reason to doubt the transfer itself. Supporting Evidence: PMID:29581287 abolishes satellite cell fusion and prevents muscle regeneration |
| GO:0043403 skeletal muscle tissue regeneration | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Tissue-level consequence of the fusion defect: without MYMX, satellite cells cannot fuse with injured myofibers, so regeneration fails and the muscle degenerates. True, but it describes the outcome of the fusion step rather than a separate MYMX activity. Reason: MYMX contributes one membrane-remodelling step; the regeneration programme involves satellite cell activation, proliferation, migration and maturation in which MYMX has no reported role. Retaining the term at non-core status records the requirement without implying that MYMX drives the whole process. Supporting Evidence: PMID:29581287 abolishes satellite cell fusion and prevents muscle regeneration |
| GO:0043403 skeletal muscle tissue regeneration | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: Sequence-similarity transfer of the same tissue-level regeneration requirement from mouse. Reason: Same reasoning as the Ensembl-projected row: a downstream tissue phenotype of the fusion block, correct but not a core function. The human protein is expressed in regenerating muscle contexts and is the functional equivalent of the mouse protein. Supporting Evidence: PMID:29581287 abolishes satellite cell fusion and prevents muscle regeneration |
| GO:0060538 skeletal muscle organ development | IBA GO_REF:0000033 | KEEP AS NON CORE | Summary: Broad developmental term reflecting the fact that MYMX nulls lack multinucleated myofibers and die perinatally. Correct but general: the specific contribution is the fusion step, already captured by GO:0007520, and differentiation and sarcomere assembly proceed normally without MYMX. Reason: Myogenic differentiation markers, sarcomere organisation and MYMK expression are unaffected in MYMX-deficient cells, so the gene is not a general regulator of muscle organ development; the organ-level phenotype is entirely downstream of failed fusion. The term is kept as non-core context rather than modified, since it is a legitimate ancestor-level statement for an essential myogenic gene. Supporting Evidence: PMID:28386024 Histological sections through limb, body wall, and diaphragm musculature revealed an absence of multinucleated myofibers in Myomixer KO embryos PMID:28569755 Disruption of myomerger in myoblast cell lines and in the mouse, through Cas9-mutagenesis, generates non-fusogenic myocytes. |
| GO:0060538 skeletal muscle organ development | IEA GO_REF:0000120 | KEEP AS NON CORE | Summary: Automatic annotation from the MYMX-family signature and mouse orthology; the same broad developmental statement as the IBA row. Reason: Accurate at the organ level - human MYMX loss of function produces a congenital myopathy with facial and axial muscle weakness - but too general to represent the gene's activity. Supporting Evidence: PMID:35642635 The premature stop codon at position 46 (referred to hereafter as R46*) results in loss of most of the hydrophobic extracellular ectodomain of MYMX, including the essential AxLyCxL motif |
| GO:0060538 skeletal muscle organ development | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: Sequence-similarity transfer of the mouse muscle-development requirement. Reason: The transfer is sound but the term is an ancestor-level description of the fusion defect's consequences rather than a statement about what MYMX does. Supporting Evidence: PMID:28569755 We show that myomerger is exclusively expressed in skeletal muscle only during developmental and regenerative myogenesis. |
| GO:0140522 fusogenic activity | IDA PMID:30197239 Myomaker and Myomerger Work Independently to Control Distinc... | NEW | Summary: GOA gives MYMX no molecular function at all, in either human or mouse, yet the protein is one of the few characterised non-viral fusogens. Its ectodomain, added in trans as a purified protein, restores fusion of MYMX-null myoblasts; it drives fusion pore formation and completion in a heterologous influenza-hemagglutinin system where MYMK is absent; and the isolated ectodomain generates positive spontaneous curvature in lipid monolayers, the physical change that resolves a hemifusion diaphragm into a pore. GO:0140522 fusogenic activity - the activity of joining two lipid bilayers to form a single membrane - is the closest existing term. Reason: Participation test: MYMX itself performs the step. The purified ectodomain acting from outside the cell is sufficient to convert hemifusion intermediates into complete fusion, and it does so in a system driven by an unrelated fusogen, so MYMX is not merely required for, or a substrate of, someone else's activity - it supplies the membrane-stressing activity that opens the pore. Comparator check: GO:0140522 and its children currently hold very few experimental annotations (for example Chlamydomonas HAP2 by IDA), and neither mouse Mymx nor Mymk carries any molecular function beyond protein binding. The absence is therefore not a convention excluding partial fusogens but a young branch of the ontology that has not been applied to the myoblast fusion machinery; no descendant of GO:0140522 fits MYMX, since the existing children specify GTPase-dependent, refolding-mediated and helical-bundle-zippering mechanisms. GO:0140912 membrane destabilizing activity was considered and rejected: its definition centres on increasing permeability and lysis, which is a side effect of MYMX overexpression rather than its function. The term is proposed at the parent level and the question of whether GO should add a pore-formation-specific child is raised in suggested_questions. Supporting Evidence: PMID:30197239 This finding demonstrates that Myomerger drives fusion pore formation and fusion completion in a Myomaker-independent manner. PMID:30197239 myoblast fusion depends on the interactions between the ectodomain of Myomerger and the plasma membrane of the myoblasts PMID:33479215 The extracellularly localized region (ectodomain) of Myomerger drives opening of a fusion pore, as demonstrated for myoblast fusion |
Loading supporting content…
Download this section (compressed HTML)Q: Should GO acquire a child of GO:0140522 fusogenic activity for proteins that execute only the hemifusion-to-pore transition rather than the whole bilayer merger? MYMX cannot fuse membranes by itself in the absence of a hemifusion intermediate, so the parent term arguably overstates its activity, while no existing child matches its mechanism.
Suggested experts: Douglas P Millay, Leonid V Chernomordik
Q: Should the MYMX molecular function be modelled as membrane-curvature generation rather than as fusogenic activity? The measured biophysical property is a positive shift in outer-leaflet spontaneous curvature, which GO currently has no molecular-function term for.
Suggested experts: Michael M Kozlov, Leonid V Chernomordik
Q: Does human MYMX have a role in adult muscle regeneration comparable to the mouse satellite-cell requirement? All regeneration annotations on the human gene are transferred from mouse; the human patients described so far present with a congenital, developmental phenotype.
Suggested experts: Eric N Olson, Douglas P Millay
Q: Is MYMX present in the ER and Golgi beyond biosynthetic transit, and is any of its activity exerted there? The ER and Golgi annotations derive from a membrane fraction that pooled all three compartments, and no imaging has resolved them.
Suggested experts: Srihari C Sampath, Douglas P Millay
Experiment: Reconstitute hemifusion between protein-free liposomes of defined composition using polyethylene glycol or a minimal viral fusion peptide, then titrate recombinant human MYMX ectodomain (residues 26-84) and measure content mixing versus lipid mixing by dual fluorescent probes. Include the R46* truncation and phosphatidylserine-free liposomes as negative controls. A cell-free demonstration would establish fusogenic activity for the human protein directly rather than by orthology.
Hypothesis: The purified MYMX ectodomain is sufficient to complete fusion of any pair of membranes already held at hemifusion, independently of cell type or of the protein that produced the hemifusion intermediate.
Type: in vitro liposome fusion assay with recombinant ectodomain
Experiment: Generate MYMX-null and MYMX R46* human iPSC-derived or primary myoblast lines, engraft them into immunodeficient mice with cardiotoxin-injured muscle, and quantify donor-derived myonuclei in regenerated fibers. Rescue with wild-type MYMX and with base-edited R46* alleles. This tests the orthology-based regeneration annotations directly in human cells.
Hypothesis: Human MYMX is required for satellite-cell-mediated regeneration of adult human muscle, not only for developmental myogenesis.
Type: xenograft regeneration assay
Experiment: Combine surface biotinylation, protease protection of intact cells, and live-cell imaging of an extracellularly tagged MYMX with brefeldin A block and retention-using-selective-hooks release, to measure what fraction of the protein is ever resident in ER or Golgi and whether trapping it there affects fusion. A negative result would justify removing the ER and Golgi annotations rather than flagging them.
Hypothesis: MYMX resides at the plasma membrane and passes through the ER and Golgi only in transit, with no activity in those compartments.
Type: trafficking and topology analysis
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)