LCI5 / EPYC1 (Chlamydomonas reinhardtii, UniProt Q94ET8) - curation notes

Identity

UniProt Q94ET8 (gene name Lci5, 321 aa, TrEMBL) is the same protein as
EPYC1, Essential Pyrenoid Component 1, locus Cre10.g436550.
PMID:27166422
PMID:28938114
The UniProt entry cross-references PDB 7JFO (EPYC1 residues 49-72 bound to Rubisco) from He et al. 2020 (PMID:33230314), and CD-CODE lists it as a pyrenoid condensate component.
The folder and gene_symbol stay as LCI5 (the UniProt gene name). EPYC1 is recorded under aliases.

Molecular function: multivalent Rubisco linker / condensate scaffold

The protein is disordered and made of repeats PMID:27166422
Four ~60-aa repeats were originally described. Later structural work defines five Rubisco-binding regions PMID:33230314
The binding site is the pair of alpha-helices on the Rubisco small subunit PMID:33230314. This explains the older result that the SSU helices are needed for a pyrenoid PMID:23112177.
Each repeat adds binding strength, so the protein is modular PMID:31504763.
Rubisco and EPYC1 alone are enough for phase separation PMID:30498228
The condensate is a codependent network, not an EPYC1 lattice PMID:33230314. Wunder et al. explicitly reject a homotypic "scaffold hypothesis" PMID:30498228. This does not argue against GO:0140693 molecular condensate scaffold activity. That GO term is defined as "binding and bringing together two or more macromolecules in contact, permitting those molecules to organize as a molecular condensate", and it does not require EPYC1 to self-assemble.
Interface mutations block binding, phase separation and pyrenoid formation PMID:33230314
EPYC1 is sufficient to condense Rubisco in a heterologous host, an Arabidopsis line with a Chlamydomonas-type SSU PMID:33298923
Condensation does not change Rubisco's catalytic activity PMID:30498228. EPYC1 is therefore not a Rubisco activator and has no catalytic role in carbon fixation.
EPYC1 and Rubisco also form complexes in the dilute phase PMID:36611062
Extant and ancestral EPYC1 homologues were tested for their effect on carboxylation in reconstituted synthetic pyrenoids PMID:42601498

Localization

The protein is in the pyrenoid matrix PMID:27166422; PMID:28938113
Part of it disperses into the stroma during division PMID:28938114
The 2006 thylakoid result needs reinterpreting. Turkina et al. reported PMID:16572472. That study predates the identification of the pyrenoid matrix and used fractionation. The pyrenoid is crossed by thylakoid-derived tubules, and EPYC1 has no transmembrane domain. Mackinder et al. selected EPYC1-like proteins in part for the "absence of transmembrane domains" (PMID:27166422). I treat the thylakoid membrane location as co-fractionation and do not propose GO:0009535. The earlier draft's NEW proposal for that term was dropped.

Phenotype (necessity evidence)

The epyc1 mutant cannot grow photoautotrophically at low CO2, and complementation rescues it PMID:27166422
The mutant mislocalizes Rubisco PMID:27166422
The canonical matrix is absent in the mutant PMID:33177094

Phosphorylation and regulation

The protein is phosphorylated in its repeats at low CO2 PMID:16572472; PMID:16572472
The interactome finds 14-3-3 proteins and a kinase with EPYC1 PMID:28938113
KEY1 is the main kinase. It targets the Rubisco-binding regions and drives dissolution of the condensate PMID:41845050; PMID:41845050
Interpretation: the low-CO2 phosphorylation reported in 2006 is regulation of the scaffold activity. It is not evidence for a separate signalling or "response to CO2" function.

Review decisions (2026-10 re-review)

GOA has no rows for Q94ET8, so every entry is a NEW proposal.

Term	Decision	Rationale
GO:0140693 molecular condensate scaffold activity	NEW (IDA, PMID:30498228), core	Reconstitution, structure and mutagenesis
GO:1990732 pyrenoid	NEW (IDA, PMID:27166422), core location	Pyrenoid proteomics and live imaging
GO:0140694 membraneless organelle assembly	NEW (IMP, PMID:27166422), core BP	Passes the participation test as the scaffold case; comparator FUS/Wnk carry 0140693+0140694
GO:0009570 chloroplast stroma	NEW (IDA, PMID:28938114), non-core	Dilute-phase pool seen during division
GO:0005515 protein binding	dropped	Uninformative; replaced by 0140693
GO:0043169 cation binding	dropped	No evidence
GO:0009535 chloroplast thylakoid membrane	dropped	2006 fractionation superseded by the pyrenoid matrix localization
GO:0015976 carbon utilization	dropped	Too broad and vague
GO:0015979 photosynthesis	dropped	Too broad; EPYC1 does not catalyse any photosynthetic step
GO:0071244 cellular response to carbon dioxide	dropped	Based only on expression induction

Terms considered and rejected:

GO:0015977 carbon fixation. Rubisco performs the carboxylation, and its activity is unchanged by condensation. EPYC1 only organises Rubisco, so the participation test fails.
GO:0046863 Rubisco activator activity. Activity is unchanged by condensation.
GO:0110102 ribulose bisphosphate carboxylase complex assembly. Holoenzyme assembly is normal without the EPYC1 interaction, as shown by the SSU helix-swap lines (PMID:27166422 discussion).

Ontology gap

GO has no "pyrenoid assembly" BP. GO:0140694 is the closest term. This is recorded as an ONTOLOGY knowledge gap.

Other files

LCI5-deep-research.md (OpenAI) is outdated. It does not recognise EPYC1 and claims that no mutant phenotype or structure exists. It is marked LOW_QUALITY in reference_review and is not used as evidence.
LCI5-pathway.md predates this re-review and still describes LCI5 as a thylakoid phosphoprotein of unknown function. It should be regenerated.