Diatom fucan polysaccharide precipitates carbon during algal blooms

The biological carbon pump, driven by growth, aggregation, and sinking of phytoplankton, exports a large fraction of fixed carbon to the deep ocean, potentially sequestering 25–40% of anthropogenic CO2 below 1000 m for millennia. Diatoms release abundant adhesive polysaccharides that can form transparent exopolymer particles (TEP) and larger marine snow, facilitating carbon export at sinking speeds of ~100 m/day. For aggregates to remain intact during descent, matrix polysaccharides must resist degradation by highly active bacterial enzymes for at least ~10 days. However, specific stable algal polysaccharides involved in particle formation have remained unidentified, partly because common analyses hydrolyze polysaccharides to monomers, destroying higher-order structures, and advanced structural methods require purified analytes. Consequently, the origin, structure, and stability of particle-forming polysaccharides remain unclear. This study aimed to identify stable adhesive glycans contributing to particle formation and carbon sequestration by tracking intact polysaccharide epitopes in dissolved and particulate pools across a diatom bloom and linking them to algal sources and bacterial degradation potential.

Prior work shows that dissolved anionic polysaccharides can spontaneously assemble into gels and TEP, promoting aggregation of phytoplankton and marine snow formation. Laminarin, a β-1,3-glucan, is a major diatom storage polysaccharide rapidly degraded by marine bacteria, notably Bacteroidetes and Gammaproteobacteria equipped with GH16/GH17/GH30/GH3 enzymes. Polysaccharides can constitute up to ~30 wt% of marine particles, yet their specific identities and sources have been ambiguous. Macroalgal fucose-containing sulphated polysaccharides (FCSP; fucoidans) are structurally complex and relatively recalcitrant, requiring numerous sulphatases and specialized hydrolases (e.g., GH107 endo-fucanases and GH29/GH95/GH141 exo-fucosidases) for degradation, often found in Verrucomicrobia. Mesocosm studies have linked secreted anionic exudates to TEP and aggregation, but the contribution of specific diatom-origin polysaccharides to such processes in situ remained unresolved.

Field sampling: Coastal North Sea (Helgoland, 54°11.3′N, 7°54.0′E) during spring 2016 (Mar 15–May 26), twice weekly. 100 L seawater was sequentially filtered (10 µm, 3 µm, 0.2 µm) to collect POM in three size fractions (>10 µm; 10–3 µm; 3–0.2 µm). Filtrate (<0.2 µm) was concentrated by tangential flow filtration (1 kDa cutoff) to obtain high molecular weight DOM (HMWDOM). DOC was measured via high temperature catalytic oxidation. HMWDOM recovery by TFF averaged 5.5% of DOM carbon. Polysaccharide extraction and microarrays: POM (as alcohol-insoluble residues) and freeze-dried HMWDOM were sequentially extracted with water, EDTA, and 4 M NaOH (with NaBH4). Carbohydrate microarrays were printed with serial dilutions of extracts and probed with 51 monoclonal antibodies/carbohydrate binding modules targeting defined epitopes. Signals were detected via alkaline phosphatase colorimetry and normalized within HMWDOM and POM datasets. Immunofluorescence microscopy: Filters containing POM (>0.2 µm) were immunolabelled with anti-fucan mAb BAM1 and FITC secondary, counterstained with DAPI, and imaged (Zeiss LSM 780 Airyscan and epifluorescence). FCSP fluorescence on diatom cells and on particle areas (no cells) was quantified; Welch’s two-sided t-tests assessed differences between beginning (March) and end (May) of bloom. Epitope detection chromatography and composition: Water extracts from POM, HMWDOM, and diatom cultures were separated by anion exchange chromatography (Hi-Trap ANX FF). Fractions were screened by ELISA with BAM1 to locate FCSP peaks. Purified FCSP fractions were desalted and acid-hydrolyzed for monosaccharide analysis by HPAEC-PAD and sulphate by ion chromatography. Diatom cultures: Monospecific lab cultures of Chaetoceros socialis (field isolate), Thalassiosira weissflogii, and Nitzschia frustulum were grown in ESAW medium. FCSP occurrence in particulate and dissolved fractions and composition were assessed as above. Metagenomics and metaproteomics: Bacterioplankton (0.2–3 µm) DNA from nine dates was shotgun-sequenced (Illumina HiSeq), assembled (metaSPAdes), and annotated for CAZymes (dbCAN/CAZy; GTDB-Tk for taxonomy). Read mapping yielded gene family abundances (RPKM). Proteins (six dates, triplicate metaproteomes) were extracted, fractionated (SDS-PAGE), LC-MS/MS analyzed (Orbitrap Elite), searched against metagenome-derived databases, and quantified via %NSAF. Focus was on laminarin-degrading GH16/17/30/3 and FCSP-degrading GH107/29/95/141 and sulphatases. Total microbial cell counts were obtained by DAPI staining and automated imaging.

• Antibody microarrays tracked 27 polysaccharide epitopes in HMWDOM and POM across the bloom. FCSP (detected by mAb BAM1) uniquely accumulated in POM for weeks to months and increased across all POM size fractions, indicating high stability; many other epitopes (e.g., arabinogalactan, homogalacturonan) rose only transiently (days to a week), while β-1,3-glucan (laminarin-like) was near detection limits in HMWDOM and transient in POM.
• Temporal dynamics indicated a dissolved-to-particulate transition for FCSP: FCSP increased in HMWDOM in March, declined to a constant level from early April, while increasing in POM from April to May, consistent with aggregation from DOM to POM.
• Quantitative ELISA showed a ~3-fold increase in FCSP concentration in POM from the beginning to end of the bloom.
• Fluorescence microscopy localized FCSP primarily on surfaces of Chaetoceros socialis cells early, later coating cells and present in particles. FCSP signal increased significantly from beginning to end both on diatom cells (P<0.0001; two-sided t-test) and in particle areas (P<0.0001).
• Anion exchange chromatography showed FCSP binds strongly to a cationic matrix and elutes at high NaCl, confirming it is an anionic polysaccharide in both HMWDOM and POM.
• Composition of purified FCSP from C. socialis and from seawater (POM and HMWDOM) showed 25–50% sulphate (per total building blocks) and was enriched in fucose, galactose, xylose, and glucuronic acid; similar composition was found for FCSP from Thalassiosira weissflogii and Nitzschia frustulum (with higher mannose in T. weissflogii FCSP).
• Metagenomes indicated abundant laminarin-degraders (Bacteroidia, Gammaproteobacteria) encoding GH16, GH17, GH30, GH3, whereas putative FCSP-degrading genes (e.g., GH107 endo-fucanases, GH29/GH95/GH141 exo-fucosidases) were at low abundance, mainly in Verrucomicrobiae/Bacteroidia.
• Metaproteomes showed strong, continuous expression of GH16 laminarinases (up to 5.8% NSAF×100) throughout the bloom, with much lower detection of GH3/GH17/GH30. No GH107 endo-fucanases were detected; only low-abundance GH29 exo-fucosidases (0.08% NSAF×100) were observed late in the bloom, indicating little to no active FCSP backbone degradation.
• Environmental context: chlorophyll a varied 2.1–11.8 mg m−3 with consecutive diatom peaks; DOC in DOM remained relatively constant (96–133 µmol L−1). Total microbial cell counts more than doubled during the bloom. Overall, FCSP is a stable, adhesive, diatom-derived polysaccharide that accumulates and aggregates, contributing to particle formation and potential carbon export, while laminarin is rapidly and actively degraded.

The study identifies a specific, structurally complex, anionic polysaccharide—fucose-containing sulphated polysaccharide (FCSP)—as a durable adhesive produced by diatoms during blooms. FCSP’s prolonged accumulation in POM and observed dissolved-to-particulate shift imply that secreted FCSP assembles and promotes aggregation (TEP-like behavior), enhancing formation of marine snow and thus facilitating the biological carbon pump. The minimal expression of FCSP-degrading enzymes compared to robust laminarinase activity explains its persistence; FCSP’s heavy sulphation and branching likely impose high enzymatic costs, deterring rapid microbial degradation. Microscopy pinpoints Chaetoceros (and other diatoms) as FCSP sources, with FCSP increasingly coating cells and particles over time. Beyond aggregation, FCSP’s stability and anionic character suggest potential biological roles such as forming a protective, mucin-like barrier on diatom surfaces, influencing microbial interactions and possibly trace metal binding. The findings provide a mechanistic link between specific diatom glycans and carbon export dynamics, highlighting a defence-aggregation tradeoff where protective exudation also promotes sinking through enhanced stickiness and gel formation.

This work reveals that globally abundant diatoms produce a fucose-containing sulphated polysaccharide that is stable against bacterial degradation, accumulates over weeks, and aggregates from dissolved to particulate phases, thereby contributing to particle formation and potential carbon sequestration. FCSP occurs on diatom cell surfaces and in particles, is highly sulphated and fucose-rich, and is secreted by multiple diatom genera. Concurrent metaproteomics demonstrate intensive laminarin turnover but little evidence for FCSP backbone degradation in situ. These insights identify a concrete molecular candidate underpinning adhesive matrices in marine snow and expand understanding of the biological carbon pump. Future directions include: resolving FCSP molecular architectures and diversity across diatom taxa; quantifying FCSP’s global distribution and flux contribution to carbon export; elucidating environmental controls on FCSP secretion and aggregation; characterizing full enzymatic pathways and organisms capable of FCSP degradation; and assessing interactions with trace metals and implications for biogeochemical cycling.

• Antibody microarray and ELISA approaches are semiquantitative and epitope-specific; they do not provide full structural elucidation and may miss polysaccharide variants not recognized by the probes.
• HMWDOM recovery via TFF averaged ~5.5% of DOM carbon, potentially biasing dissolved pool representation.
• Gene presence in metagenomes indicates potential, not activity; lack of metaproteomic detection of some enzymes could reflect low abundance, detection limits, or transient expression.
• Quantitative ELISA used a macroalgal fucoidan standard, which may not perfectly match diatom FCSP epitopes, affecting absolute concentration estimates (fold-change interpretation is more robust).
• Site-specific coastal time series (North Sea) may limit generalizability; lab diatom cultures were non-axenic and grown under specific conditions.
• Independent normalization of HMWDOM and POM microarray datasets precludes direct comparison of absolute abundances between pools.