The Problem

GO:0051082 "unfolded protein binding" and GO:0031249 "denatured protein binding" proposed for obsoletion (go-ontology#30962)

These terms conflate mechanistically distinct activities under a single "binding" label:

• ATP-dependent foldases that actively refold clients
• ATP-independent holdases that prevent aggregation in situ
• Disaggregases that solubilize existing aggregates
• Sensors that recognize misfolded proteins for degradation
• Co-chaperones that modulate chaperone ATPase cycles

The Holdase-Foldase Network

Conserved across all domains of life: stress unfolds proteins; sHSPs (holdases) capture them to prevent aggregation; clients are handed off to Hsp70 (foldase) for ATP-dependent refolding, or to Hsp100 (disaggregase). Eukaryotes also use Hsp90 and Hsp60/CCT.

Figure 3 from Luo et al. (2022) Front Mol Biosci 9:832160, CC-BY 4.0

The Proteostasis Triage Decision

(A) HSP70-HOP-HSP90 axis: fold-forward pathway to native protein.
(B) HSP70-CHIP axis: ubiquitination and proteasomal degradation.

Co-chaperones (DNAJ, NEFs, BAG1/3) tune the fold vs degrade decision.

Figure 2 from Radons (2020) Cells 9:587, CC-BY 4.0

Recent Literature Context

Key reviews informing this reclassification:

• Mitra et al. (2022) Annu Rev Biophys 51:409 — ATP-independent chaperones; force-clamp experiments show context-dependent holdase/foldase switching
• Bhattacharjee et al. (2025) J Biosci — Co-chaperones fine-tune HSP70: fold, hold, or degrade; O-GlcNAc modification of HSP27 promotes BAG3-mediated refolding
• Boopathy (2024) — HUWE1 + HSP70 cooperate to clear nuclear inclusions, linking holdase activity to E3-mediated clearance
• Le et al. (2024) Mol Cells — UBR1/UBR2 as ER-stress sensors: substrate load stabilizes these E3s to enhance PQC

Scope of This Review

All experimental annotations to GO:0051082/GO:0031249 across all species.

Mechanism Classes

Decision Rules

Impact: Human Genes (n=33)

Plus 3 co-annotations to GO:0140545 (disaggregase): HSPA1A, HSPA1B, HSPA8

Before/After Examples

Critical Finding: The Holdase Gap

GO:0140309 "unfolded protein carrier activity" does not fit most holdases.

• Created Nov 2025 for TIM carrier-holdases (Tim9-Tim10) in go-ontology#30552
• Definition requires escort "between two different cellular components"
• "holdase" is a BROAD synonym (not exact) on GO:0140309
• Val acknowledged a general holdase term was needed but deferred it

7 human genes + HSPH1 are in-situ holdases that prevent aggregation without inter-compartment escort. They cannot use GO:0140309.

Proposed: "Holdase Chaperone Activity"

Definition: Binding to an unfolded or misfolded protein to prevent its aggregation without actively catalyzing refolding. The holdase maintains the client protein in a soluble, folding-competent state.

Parentage: Direct child of GO:0003674 (molecular_function).
GO:0140309 (carrier-holdase) becomes a child of this new term.

Affected genes: CRYAA, CRYAB, HSPB6, CLU, SCG5, DNAJB6, DNAJB8, HSPH1

GO:0051082 obsoletion must be blocked until this NTR exists.

Other Ontology Gaps

Misfolded protein sensor activity

• Recognition of misfolded conformation for QC degradation
• CHIP/STUB1 ubiquitinates chaperone-bound clients; UBR1/UBR2 recognize N-degrons (Le et al. 2024)
• Affects: SYVN1, SAN1 (yeast), Fbxo2 (mouse)

Co-chaperone MF representation

• GO:0003767 "co-chaperone activity" deliberately obsoleted
• No MF term for J-domain co-chaperone function (ATPase activation + substrate delivery)
• GO:0044183 used as pragmatic interim
• Affects all J-domain proteins across species

Cross-Species Validation

The same mechanism classes apply universally across 17 species.

Notable Cross-Species Findings

• SlrP (S. typhimurium): misannotation removed — T3SS effector E3 ligase, not a chaperone
• CnoX (E. coli): redox-activated holdase — becomes active under oxidative stress when Cys residues are oxidized
• Peroxiredoxins (TSA1, pmp20, tpx1): dual-function — peroxidase at low stress, holdase at high stress (overoxidation switch)
• Assembly factors (ATP10, PET100, COX20): single-client chaperones, not general UPB — all OVER_ANNOTATED
• IRE1 (yeast + T. reesei): UPR sensor, not chaperone — cross-kingdom confirmation

What We Need from GO Editors

1. Holdase NTR (BLOCKING): Create "holdase chaperone activity" for in-situ holdases
2. Block GO:0051082 obsoletion until holdase NTR exists (7 genes have no replacement)
3. Preferred labels: "foldase" exact synonym on GO:0044183; "holdase" exact on new term
4. Co-chaperone MF gap: How should J-domain function be annotated?
5. HSP70 dual annotation: Confirm foldase + holdase per experimental context
6. Misfolded protein sensor: New term for SYVN1, SAN1, Fbxo2?