eryCII (A4F7P2, SACE_0725) — review notes

Part of the BGC exemplar curation project (projects/BGC.md). MIBiG BGC0000055
(Saccharopolyspora erythraea, erythromycin PKS). GenBank CAM00066.1 → UniProt
A4F7P2 (ERYC2_SACEN), gene eryCII / SACE_0725. Also relevant to
projects/PSEUDOENZYMES.md.

Function — a P450 PSEUDOENZYME that activates a glycosyltransferase

EryCII is a cytochrome-P450-family homologue that has lost catalytic competence
and instead serves as the activating partner of the desosaminyl transferase
EryCIII in erythromycin biosynthesis.

• UniProt A4F7P2 CAUTION: "Although related to the cytochrome P450 family, lacks the
  heme-binding sites." (ECO:0000305) — note this is a curator inference, so it needs
  corroboration (below).
• Beyond the assertion — primary structural evidence (PMID:22056329 full text):
• "...lacks the essential cysteine residue that coordinates the heme iron."
• "the conserved cysteine is absent (Supplementary Fig. S5); thus, they are not
  active P450 enzymes." (structure-based alignment vs P450 Cyp125)
• "Having lost the heme group in the central core, the protein may be more
  conformationally dynamic..."
• PDB 2YJN is apo: the RCSB entry has zero non-polymer entities (no heme ligand).
• Activation mechanism: EryCIII activity "is almost completely rescued by the
  addition of soluble EryCII... able to induce a catalytically active structure in
  EryCIII."

So the no-heme / no-catalysis conclusion rests on (1) the absent conserved Cys in a
structure-based alignment, and (2) an apo crystal structure — not merely UniProt.
- UniProt FUNCTION: "Involved in the erythromycin biosynthesis pathway. Acts by
forming a complex and stabilizing the desosaminyl transferase EryCIII."
(ECO:0000269|PubMed:22056329); SUBUNIT: "Heterotetramer composed of EryCII and EryCIII."
- Structure (PDB 2YJN): "EryCIII, in concert with its partner EryCII, attaches a
nucleotide-activated sugar to the macrolide scaffold... a heterotetramer...
EryCII stabilizes EryCIII and also functions as an allosteric activator of the GT."
PMID:22056329
- EryCII is required for EryCIII catalysis (not merely an enhancer): purified
EryCIII "was inactive under all conditions tried"; combining EryCIII + EryCII
extracts "restored" GT activity [PMID:22056329 full text]. Stoichiometry is
α2β2 (dimer of heterodimers, ~188 kDa); activation is allosteric via EryCII's
N-terminal A″ helix, which contacts EryCIII's acceptor domain.

Annotation issues — domain-based over-annotation of a pseudoenzyme

All four GOA molecular-function terms are IEA from the P450 sequence signature and are
incorrect for this catalytically dead homologue:

• GO:0004497 monooxygenase activity → REMOVE
• GO:0005506 iron ion binding → REMOVE
• GO:0016705 oxidoreductase activity (paired donors, O2) → REMOVE
• GO:0020037 heme binding → REMOVE (UniProt: "lacks the heme-binding sites")

Accurate roles to assert instead: GO:0008047 enzyme activator activity (allosteric
activator of EryCIII; added as a NEW annotation), GO:0050821 protein stabilization
(stabilizes EryCIII's fold/quaternary structure), GO:1901115 erythromycin
biosynthetic process (BP), GO:0032991 protein-containing complex (EryCII-EryCIII
heterotetramer).

"Conformational chaperone"?

Mechanistically EryCII templates/stabilizes the catalytically competent conformation of
EryCIII, which is chaperone-like. But it is not annotated as a protein folding chaperone (GO:0044183): a chaperone acts transiently and is released, and its client
stays active afterwards; EryCII is a stably bound stoichiometric subunit and EryCIII
is inactive without it (PMID:22056329). So the stabilizing aspect is captured by
protein stabilization (GO:0050821) and the activating aspect by enzyme activator activity (GO:0008047) - not a chaperone MF. (The older fold-and-release / "does not
bind tightly" model, ref 13, was superseded by the stable-complex structure.)

Capturing absence of function (NOT annotations)

A "NEW NOT" (negated: true + action: NEW) can only be used for a term NOT already in
GOA. Here the erroneous P450 functions (monooxygenase, heme binding, iron binding,
oxidoreductase) are already present in GOA, so they are handled by REMOVE; there is no
review action to convert an existing positive annotation into a NOT.

Predicted-complex evidence (BGC project)

Moriwaki et al. (bioRxiv 2025.10.26.684697) predict the EryCII-EryCIII interaction
(BGC0000055; CAM00066.1/CAM00067.1) at ipTM 0.92, matching PDB 2YJN (hetero 4-mer).

References

• PMID:22056329 — Moncrieffe et al. 2012, J Mol Biol (EryCIII·EryCII structure, PDB
  2YJN; EryCII stabilizes and allosterically activates EryCIII). VERIFIED.
• UniProt A4F7P2 (CAUTION: lacks heme-binding sites).