Cosmopolitan marine bacteria facilitate a vast phytoplankton-derived sulfonate-based carbon flow through sulfoquinovosidases

Ma, Xiaoyi; Wang, Huanyu; Dong, Changjie; Liu, Le; Qiu, Xuanyun; Chen, Xiaofeng; Chen, Quanrui; Wang, Liang; Zhang, Yao; Jiao, Nianzhi; Williams, Spencer J.; Tang, Kai

doi:10.1038/s41467-025-66898-z

Download PDF

Article
Open access
Published: 05 December 2025

Cosmopolitan marine bacteria facilitate a vast phytoplankton-derived sulfonate-based carbon flow through sulfoquinovosidases

Nature Communications volume 17, Article number: 209 (2026) Cite this article

2534 Accesses
1 Citations
5 Altmetric
Metrics details

Subjects

Abstract

Sulfoquinovose (SQ) and sulfoquinovosyl glycerol (SQGro) are derived from abundant membrane sulfolipids termed sulfoquinovosyl diacylglycerols (SQDG) and produced by photosynthetic organisms, serving as sources of carbon and sulfur for bacteria. The conversion processes of these sulfoquinovosyl compounds within marine ecosystems, and their quantitative contributions to the marine organic matter pool, are poorly understood. Here, we identify Alteromonas macleodii, a marine bacterium capable of metabolizing SQ and SQGro through a sulfoquinovosidase. This enzyme converts SQGro to SQ and is a member of a clade within glycoside hydrolase family 31, distinct from other sulfoquinovosidases. The ubiquitous presence of sulfoquinovosidases and their transcripts throughout marine environments implicates active metabolism of sulfoquinovose glycosides, particularly in the sunlit surface ocean. We further demonstrate that marine algae produce significant quantities of cellular SQGro, and we estimate the annual turnover of SQGro using field samples from coastal and open ocean environments. Together with SQDG and SQ, these sulfoquinovosyl compounds constitute a substantial portion of the marine organic carbon turnover, estimated at around 1.5 petagrams of carbon per annum. These findings reveal a vast, previously unappreciated pool of organosulfonates within the microbial food web that contributes significantly to the marine carbon and sulfur cycles.

Sulfoquinovose is a widespread organosulfur substrate for Roseobacter clade bacteria in the ocean

Article 02 January 2023

Response to substrate limitation by a marine sulfate-reducing bacterium

Article 20 July 2021

Functional annotation and importance of marine bacterial transporters of plankton exometabolites

Article Open access 25 April 2023

Introduction

Organic matter flows through the marine microbial food web, facilitating nutrient and energy exchange between autotrophs and heterotrophs, and sustaining ecosystem function¹. Annual net primary production consists of 1.3 billion tons of sulfur fixed by phytoplankton in the global ocean, and forms the foundation of the ocean’s organic sulfur pool². Organosulfur compounds provide marine microorganisms with protein building blocks (cysteine and methionine)³, photosynthetic membrane lipids⁴, and molecules that play crucial roles as cell osmolytes⁵, redox balancers⁶, predator deterrents⁷, and hydrostatic pressure protectants⁸. Bacteria can catabolize various organosulfur compounds from the exudates, detritus, and viral lysates of phytoplankton. The canonical elemental ratio of sulfur to carbon in marine phytoplankton biomass is approximately 1:95^9,10. Organosulfur compounds produced by phytoplankton, such as dimethylsulfoniopropionate (DMSP), provide a vital sulfur source for bacteria¹⁰. Furthermore, these compounds also supply carbon for heterotrophic bacteria, as exemplified by molecules like organosulfonates^11,12. Abundant organosulfonates serve as ecologically important metabolic currencies between phytoplankton and heterotrophic bacteria, forming a metabolic network in the surface ocean¹³. Consequently, organosulfur compounds function as key molecular links between the marine sulfur and carbon cycles.

Marine organosulfonates include a range of simple C2 (e.g., taurine, isethionate) and C3 carbon species (e.g., 2,3-dihydroxypropane-1-sulfonate (DHPS))^14,15, as well as the C6 sulfosugar sulfoquinovose (SQ; 6-deoxy-6-sulfo-D-glucose) (Supplementary Fig. 1a)¹⁶. Marine algae generate SQ (10–27,000 μM), which is metabolized by marine Roseobacteraceae and Oceanospirillaceae bacteria through three distinct but related pathways: the sulfolytic SQ monooxygenase (sulfo-SMO), a variant of SQ Embden−Meyerhof−Parnas (sulfo-EMP2), and the sulfolytic SQ dioxygenase (sulfo-SDO) pathways^17,18,19. Sulfoquinovosyl diacylglycerol (SQDG) (Supplementary Fig. 1a), a precursor of SQ, is widespread in the membranes of various photosynthetic organisms, where it plays a crucial role in chloroplast and photosynthesis development and function^20,21,22. Its annual production in the biosphere has been estimated to be on the order of 10 billion tons²³. Available evidence suggests that SQDG biosynthesis involves UDP-SQ synthase and UDP-SQ: diacylglycerol glycosyltransferase, however, these enzymes have not been studied within marine algae. Marine diatoms and cyanobacteria, responsible for approximately 40% and 25% of ocean net primary productivity, respectively^24,25, can increase their SQDG production to compensate for phospholipid deficiencies in oligotrophic environments^26,27. In the diatom Thalassiosira pseudonana, SQDG contributes around 9% of its cellular sulfur pool¹⁴. For cyanobacteria, SQDG content varies between 50% and 60% of their total polar membrane lipids²⁶. SQDG can be detected throughout the water column, from the upper ocean to the deep sea, and is a potential vector for vertical carbon transport in the ocean^28,29.

Sulfoquinovosyl glycerol (SQGro) (Supplementary Fig. 1a), the delipidated form of SQDG, was first described in the freshwater alga Chlorella ellipsoidea³⁰. Terrestrial and gut microorganisms harbor sulfoquinovosidases belonging to subfamily 13 of glycoside hydrolase (GH) family 31 that can hydrolyse SQGro into SQ³¹. These enzymes have not been found in marine bacteria. However, the recent discovery of a dedicated family of sulfoquinovosidases (family GH188), which also hydrolyze SQGro to release SQ, revealed that these enzymes are widespread in marine Roseobacteraceae and marine algae¹⁸. However, SQGro itself has not been identified in the ocean. Furthermore, the role of sulfoquinovosyl compounds in marine carbon flux remains poorly understood, particularly regarding their conversions and contributions to the organic matter pool.

In this work, we identify common marine bacteria capable of metabolizing SQ and SQGro, and characterize a previously undescribed sulfoquinovosidase defining an unexpected trait within a distinct clade of GH31 sulfoquinovosidases. We quantify the distribution of sulfoquinovosyl compounds in representative algae, showing that they are ubiquitous and can be metabolized by GH188 sulfoquinovosidases. By analyzing field samples, we estimate the total amount of marine sulfoquinovosyl compounds. These findings, combined with metagenomic and transcriptomic data from the ocean, suggest a potentially vast flow of carbon based on organosulfonates within the microbial food web, that likely fuels the marine carbon cycle.

Results

Identification of SQ-catabolizing bacteria within diatom-associated microbial communities

DNA stable-isotope probing (DNA-SIP) analysis of bacteria isolated from a non-axenic culture of Thalassiosira pseudonana 220 showed that bacteria belonging to the Alteromonadaceae and Roseobacteraceae families incorporated ¹³C from ¹³C₆-SQ added to the media (Fig. 1a, b), indicating the capacity of some genera within these two families to metabolize SQ. Within Alteromonadaceae, only the genus Alteromonas demonstrated a 16S rRNA gene abundance exceeding 1% of the total community and showed significant ¹³C assimilation, with an atom-fraction excess of ¹³C up to 0.62 (Supplementary Data 1). PacBio sequencing results indicated that the exogenously added ¹³C₆-SQ stimulated the growth of Alteromonadaceae with an increasing population fraction from 0.4% to 40.9%, wherein it became the dominant taxa in the system (Fig. 1c). A. macleodii exhibited the highest abundance, accounting for 74% of the Alteromonas species and 30.4% of the total bacterial 16S rRNA gene sequences (Fig. 1d), suggesting that this species may play a key role in SQ metabolism. In contrast, the Roseobacteraceae demonstrated greater diversity, and included Roseobacter denitrificans and Dinoroseobacter shibae (Fig. 1e). Our previous research showed that Roseobacter denitrificans OCh 114 and Dinoroseobacter shibae DFL 12 are SQ-metabolizing species¹⁹.

**Fig. 1: DNA-stable isotope probing reveals different taxa incorporate ¹³C from ¹³C₆-SQ.**

We studied A. macleodii ATCC 27126 and demonstrated its ability to utilize SQ as a growth substrate. After 36 h of cultivation, the strain consumed approximately 64% of the SQ in the media, accompanied by the production of sulfate and sulfite (Fig. 1f). Comparative proteomic analysis revealed a cluster of eight proteins in A. macleodii ATCC 27126 that were present at significantly increased abundance during growth with SQ compared to without (locus tags: MASE_17990-18035; fold-change > 2, p < 0.05) (Fig. 1g). Among the identified proteins, MASE_18005 (named AmSqoD) was 58-fold more abundant during growth with SQ (Supplementary Data 2). It exhibited 34% sequence identity to MaSqoD from Marinobacterium aestuarii ST58-10, an Fe²⁺ dependent α-ketoglutarate (α-KG): SQ dioxygenase in the sulfo-SDO pathway (Fig. 1i)¹⁷. Phylogenetic analysis revealed that AmSqoD and MaSqoD cluster together, suggesting a close evolutionary relationship (Supplementary Fig. 1b and Supplementary Data 3). Furthermore, they share similar predicted 3D structures with a root-mean-square deviation of 1.3 Å and possess conserved catalytic and binding residues (H121, E123, and H313 in AmSpoD, H115, E117, and H308 in MaSqoD) as determined by mutation analysis (Supplementary Fig. 1c, d). Incubation of SQ and α-KG with recombinantly expressed and purified AmSqoD led to the release of sulfite as detected by a colorimetric Fuchsin assay (Fig. 1h), indicating that AmSqoD can catalyze the cleavage of SQ¹⁷. The product of the reaction was derivatized with 2,4-dinitrophenylhydrazine (DNPH) and then analyzed by LC-MS, which revealed a peak at m/z (-ve) 357.0318, indicating the formation of the DNP-hydrazone of 6-dehydroglucose (Supplementary Fig. 1e, f). The relevant gene cluster in A. macleodii ATCC 27126 exhibits distinct features for SQ metabolism compared to M. aestuarii ST58-10 (Fig. 1i). It includes MASE_17995 and MASE_17990, which were present at higher abundance in the comparative proteomics data and are predicted to encode a two-component sulfite dehydrogenase, and MASE_18000, predicted to encode a sulfate transporter (Fig. 1i). Additionally, a TonB-dependent receptor (MASE_18030), potentially facilitating SQ uptake, exhibits a dramatic 107-fold higher abundance (Supplementary Data 2). These findings reveal A. macleodii ATCC 27126 as a marine bacterium capable of metabolizing SQ. Evolutionary analysis indicated that most of the A. macleodii strains (22/24) contain the sulfo-SDO pathway responsible for SQ metabolism (Supplementary Fig. 2).

Growth of marine bacteria on SQ glycosides: discovery of a novel sulfoquinovosidase in A. macleodii ATCC 27126

R. denitrificans OCh 114 can grow using SQGro (Fig. 2a). This bacterium encodes a family GH188 protein located upstream of the previously identified sulfo-SMO gene cluster of R. denitrificans OCh 114¹⁹. Analysis of the recombinant protein, RdSqgA, revealed it to be an NAD⁺-dependent sulfoquinovosidase that acts on SQGro to form SQ (Supplementary Fig. 3a, b). Growth experiments conducted with A. macleodii ATCC 27126 showed that it consumed 77% of SQGro within 36 h (Fig. 2b), producing sulfite and sulfate, suggesting it contains an SQase capable of cleaving SQGro to give SQ, which is metabolized through a sulfo-SDO pathway. Proteins within the SQ-metabolizing gene cluster of A. macleodii ATCC 27126 that are of higher abundance when cultured on SQ are also significantly increased in abundance (2-223 fold change, p < 0.01) when cultured in the presence of SQGro (Fig. 2c and Supplementary Data 2). MASE_18025 (named AmSqgH) shares low protein sequence identities of 27.9% with sulfoquinovosidase EcYihQ from Escherichia coli K-12 and 28% with sulfoquinovosdiase AtSQase from Agrobacterium fabrum C58^31,32. LC-MS analysis of the reaction mixture containing SQGro and recombinant AmSqgH revealed a peak at m/z 243.0170 that co-eluted with authentic SQ (Fig. 2d, e). Recombinant AmSqgH was highly active on p-nitrophenyl sulfoquinovoside (PNPSQ) (Fig. 2f), with Michaelis–Menten parameters k_cat = 119 ± 19 s⁻¹, K_M = 13.0 ± 2.2 mM, and k_cat/K_M = 9170 ± 1490 M⁻¹ s⁻¹ (Fig. 2g and Supplementary Fig. 3c). These results establish that AmSqgH is a sulfoquinovosidase.

**Fig. 2: SQGro is metabolized by R. *denitrificans* OCh 114 and *A. macleodii* ATCC 27126.**

AmSqgH, EcYihQ, and AtSQase share a common (a,b)₈-barrel fold with an additional β-sheet domain³³. However, EcYihQ and AtSQase have additional connecting β-sheets absent in AmSqgH (Fig. 2h). Despite sharing a conserved catalytic dyad (Asp-Asp) comprising the catalytic nucleophile and general acid-base for glycoside cleavage, 3D modelling, docking and mutational analyses reveal significant divergence in the amino acid residues responsible for substrate binding (Supplementary Fig. 3d). The important Q/E-RWY motif (Q288(E270), R301/R283, W304/W286 and Y508/Y491) found in EcYihQ and AtSQase for the structural recognition of the side chains and sulfonate group of SQGro is not conserved in AmSqgH (Supplementary Fig. 3d). AlphaFold2 analysis predicts that AmSqgH uses a distinct set of residues (R302, T377, K412, F513) to recognize and bind SQGro (Supplementary Fig. 3d, e). AmSqgH and RdSqgA, like EcYihQ, exhibit SQDG cleavage activity (Supplementary Fig. 3f, g). The absence of a signal peptide in these sulfoquinovosidases suggests an intracellular localization. This is consistent with the observation that R. denitrificans OCh 114 cannot utilize exogenously added SQDG (Supplementary Fig. 3h), suggesting that SQDG cannot be directly imported into the cell. However, the presence of a TRAP transporter and substrate binding protein in the gene cluster suggests that R. denitrificans OCh 114 can import SQ/SQGro (Supplementary Fig. 3i–k). Therefore, SQGro is likely the natural substrate for these bacterial intracellular sulfoquinovosidases.

AmSqgH homologs (SqgH) were identified in marine gammaproteobacteria, including Alteromonadaceae, Pseudoalteromonadaceae, Colwelliaceae, Arenicellaceae, and Cellvibrionaceae, and share similar predicted structures and possess conserved catalytic and binding residues (Fig. 3, Supplementary Fig. 4a, and Supplementary Data 4). These gammaproteobacteria harbor gene clusters containing both sqgH and sulfo-SDO pathway genes (Supplementary Fig. 4b). The gene cluster compositions closely resemble that found in A. macleodii ATCC 27126, although some, such as those in Pseudoalteromonas, lack genes associated with sulfite metabolism (Supplementary Fig. 4b). The metabolic potential encoded by these gene clusters is exemplified by Alteromonas stellipolaris R10SW13 and Pseudoalteromonas marina ECSMB 14103, which are both capable of metabolizing SQGro and SQ (Supplementary Fig. 4c–f). A. macleodii F12, which lacks this gene cluster or any others associated with SQ metabolism, exhibited no evidence of SQGro or SQ metabolism in feeding experiments (Supplementary Fig. 4g, h).

**Fig. 3: Phylogenetic analysis of GH31 subfamily distribution.**

SqgH belongs to a distinct clade of family GH31

Phylogenetic analysis of AmSqgH and its close homologs revealed that they form a distinct clade separate from the previously classified 20 subfamilies within GH31 (Fig. 3). Previously characterized GH31 sulfoquinovosidases, such as EcYihQ and AtSQase, are members of subfamily GH31_13 and are associated with various SQ degradation pathways, excluding the sulfo-SDO pathway (Fig. 3). Despite their close phylogenetic relationship to the SqgH clade, GH31_3 and GH31_6 subfamily proteins exhibit no detectable activity on PNPSQ (Fig. 3). GH31_3 possesses activity towards 4-nitrophenyl α-D-xylopyranoside (PNPXyl) or 4-nitrophenyl α-D-glucopyranoside (PNPGlc)³⁴, while AmSqgH, similar to GH31_13 members (EcYihQ and AtSQase)³³, shows no detectable activity towards these two substrates (Fig. 3). The SqgH clade and GH31_13 subfamily appear to be examples of convergent evolution, with each independently arising from a common ancestor within family GH31, and developing the catalytic capacity for SQGro in marine and non-marine environments, respectively. While phylogenetic analysis of the AmSqgH homologues highlights that they form a clade distinct from subfamilies 3 and 6, application of the CAZy classification methodology based on hidden Markov models results in formal assignment into subfamily 3 (Supplementary Data 5).

Ubiquitous distribution of sulfoquinovosidase and SQ-catabolizing genes in marine environments

Gene abundances of sqgH and sqgA homologs were investigated in sequences stemming from the Tara Oceans expeditions³⁵ and coastal stations (Bohai Sea and Jiulongjiang Estuary)³⁶ and were compared with those for sqoD in sulfo-SDO and smoC in sulfo-SMO (encoding SQ monooxygenase). The sqgH, sqoD, sqgA, and smoC genes are widely expressed in the global ocean (Fig. 4a and Supplementary Data 6–8). The abundance of prokaryotic sqgH transcripts (median: 0.09%) was almost equal to that of sqoD (median: 0.11%), while sqgA transcript abundance (median: 0.88%) was slightly lower than that of smoC (median: 1.60%) (Fig. 4a). This suggests that prokaryotic sqgH and sqoD, as well as sqgA and smoC, collaborate for heterotrophic metabolism of SQGro and SQ in the ocean. A positive relationship between chlorophyll a concentration and abundance was found for sqgH, sqgA, sqoD, and smoC, in both metagenomes and metatranscriptomes (Fig. 4b, c), suggesting that active bacterial metabolism of SQGro occurs in the global ocean, particularly in regions of high photosynthetic productivity. Regions with high chlorophyll a values (>2 mg/m³) coincided with high-latitude zones (60–90° N and 60–90° S), such as the Arctic and Southern Oceans (Fig. 4d, e and Supplementary Data 7). These areas harbor thriving communities of photosynthetic organisms, potentially acting as abundant sources of SQGro for epipelagic consumers. Among all Tara Oceans stations, bacterial sqgH metagenomics, transcript and sqoD transcript abundances were slightly lower in the epipelagic zone (depth, <200 m) compared to the mesopelagic zone (depth, 200–1000 m) (Fig. 4f). The metagenomics abundance of bacterial sqgA and smoC, along with their transcript levels, was significantly higher in the epipelagic zone compared to the mesopelagic zone (Mann–Whitney test, p < 0.001) (Fig. 4f). This indicated that the metabolic activity for SQGro and SQ is widely distributed across the epipelagic and mesopelagic ocean. Additionally, sqgA genes are also present marine gammaproteobacteria, particularly Oceanospirillaceae (Supplementary Fig. 5). These SqgA proteins exhibit sulfoquinovosidase activity (Supplementary Fig. 6a–c). However, transcripts of sqgA and the sulfo-SDO pathway gene sqoD from Oceanospirillaceae were rarely detected in Tara Oceans samples. These findings suggest that Alteromonadaceae and Roseobacteraceae bacteria are the principal consumers of SQGro and SQ in the global ocean.

**Fig. 4: Distribution and expression of bacterial *sqgH*, *sqoD*, *sqgA*, and *smoC* genes in the global ocean.**

Abundances of sulfoquinovosyl compounds in culturable algae

SQDG, SQGro, and SQ were detected within all investigated cyanobacteria and algae, including diatoms, coccolithophore, flagellates, dinoflagellates, green algae, Prochlorococcus, and Synechococcus (Table 1). SQGro concentrations in cultured algae ranged from 40 to 1870 μM, while SQ concentrations ranged from 8 to 242 μM (Table 1). The lowest cellular concentrations of SQ and SQGro were measured in the dinoflagellate Alexandrium tamarense, whereas the highest concentrations were measured in the diatom Skeletonema costatum. Among algae, diatoms exhibited SQGro concentrations nearly tenfold higher than SQ. Coccolithophore and flagellates of haptophyta species also showed significantly higher SQGro than SQ (Table 1). In contrast, cyanobacteria exhibit higher intracellular SQ concentrations than SQGro (Table 1). SQDG was present at millimolar cellular concentrations in all investigated algae, with concentrations two to three orders of magnitude higher than SQGro (Table 1). SQGro was detected in extracellular metabolite pools of all investigated marine algae cultures (Table 1), suggesting that this compound can be released from photosynthetic organisms into seawater in marine environments.

Table 1 Analysis of organic sulfonate metabolites in pure plankton cultures

Full size table

Abundance of organosulfonates in the ocean

We conducted field investigations on sulfoquinovosyl compounds and DHPS from marine particulates (Fig. 5a). These compounds have a widespread distribution throughout the euphotic layer in coastal and open ocean environments (Fig. 5b and Supplementary Data 9). The concentrations observed are positively correlated with chlorophyll a: SQDG (Pearson’s r = 0.77, p = 7.5 × 10⁻⁸), SQGro (r = 0.61, p = 1.4 × 10⁻⁴), and SQ (r = 0.70, p = 1.5 × 10⁻⁵) and DHPS (r = 0.35, p = 6 × 10⁻²), respectively (Fig. 5c–f). Particulate organic carbon (POC) also showed a positive correlation with chlorophyll a (r = 0.82, p = 3.9 × 10⁻⁹) (Fig. 5g). This indicates a primarily photosynthetic origin for these organosulfur compounds in the ocean. Geographical patterns indicate that coastal environments exhibit significantly higher concentrations of sulfoquinovosyl compounds compared to the open ocean (Supplementary Data 9). SQDG was present at significantly higher concentrations than SQGro (Tukey’s test, p = 3.3 × 10⁻¹²) and SQ (p = 3.0 × 10⁻¹⁰) (Fig. 5b and Supplementary Data 9). The concentrations of SQGro and SQ are comparable to that of DHPS, widely recognized as an important species in the marine organosulfonate metabolic network (Fig. 5b). SQDG in all survey stations was mainly composed of medium-chain fatty acids with a total of 28–32 carbon atoms in the fatty acid chains (Supplementary Fig. 7a). The median molar C:S ratio in the marine environment reaches up to 42:1 (Supplementary Fig. 7a). Compared with SQ, SQGro, and DHPS (median: 0.11%, 0.01%, 0.01%), SQDG contributed more to POC (median: 5.5%), which was similar to the proportion of 4.2% POC (mean value) contributed by SQDG in the Northeast Atlantic Ocean (Supplementary Data 9)²⁹. While these sulfoquinovosyl compounds represent only 5.3% of coastal POC, their contribution can reach up to 7.4% of POC in the open ocean, potentially enhancing surface water carbon export in the global open ocean.

**Fig. 5: Organosulfonates are abundant in the marine environment.**

Discussion

Bacteria possessing sqgH or sqgA sulfoquinovosidase genes are primarily found in marine environments, including coastal regions, the open ocean, and in association with phytoplankton (Supplementary Fig. 5 and Supplementary Data 10). Our ¹³C₆-SQ incubation experiments suggest that SQ catabolism occurs within phycosphere-associated microbial communities. Among identified Roseobacteraceae genera containing sqgA sulfoquinovosidase genes, 23 out of 94 were isolated from niches associated with phytoplankton, such as diatoms, dinoflagellates, and red algae (Supplementary Fig. 5 and Supplementary Data 10). Besides inhabiting the phycosphere, Alteromonadaceae species are also known to interact closely with cyanobacteria, influencing their growth, survival, and metabolic processes³⁷. Alteromonadaceae and Roseobacteraceae bacteria are often opportunistic, utilizing phytoplankton-derived carbon sources abundant in surface waters^38,39. Analysis of cultures with added SQ showed a shift in dominant populations, with Alteromonadaceae surpassing the previously dominant Roseobacteraceae populations (Fig. 1c). Alteromonas bacteria are ubiquitous in marine environments, thriving in waters spanning from polar to tropical regions and from the surface to the deep ocean⁴⁰. They assist in remineralizing sinking marine particles, such as algal aggregates, and contribute significantly to global carbon cycling⁴¹. Additionally, Pseudoalteromonas is commonly found within marine particle-associated microbial communities⁴². Consequently, these organisms have potential to utilize sulfoquinovosyl compounds present in POC. Sulfoquinovosyl compounds on POC exhibited a dramatic decline with increasing seawater depth in the Central North Pacific (Supplementary Fig. 7b), indicating that most sulfoquinovosyl compounds are readily metabolized by bacteria in the sunlit surface ocean.

This study provides updated estimates of the total annual production of SQDG, SQ, SQGro, and DHPS in the sunlit surface ocean. Based on the correlation between sulfoquinovosyl compounds and chlorophyll a and assuming steady-state kinetics, using the median values, their relative contributions to the POC pool were combined with net primary productivity (53 Pg C/year (yr), 1 Pg = 1 petagram = 10¹⁵ × g) and the proportion of algal POC used for heterotrophic consumption (50%) to estimate the total amounts of these organosulfonates^43,44. Previous studies considered the variability in bacterial assimilation of organic matter (ranging from 30% for glycolate to 80% for amino acids, with an average of 50%) and combined with bacterial assimilation of marine fixed carbon of 5–25%, leading to an estimated 10–50% range of marine fixed carbon consumed heterotrophically in both coastal and open ocean environments^45,46,47. Our estimation adopts a conservative 50% value, consistent with previous estimates evaluating laminarin abundance in the ocean⁴⁷. This approach yielded annual production estimates of 1.5 Pg C/yr for SQDG (median, 25% quartile, 0.5 Pg C/yr; 75% quartile, 2.1 Pg C/yr), 28.4 Tg C/yr for SQ (25% quartile, 8.7 Tg C/yr; 75% quartile, 72.1 Tg C/yr) (1 Tg = 1 teragram = 10¹² × g), 3.9 Tg C/yr for SQGro (25% quartile, 1.3 Tg C/yr; 75% quartile, 7.3 Tg C/yr), and 3.0 Tg C/yr for DHPS (25% quartile, 1.4 Tg C/yr; 75% quartile, 8.6 Tg C/yr) in the ocean. Previous estimates of SQDG were based on terrestrial sources and relied on various assumptions, including the sulfur content of sulfolipids being equal to half the sulfur content of total leaf protein²³. This estimation method based on sulfur content is not applicable in the ocean, as we show that the sulfur content of SQDG in marine algae varies greatly from 0.5 to 19% of the total organic sulfur (Supplementary Data 11). An earlier estimate of SQDG production in the biosphere was around 36 Pg/yr (corresponding to roughly 22 Pg C/yr)²³. However, this value seems excessively high considering that the total net primary production on Earth is only around 105 Pg C/yr, with roughly equal contributions from land and ocean²⁴.

Our estimates suggest that SQDG may be the most abundant organosulfonate in the ocean. These vast quantities of SQDG serve as a source of SQGro and SQ through biodegradation processes. A lipase capable of degrading SQDG to SQGro was described in the alga Scenedesmus obliquus⁴⁸. Similarly, a galactolipase in Chlorella kessleri exhibits broad substrate specificity, allowing it to hydrolyze the fatty ester bonds in galactolipids and SQDG⁴⁹. Recent studies have reported that marine bacteria isolated from marine particles can degrade SQDG⁵⁰, suggesting the involvement of unidentified extracellular enzymes that might generate SQGro or SQ and allow utilization of the fatty acid content of SQDG by diverse bacteria. This study extends the known presence of SgqA from bacteria to eukaryotic algae through functional validation¹⁸. Phylogenetic analysis of SqgA sequences indicates that diatom SqgA cluster with gammaproteobacterial homologs, and other algal SqgA also cluster with gammaproteobacterial Marinomonas hwangdonensis and Granulosicoccus antarcticus (Supplementary Fig. 8). Further analysis strengthens the hypothesis that the eukaryotic sqgA originated from gammaproteobacteria. In contrast to earlier findings that detected YihQ homologues in a limited number of algal species⁵¹, this study reveals the widespread presence of SqgA across various marine algal groups, including diatoms, coccolithophores, dinoflagellates, and red algae. Although these homologs exhibited low amino acid identity (31-47%) with RdSqgA (Supplementary Fig. 8 and Supplementary Data 12), they maintain the same conserved catalytic acid/base and nucleophilic residues (K98, H183, Y127, R156, L160, K162, and E179 in RdSqgA) and exhibit an overall 3D structure consistent with the crystal structure of the RdSqgA, which was resolved at a diffraction resolution of 1.8 Å. These homologs have sulfoquinovosidase activity, allowing the liberation of SQ from marine algae derived SQDG or SQGro (Supplementary Figs. 9–11). By contrast, we could not identify any SqgH clade sulfoquinovosidases with conserved residues in marine algae (Supplementary Fig. 12). Metagenomic data and transcripts from eukaryotic organisms in Tara Oceans database show that they express sqgA, with higher expression levels observed in coastal regions (Supplementary Fig. 13a, b and Supplementary Data 13 and 14). In these regions, the Roseobacteraceae typically comprise upwards of 20% of coastal bacterial taxa^52,53. However, the absence of SqgA homologues in algae like Chlorella species and cyanobacteria Prochlorococcus and Synechococcus, along with the lack of other SQGro hydrolyzing enzymes from family GH31, suggests the existence of alternative pathways for SQDG degradation in these organisms.

This work reveals that SQGro and SQ concentrations within diatoms and coccolithophore are one-to-two orders of magnitude lower than that previously reported for DHPS¹⁴. The concentrations of DHPS in marine algal cells are comparable to that of another ecologically important organosulfur compound, DMSP¹³. However, DHPS has a limited distribution, and is found only in diatoms and coccolithophores¹⁴. Although SQGro has the same order of magnitude of abundance as DHPS in marine environment (Fig. 5), sulfoquinovosyl compounds are more widely distributed across algae (Table 1).

While our work shows that the annual production of SQDG is comparable to that of oceanic DMSP (3.8 Pg C/yr), SQDG has a higher ratio of carbon to sulfur compared to DMSP⁵⁴. This is reflected in the total sulfur content of SQDG (102 Tg S/yr), which is 20-fold less than that of DMSP (2.0 Pg S/yr)². In aerobic ocean environments, marine bacteria rely on DMSP for 30-100% of their sulfur needs^55,56, and while they can also utilize sulfite derived from sulfoquinovose for biosynthesis of essential sulfur-containing molecules (e.g., cysteine, methionine, cofactors, and vitamins), most of this sulfite is oxidized to sulfate for energy generation and to avoid toxicity^19,57. In contrast, the sugar and glycerol moieties, along with the fatty acids liberated from sulfoquinovosyl compound degradation, provide community carbon sources and supply metabolic needs^57,58. Consequently, sulfoquinovosyl compounds primarily contribute carbon to the microbial food web, with minimal contribution of their sulfur. Marine bacteria similarly catabolize low carbon-sulfur ratio organosulfonates (e.g., taurine, DHPS, sulfolactate, isethionate) into usable carbon (C2 and C3 molecules) and sulfite^14,59,60. This “sulfonate-based network” allows utilization of carbon from algae for energy and central carbon metabolism by cleaving the carbon-sulfur bond in these compounds^13,61. The sulfite released is then mineralized into sulfate.

Marine algae, major producers of the sulfoquinovosyl compounds SQ and SQDG, also harbor significant quantities of SQGro. Collectively, these compounds constitute a significant fraction of the oceanic organic matter pool, estimated at about 1.5 Pg carbon annually. Marine bacteria metabolizing SQ and SQGro are a key microbial link in the sulfonate-based carbon flow, and their widespread expression of diverse sulfoquinovosidases points to their active participation in the hydrolysis of sulfoquinovose glycosides, enabling catabolism of the released sulfosugar and fueling the oceanic microbial food web (Supplementary Fig. 14). Collectively, these findings underscore a potentially critical link between these abundant organosulfonate species and the essential carbon cycle that sustains marine ecosystems.

Methods

DNA-SIP experiments

¹³C₆-SQ was synthesized using ¹³C₆-glucose as raw material according to the method of Abayakoon et al.⁶². T. pseudonana 220 (CCMA-220) was from the Center for Collections of Marine Algae (Xiamen University, China), and was cultured in f/2 (with Na₂SiO₃) medium at 20 °C with an irradiance of 100 μmol photons m⁻² s⁻¹ until the algal cell abundance was about 10⁵ cells/mL. The attached bacteria were separated by gentle ultrasound, and the T. pseudonana was removed by filtration through a sterile 3-μm polycarbonate filter (47 mm, Millipore, USA). The filtrate was filtered through a sterile 0.2-µm polycarbonate membrane (47 mm, Merck Millipore). The microbial cells collected on the filter membrane were re-inoculated into 1L of sterile aged seawater sourced from the Xiamen offshore (water depth: ~1 m), which had been filtered through a 0.2-µm filter membrane and sterilized by high-pressure steam. The seawater contained 100 μM ¹³C₆-SQ for the experimental group and 100 μM ¹²C-SQ for the control group, and was incubated in the dark at 25 °C in a static state. Three replicates were set up for both the experimental and control groups. A volume of 1 mL samples was collected every 24 h to quantify the changes in SQ. After 3.5 days, the concentration of SQ in the system decreased by 87%. Microbial cells incubated for 3.5 days were collected through a polycarbonate membrane for subsequent DNA-SIP and bacterial community structure analysis.

DNA Extraction and density centrifugation

DNA from microbial cells captured on filter membranes was extracted using the FastDNA Spin Kit for Soil following the manufacturer’s instructions (MP Biomedicals, USA). DNA concentrations were quantified fluorometrically using the Qubit dsDNA Assay Kit and Qubit 2.0 Fluorometer (both from Invitrogen, USA). Approximately 2 μg of DNA from each sample was added to a CsCl solution to a final density of 1.723–1.725 g·mL⁻¹ and transferred into 5.1 mL Quick-Seal™ centrifuge tubes (Backman Coulter, USA). Ultracentrifugation was conducted using an Optima L-100 XP ultracentrifuge (Beckman Coulter, USA) equipped with a VTi 65.2 vertical rotor at 140,000 × g for 72 h. Following centrifugation, mineral oil was layered onto the gradient using a syringe pump (Braintree Scientific, USA). The gradient was divided into 15 fractions, and the refractive index of each fraction was measured using a refractometer (Brix/Rl-Chek, Reichert) for density calculation⁶³. DNA was precipitated by adding two volumes of PEG solution (30% PEG 6000, w/v, 1.6 M NaCl, and 20 μg of glycogen) and then resuspended in 35 μL of TE (10 mM Tris-HCl, 1 mM EDTA, pH 8.0). These purified DNA served as templates for qPCR and 16S rRNA gene PCR for Illumina sequencing.

The qPCR of density fractions

The abundance of bacterial 16S rRNA gene in each gradient fraction was quantified using qPCR with bacterial universal primers 338 F and 518 R on a CFX 96™ real-time system (BIO-RAD)⁶⁴. Reaction conditions and cycling parameters followed the protocol described by Deng et al. ⁶⁵. Each 20 μL reaction contained 1 μL of template DNA, 0.5 μM of each primer, 10 μL of SYBR Premix Ex Taq™ II (Takara Bio, Japan), and 5 μg of bovine serum albumin. The amplification protocol consisted of an initial enzyme activation at 95 °C for 105 s, followed by 40 cycles of denaturation at 95 °C for 15 s, annealing at 55 °C for 30 s, and extension at 72 °C for 30 s. All reactions were run in triplicate, and non-template controls were included in each run to ensure the absence of contamination, and the standard curve was generated using a purified target DNA fragment from E. coli P10⁶⁵.

Illumina DNA sequencing

Based on the abundance of 16S rRNA genes across density fractions, DNA from 10 selected fractions per sample was used for amplicon sequencing. The V3–V4 hypervariable regions of the 16S rRNA genes were amplified using barcoded universal primers 343 F and 798 R on an ABI GeneAmp® 9700 thermal cycler (Thermo, USA) (Supplementary Data 15)⁶⁶. PCR products were visualized by 2% agarose gel electrophoresis, purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, USA), and quantified with a Qubit®3.0 fluorimeter (Life Invitrogen). Twenty-four amplicons with distinct barcodes were pooled in equimolar amounts to construct an Illumina paired-end (PE) sequencing library. Sequencing was performed on the Illumina PE250 platform at Shanghai Biozeron Biotechnology Co., Ltd (Shanghai, China).

Raw sequencing data in FASTQ format were preprocessed using Trimmomatic⁶⁷, with demultiplexing performed based on sample-specific barcode sequences. Low-quality reads—defined as those with an average quality score below 20 within a 10 bp sliding window were discarded. Paired-end reads with overlapping regions longer than 10 bp were assembled based on sequence overlap. Quality-filtered sequences were then processed using the QIIME pipeline⁶⁸. Following merging and chimera removal, amplicon sequence variants (ASVs) were identified, and taxonomic classification of each ASV was performed using the UCLUST algorithm (https://github.com/topics/uclust) against the SILVA reference database (https://www.arb-silva.de/), applying a confidence threshold of 80%.

PacBio sequencing

Full-length 16S rRNA gene sequencing of bacterial communities cultured before and after the addition of ¹³C₆-SQ was performed using the PacBio Sequel II platform. The V1–V9 regions of the bacterial 16S rRNA gene were amplified using universal primers 27 F and 1492 R (Supplementary Data 15)⁶⁹, each tagged with sample-specific PacBio barcode sequences. PCR products were visualized on 2% agarose gels, purified using the AxyPrep DNA Gel Extraction Kit (Axygen, USA), and quantified using a Qubit®3.0 fluorimeter (Life Invitrogen). SMRTbell libraries were prepared from the amplified DNA using SMRTbell prep kit 3.0 (PacBio, USA). Purified SMRTbell libraries from the pooled and barcoded samples were sequenced on a single PacBio Sequel II system at Shanghai Biozeron Biotechnology Co., Ltd (Shanghai, China).

Raw PacBio reads were processed using SMRT Link Analysis Software v11.0. Demultiplexing and removal of barcode and primer sequences were performed using the lima pipeline (Pacific Biosciences, https://lima.how/). Quality-filtered sequences were further analyzed using QIIME⁶⁸. Trimming and filtering were applied with a maximum expected error rate of two per read (maxEE = 2). ASVs were assigned taxonomic classifications using the UCLUST algorithm (https://github.com/topics/uclust) against the SILVA reference database (https://www.arb-silva.de/) using confidence threshold of 80%.

Determination of ¹³C-labeling between unlabeled and labeled density fractions

The distribution curves of bacterial 16S rRNA gene abundance along the CsCl density gradient were analyzed based on the density, 16S rRNA gene copy numbers obtained by qPCR and bacterial 16S rRNA gene abundance of the CsCl fractions from each sample (Supplementary Data 16 and 17). The density of labeled (W_LABi) and unlabeled (W_LIGHTi) DNA is calculated for each taxon by integrating the 16S rRNA gene abundances of ASVs within the same genus and the 16S rRNA gene copy numbers obtained via qPCR according to Hungate’s et al.⁷⁰. To minimize biases introduced by baseline variability, only the major peak of DNA was integrated. All fractions that contained >20% of the maximum observed quantity were hence integrated according to Connelly et al.⁷¹. Only genera with 16S rRNA gene abundances exceeding 1% of the total abundance were retained for analysis (Supplementary Data 1).

Bacterial and phytoplankton culture

A. macleodii ATCC 27126 was obtained from the Collection Center of Marine Bacteria (Xiamen University, China). R. denitrificans OCh 114 (DSM 7001) was obtained from the German Collection of Microorganisms and Cell Cultures GmbH (Braunschweig, Germany). R. denitrificans OCh 114, A. stellipolaris R10SW13, P. marina ECSMB 14103, and A. macleodii F12 were cultured in modified sterile artificial seawater culture medium (pH 8.2, with 1 g/L yeast extract) containing either 222 μM SQGro or 333 μM SQ. Additional culture experiments with SQDG (46.5 µM) were performed for strain R. denitrificans OCh 114. SQDG was first dissolved in methanol, filtered, and added to a medium containing 1% Tween-20 (final volume of methanol is 1% of the medium). An aliquot of 2% of a pre-adapted culture of bacteria was used to inoculate a 5 mL culture and incubated at 28 °C for growth. Bacterial density (A₆₀₀) was measured using a plate reader (Tecan M200, Männedorf, Switzerland). All bacteria were cultured separately. Samples from each culture were collected and stored at −80 °C for liquid chromatography-mass spectrometry (LC-MS) analysis.

All algae (Table 1) were from the Center for Collections of Marine Algae (Xiamen University, China). Diatoms (f/2 media with Na₂SiO₃), dinoflagellates, coccolithophore, flagellates and cholorophyta (f/2 media) were cultured at 20 °C in an incubator with an irradiance of 100 μmol photons m⁻² s⁻¹⁷². Pro99 media and SN media were used for cyanobacteria cultivation in an incubator with an irradiance of 50 μmol photons m⁻² s⁻¹ at 20 °C^73,74. Accuri C6 flow cytometer (BD Biosciences, New Jersey, USA) and a microscope was used to track the cell abundance of algae. Each culture contained three independent replicates. Algal cells were collected in exponential phase through 0.7-μm GF/F filter membrane (47 mm, Whatman, USA) for SQGro, SQ, SQDG, and total organic sulfur detection. The liquid of the culture filtered by 0.2-μm polycarbonate membrane (Merck Millipore) was used for the detection of extracellular SQGro. All samples were stored at −80 °C until further extraction. Total organic sulfur of the cultured samples was detected by elemental analyzer (vario EL Cube, Elementar, Germany). Before analysis, the sample was washed with 3.2% ammonium formate to remove sulfate⁷⁵.

Proteomic analysis of A. macleodii ATCC 27126

A. macleodii ATCC 27126 was cultured in modified sterile artificial seawater culture medium (pH 8.2, supplemented with 1 × g/L yeast extract) containing either 333 μM SQ or 222 μM SQGro. Medium containing only 1 × g/L yeast extract was used as a control. Triplicate samples from each growth condition were harvested after centrifugation and digested with trypsin⁷⁶. The peptides from the digested samples were desalted using a C18 cartridge. After freeze-drying, the peptides were re-dissolved in 40 µL of 0.1% formic acid solution. Data independent acquisition (DIA) mass spectrometry was performed using an Astral high-resolution mass spectrometer (Thermo Scientific). Data were analyzed using DIA-NN software⁷⁷ with the following parameters: trypsin as the enzyme, a maximum of one missed cleavage site, carbamidomethyl (C) as a fixed modification, and oxidation (M) and acetylation (protein N-terminus) as dynamic modifications. Proteins identified by database retrieval were filtered using a false-discovery rate threshold of <1%. Differential expression analysis was performed using a volcano plot generated with the Chiplot online tool (https://www.chiplot.online) based on fold change and p-value (t-test) for each comparison group.

Gene synthesis, expression, and purification

All genes were synthesized by Sangon Biotech (Sangon, Shanghai, China). The synthesized genes were cloned into the plasmid pET28a and heterologously expressed in E. coli BL21(DE3). These genes include AmsqoD (gene locus tag: MASE_18005) and AmsqgH (gene locus tag: MASE_18025) of A. macleodii ATCC 27126, GH31_3 (gene locus tag: Ga0198492_01_4502455_4504773) of E. coli K-12, GH31_6 (gene locus tag: RA224_28300) of Achromobacter aegrifaciens, RdsqgA (gene locus tag: RD1_2308) and RddctP (gene locus: RD1_2315) of R. denitrificans OCh 114, GasqgA (gene locus tag: Ga0210579_114280) of G. antarcticus IMCC3135, MasqgA (gene locus tag: Ga0133504_112207) of M. aestuarii ST58-10, MhsqgA (gene locus tag: Ga0365646_2468) of M. hwangdonensis HDW-15, CtsqgA (gene locus tag: CTEN210_14965) of Chaetoceros tenuissimus NIES-3715, PpsqgA (gene locus tag: FVE85_4141) of Porphyridium purpureum CCMP 1328, SpsqgA+SpsqoD (gene locus tag: SPIL2461_LOCUS1475) of S. pilosum, AtsqgA (gene locus tag: AT4G17370) of Arabidopsis thaliana. Mutagenesis of AmSqoD, AmSqgH, and RdSqgA was performed using the Mut Express II Fast mutagenesis kit V2 (Vazyme, Nanjing, China). Primers for mutagenesis are presented in Supplementary Data 15. All the bacteria were cultured to OD₆₀₀ of approximately 0.6 in LB liquid medium with kanamycin resistance, and then transferred to a larger volume at 37 °C, 180 rpm to an OD₆₀₀ of about 0.8. Subsequently, the bacteria were cooled to 16 °C, then isopropyl β-D-1-thiogalactopyranoside was added to a final concentration of 0.5 mM. Incubation was continued at 180 rpm for 16–18 h at 16 °C, and the bacteria were collected by centrifugation at 12,000 × g for 10 min. Bacteria were resuspended in buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl) and fragmented on ice with an ultrasonic disruptor (SM-650D), then centrifuged at 4 °C, 12,000 × g for 30 min. The supernatant was filtered through a 0.22-μm membrane and purified using AKTA pure (Cytiva, MA, USA) in combination with a HisTrap HP column (5 mL, Cytiva). Samples were loaded using 2% elution buffer (10 mM Tris-HCl, 150 mM NaCl, 500 mM imidazole, pH 8.0) at a flow rate of 1 mL/min, followed by elution using an increasing gradient of elution buffers (4, 8, 10, 50%). The purified proteins were examined by SDS-PAGE on a commercial gel (precast protein plus gel, Hepes-Tris, 4–20%). Fractions containing the target protein were subjected to a desalting column (5 mL, Cytiva) with phosphate buffer (10 mM, pH 8.0), filtered with 0.2-μm membrane, and stored at −80 °C until further use.

Enzymatic activity assays of AmSqoD, AmSqgH, and RdSqgA

The activity of AmSqoD towards SQ was measured using a colorimetric fuchsin assay¹⁷. A negative control omitting AmSqoD was also performed. The 6-dehydroglucose generated in the reaction was derivatized with DNPH and detected according to the following method. Mobile phase A was 0.1% formic acid in water and mobile phase B was 100% acetonitrile. The column was equilibrated with 95% A and 5% B for 5 min. The gradient conditions were: 5–50% B, 20 min; 50–95% B, 2 min; maintain 95% B, 2 min; 95–5% B, 1 min; maintain 5% B, 5 min. AmSqgH activity assays were conducted using PNPSQ⁷⁸, PNPGlc, and PNPXyl as substrates. Reaction mixtures (200 µL) contained 10 mM phosphate buffer (pH 7.5), 10 mM PNPSQ (PNPGlc or PNPXyl), and 2 µM AmSqgH/GH31_3/GH31_6 with enzyme omitted in the control. Reactions were monitored at 405 nm using a plate reader (Tecan M200, Männedorf, Switzerland). The k_cat /K_M values of pH dependence for AmSqgH were measured for PNPSQ hydrolysis in 10 mM phosphate buffer at a range of pH values (5.5, 6.0, 6.5, 7, 7.5, 8.0) and 10 mM Tris-HCl buffer (pH 9.0). AmSqgH and RdSqgA activities were measured using SQGro or SQDG as a substrate. A 100 μL reaction containing 20 μM SQGro, 5 μM AmSqgH, or RdSqgA (add an additional 10 mM NAD⁺) in buffer (200 mM NH₄HCO₃, pH 7.5 for AmSqgH, pH 7.0 for RdSqgA) was incubated overnight at room temperature, respectively. Negative controls omitting enzyme were also performed. Reaction mixtures were dried, resuspended in CHCl₃: MeOH: H₂O (4: 6: 1, v/v/v), filtered through a 0.22-µm nylon membrane filter prior to LC-MS analysis. Measurement of AmSqgH and RdSqgA activity with SQDG was conducted according to the method of Speciale et al.³¹. SQDG (> 99%, 20 μL, 1 mg/ml in CHCl₃: MeOH, 1: 1) in 200 mM NH₄HCO₃, 5% DMSO, pH 7.5 was incubated with 5 μM AmSqgH or RdSqgA (add an additional 10 mM NAD⁺) overnight at room temperature, respectively. Reaction mixtures were dried, resuspended in CHCl₃: MeOH: H₂O (4: 6: 1, v/v/v) and filtered prior to LC-MS analysis.

Surface plasmon resonance (SPR) measurement

SPR analyses for substrate binding protein (RdDctP) was performed using Biacore T200 optical biosensors equipped with research-grade CM5 sensor chips (Biacore AB, Uppsala, Sweden). RdDctP was used at a concentration >1 mg/mL. Prior to analysis, a two-channel coupling was performed, with chip 1 channel as a closed coupling and 2 channels as a ligand coupling to eliminate refractive index variation, feed noise, and data drift. Pre-prepared 70% glycerol was normalized to the chip. Sodium acetate buffer (10 mM, pH 4.5) was selected as the buffer for the next step according to the adsorption of ligands to the chip under different pH conditions. A series of concentrations of SQGro or SQ substrate were injected into the flow system and analyzed, respectively, and glucose as a negative control substrate.

Field sample collection

Seven separate cruises samples from seven oceanic regions were collected from July 2022 to August 2023. Samples from the Bohai Sea and the Yellow Sea were collected in July 2022, samples from the East China Sea were collected in August 2022, samples from the Jiulongjiang Estuary were collected in February 2023, samples from the Dongshan Bay were collected in July 2023, samples from the South China Sea were collected in August 2023, and samples from Central North Pacific were collected in July 2023. With the exception of samples collected from 3 to 500 meters depth in the central North Pacific, all other stations sampled water from depths of 3 to 5 meters. In each site, three parallel water samples were collected. Prior to sampling, seawater (1–3 L) was pre-filtered by a 200-µm a nylon screen and then was collected for particulate organosulfonates (SQDGs, SQ, SQGro, and DHPS) and POC samples using through 0.7-µm GF/F filter membrane (47 mm, Whatman, USA), which was pre-combusted at 480 °C for 5 h to remove carbon contamination. After sampling, the filters were immediately wrapped in pre-combusted aluminum foil and kept frozen at −20 °C until further extraction and analysis. Chlorophyll a content in the field were measured using Seabird SBE recorders (Sea-Bird Electronics Inc., USA). POC of the field samples was detected by elemental analyzer (vario EL Cube, Elementar, Germany) after 24 h of hydrochloric acid fumigation to remove inorganic carbon⁷⁹. The base map was generated with Ocean Data View (v5.7.2) software.

Quantification of organosulfonates from cultured phytoplankton samples and field samples

For determination of intracellular metabolites SQGro and SQ from algae, the sample filters were cut into pieces and added to bead beating tubes containing 1 mL of a solvent mixture (MeOH: ACN: H₂O, 2:2:1, v/v/v) along with 500 µL of silica beads, isotope-labelled taurine-D4 (Cambridge Isotope Laboratories, USA) as an internal standard was added¹⁹. Three biological replicates of each sample were processed simultaneously. The FastPrep-24 homogenizer (MP Biomedicals, CA, USA) was used for grinding and homogenizing samples. Finally, the samples were centrifuged at 12,000 × g for 10 min and the extracted supernatants were filtered through a 0.22-µm membrane filter to a clean 1 mL glass vial for HPLC/MS analysis. The extraction steps of SQGro, SQ, DHPS in field samples are consistent with the above treatment.

The extraction of SQDG is based on the method of Bligh and Dyer^80,81, with minor modifications. The collected filters of algae samples, along with the field sample filters, were cut into pieces and transferred to bead beating tubes containing 960 μL mixed solvent of MeOH: CH₂Cl₂: PBS (2:1:0.8, v/v/v) and approximately 500 μL of silica beads. 3-sn-Phosphatidylcholine (C₁₆ PAF) (Merck, Germany) was added as an internal standard²⁸. Then, the samples were shaken on a FastPrep-24 Homogenizer (MP Biomedicals, CA, USA) for three repetitions, then centrifuged for 10 min at 12,000 × g. Sufficient CH₂Cl₂: PBS (1:1, v/v) was added to the collected supernatant to make the final solvent volume ratio MeOH: CH₂Cl₂: PBS (1:1:0.9, v/v/v). The samples were centrifuged at 5000 × g for 5 min at 4 °C, to allow phase separation. The CH₂Cl₂ layer, containing total lipids, was collected, and the MeOH: PBS was extracted twice with CH₂Cl₂. The extracts were dried under a flow of nitrogen. Finally, the sample were resuspended in CH₂Cl₂: MeOH (1:4, v/v) mixture and filtered through a 0.2-µm membrane filter to a clean 1 mL glass vial for subsequent analysis.

The concentration of SQGro, SQ, and DHPS of all samples were quantified using a Q Exactive high-performance liquid chromatography system (Thermo Fisher Scientific, Waltham, USA) with an ACQUITY UPLC BEH HILIC Column (130 Å, 1.7 µm, 2.1 mm × 100 mm, Waters, USA). A flow rate of 0.25 mL/min was used with two mobile phases A and B, A was acetonitrile and B was 10 mM ammonium acetate. The gradient was as follows: 0 min: 10% B, 15 min: 35% B, 16 min: 60% B, 20 min: 60% B, followed by 10 min of re-equilibration with 10% B and the injection volume was 2 μL. Quantification of SQGro and SQ was achieved with isotope-labelled taurine-D4 as an internal standard. One quality control sample was set for every 10 experimental samples. Data analyses were performed with Xcalibur software version 2.0.3 (Thermo Fisher Scientific).

The concentration of SQDG of all samples was quantified using an Agilent 1290 Infinity II series UHPLC System (Agilent) coupled to an Agilent 6460 triple quadrupole mass spectrometer (Agilent, CA, USA) via electrospray ionization (ESI), with an ACQUITY UPLC BEH C18 Column (130 Å, 1.7 µm, 2.1 mm × 50 mm, Waters, USA), flow rate of 0.25 mL/min with two mobile phases A and B, A was MeOH: H₂O (85:15, v/v) and B was isopropanol, both containing 10 mM ammonium acetate. Mobile phases were proportioned as the following gradients: 0 min: 20% B, 1 min: 20% B, 8 min: 40%, 9 min: 80% B,11 min: 20% B, 14 min: 20% B, 17 min: 20% B, 20 min: 20% B. Column temperature was kept at 40 °C and the injection volume was 5 μL. Data analyses were performed with the Agilent MassHunter Qualitative Analysis Software (B.07.00, Agilent Technologies, USA). One quality control sample was set for every 10 experimental samples. Quantitation was titrated using external calibration curves of commercially purchased standard SQDG (Larodan, Sweden) and normalized to the recovery of the internal standard.

Estimation of annual production of sulfoquinovosyl compounds

We estimated the annual production of sulfoquinovosyl compounds in the ocean based on their correlation with chlorophyll a concentrations and the assumption of a steady-state kinetic process. First, we calculated the proportion of sulfoquinovosyl carbon relative to POC, focusing specifically on SQDG molecules. Given variability in fatty acid chain composition, SQDGs in our dataset contained between 37 and 48 carbon atoms. Using the carbon content and the amount of individual SQDG lipoforms, we estimated the carbon contribution of SQDGs to the POC pool. This value was then multiplied by the global net primary productivity of 53 Pg C/yr⁴³. To account for heterotrophic consumption ratio of algae-derived organic matter, we applied a 50% correction factor based on the study of Azam et al. (1982), resulting in an estimate of the annual production of sulfoquinovosyl compounds⁴⁴.

Bioinformatic analysis

The online AlphaFold2 server was used to predict the protein structures of AmSqgH, AmSqoD and the eukaryotic proteins CtSqgA, PpSqgA, AtSqgA, and SpSqgA⁸². Crystals of RdSqgA were grown by the sitting-drop vapor diffusion method. The structure was determined by molecular replacement with an AlphaFold2 model (UniProt ID: Q167F5) and refined to a resolution of 1.8 Å (Supplementary Data 18). Molecular docking of SQGro into the catalytic domains of AmSqgH and RdSqgA was carried out using AutoDock 4. Gasteiger charges and all hydrogen atoms were added to the protein structures prior to docking. A genetic algorithm was employed for molecular docking, with the maximum number of evaluations set to medium. The conformations with the lowest binding energy were retained for further analysis and used to guide site-directed mutagenesis to verify their functions. Protein-ligand interaction patterns between SQGro and key residues were analyzed using the Protein-Ligand Interaction Profiler online tool. For RdSqgA, molecular dynamics simulations were performed using the NAMD program on a Dell EMC PowerEdge R940xa server⁸³. The complex was placed in an orthogonal box containing 9440 water molecules, 34 sodium ions, and 25 chloride ions. The sodium and chloride ions diffuse randomly under periodic boundary conditions. The solvated system underwent energy minimization over 1 ns, followed by 50 ns of molecular dynamics at 300 K. Molecular structures were visualized with Pymol 2.4 (Schrodinger, NY, USA). The representative electron density map for crystal structure is shown in Supplementary Fig. 9. The refinement statistics are summarized in Supplementary Data 18.

Alteromonas genomes were obtained from GTDB v220. A phylogenetic tree was constructed using GTDB-Tk based on the marker genes of 120 bacteria strains⁸⁴, with Pseudoalteromonas caenipelagi serving as the outgroup. Genes involved in the sulfo-SDO pathway were identified using the Basic Local Alignment Search Tool (BLAST)⁸⁵, and genomes containing a contiguous cluster of the target protein-coding genes and retain the genomes that contain contiguous protein-coding genes (WP_014951127.1, WP_014951128.1, WP_014951129.1, WP_014951130.1, WP_014951131.1, WP_014951132.1, WP_014951133.1, WP_014951134.1, WP_014951135.1, and WP_014951136.1) were retained for further analysis. Representative homologs and paralogs of SqoD, SqgH, and SqgA were obtained from the NCBI reference sequence database. SQGro and SQ gene clusters were identified using the cblaster tool within the NCBI reference sequence database⁸⁶. Representative sequences of all GH31 subfamilies were retrieved from the CAZy database (http://www.cazy.org/). Homologous eukaryotic SqgA sequences were obtained from UniProt and IMG databases (https://www.uniprot.org/uniprot and https://img.jgi.doe.gov). The optimal model for phylogenetic analysis was determined using ModelFinder⁸⁷. Phylogenetic trees for SqoD, SqgH, and eukaryotic SqgA were constructed using IQ-TREE under the WAG + G4 model with 1000 replicates⁸⁸. A phylogenetic tree was constructed for bacterial genera containing sqgA using RAxML with the LG + IU+G4m model and 1000 bootstrap replicates⁸⁹, based on a concatenated alignment of 11 single-copy orthologous genes (COG0087, COG0222, COG0292, COG0256, COG0049, COG0096, COG0051, COG0098, COG0085, and COG0691) shared among bacterial strains. The phylogenetic trees were visualized with Evolview v3 program⁹⁰. Amino acid sequence alignment of RdSqgA and a eukaryotic representative was performed using MAFFT⁹¹, and the results were visualized with ESPript 3 (https://espript.ibcp.fr)⁹².

Sequences of SqgH and SqoD from A. macleodii ATCC 27126 and SqgA and SmoC from R. denitrificans OCh 114 were used as BLAST queries against the Tara Oceans metagenome (OM-RGC_v2_metaG), metatranscriptome (OM-RGC_v2_metaT), and Bohai Sea and Jiulongjiang Estuary metatranscriptomes³⁶. Eukaryotic SqgA sequences were obtained from the Tara Oceans MATOU database³⁵. An identity cutoff of 50% and 54% was applied to prokaryotic SqgH and SqgA, respectively, while a 60% cutoff was used for SqoD and SmoC. These thresholds were determined by analyzing the sequence alignments of these genes in various bacteria possessing the SQ metabolic gene cluster in this study. Additionally, a bit score cutoff of 204 was used for eukaryotic SqgA, consistent with our previous research¹⁹. Prokaryotic gene abundances were normalized to the median abundance of 10 universal single-copy genes (COG0012, COG0016, COG0018, COG0172, COG0215, COG0495, COG0525, COG0533, COG0541, COG0552) for both metagenome and metatranscriptome analyses^93,94. Eukaryotic gene abundances were normalized to the median abundance of 10 single-copy marker genes (COG0522, COG0190, COG0097, COG0092, COG0191, COG0194, COG0187, COG0093, COG0096, COG0184)⁹³. The relationship between Tara metadata and log-transformed genes relative abundance was analyzed by Spearman correlation in IBM SPSS Statistics 26. Significance was corrected using the Benjamini–Hochberg approach⁹⁵. Finally, data visualization and analysis were performed using Origin 2021 (OriginLab, MA, USA). The base maps were created using Tableau Desktop (v2019.4).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

Data supporting the findings of this work are available within this paper and its Supplementary Information files. The raw sequencing data of the bacterial 16S rRNA genes generated in this study have been deposited in the NCBI/Sequence Read Archive (SRA) database under accession PRJNA1310569. The Proteomic data generated in this study have been deposited in the PRotromics IDEntifications (PRIDE) under accession PXD058686 and PXD058710. Biomolecular mass spectrometry data generated in this study have been deposited in the Figshare database (https://doi.org/10.6084/m9.figshare.29984182.v1). The model of structure generated in this study have been deposited in the Figshare database (https://doi.org/10.6084/m9.figshare.30008725). Crystallographic data for the structures reported in this article have been deposited in Protein Data Bank database (https://www.rcsb.org/) under deposition numbers 8ZN9. Source data are provided with this paper.

References

Cherabier, P. & Ferrière, R. Eco-evolutionary responses of the microbial loop to surface ocean warming and consequences for primary production. ISME J. 16, 1130–1139 (2022).
Article PubMed Google Scholar
Ksionzek, K. B. et al. Dissolved organic sulfur in the ocean: biogeochemistry of a petagram inventory. Science 354, 456–459 (2016).
Article PubMed Google Scholar
Wu, B. et al. Microbial sulfur metabolism and environmental implications. Sci. Total. Environ. 778, 146085 (2021).
Article PubMed Google Scholar
Goddard-Borger, E. D. & Williams, S. J. Sulfoquinovose in the biosphere: occurrence, metabolism and functions. Biochem. J. 474, 827–849 (2017).
Article PubMed Google Scholar
Kiene, R. P., Linn, L. J. & Bruton, J. A. New and important roles for DMSP in marine microbial communities. J. Sea Res. 43, 209–224 (2000).
Article Google Scholar
Loi, V. V., Rossius, M. & Antelmann, H. Redox regulation by reversible protein S-thiolation in bacteria. Front. Microbiol. 6, 187 (2015).
Article PubMed PubMed Central Google Scholar
Teng, Z. J. et al. Acrylate protects a marine bacterium from grazing by a ciliate predator. Nat. Microbiol. 6, 1351–1356 (2021).
Article PubMed Google Scholar
Sunda, W., Kieber, D. J., Kiene, R. P. & Huntsman, S. An antioxidant function for DMSP and DMS in marine algae. Nature 418, 317–320 (2002).
Article PubMed Google Scholar
Ho, T.-Y. et al. The elemental composition of some marine phytoplankton. J. Phycol. 39, 1145–1159 (2003).
Article Google Scholar
Moran, M. A. & Durham, B. P. Sulfur metabolites in the pelagic ocean. Nat. Rev. Microbiol. 17, 665–678 (2019).
Article PubMed Google Scholar
Landa, M. et al. Sulfur metabolites that facilitate oceanic phytoplankton-bacteria carbon flux. ISME J. 13, 2536–2550 (2019).
Article PubMed PubMed Central Google Scholar
Tang, K. & Liu, L. Bacteria are driving the ocean’s organosulfur cycle. Trends Microbiol. 31, 772–775 (2023).
Article PubMed Google Scholar
Durham, B. P. et al. Cryptic carbon and sulfur cycling between surface ocean plankton. Proc. Natl. Acad. Sci. USA 112, 453–457 (2015).
Article PubMed Google Scholar
Durham, B. P. et al. Sulfonate-based networks between eukaryotic phytoplankton and heterotrophic bacteria in the surface ocean. Nat. Microbiol. 4, 1706–1715 (2019).
Article PubMed Google Scholar
Cook, A. M., Denger, K. & Smits, T. H. Dissimilation of C3-sulfonates. Arch. Microbiol. 185, 83–90 (2006).
Article PubMed Google Scholar
Snow, A. J. D., Burchill, L., Sharma, M., Davies, G. J. & Williams, S. J. Sulfoglycolysis: catabolic pathways for metabolism of sulfoquinovose. Chem. Soc. Rev. 50, 13628–13645 (2021).
Article PubMed Google Scholar
Ye, Z., Wei, Y., Jiang, L. & Zhang, Y. Oxygenolytic sulfoquinovose degradation by an iron-dependent alkanesulfonate dioxygenase. iScience 26, 107803 (2023).
Article PubMed PubMed Central Google Scholar
Kaur, A. et al. Widespread family of NAD⁺-dependent sulfoquinovosidases at the gateway to sulfoquinovose catabolism. J. Am. Chem. Soc. 145, 28216–28223 (2023).
Article PubMed PubMed Central Google Scholar
Liu, L. et al. Sulfoquinovose is a widespread organosulfur substrate for Roseobacter clade bacteria in the ocean. ISME J. 17, 393–405 (2023).
Article PubMed PubMed Central Google Scholar
Anesi, A. et al. Comparative analysis of membrane lipids in psychrophilic and mesophilic freshwater dinoflagellates. Front. Plant. Sci. 7, 524 (2016).
Article PubMed PubMed Central Google Scholar
Yongmanitchai, W. & Ward, O. P. Positional distribution of fatty acids, and molecular species of polar lipids, in the diatom Phaeodactylum tricornutum. J. Gen. Microbiol. 139, 465–472 (1993).
Article PubMed Google Scholar
Goss, R. et al. The lipid dependence of diadinoxanthin de-epoxidation presents new evidence for a macrodomain organization of the diatom thylakoid membrane. J. Plant Physiol. 166, 1839–1854 (2009).
Article PubMed Google Scholar
Harwood, J. L. & Nicholls, R. G. The plant sulpholipid-a major component of the sulphur cycle. Biochem. Soc. Trans. 7, 440 (1979).
Article PubMed Google Scholar
Field, C. B. et al. Primary production of the biosphere: integrating terrestrial and oceanic components. Science 281, 237–240 (1998).
Article PubMed Google Scholar
Flombaum, P. et al. Present and future global distributions of the marine Cyanobacteria Prochlorococcus and Synechococcus. Proc. Natl. Acad. Sci. USA 110, 9824–9829 (2013).
Article PubMed PubMed Central Google Scholar
Van Mooy, B. A. S., Rocap, G., Fredricks, H. F., Evans, C. T. & Devol, A. H. Sulfolipids dramatically decrease phosphorus demand by picocyanobacteria in oligotrophic marine environments. Proc. Natl. Acad. Sci. USA 103, 8607–8612 (2006).
Article PubMed PubMed Central Google Scholar
Van Mooy, B. A. S. et al. Phytoplankton in the ocean use non-phosphorus lipids in response to phosphorus scarcity. Nature 458, 69–72 (2009).
Article PubMed Google Scholar
Flores, E. et al. Bacterial and eukaryotic intact polar lipids point to in situ production as a key source of labile organic matter in hadal surface sediment of the Atacama Trench. Biogeosciences 19, 1395–1420 (2022).
Article Google Scholar
Gašparović, B. et al. Particulate sulfur-containing lipids: production and cycling from the epipelagic to the abyssopelagic zone. Deep Sea Res. Pt. I. 134, 12–22 (2018).
Article Google Scholar
Shibuya, I., Yagi, T. & Benson, A. A. Sulfonic acids in algae. In Microalgae and Photosynthetic Bacteria. (eds Japanese Society of Plant Physiologists) 627–636 (Tokyo University Press, 1963).
Speciale, G., Jin, Y., Davies, G. J., Williams, S. J. & Goddard-Borger, E. D. YihQ is a sulfoquinovosidase that cleaves sulfoquinovosyl diacylglyceride sulfolipids. Nat. Chem. Biol. 12, 215–217 (2016).
Article PubMed Google Scholar
Sharma, M. et al. Oxidative desulfurization pathway for complete catabolism of sulfoquinovose by bacteria. Proc. Natl. Acad. Sci. USA 119, e2116022119 (2022).
Article PubMed PubMed Central Google Scholar
Abayakoon, P. et al. Structural and biochemical Insights into the function and evolution of sulfoquinovosidases. ACS Cent. Sci. 4, 1266–1273 (2018).
Article PubMed PubMed Central Google Scholar
Arumapperuma, T. et al. A subfamily classification to choreograph the diverse activities within glycoside hydrolase family 31. J. Biol. Chem. 299, 103038 (2023).
Article PubMed PubMed Central Google Scholar
Villar, E. et al. The Ocean Gene Atlas: exploring the biogeography of plankton genes online. Nucleic Acids Res. 46, W289–W295 (2018).
Article PubMed PubMed Central Google Scholar
Han, Y. et al. Transcriptomic evidences for microbial carbon and nitrogen cycles in the deoxygenated seawaters of Bohai Sea. Environ. Int. 158, 106889 (2022).
Article PubMed Google Scholar
Hou, S. et al. Benefit from decline: the primary transcriptome of Alteromonas macleodii str. Te101 during Trichodesmium demise. ISME J. 12, 981–996 (2018).
Article PubMed PubMed Central Google Scholar
Pedler, B. E., Aluwihare, L. I. & Azam, F. Single bacterial strain capable of significant contribution to carbon cycling in the surface ocean. Proc. Natl. Acad. Sci. USA 111, 7202–7207 (2014).
Article PubMed PubMed Central Google Scholar
López-Pérez et al. Genomes of surface isolates of Alteromonas macleodii: the life of a widespread marine opportunistic copiotroph. Sci. Rep. 2, 696 (2012).
Article PubMed PubMed Central Google Scholar
García-Martínez, J., Acinas, S. G., Massana, R. & Rodríguez-Valera, F. Prevalence and microdiversity of Alteromonas macleodii-like microorganisms in different oceanic regions. Environ. Microbiol. 4, 42–50 (2002).
Article PubMed Google Scholar
Roth Rosenberg, D. et al. Particle-associated and free-living bacterial communities in an oligotrophic sea are affected by different environmental factors. Environ. Microbiol. 23, 4295–4308 (2021).
Article PubMed Google Scholar
Mazzotta, M. G. et al. Characterization of the metalloproteome of Pseudoalteromonas (BB2-AT2): biogeochemical underpinnings for zinc, manganese, cobalt and nickel cycling in a ubiquitous marine heterotroph. Metallomics 13, mfab060 (2021).
Article PubMed PubMed Central Google Scholar
Johnson, K. S. & Bif, M. B. Constraint on net primary productivity of the global ocean by Argo oxygen measurements. Nat. Geosci. 14, 769–774 (2021).
Article Google Scholar
Fuhrman, J. A. & Azam, F. Thymidine incorporation as a measure of heterotrophic bacterioplankton production in marine surface waters: evaluation and field results. Mar. Biol. 66, 109–120 (1982).
Article Google Scholar
Williams, P. J. L. B. Heterotrophic Utilization of dissolved organic compounds in the Sea I. Size distribution of population and relationship between respiration and incorporation of growth substrates. J. Mar. Biol. Assoc. Uk. 50, 859–870 (1970).
Article Google Scholar
Wright, R. T. & Shah, N. M. The trophic role of glycolic acid in coastal seawater. I. Heterotrophic metabolism in seawater and bacterial cultures. Mar. Biol. 33, 175–183 (1975).
Article Google Scholar
Becker, S. et al. Laminarin is a major molecule in the marine carbon cycle. Proc. Natl. Acad. Sci. USA 117, 6599–6607 (2020).
Article PubMed PubMed Central Google Scholar
Yagi, T. & Benson, A. A. Plant sulfolipid. V. Lysosulfolipid formation. Biochim. Biophys. Acta 57, 601–603 (1962).
Article PubMed Google Scholar
Hashiro, S., Fujiuchi, K., Sugimori, D. & Yasueda, H. A novel galactolipase from a green microalga Chlorella kessleri: Purification, characterization, molecular cloning, and heterologous expression. Appl. Microbiol. Biotechnol. 102, 1711–1723 (2018).
Article PubMed PubMed Central Google Scholar
Behrendt, L. et al. Microbial dietary preference and interactions affect the export of lipids to the deep ocean. Science 385, eaab2661 (2024).
Article PubMed Google Scholar
Scholz, S., Serif, M., Schleheck, D., Sayer, M. D. J. & Frithjof Christian, K. Sulfoquinovose metabolism in marine algae. Botanica Mar. 64, 301–312 (2021).
Article Google Scholar
Liang, K. Y. H. F., Orata, D., Boucher, Y. F. & Case, R. J. Roseobacters in a sea of poly- and paraphyly: whole genome-based taxonomy of the family Rhodobacteraceae and the proposal for the split of the “Roseobacter clade” into a novel family, Roseobacteraceae fam. nov. Front. Microbiol. 12, 683109 (2021).
Article PubMed PubMed Central Google Scholar
Moran, M. A. et al. Ecological genomics of marine Roseobacters. Appl. Environ. Microbiol. 73, 4559–4569 (2007).
Article PubMed PubMed Central Google Scholar
Galí, M., Devred, E., Levasseur, M., Royer, S.-J. & Babin, M. A remote sensing algorithm for planktonic dimethylsulfoniopropionate (DMSP) and an analysis of global patterns. Remote Sens. Environ. 171, 171–184 (2015).
Article Google Scholar
Kiene, R. P. & Linn, L. J. Distribution and turnover of dissolved DMSP and its relationship with bacterial production and dimethylsulfide in the Gulf of Mexico. Limnol. Oceanogr. 45, 849–861 (2000).
Article Google Scholar
Yang, G. P., Levasseur, M., Michaud, S. & Scarratt, M. Biogeochemistry of dimethylsulfide (DMS) and dimethylsulfoniopropionate (DMSP) in the surface microlayer and subsurface water of the western North Atlantic during spring. Mar. Chem. 96, 315–329 (2005).
Article Google Scholar
Chen, X., Liu, L., Gao, X., Dai, X. & Tang, K. Metabolism of chiral sulfonate compound 2,3-dihydroxypropane-1-sulfonate (DHPS) by Roseobacter bacteria in marine environment. Environ. Int. 157, 106829 (2021).
Article PubMed Google Scholar
Black, P. N. & DiRusso, C. C. Transmembrane movement of exogenous long-chain fatty acids: Proteins, enzymes, and vectorial esterification. Microbiol. Mol. Biol. Rev. 67, 454–472 (2003).
Article PubMed PubMed Central Google Scholar
Denger, K. & Cook, A. M. Racemase activity effected by two dehydrogenases in sulfolactate degradation by Chromohalobacter salexigens: purification of (S)-sulfolactate dehydrogenase. Microbiology 156, 967–974 (2010).
Article PubMed Google Scholar
Styp von Rekowski, K., Denger, K. & Cook, A. M. Isethionate as a product from taurine during nitrogen-limited growth of Klebsiella oxytoca TauN1. Arch. Microbiol. 183, 325–330 (2005).
Article PubMed Google Scholar
Durham, B. P. et al. Recognition cascade and metabolite transfer in a marine bacteria-phytoplankton model system. Environ. Microbiol. 19, 3500–3513 (2017).
Article PubMed Google Scholar
Abayakoon, P. et al. Comprehensive synthesis of substrates, intermediates, and products of the sulfoglycolytic Embden-Meyerhoff-Parnas pathway. J. Org. Chem. 84, 2901–2910 (2019).
Article PubMed Google Scholar
Buckley, D. H., Huangyutitham, V., Hsu, S. & Nelson, T. A. Stable isotope probing with ¹⁵N achieved by disentangling the effects of genome G+C content and isotope enrichment on DNA density. Appl Environ. Microbiol. 73, 3189–3195 (2007).
Article PubMed PubMed Central Google Scholar
Park, J.-W. & Crowley, D. E. Normalization of soil DNA extraction for accurate quantification of target genes by Real-Time PCR and DGGE. Biotechniques 38, 579–586 (2005).
Article PubMed Google Scholar
Deng, W. et al. Strategies of chemolithoautotrophs adapting to high temperature and extremely acidic conditions in a shallow hydrothermal ecosystem. Microbiome 11, 270 (2023).
Article PubMed PubMed Central Google Scholar
Zhu, C. et al. Analysis of microbial diversity and composition in small intestine during different development times in ducks. Poult. Sci. 99, 1096–1106 (2020).
Article PubMed PubMed Central Google Scholar
Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120 (2014).
Article PubMed PubMed Central Google Scholar
Caporaso, J. G. et al. QIIME allows analysis of high-throughput community sequencing data. Nat. Methods 7, 335–336 (2010).
Article PubMed PubMed Central Google Scholar
Buetas, E. et al. Full-length 16S rRNA gene sequencing by PacBio improves taxonomic resolution in human microbiome samples. BMC Genomics 25, 310 (2024).
Article PubMed PubMed Central Google Scholar
Hungate, B. A. et al. Quantitative microbial ecology through stable isotope probing. Appl. Environ. Microb. 81, 7570–7581 (2015).
Article Google Scholar
Connelly, T. L., Baer, S. E., Cooper, J. T., Bronk, D. A. & Wawrik, B. Urea uptake and carbon fixation by marine pelagic bacteria and archaea during the Arctic summer and winter seasons. Appl. Environ. Microbiol. 80, 6013–6022 (2014).
Article PubMed PubMed Central Google Scholar
Lin, S. Algal culturing techniques. J. Phycol. 41, 906–908 (2010).
Article Google Scholar
Guillard, R. Culture of Phytoplankton for Feeding Marine Invertebrates (Springer US, Boston, MA, 1975).
Moore, L. R. et al. Culturing the marine cyanobacterium Prochlorococcus. Limnol. Oceanogr.-Meth. 5, 353–362 (2007).
Article Google Scholar
Chen, C. T. A., Lin, C. M., Huang, B. T. & Chang, L. F. Stoichiometry of carbon, hydrogen, nitrogen, sulfur and oxygen in the particulate matter of the western North Pacific marginal seas. Mar. Chem. 54, 179–190 (1996).
Article Google Scholar
Wiśniewski, J. R., Zougman, A., Nagaraj, N. & Mann, M. Universal sample preparation method for proteome analysis. Nat. Methods 6, 359–362 (2009).
Article PubMed Google Scholar
Demichev, V. et al. DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput. Nat. Methods 17, 41–44 (2020).
Article PubMed Google Scholar
Zhang, Y. et al. Concise synthesis of sulfoquinovose and sulfoquinovosyl diacylglycerides, and development of a fluorogenic substrate for sulfoquinovosidases. Org. Biomol. Chem. 18, 675–686 (2020).
Article PubMed Google Scholar
Zhang, J. et al. Carbon isotopic constraints on Basin-Scale vertical and lateral particulate organic carbon dynamics in the Northern South China Sea. J. Geophys. 127, e2022JC018830 (2022).
Google Scholar
Bligh, E. G. & Dyer, W. J. A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 37, 911–917 (1959).
Article PubMed Google Scholar
Cantarero, S. I. et al. Size-Fractionated contribution of microbial biomass to suspended organic matter in the Eastern Tropical South Pacific Oxygen Minimum Zone. Front. Mar. Sci. 7, 540643 (2020).
Article Google Scholar
Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021).
Article PubMed PubMed Central Google Scholar
Phillips, J. C. et al. Scalable molecular dynamics with NAMD. J. Comput. Chem. 26, 1781–1802 (2005).
Article PubMed PubMed Central Google Scholar
Chaumeil, P.-A., Mussig, A. J., Hugenholtz, P. & Parks, D. H. GTDB-Tk: a toolkit to classify genomes with the Genome Taxonomy Database. Bioinformatics 36, 1925–1927 (2020).
Article Google Scholar
Altschul, S. F. Basic local alignment search tool (BLAST). J. Mol. Biol. 215, 403–410 (2012).
Article Google Scholar
Gilchrist, C. L. M. et al. cblaster: a remote search tool for rapid identification and visualization of homologous gene clusters. Bioinform. Adv. 1, vbab016 (2021).
Article PubMed PubMed Central Google Scholar
Kalyaanamoorthy, S., Minh, B. Q., Wong, T. K. F., von Haeseler, A. & Jermiin, L. S. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat. Methods 14, 587–589 (2017).
Article PubMed PubMed Central Google Scholar
Nguyen, L. T., Schmidt, H. A., von Haeseler, A. & Minh, B. Q. IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol. Biol. Evol. 32, 268–274 (2015).
Article PubMed Google Scholar
Stamatakis, A. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30, 1312–1313 (2014).
Article PubMed PubMed Central Google Scholar
Subramanian, B., Gao, S., Lercher, M. J., Hu, S. & Chen, W. H. Evolview v3: a webserver for visualization, annotation, and management of phylogenetic trees. Nucleic Acids Res 47, W270–W275 (2019).
Article PubMed PubMed Central Google Scholar
Rozewicki, J., Li, S., Amada, K. M., Standley, D. M. & Katoh, K. MAFFT-DASH: integrated protein sequence and structural alignment. Nucleic Acids Res. 47, W5–W10 (2019).
PubMed PubMed Central Google Scholar
Robert, X. & Gouet, P. Deciphering key features in protein structures with the new ENDscript server. Nucleic Acids Res. 42, W320–W324 (2014).
Article PubMed PubMed Central Google Scholar
Milanese, A. et al. Microbial abundance, activity and population genomic profiling with mOTUs2. Nat. Commun. 10, 1014 (2019).
Article PubMed PubMed Central Google Scholar
Chen, T. et al. Virus–pathogen interactions improve water quality along the Middle Route of the South-to-North Water Diversion Canal. ISME J. 17, 1719–1732 (2023).
Article PubMed PubMed Central Google Scholar
Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B Stat. Method. 57, 289–300 (1995).
Article MathSciNet Google Scholar
Wei, Y., Tong, Y. & Zhang, Y. New mechanisms for bacterial degradation of sulfoquinovose. Biosci. Rep. 42, BSR20220314 (2022).
Article PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank Prof Kunshan Gao (Xiamen University, China) and Center for Collections of Marine Algae (Xiamen University, China) for providing algae strains. We thank Prof Shady A. Amin (New York University Abu Dhabi, United Arab Emirates) for providing the A. macleodii F12. We thank all participants of the cruise from the Third Institute of Oceanography, and the Institute of South China Sea Marine Survey and Technology Center for their assistance in collecting samples in the Central North Pacific. This work was supported by the National Key Research and Development Program of China (2020YFA0608300), and the National Natural Science Foundation of China project (92251306, 42188102, 42276120, 42141003), Australian Research Council (DP240100126, DP250100819), the National Program on Global Change and Air-Sea Interaction (GASI-01-CNPAC-CJ07). Data and samples were collected onboard of R/V ' Lanhai 101 & apos; implementing the open research cruise NORC2022-01 supported by NSFC Shiptime Sharing Project (42149901). This paper contributes to the science plan of the Global Ocean Negative Carbon Emissions (Global ONCE) Program.

Author information

These authors contributed equally: Xiaoyi Ma, Huanyu Wang, Changjie Dong.

Authors and Affiliations

State Key Laboratory of Marine Environmental Science, Fujian Key Laboratory of Marine Carbon Sequestration, College of Ocean and Earth Science, Xiamen University, Xiamen, China
Xiaoyi Ma, Huanyu Wang, Changjie Dong, Le Liu, Xuanyun Qiu, Quanrui Chen, Yao Zhang, Nianzhi Jiao & Kai Tang
Technical Innovation Center for Utilization of Marine Biological Resources, Third Institute of Oceanography, Ministry of Natural Resources, Xiamen, China
Xiaofeng Chen & Liang Wang
School of Chemistry and Bio21 Molecular Science and Biotechnology Institute, University of Melbourne, Parkville, VIC, Australia
Spencer J. Williams

Authors

Xiaoyi Ma
View author publications
Search author on:PubMed Google Scholar
Huanyu Wang
View author publications
Search author on:PubMed Google Scholar
Changjie Dong
View author publications
Search author on:PubMed Google Scholar
Le Liu
View author publications
Search author on:PubMed Google Scholar
Xuanyun Qiu
View author publications
Search author on:PubMed Google Scholar
Xiaofeng Chen
View author publications
Search author on:PubMed Google Scholar
Quanrui Chen
View author publications
Search author on:PubMed Google Scholar
Liang Wang
View author publications
Search author on:PubMed Google Scholar
Yao Zhang
View author publications
Search author on:PubMed Google Scholar
Nianzhi Jiao
View author publications
Search author on:PubMed Google Scholar
Spencer J. Williams
View author publications
Search author on:PubMed Google Scholar
Kai Tang
View author publications
Search author on:PubMed Google Scholar

Contributions

K.T. conceived the study, analyzed data, and wrote the manuscript. X.M. H.W., and C.D. performed the experiments, analyzed data, wrote, and edited the manuscript. L.L. analyzed data, wrote, and edited the manuscript. X.C. and L.W. collected samples from Central North Pacific. X.Q. and Q.C. performed the experiments. Y.Z., S.J.W., and N.J. revised the manuscript and provided constructive advice.

Corresponding authors

Correspondence to Spencer J. Williams or Kai Tang.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Ömer Coskun and the other anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Descriptions of Additional Supplementary Files

Supplementary Data 1-19

Reporting Summary

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Ma, X., Wang, H., Dong, C. et al. Cosmopolitan marine bacteria facilitate a vast phytoplankton-derived sulfonate-based carbon flow through sulfoquinovosidases. Nat Commun 17, 209 (2026). https://doi.org/10.1038/s41467-025-66898-z

Download citation

Received: 26 November 2024
Accepted: 17 November 2025
Published: 05 December 2025
Version of record: 07 January 2026
DOI: https://doi.org/10.1038/s41467-025-66898-z