RBMP2 (Cre09.g416850; UniProt A0A2K3DG19) notes

Deep research

Not run. The falcon provider returned HTTP 402 and perplexity was unavailable in this
session, so this review rests on the cached primary literature, a PubMed search, and a
local sequence analysis (RBMP2-bioinformatics/). No -deep-research-*.md file exists.

Identity and record

• UniProt A0A2K3DG19 is an unreviewed TrEMBL entry, "Rhodanese domain-containing
  protein", 1690 aa, ORF CHLRE_09g416850v5. Its only feature annotations are a PROSITE
  PS50206 rhodanese domain (775-920) and MobiDB disordered regions. PANTHER places it in
  PTHR24216 ("paxillin-related"), which looks like a low-complexity artefact rather than a
  meaningful family call.
• GOA has 0 annotations for this protein.
• Named by Meyer et al. 2020. PMID:33177094
• PubMed search "RBMP2 AND (pyrenoid OR Chlamydomonas)" (Oct 2026) returned one hit,
  the functional preprint PMID:42427617, which was then cached.

Domain architecture

• Wu et al. 2026: PMID:42427617
• Six RBMs, consistent with a review. PMID:35961043
• Local check (file:CHLRE/RBMP2/RBMP2-bioinformatics/RESULTS.md): six matches to the
  [D/N]W[R/K]XX[L/I/V/A] consensus at 1126, 1265, 1345, 1592, 1638 and 1685 (the last at the
  exact C-terminus, ending in Val). Hydrophobic segments at 684-714 (strong) and 961-981
  (weaker) flank the rhodanese domain.

Rhodanese domain: non-catalytic

• The whole protein has a single cysteine (Cys498, in the low-complexity N-terminal half).
  The rhodanese domain (775-920) has none. A crude loop scan puts Asp870 (DSYGVE) in the
  position of the catalytic cysteine.
• Wu et al.'s alignment agrees. PMID:42427617
• Consequence: no sulfurtransferase (thiosulfate-cyanide sulfurtransferase) or
  CDC25-like phosphatase activity can be inferred. The InterPro rhodanese hit should not
  be turned into an MF annotation by any pipeline. Its role is structural or interaction
  based. PMID:42427617

Localization

• Pyrenoid tubules, Venus fusion. PMID:33177094
• Central reticulated region, unlike RBMP1. PMID:33177094
• Confirmed with expansion microscopy, and the region expands on overexpression. PMID:42427617
• PMID:42465278 (ExM of pyrenoid membranes) does not mention RBMP2. It shows CAH3
  concentrated in the reticulated region. PMID:42465278

Rubisco binding

• AP-MS (Meyer's ref 18). PMID:33177094
  The cached text of PMID:28938113 (the spatial interactome, presumably ref 18) does not
  name RBMP2 or its locus, so I did not use it as support.
• Peptide arrays tiling RBMP2 were probed with Rubisco. PMID:33177094
• The SAGA1 antibody did not detect RBMP2 on blots, which the authors put down to its
  C-terminal Val.
• GO has no Rubisco-binding MF term, and "protein binding" is uninformative, so no MF was
  made from this.

Tether hypothesis vs membrane shaping

• Original model: RBMs on RBMP1/RBMP2 recruit Rubisco to the tubules. PMID:33177094
• RBMP1/BST4 was later shown not to be needed for tubule-matrix contact. The paper raises
  redundancy in general terms but does not name RBMP2. PMID:39240724
• Wu et al. 2026 (bioRxiv, not yet peer reviewed) is the first functional study of RBMP2:
• Null phenotype: PMID:42427617
• Rescue: PMID:42427617
• Tubules still enter the matrix without RBMP2, so it is not needed for the
  tubule-matrix interface as such. PMID:42427617
• Overexpression is sufficient to enlarge the region. PMID:42427617
• Rhodanese domain: PMID:42427617
• MCP and TM: PMID:42427617
• RBMs dispensable: PMID:42427617
• LCI16 binds (IP, AlphaFold model with the rhodanese domain) but is not needed. PMID:42427617
• Conclusion: RBMP2 is a membrane-shaping factor for the CAH3-hosting reticulated region,
  not an established tether. A tether contribution by its RBMs, redundant with other RBM
  proteins, is still untested.

GO decisions

• NEW CC GO:0160223 pyrenoid tubule (IDA, PMID:33177094; also PMID:42427617).
• NEW BP GO:0010027 thylakoid membrane organization (IMP, PMID:42427617). This passes the
  participation test: RBMP2 is in the remodelled membrane, its MCP/TM regions are needed
  for the shape change, and overexpression enlarges the region. It matches the term used
  for SAGA1 and MITH1. GO:0097749 membrane tubulation was considered, but the tubules
  form without RBMP2, so it is not the tubulation factor. What RBMP2 does is extend them
  and convert them into the reticulated region.
• No MF asserted. GO:0180020 membrane bending activity is the candidate if bending is
  shown in vitro. GO:0043495 protein-membrane adaptor activity (used for SAGA1/MITH1) is
  not supported, because the RBMs are dispensable. Sulfurtransferase or phosphatase
  activity is excluded on sequence grounds.

Module (pyrenoid_ccm, node tubule_matrix_tethering)

• Proposed annoton: participant UniProtKB:A0A2K3DG19; function given as free text
  only (no GO id; "membrane-shaping protein, activity unknown"); process GO:0010027;
  location GO:0160223. The role is to build the reticulated region that positions CAH3.
  The node text "RBMP2 ... untested" should be updated: RBMP2 has now been tested and
  shapes membranes rather than tethering them.