# MVB12A review notes

## Scope

MVB12A is reviewed in the PN ESCRT-I branch. PN entries without PMIDs were used as context only. Unlike VPS37C/D, MVB12A has direct ESCRT-I composition, membrane-binding, EGFR, and HIV-budding evidence. The review keeps viral budding and virus maturation as supported non-core contexts for proteostasis, while treating endosomal ESCRT-I cargo sorting as the core cellular function.

## Evidence synthesis

MVB12A is a metazoan fourth subunit of ESCRT-I. UniProt describes it as a "Component of the ESCRT-I complex" and says it is "Required for the sorting of endocytic ubiquitinated cargos into multivesicular bodies" [file:human/MVB12A/MVB12A-uniprot.txt, "Component of the ESCRT-I complex"; file:human/MVB12A/MVB12A-uniprot.txt, "Required for the sorting of endocytic"]. UniProt also states that human ESCRT-I "consists of TSG101, VPS28, a VPS37 protein" plus MVB12A or MVB12B, and that MVB12A interacts with TSG101, VPS28, VPS37B, VPS37C, VPS37D, and CEP55 [file:human/MVB12A/MVB12A-uniprot.txt, "which consists of TSG101, VPS28, a VPS37"; file:human/MVB12A/MVB12A-uniprot.txt, "Interacts with TSG101"; file:human/MVB12A/MVB12A-uniprot.txt, "Interacts with VPS28"; file:human/MVB12A/MVB12A-uniprot.txt, "Interacts with VPS37B"].

The main ESCRT-I/MVB12 paper directly supports MVB12A as an ESCRT-I subunit and viral-budding regulator. It reports that MVB12A and MVB12B "constitute the fourth class of metazoan ESCRT-I subunits", that soluble human ESCRT-I complexes contain one copy of each subunit type, and that MVB12 subunits associate with the TSG101-VPS37 core [PMID:18005716, "constitute the fourth class of metazoan ESCRT-I subunits"; PMID:18005716, "one copy of each of the four subunit types"; PMID:18005716, "associate with the core region of the binary TSG101-VPS37 complex"]. The same abstract says MVB12 depletion or overexpression inhibits HIV-1 infectivity and causes viral assembly defects [PMID:18005716, "MVB12 depletion and overexpression inhibit HIV-1 infectivity"; PMID:18005716, "aberrant virion morphologies and altered viral Gag protein processing"]. This supports viral budding/virus maturation as real but non-core cellular contexts.

MVB12A has direct EGFR/CIN85/CD2AP evidence. The CFBP paper reports that MVB12A/CFBP is phosphorylated on Tyr204 after EGF stimulation, binds CIN85/CD2AP family proteins through a proline-rich motif, and accelerates EGF receptor down-regulation by facilitating Cbl recruitment to the CD2AP/CIN85 complex [PMID:16895919, "phosphorylated at tyrosine 204 upon EGF stimulation"; PMID:16895919, "CIN85/CD2AP family was identified as a binding partner"; PMID:16895919, "accelerated the EGF receptor's down-regulation"]. A later MVB12A/B paper says Tyr204 phosphorylation of MVB12A affects CD2AP binding and regulates the amount of EGF receptor bound to ESCRT-I [PMID:20654576, "Tyr204 phosphorylation of MVB12A"; PMID:20654576, "affects binding to CD2AP"; PMID:20654576, "amounts of EGF receptor bound to ESCRT-I"]. This supports receptor catabolic process and EGFR-signaling regulation as part of MVB12A's endosomal sorting biology.

MVB12A has informative molecular-function evidence beyond generic protein binding. The MABP structural paper reports that MVB12A and MVB12B MABP domains bind acidic-lipid liposomes in vitro, can autonomously localize to puncta/plasma membrane, and provide ESCRT-I with a way to detect acidic phospholipids and protein ligands [PMID:22232651, "MABP domains of the MVB12A and B subunits"; PMID:22232651, "bind in vitro to liposomes containing acidic lipids"; PMID:22232651, "coincidence detector for acidic phospholipids and protein ligands"]. This argues that broad `lipid binding` should be modified to `phospholipid binding` (GO:0005543). The GOA/UniProt `ubiquitin binding` row cites PMID:20654576, but the cached abstract does not expose the underlying ubiquitin-binding experiment; that row should remain undecided pending full-text confirmation. Generic `protein binding` rows should be replaced where possible by ESCRT-I complex membership or SH3 domain binding context rather than accepted as molecular function.

The MVB12A-containing structural ESCRT-I paper directly supports MVB12A incorporation into the human ESCRT-I headpiece and the ESCRT-I membrane-remodeling mechanism. PMID:32424346 determined a headpiece "comprising TSG101-VPS28-VPS37B-MVB12A", found that a 22-amino-acid MVB12A fragment pulls down the ESCRT-I headpiece, and concludes that ESCRT-I has an essential scaffolding/mechanical role [PMID:32424346, "comprising TSG101-VPS28-VPS37B-MVB12A"; PMID:32424346, "22-amino acid fragment of MVB12A"; PMID:32424346, "ESCRT-I is not merely a bridging adaptor"]. Because the autophagosome-closure assay in that paper tests VPS28 helical-interface mutants rather than MVB12A perturbation, `membrane fission` can be retained as complex-level ESCRT-I context, but broad `macroautophagy` should not be treated as a core MVB12A function without MVB12A-specific phagophore-closure evidence.

The broad 2010 ESCRT review is not MVB12A-specific for autophagy. It says ESCRT-III-mediated neck cleavage is crucial for MVBs, viral budding, cytokinesis, and "probably, autophagy", and notes that direct ESCRT neck closure in autophagy remained unresolved [PMID:20588296, "viral budding, cytokinesis and, probably, autophagy"; PMID:20588296, "direct neck closure reaction in autophagy"]. The direct mammalian phagophore-closure paper identifies VPS37A as the ESCRT-I subunit needed for phagophore closure [PMID:31519728, "identify the ESCRT-I subunit VPS37A as a critical component"; PMID:31519728, "required for autophagosome completion"]. Thus a direct MVB12A macroautophagy/autophagosome assembly annotation should not be added.

Nucleus, cytosol/cytoplasm, centrosome, vesicle, and extracellular exosome rows are supported localization/context rows but are not the core proteostasis function. The core function remains ESCRT-I-dependent endosomal sorting of ubiquitinated cargo and associated membrane/cargo recognition.

## Falcon

Falcon deep research was started for MVB12A on 2026-06-02 but timed out after 600 seconds and did not produce a usable `MVB12A-deep-research-falcon.md` report. The review therefore relies on the local UniProt, GOA, cached-publication, Reactome, and PN-context evidence summarized above.

## Description cleanup note

The YAML `description` field was revised to keep it as a standalone biological summary. Project-specific curation framing moved here instead.

- Moved out of the YAML description: viral budding and virus maturation were described as direct ESCRT contexts but non-core for this proteostasis review.