SAGA2 (StArch Granules Abnormal 2; locus Cre09.g394621) is a large (~190 kDa) nuclear-encoded, chloroplast-targeted pyrenoid protein of Chlamydomonas reinhardtii and a homologue of SAGA1 (about 30% identical). UniProt holds two TrEMBL entries for this locus, A0A2K3DEH1 (1,816 aa) and A0A2K3DEF8 (1,748 aa), which are alternative gene-model products of the same gene. SAGA2 has two consecutive N-terminal CBM20 starch-binding domains, a long coiled-coil middle region and extended disordered segments that carry several copies of the pyrenoid Rubisco-binding motif. Rubisco binds SAGA2 motif peptides in vitro, and SAGA2 binds the Rubisco large and small subunits and the linker EPYC1 in yeast two-hybrid assays. A purified construct of its two CBM20 domains binds starch granules and the starch precursor maltoheptaose and the mimic beta-cyclodextrin in vitro. In the cell SAGA2 forms a fairly uniform shell at the interface between the Rubisco matrix and the starch sheath, and it is depleted from the gaps where thylakoid tubules enter the matrix, which are occupied by SAGA1. The saga2 single mutant grows normally at limiting CO2, has a single pyrenoid with normal tubules, and shows reduced starch coverage of the matrix early in pyrenoid formation that recovers later. SAGA2 acts redundantly with SAGA1 to place the starch sheath on the pyrenoid. The saga1 saga2 double mutant lacks a starch sheath and its starch is spread through the stroma, and either protein restores the sheath. Expressed in Arabidopsis with an EPYC1-Rubisco condensate, SAGA2 increases condensate-associated starch. SAGA2 is proposed to tether starch granules to the matrix and perhaps to prime granule initiation at the matrix surface. It does not take part in forming the pyrenoid tubules.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0007034 vacuolar transport | IBA GO_REF:0000033 | REMOVE | Summary: Propagated from the PANTHER PTHR15048 (STARCH-BINDING DOMAIN-CONTAINING PROTEIN 1) node seeded by mammalian STBD1, an ER-anchored glycogen-binding protein that delivers glycogen to lysosomes (glycophagy). SAGA2 is a chloroplast pyrenoid protein with no role in vacuolar transport. Reason: The family node is defined by animal STBD1, whose vacuolar (lysosomal) transport role depends on an N-terminal membrane anchor and on glycogen delivery to lysosomes. SAGA2 has diverged in a way that removes it from this function. It has no predicted transmembrane segment, its transit peptide targets it to the chloroplast stroma (PMID:38241434), and it acts at the matrix-starch sheath interface of the pyrenoid inside the chloroplast (PMID:33177094, PMID:42090253). No Chlamydomonas data link SAGA2 to any vacuolar process. The shared feature with STBD1 is only the CBM20 module. Propagation Review Root cause: PROPAGATION BAD Failure modes: FUNCTIONAL DIVERGENCE COMPARTMENT OR COMPLEX MISMATCH Sources checked: PANTHER:PTN000386078 · PTN000386078 SUPPORTS SOURCE BUT NOT TARGET Node seeded by mammalian STBD1 (UniProtKB:O95210, MGI:1261768), an ER-anchored glycogen-binding protein that targets glycogen to lysosomes. SAGA2 shares only the CBM20 module and is a chloroplast-stromal pyrenoid protein with no membrane anchor. Supporting Evidence: PMID:38241434 the native SAGA2 transit peptide could target SAGA2::mNeon to the chloroplast stroma PMID:33177094 We observed that SAGA2 also localized to that interface but appeared to cover the surface of the matrix more homogeneously than SAGA1. |
| GO:0016020 membrane | IBA GO_REF:0000033 | REMOVE | Summary: Propagated from STBD1, which is anchored in the ER membrane by an N-terminal hydrophobic segment. SAGA2 lacks any transmembrane segment and sits at the starch-matrix interface of the pyrenoid, away from membranes. Reason: Membrane association in STBD1 depends on its N-terminal hydrophobic anchor, which SAGA2 does not have (UniProt lists no transmembrane region). SAGA2 is located at the matrix-starch interface and is depleted where thylakoid tubules enter the matrix (PMID:42090253). Hennacy et al. note SAGA2 localizes away from membranes (PMID:39548241). Target-specific evidence therefore argues against inheriting the family membrane location. The correct location is pyrenoid (GO:1990732), added below. Propagation Review Root cause: PROPAGATION BAD Failure modes: FUNCTIONAL DIVERGENCE COMPARTMENT OR COMPLEX MISMATCH Sources checked: PANTHER:PTN000386078 · PTN000386078 SUPPORTS SOURCE BUT NOT TARGET Node seeded by mammalian STBD1 (UniProtKB:O95210, RGD:1311800), an ER-anchored glycogen-binding protein that targets glycogen to lysosomes. SAGA2 shares only the CBM20 module and is a chloroplast-stromal pyrenoid protein with no membrane anchor. Supporting Evidence: PMID:42090253 Our results indicate that whereas SAGA1 is enriched at the starch–tubule–matrix junctions of the pyrenoid, SAGA2 is depleted from these regions PMID:39548241 but it is unlikely to contribute to matrix-traversing membrane biogenesis based on the growth phenotypes of the saga2 mutant |
| GO:0030246 carbohydrate binding | IEA GO_REF:0000002 | ACCEPT | Summary: SAGA2 has two CBM20 domains, and the purified tandem domain binds starch, beta-cyclodextrin and maltoheptaose. Carbohydrate binding is correct but general. Reason: The InterPro carbohydrate-binding-like fold mapping is correct and is confirmed by direct binding data (PMID:42090253). The more specific child term GO:2001070 starch binding is also annotated, so this broad IEA is harmless. Same decision as for SAGA1. Supporting Evidence: PMID:42090253 Binding of 10×His-mVenus-2×SAGA1-CBM20 and the 10×His-mVenus-SAGA2-2×CBM20 to starch was decreased in the presence of either β-cyclodextrin or maltoheptaose, suggesting that each protein bound to all three molecules. |
| GO:2001070 starch binding | IEA GO_REF:0000002 | ACCEPT | Summary: The CBM20 (IPR002044) mapping to starch binding is correct. The purified construct of both native SAGA2 CBM20 domains binds starch granules in vitro, and SAGA2 rings starch granules when expressed in Arabidopsis. Reason: Confirmed experimentally. A 10xHis-mVenus-SAGA2-2xCBM20 construct containing both native CBM20 domains bound waxy maize starch, the mVenus control did not, and beta-cyclodextrin and maltoheptaose competed (PMID:42090253). This is domain-level IDA evidence rather than full-length protein. The in vivo phenotype agrees, since SAGA2 is needed for normal early starch coverage of the matrix and, with SAGA1, for any starch sheath (PMID:42090253). Starch binding is the core molecular function of SAGA2. Supporting Evidence: PMID:42090253 We observed that both the 10×His-mVenus-2×SAGA1-CBM20 and the 10×His-mVenus-SAGA2-2×CBM20 proteins, but not a 10×His-mVenus control protein, bound to starch PMID:42090253 SAGA2 natively has two consecutive CBM20 domains, both of which were included in the SAGA2 construct for this experiment PMID:38241434 Together, our imaging data suggested that SAGA2 interacts differently with the condensate compared to SAGA1 but is also able to recruit starch to the matrix. |
| GO:1990732 pyrenoid | IDA PMID:33177094 Assembly of the algal CO(2)-fixing organelle, the pyrenoid, ... | NEW | Summary: SAGA2-Venus localizes in a uniform ring at the interface between the Rubisco matrix and the starch sheath of the pyrenoid, depleted at tubule entry sites. Reason: Live-cell fluorescence of SAGA2-Venus in Chlamydomonas (PMID:33177094), confirmed by super-resolution imaging of the complementing SAGA2-Venus in the saga2 mutant (PMID:42090253). This is where SAGA2 acts. GO has no term for the starch sheath or the matrix-starch interface, so pyrenoid is the most specific correct term. It replaces the IBA membrane location. Supporting Evidence: PMID:33177094 We observed that SAGA2 also localized to that interface but appeared to cover the surface of the matrix more homogeneously than SAGA1. PMID:42090253 Our results indicate that whereas SAGA1 is enriched at the starch–tubule–matrix junctions of the pyrenoid, SAGA2 is depleted from these regions |
Loading supporting content…
Download this section (compressed HTML)Q: Should GO add a process term for starch sheath assembly or localization at the pyrenoid (for example "pyrenoid starch sheath assembly"), so that SAGA1 and SAGA2 can be annotated to the process they carry out? The closest existing term, GO:0033037 polysaccharide localization, is used almost only for glycosaminoglycans.
Suggested experts: Jonikas MC
Q: Do SAGA1 and SAGA2 prime starch granule initiation at the matrix surface (GO:0062052 starch granule initiation), as their maltoheptaose binding and the stromal starch in the double mutant suggest, or do they only capture and tether granules made elsewhere?
Q: Which gene model is correct for Cre09.g394621? UniProt has A0A2K3DEH1 (1,816 aa) and A0A2K3DEF8 (1,748 aa) for the same locus, and published papers describe a ~190 kDa protein with four Rubisco-binding motifs.
Q: What is the in vivo role of the SAGA2-EPYC1 interaction seen in yeast two-hybrid, given that EPYC1 binds SAGA2 through the same helices that bind Rubisco?
Experiment: Complement saga1;saga2 with SAGA2 carrying point mutations in the conserved aromatic starch-binding residues of one or both CBM20 domains. Score SAGA2 localization, starch sheath coverage at 3 h and 21 h after transfer to low CO2, and starch distribution.
Hypothesis: Starch binding by the SAGA2 CBM20 domains is required for its role in early starch sheath coverage.
Type: genetic complementation and electron microscopy
Experiment: Express SAGA2 with all Rubisco-binding motifs mutated in saga1;saga2 and test whether it still localizes to the pyrenoid and restores a sheath, or instead coats stromal starch granules.
Hypothesis: The Rubisco-binding motifs anchor SAGA2 to the matrix.
Type: mutagenesis and live-cell imaging
What is not known — curated, literature-grounded statements of the open unknowns (the inverse of core functions).
Gap: GO has no process term for placing or assembling the pyrenoid starch sheath, so the process SAGA2 contributes to cannot yet be annotated.
OPEN ONTOLOGY
What is known: SAGA2 binds starch and is required, redundantly with SAGA1, for the starch sheath to form on the pyrenoid matrix.
Provenance (the field's own admissions):
Gap: Whether SAGA2 initiates starch granules or only tethers them is unresolved.
OPEN BIOLOGY
What is known: SAGA2 domains bind starch, beta-cyclodextrin and maltoheptaose in vitro.
Provenance (the field's own admissions):
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)