# CHMP4B curation notes

## 2026-06-02 review

Falcon deep research was started with `just deep-research-falcon human CHMP4B`; the wrapper printed `Timeout: 600s`, but it produced `CHMP4B-deep-research-falcon.md` with a completed report and 22 citations. I used it as synthesis context only where it agreed with cached primary evidence, especially its summary that CHMP4B is an ESCRT-III membrane-remodeling scaffold rather than an enzyme [file:human/CHMP4B/CHMP4B-deep-research-falcon.md "polymerizes on membranes as part of **ESCRT-III** to drive **reverse-topology membrane constriction and scission**"].

CHMP4B encodes hSnf7-2, a Snf7-family core ESCRT-III subunit. UniProt summarizes it as a probable ESCRT-III core component involved in MVB formation, endosomal cargo sorting, membrane fission, cytokinesis, nuclear envelope sealing, and exosomal release [UniProt:Q9H444 "Probable core component of the endosomal sorting required for transport complex III (ESCRT-III)"; UniProt:Q9H444 "involved in multivesicular bodies (MVBs) formation and sorting of endosomal cargo proteins into MVBs"; UniProt:Q9H444 "believed to mediate the necessary vesicle extrusion and/or membrane fission activities"].

The strongest direct mechanistic evidence for the molecular function is ESCRT-III filament formation and membrane curvature. Hanson et al. explicitly examined hSnf7-1/CHMP4A and hSnf7-2/CHMP4B and found that both assemble into approximately 5 nm curved filaments, create circular arrays, and promote/stabilize negative membrane curvature and outward budding [PMID:18209100 "hSnf7-1 (CHMP4A) and hSnf7-2 (CHMP4B)"; PMID:18209100 "assemble into regular approximately 5-nm filaments that curve and self-associate"; PMID:18209100 "form novel membrane-attached filaments that can promote or stabilize negative curvature and outward budding"]. This supports adding membrane bending activity, and supports retaining protein polymerization and membrane fission while replacing generic identical protein binding/homodimerization with more informative terms.

CHMP4B has direct evidence in endosomal/MVB sorting. The original ALIX paper identified CHMP4B as a human Snf7 homolog involved in MVB sorting, found ALIX-CHMP4B interaction, and localized CHMP4B to early/late endosomal puncta [PMID:12860994 "The ALG-2-interacting protein Alix associates with CHMP4b, a human homologue of yeast Snf7 that is involved in multivesicular body sorting"; PMID:12860994 "CHMP4b transiently expressed in HeLa cells mainly exhibited a punctate cytoplasmic pattern"; PMID:12860994 "partly overlapped the distributions of early and late endosome markers"]. CHMP6 work provides additional support for CHMP4B interaction in mammalian ESCRT-III assembly and endosomal cargo sorting [PMID:15511219 "interacts directly with another ESCRT-III component CHMP4b/Shax"; PMID:15511219 "CHMP6 acts as an acceptor for ESCRT-II on endosomal membranes and regulates cargo sorting"].

CHMP4B is stronger than CHMP4A for several ESCRT output contexts. In the CHMP4C abscission checkpoint paper, depletion of CHMP4B, but not CHMP4A or CHMP4C, prevented MHC-I degradation, inhibited HIV-1 release, and blocked completion of cytokinesis; GFP-CHMP4B localized transiently to midbody arms just before abscission [PMID:22422861 "depletion of CHMP4B prevented MHC-I degradation, whereas depletion of CHMP4A or CHMP4C had little effect"; PMID:22422861 "only depletion of CHMP4B inhibited this ESCRT-dependent process"; PMID:22422861 "CHMP4B was again the sole paralog required"; PMID:22422861 "GFP-CHMP4B localized transiently to the midbody arms immediately"]. These are valid non-core ESCRT scission roles in this PN-focused review.

Autophagy annotations are retained as PN-relevant endolysosomal/autophagy context rather than treated as the only core function. General ESCRT/MVB work shows ESCRT depletion inhibits autophagic degradation and causes autophagosome/amphisome accumulation [PMID:17984323 "autophagic degradation is inhibited in cells depleted of ESCRT subunits"; PMID:17984323 "Functional multivesicular bodies are required for autophagic clearance"]. Human neuron work directly implicates hSnf7-2/CHMP4B: hSnf7-2 is abundant in human neurons, required for neuronal survival, preferentially associates with CHMP2B, regulates transmembrane cargo turnover, and hSnf7-2 knockdown causes GFP-LC3 autophagosome accumulation [PMID:21975012 "hSnf7-1 and hSnf7-2 are most abundantly expressed in human neurons"; PMID:21975012 "hSnf7-2 are required for the survival of human neurons"; PMID:21975012 "complexes containing hSnf7-2 and CHMP2B were more abundant"; PMID:21975012 "SiRNA knockdown of hSnf7-2 in human neurons leads to autophagosome accumulation"]. The STX13 paper is useful PN context for ESCRT-III dysfunction and phagophore/autophagosome maturation, but it is mainly about STX13/Vti1a and mutant CHMP2B, not a direct CHMP4B assay [PMID:24095276 "STX13 is a genetic modifier of ESCRT-III dysfunction and participates in the maturation of phagophores into closed autophagosomes"].

Other retained non-core contexts include plasma membrane repair [PMID:24482116 "ESCRT proteins were recruited within seconds to plasma membrane wounds"; PMID:24482116 "repair of certain wounds is ensured by ESCRT-mediated extracellular shedding"], nuclear envelope sealing/reassembly with SPAST [UniProt:Q9H444 "promotes nuclear envelope sealing and mitotic spindle disassembly during late anaphase"; PMID:26040712 "Spastin and ESCRT-III coordinate mitotic spindle disassembly and nuclear envelope sealing"], viral budding/HIV release [PMID:24878737 "ESCRT machinery along with the AAA+ ATPase Vps4 drive membrane scission"; PMID:24878737 "viral budding and cytokinesis"], and lens transparency as a disease/phenotype context rather than core molecular function [PMID:17701905 "CHMP4B, which encodes a key component of the endosome sorting complex required for transport-III (ESCRT-III) system"; PMID:17701905 "plays a vital role in the maintenance of lens transparency"].

Several annotations are over-annotated or unsupported. Generic protein binding rows do not capture CHMP4B function and are marked over-annotated. Cadherin binding and vesicle annotations from the E-cadherin proximity proteomics study are broad high-throughput proximity/localization findings, not core CHMP4B molecular function [PMID:25468996 "proximity biotinylation and quantitative proteomics to identify 561 proteins in the vicinity"]. The ER membrane targeting annotation from PMID:21975012 is removed because the paper supports CHMP4B/Snf7-2 in neuronal cargo turnover/autophagosome accumulation, not post-translational targeting to the ER membrane.
