henhaobio
Sign in

bioRxivCell BiologyPreprint

A spatial atlas of the seaweed CO2-fixation machinery reveals a unique Rubisco condensation mechanism

James BarrettiD, Clément DégutiD, Alejandro Sebiani CalvoiD, Mike Hodgkinson, Hannah Derry, Aidan SalteriD, Emma Coulembier VandelannooteiD, Rebecca BeveridgeiD, John H. F. BothwelliD, Thomas B. JacobsiD, Olivier De ClerckiD, James N. BlazaiD, Michael J. PleviniD, Luke C. M. MackinderiD, Jonas BlommeiD

PDFDOICodeData
In-depth full-text digest available · Jump to digest ↓

Abstract

Seaweeds (macroalgae) are important primary producers that sustain food webs in coastal ecosystems. Most algae accelerate inorganic carbon assimilation by actively concentrating CO2 in a Rubisco-rich specialized organelle called the pyrenoid. However, the molecular composition of this pathway is unknown in seaweeds. Here, we investigated the intracellular localization of 160 proteins associated with CO2 acquisition in the green seaweed Ulva (Sea lettuce). We assign 68 proteins to different pyrenoid subdomains and identify a consensus Ulva Rubisco binding motif revealing the molecular logic of the Ulva pyrenoid. We reveal Seaweed Ulva Pyrenoid Assembly 1 (SUPA1) as the core pyrenoid assembly factor. We show that Rubisco condensation is driven by a unique mechanism: the helical folding of SUPA1 motifs upon Rubisco binding, combined with steric hindrance that halves the available Rubisco binding sites from eight to four. Our data gives an unprecedented sub-cellular spatial understanding on seaweed carbon fixation and provides insights into the evolution of this important pathway in the global carbon cycle.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (16 Jul 2026) · CC BY 4.0

Fluorescent tagging of 160 proteins in the seaweed Ulva maps its carbon-concentrating pathway, and identifies SUPA1, a Rubisco linker that binds only four of eight sites.

The molecular makeup of the pyrenoid-based CO2-concentrating pathway was unknown in seaweeds, and the most extensive spatial mapping so far had been done in the unicellular Chlamydomonas. This work provides a protein localization atlas in Ulva and shows that its pyrenoid assembly differs from the Chlamydomonas linker system.

Key results

  1. ShownA tagging screen in Ulva localized 160 pCCM-associated proteins to 34 subcellular regions, with a higher construct success rate than comparable Chlamydomonas screens.

    201 of 309 genes (65%) were tagged with YFP; 160 gave localization data (80% of constructs, vs 33-49% in Chlamydomonas). 57% of transformants per construct expressed the fusion vs <15% in Chlamydomonas. 3-29 transformants imaged per construct (average 10). · Large-scale fluorescent tagging assigns 160…

  2. ShownMany proteins localize to pyrenoid subdomains, but orthologs of most core Chlamydomonas pyrenoid assembly proteins were not found in Ulva, suggesting that pyrenoid assembly differs.

    68 proteins localized to the pyrenoid, 15 exclusively. No orthologs found for EPYC1, SAGA1-2, RBMP2 or MITH1; only the EPOS1 ortholog (UM149_0033) was found. Only eight proteins shared the same exclusive pyrenoid subdomain in Ulva and Chlamydomonas. · Core Chlamydomonas pyrenoid structural…; Proteins associate to Ulva pyrenoid subdomains

  3. ShownSUPA1 (UM120_0037) is the pyrenoid matrix protein that condenses Rubisco: its knockout disperses Rubisco, and SUPA1 plus Rubisco form condensates in vitro.

    One of 8 FLIPPer candidates was detected by MS (54th most abundant of 2895 proteins). Rubisco partitioning in the pyrenoid in RbcS1-YFP knockout lines: median 28% (site 1, 4 individuals) and 30% (site 2, 7 individuals) vs 86% in control. FRAP T0.5 was 91.32 ± 9.3 s for RbcS1 and 111.4 ± 9.3 s for SUPA1. · Identification of the Rubisco-condensing Ulva…; SUPA1 condenses Rubisco into dynamic condensates

  4. SuggestedA 52-residue SUPA1 R4 fragment appears to fold into a short helix on binding Rubisco, and overlap of adjacent grooves is interpreted as allowing at most four of the eight sites to be occupied.

    Cryo-EM 1.8-2.2 Å (R4 region ~2.0 Å). ITC gave KD = 0.85 ± 0.15 μM with N = 2.8 R4 per Rubisco. Native MS at 8:1 R4:Rubisco showed only 2, 3 or 4 bound R4 and no higher species. K164A/R165A abolished detectable ITC binding. · Structural basis of SUPA1 binding to Rubisco; Perspective

  5. ShownA SUPA1-derived Rubisco-binding motif occurs in four additional proteins, all of which localized to the pyrenoid. Within the Ulvophyceae, SUPA1 homologs were found only in the Ulvales order, plus fragments in some red algal transcriptome reads.

    MEME/MAST search found 4 additional RBM proteins (UM003_0241, UM004_0280, UM016_0026, UM047_0002), 3 with 7 or more RBMs and one with a single RBM. Imaging data from all 4 showed pyrenoid localization. SUPA1 homologs were seen in the Ulvales families only (Table S6). · Pyrenoid structural components contain a…; Conservation of SUPA1

Methods, data and reproducibility

Orthology- and physicochemical-based selection of pCCM genes, C-terminal YFP tagging of 201 genes and confocal localization in stably transformed Ulva. SUPA1 was then tested by CRISPR knockout, FRAP, in vitro droplet assays, cryo-EM, ITC and native MS.

  • YFP-tagged Ulva transformant lines; pCCM-associated genes tagged and imaged by confocal microscopy · Ulva compressa/mutabilis · n = 201 genes tagged; 160 localized; 3-29 transformants per construct
  • Whole-cell and crude pyrenoid-enrichment proteomics (DIA mass spectrometry) · Ulva sp. · n = triplicate samples · MSV000102502; PXD081181
  • Cryo-EM of Ulva Rubisco bound to SUPA1 repeat R4 (52-residue fragment) · Ulva intestinalis · n = 3,200 micrographs; 462,744 particles in final refinement · EMD-58973; PDB 32LL
  • CRISPR SUPA1 knockout lines in an RbcS1-YFP background · Ulva compressa/mutabilis · n = 4 and 7 individuals at two target sites
  • In vivo FRAP of RbcS1-YFP and SUPA1-YFP lines · Ulva compressa/mutabilis · n = n = 19, 20
  • Public RNA-seq reads mapped to RbcS1 and RbcS2 coding sequences · Ulva · ERR2722103; ERR2722112-ERR2722114
Methods & tools
PLAZA Integrative Orthology Viewer, FLIPPer, MEME suite / MAST, PB-Chlamy, PredAlgo, TargetP, GreenGate / Golden Gate cloning, Cas9 RNP CRISPR targeted insertion, Spectronaut, limma via FragPipe-Analyst, RELION 5, AlphaFold3, Phenix / ISOLDE / Moorhen, MsMBG-Fit
Software & environment
RELION 5; Geneious Prime 2023.0.1; BWA-0.7.18; SAMtools-1.21; SRA Toolkit-3.2.0; Volocity V2.1; FluoView FV10-ASW 4.2.
Reproducibility gaps
Per-protein localization calls, ortholog lists and proteomics tables are in supplementary tables, not the main text. Raw confocal image deposition is not stated.
Limitations (5)
  • A constitutive promoter and a single developmental time point were used, so localization artefacts or incomplete observations cannot be excluded. Flagellar localization could not be assessed. · stated by the authors
  • The dataset relies heavily on sequence-similarity-based ortholog identification, so Ulva-specific pCCM components may be missing; unbiased interaction methods could reveal further ones. · stated by the authors
  • The apparent continuous cryo-EM density across adjacent sites was interpreted as an artefact of averaging partially occupied sites that classification could not separate, so site occupancy rests partly on interpretation. · stated by the authors
  • Structure, ITC and native MS used only a 52-residue fragment of repeat R4, not full-length SUPA1. Reduced stoichiometry in the full multivalent linker in vivo is inferred rather than measured. · noted in this summary
  • C-terminal YFP fusions under a strong promoter may alter localization or abundance. Roles proposed for RBM proteins and kinases rest on localization alone, with no knockout tests. · noted in this summary

Read this if you study pyrenoid or CO2-concentrating mechanism evolution, algal protein-tagging screens, or how Rubisco linker proteins condense Rubisco.

Citations
1

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISMeyer '20Atkinson '20Elad '25Barrett '24Chen '25Nam '23
71% semantic matchCell Biology
M. T. Meyer, A. K. Itakura, W. Patena+7 more·bioRxiv·Aug 16, 2020