Forced a current GOA/UniProt refresh before the IBA pass because the cached review and local GOA snapshot lagged IntAct, InterPro2GO, and SGD. The old cached GOA snapshot had 213 live rows, including 167 GO:0005515 protein binding rows; the 2026-07/2026-10 live export has 127 rows, including 82 GO:0005515 rows. The refresh also retired the old broad GO:0051082 unfolded protein binding rows, removed the obsolete keyword-style GO:0000166 nucleotide binding row, and shifted the ATP-binding InterPro2GO assertion from GO_REF:0000120 to the current GO_REF:0000002 provenance.
The live PAINT rows still place the core ATP binding, ATP hydrolysis, and protein-folding assertions at PANTHER:PTN000163527 and the cytosolic-Hsp90 localization/client-stabilization/heat-response assertions at PANTHER:PTN000163629. These are plausible ancestral placements for a cytosolic eukaryotic Hsp90; the target appearing in the old WITH/FROM lists is descendant evidence, not circular support. The plasma-membrane and perinuclear-cytoplasm transfers overstate localization for abundant soluble yeast Hsp82 and are retained only as over-annotated compartment transfers.
Current IntAct-backed GO:0005515 protein binding rows were generally changed to REMOVE: the interactions can be real evidence, especially for co-chaperones such as Aha1, Sti1, Tah1, Sba1, Cpr6/Cpr7, and Ppt1, but GO:0005515 adds no function beyond Hsp82 being a chaperone with a large co-chaperone/client network. I revisited all live GO:0005515 rows whose partner is the TPR co-chaperone Sti1, Ppt1, or Cns1. Three rows have enough cached, paper-specific support to narrow the generic term to GO:0030911 TPR domain binding: the Ppt1 phosphatase paper directly states that Hsp90 must be attached to Ppt1 through the Ppt1 TPR domain PMID:16407978, the R2TP paper used Sti1 TPR2a as a positive control for binding to the Hsp82 C-terminal fragment PMID:18268103, and the Cns1 paper showed interaction between Cns1 and the Hsp90 carboxy terminus PMID:9819422. High-throughput AP-MS/two-hybrid rows for the same co-chaperones were left REMOVE, because those cited papers support a physical hit but do not themselves map the TPR-domain contact.
The new DisProt GO:0019899 enzyme binding row from the FKBP38 paper was narrowed to GO:0030911 TPR domain binding: the experiment used yeast Hsp90's C-terminal MEEVD-containing fragment and showed that a K307E mutation in the FKBP8 TPR carboxylate clamp disrupts binding PMID:28278223. The live GO:0030911 row from the Tah1/R2TP structure is accepted because Tah1 uses a minimal TPR domain to tether Hsp90 to R2TP PMID:24012479.
The live SGD Tom70 row to GO:0070585 protein localization to mitochondrion was removed. Its cached abstract explicitly contrasts mammalian Hsp90-dependent preprotein delivery with yeast import, where Tom70 uses Hsp70 rather than Hsp90 PMID:12526792.
Newer-paper search found Reidy and Masison 2025, now cached as PMID:40691164, which supports a structural role for ATP in closing the yeast Hsp90 clamp PMID:40691164. A follow-up 2026-10 PubMed search also found 2024 papers on cycle-dependent Hsp90 client interactions and dominant-negative Hsp90 mutations; they reinforce selective client maturation and coordinated closing/reopening but do not change the live GOA decisions.
Fresh Falcon research reconfirmed that P02829 is the stress-inducible cytosolic Hsp90 paralog Hsp82, distinct from the constitutively expressed Hsc82 despite approximately 97% amino-acid identity [file:yeast/HSP82/HSP82-deep-research-falcon.md "Yeast has two closely related cytosolic Hsp90 isoforms, the constitutive Hsc82 and stress-inducible Hsp82, sharing ~97% amino-acid identity but exhibiting measurable functional differences."]. The most informative core molecular function is ATP-dependent protein folding chaperone activity: Hsp82 uses an ATP-regulated dimeric conformational cycle and cochaperones to promote late-stage folding, activation, and stability of selected client proteins, rather than acting as a generic unfolded-protein binder or generic protein-binding hub [file:yeast/HSP82/HSP82-deep-research-falcon.md "Hsp82 (Hsp90) is an ATP-dependent molecular chaperone that assists the late-stage folding, activation, and stability of a large set of specific “client” proteins (substrates), including many signaling regulators (notably protein kinases and transcription factors)."] [file:yeast/HSP82/HSP82-deep-research-falcon.md "Client loading, conformational closing, ATP hydrolysis/nucleotide exchange, and reopening are regulated by co-chaperones that bind distinct surfaces and bias the timing of transitions."].
ATP hydrolysis is a genuine biochemical activity of Hsp82, but the review should not describe catalytic rate as the sole determinant of biological function. Current synthesis emphasizes nucleotide-coupled conformational timing and the residence time of particular states [file:yeast/HSP82/HSP82-deep-research-falcon.md "A central current concept is that dwell time in specific conformations can be more functionally determinative than absolute ATPase rate, emphasizing kinetic regulation of the cycle in vivo."]. This supports retaining GO:0016887 while interpreting it as regulation of the client-folding cycle, alongside the more functionally descriptive GO:0140662.
Hsp82 is specialized for stress without acquiring a different fundamental chaperone mechanism. It is low-abundance under nonstress conditions, induced at elevated temperature, and more thermally stable than Hsc82 [file:yeast/HSP82/HSP82-deep-research-falcon.md "Hsp82 is more thermally stable than Hsc82 (Tm ~60.4°C vs 57.1°C)."] [file:yeast/HSP82/HSP82-deep-research-falcon.md "These findings support annotating Hsp82 as the heat-inducible, stress-resilient cytosolic Hsp90 isoform that preserves the core Hsp90 mechanism while shifting stability/dynamics."]. Heat response is therefore a biologically meaningful paralog specialization, while the underlying foldase activity remains its core molecular function.
The primary localization is cytosolic during vegetative growth. Nuclear accumulation is condition dependent during glucose exhaustion/quiescence and sporulation rather than evidence that Hsp82 is constitutively nuclear [file:yeast/HSP82/HSP82-deep-research-falcon.md "Primary localization: Cytosolic Hsp90 (Hsp82/Hsc82) is broadly distributed during vegetative growth."] [file:yeast/HSP82/HSP82-deep-research-falcon.md "Condition-dependent nuclear accumulation: A key cell-biological finding is that Hsp90 (and the cochaperone Sba1/p23) accumulates in the nucleus in quiescent cells (glucose exhaustion) and in sporulating diploids"].
Client-dependent pathway annotations need to remain distinct from core function. The R2TP paper directly shows that Hsp90 and Tah1 stabilize Pih1 and thereby affect box C/D snoRNP accumulation, supporting a real but secondary pathway consequence PMID:18268103. By contrast, the split-GFP paper used abundant cytosolic Hsp82 as a negative-control protein that produced little or no spurious signal at the septin collar; it supports cytosolic accessibility, not septin-collar residency PMID:27385335.