Multi-omics and cultivation reveal laminarin-degrading PVC bacteria in the deep sea Authors: Rikuan Zheng, Chong Wang, Chaomin Sun Published: 2026 ABSTRACT The deep sea is home to a vast and largely unexplored microbial biosphere, along with large amounts of complex organic matter (COM). However, the functional capacity of the deep-sea microbiome to metabolize organic matter across diverse regions remains poorly understood. Here, we combine 16S rRNA gene amplicon sequencing, metagenomics, and metatranscriptomics to comprehensively characterize prokaryotic communities across different years (2018 and 2022) and habitats (cold seeps, hydrothermal vents, and seamounts). Our results reveal spatio-temporal community heterogeneity driven by geochemical gradients, alongside a widespread genetic potential for organic matter metabolism. Notably, the PVC (Planctomycetota-Verrucomicrobiota-Chlamydiota) superphylum exhibits extensive polysaccharide degradation capabilities, exemplified by the isolation of Planctomycetota strain WC338 and Lentisphaerota strain WC36 via laminarin enrichment. Growth experiments and transcriptomics confirm their strict laminarin dependence and characterize the underlying catabolic machinery—specifically, the deployment of different glycoside hydrolase (GH) families, which are broadly distributed and prevalent across the PVC superphylum. Furthermore, we demonstrate that laminarin acts as an effective selective substrate for enriching and isolating the deep-sea PVC superphylum bacteria. Collectively, these findings reveal that PVC bacteria—an overlooked group in laminarin degradation—possess specialized adaptations for polysaccharide breakdown and actively participate in laminarin turnover in deep-sea environments. Introduction The deep seafloor contains large amounts of organic matter—such as proteins, polysaccharides, nucleic acids, lipids, and humic acid fractions—that sustain diverse heterotrophic microbial communities1–3. Among the myriad components of polysaccharides, laminarin stands out as one of the most abundant polysaccharides globally, comprising approximately 26% of marine particulate organic carbon (POC)4,5. Structurally, laminarin is characterized by a β−1,3-D-glucan backbone, often with β−1,6-D-glucan branches, making it an important carbon source within marine algae and phytoplankton6,7. The rapid degradation of laminarin, particularly during algal blooms, is facilitated by abundant heterotrophic glycoside hydrolases and sugar transporters8. Key enzymes involved in laminarin hydrolysis are classified within the Carbohydrate-Active Enzyme (CAZyme) database. Specifically, glycoside hydrolase families GH16, GH17, GH55, GH64, GH81, and GH128 are typically classified as endo-β-1,3-glucanases, whereas GH3 and GH30 possess exo-β-1,3-glucanase activities. Evidence of laminarinase activity is prevalent across various marine strata, from surface waters to deeper water columns and sediments9,10, strongly suggesting widespread distribution of laminarin-degrading bacteria and their associated enzymes throughout the world’s oceans. Despite this ubiquity, detailed investigations into the ability of deep-sea microorganisms to degrade laminarin remain remarkably scarce. There are currently no reports documenting laminarin degradation by purely cultured deep-sea bacteria, particularly those belonging to the PVC superphylum, which constitutes a highly abundant and often dominant phylogenetic lineage in deep-sea environments11. Addressing this void requires characterizing the laminarin degradation potential of deep-sea microbes, with a focus on the understudied PVC superphylum, to understand their role in utilizing this widespread carbon source. The PVC superphylum—comprising Planctomycetota, Verrucomicrobiota, Chlamydiota, and related phyla—represents a bacterial lineage with unique cellular traits reminiscent of eukaryotes12–14. Among these, Planctomycetota are the most extensively studied and are ubiquitously distributed across diverse environments15–19. They are also hypothesized to dominate deep-sea sediment communities20, reaching relative abundances of up to 28% in regions such as the Gulf of Mexico11. Metagenomic and distributional evidence has suggested their apparent involvement in nitrogen cycling and the breakdown of complex organic matter delivered as marine snow11,21,22. However, to date, only two Planctomycetota strains have been successfully isolated from deep-sea environments23–25, hindering comprehensive assessment of their physiology and metabolic capabilities26. Despite their apparent ecological importance, our understanding of Planctomycetota function in the deep sea remains limited. Obtaining deep-sea isolates from the PVC superphylum is therefore essential for elucidating their role in degrading complex polysaccharides such as laminarin. In this study, we integrated multi-omics analyses with targeted pure-culture isolation from diverse deep-sea environments sampled in 2018 and 2022. We reveal that resident microbial communities harbor a widespread genetic potential for organic matter degradation. Critically, our findings reveal that novel isolates of Planctomycetota and Lentisphaerota possess a previously unrecognized capacity for laminarin degradation, supported by the conservation and high expression of key laminarinase enzymes. Furthermore, we demonstrate that laminarin serves as an effective substrate for enriching and isolating a broad range of PVC superphylum bacteria. Together, these metabolic traits suggest that these lineages may play an underappreciated role in the turnover of marine polysaccharides, with potential implications for carbon cycling in deep-sea environments. Results and discussion Microbial community diversity across deep-sea sediment habitats The deep-sea environment hosts a vast, largely unexplored microbial biosphere, where prokaryotes, endowed with diverse metabolic capabilities, are important drivers of organic matter cycling27,28. To analyze microbial diversity and composition in different deep-sea sediments, we previously employed 16S rRNA gene amplicon sequencing on sediment samples collected from seven distinct locations and years: 2018 Site F cold seep, 2018 Site H1 hydrothermal vent, 2018 Site H2 hydrothermal vent, 2022 Site F cold seep 1, 2022 Site F cold seep 2, 2022 Haima cold seep, and 2022 Zhenbei seamount (Fig. 1a and Supplementary Data 1)29. As shown in Supplementary Fig. 1, the microbial communities in the 2018 cold seep and hydrothermal vent samples were predominantly composed of the phyla Pseudomonadota, Chloroflexota, Planctomycetota, Bacteroidota, Acidobacteriota, and Actinomycetota. Notably, Pseudomonadota dominated the cold seep communities, reaching up to 58.6% relative abundance. In contrast, this phylum was less prevalent at the two hydrothermal vent sites (21.9% at H1 and 32.1% at H2), where Thermodesulfobacteriota (24.0% and 11.8%) and Campylobacterota (18.9% and 22.1%) were more abundant. Both of these phyla were rare in cold seep sediments (0.2% and 0.3%, respectively). This observed difference in community composition between cold seeps and hydrothermal vents may reflect varying geochemical gradients and nutrient availability27,30. As shown in Supplementary Fig. 2, the microbial communities in the 2022 cold seep and seamount samples were dominated by the phyla Chloroflexota, Thermodesulfobacteriota, Pseudomonadota, and Atribacterota. The abundance of Thermodesulfobacteriota (22.5% on average) and Atribacterota (19.4% on average) was higher in the cold seep sites compared to the seamount site (0.18% and 0%, respectively). This contrasts with the 2018 cold seep communities, which were dominated by Pseudomonadota (up to 58.6%) and showed lower relative abundances of Chloroflexota and Thermodesulfobacteriota. Thermodesulfobacteriota, a key group of sulfate-reducing bacteria, are thus expected to thrive in these sulfate-rich environments, which also support the activity of anaerobic methane oxidizers (ANMEs). The microbial communities in these cold seep sediments likely govern the dominant biogeochemical processes through the synergistic microbial sulfate reduction and anaerobic methane oxidation31. These findings highlight the substantial heterogeneity in microbial community composition across deep-sea sediment samples, both temporally and spatially.Fig. 1Sample location and phylogenetic analysis of deep-sea metagenome-assembled genomes (MAGs).a Geographic distribution of sampling sites across seven deep-sea environments (2018 Site F cold seep, 2018 Site H1 hydrothermal vent, 2018 Site H2 hydrothermal vent, 2022 Site F cold seep 1, 2022 Site F cold seep 2, 2022 Haima cold seep and 2022 Zhenbei seamount). Image generated with Ocean Data View (ODV; Schlitzer, Reiner, https://odv.awi.de, 2021). b Phylogenetic tree of MAGs. Innermost ring: phylum-level clades (colored). Second ring: genome size. Third and fourth rings: contamination and completeness levels. Outermost ring: Bacterial (light yellow) and Archaeal (light blue) MAGs. Source data are provided as a Source Data file. Sample location and phylogenetic analysis of deep-sea metagenome-assembled genomes (MAGs). a Geographic distribution of sampling sites across seven deep-sea environments (2018 Site F cold seep, 2018 Site H1 hydrothermal vent, 2018 Site H2 hydrothermal vent, 2022 Site F cold seep 1, 2022 Site F cold seep 2, 2022 Haima cold seep and 2022 Zhenbei seamount). Image generated with Ocean Data View (ODV; Schlitzer, Reiner, https://odv.awi.de, 2021). b Phylogenetic tree of MAGs. Innermost ring: phylum-level clades (colored). Second ring: genome size. Third and fourth rings: contamination and completeness levels. Outermost ring: Bacterial (light yellow) and Archaeal (light blue) MAGs. Source data are provided as a Source Data file. Metagenomic sequencing of the seven deep-sea sediments yielded 238 high- to medium-quality (completeness ≥50%, contamination ≤10%) microbial metagenome-assembled genomes (MAGs)32 after de novo assembly and binning. These 193 bacterial and 45 archaeal MAGs spanned 37 known and eight unclassified phyla (Fig. 1b and Supplementary Data 2). Bacterial MAGs were predominantly derived from dominant lineages, with significant representation from Pseudomonadota (n = 49), Thermodesulfobacteriota (n = 22), Bacteroidota (n = 20), Chloroflexota (n = 19), Patescibacteriota (n = 9), Calditrichota (n = 8), Bipolaricaulota (n = 4), Planctomycetota (n = 4), Acidobacteriota (n = 4), and Nitrospinota (n = 4). These dominant bacterial phyla were present across all seven sediments, consistent with the amplicon sequencing results (Supplementary Figs. 1a and 2). Within the Archaea domain, the 45 MAGs were predominantly derived from Halobacteriota (n = 12), Asgardarchaeota (n = 11), Thermoplasmatota (n = 7), Nanobdellota (n = 6), Hydrothermarchaeota (n = 4), Thermoproteota (n = 2), Aenigmatarchaeota (n = 1), along with two unclassified archaeal MAGs. These MAGs spanned a wide range of taxonomic lineages and enabled a comprehensive assessment of the genomic potential for COM degradation across different deep-sea habitats. Deep-sea microbes possess substantial potential for COM metabolism To estimate the genomic potential for degrading COM, we screened these MAGs for genes encoding enzymes involved in the breakdown of proteins, polysaccharides, nucleic acids, and lipids. Initially, the majority of MAGs encoded extracellular peptidases, predominantly found in dominant phyla such as Pseudomonadota, Thermodesulfobacteriota, Bacteroidota, and Krumholzibacteriota (Fig. 2 and Supplementary Data 3). Analysis of their protein transport systems indicated that most MAGs possess transporters for polar amino acids, branched-chain amino acids, peptides, and oligopeptides. The presence of genes encoding amino acid degradation enzymes (gdh, ior, kor, por, and vor) suggests that these MAGs can likely utilize environmental proteins as energy sources. Remarkably, a significant proportion of our recovered MAGs, particularly from these cold seep, hydrothermal vent, and seamount sediments, encodes most amino acid transport and degradation pathways, contrasting sharply with previous reports of widespread absence of such pathways in MAGs from gas hydrate-bearing subseafloor and the Helgoland mud33,34.Fig. 2Distribution of key catabolic genes across deep-sea MAGs reveals diverse degradation capabilities of complex organic matter.Key genes responsible for the catabolism of proteins, polysaccharides, nucleic acids, and lipids were detected in MAGs. Color intensity indicates gene prevalence within each cluster: dark shading denotes presence in >50% of MAGs, and light shading indicates presence in <50% of MAGs. The total number of MAGs per cluster is shown in parentheses. A: extracellular peptidases; B: glutamate/aspartate transport system; C: general L-amino acid transport system; D: glutamate transport system; E: cystine transport system; F: arginine/ornithine transport system; G: putative polar amino acid transport system; H: branched-chain amino acid transport system; I: peptides/nickel transport system; J: oligopeptide transport system; K: arginine/lysine/histidine/glutamine transport system; L: gdh; M: iorA; N: iorB; O: korA; P: korB; Q: korC; R: korD; S: porA; T: porB; U: porC; V: porD; W: vorA; X: vorB; Y: extracellular CAZymes; Z: maltose/maltodextrin transport system; A1: putative multiple sugar transport system; B1: ribose transport system; C1: putative chitobiose transport system; D1: glucose/mannose transport system; E1: PTS system, D-glucosaminate-specific II component; F1: glycolysis; G1: cytosine/cytidine deaminase; H1: extracellular nuclease; I1: guanine/adenine/adenosin deaminase; J1: nucleobase transport; K1: nucleoside phosphorylase; L1: nucleoside transport; M1: pentose phosphate pathway; N1: phospholipid transport system; O1: fatty acid transport; P1: fadD; Q1: acd and bcd; R1: crt and paaF; S1: fadN and paaH; T1: fadA and atoB. Pathway identification was performed with the software METABOLIC. All annotations are provided in Supplementary Data 3. Source data are provided as a Source Data file. Distribution of key catabolic genes across deep-sea MAGs reveals diverse degradation capabilities of complex organic matter. Key genes responsible for the catabolism of proteins, polysaccharides, nucleic acids, and lipids were detected in MAGs. Color intensity indicates gene prevalence within each cluster: dark shading denotes presence in >50% of MAGs, and light shading indicates presence in <50% of MAGs. The total number of MAGs per cluster is shown in parentheses. A: extracellular peptidases; B: glutamate/aspartate transport system; C: general L-amino acid transport system; D: glutamate transport system; E: cystine transport system; F: arginine/ornithine transport system; G: putative polar amino acid transport system; H: branched-chain amino acid transport system; I: peptides/nickel transport system; J: oligopeptide transport system; K: arginine/lysine/histidine/glutamine transport system; L: gdh; M: iorA; N: iorB; O: korA; P: korB; Q: korC; R: korD; S: porA; T: porB; U: porC; V: porD; W: vorA; X: vorB; Y: extracellular CAZymes; Z: maltose/maltodextrin transport system; A1: putative multiple sugar transport system; B1: ribose transport system; C1: putative chitobiose transport system; D1: glucose/mannose transport system; E1: PTS system, D-glucosaminate-specific II component; F1: glycolysis; G1: cytosine/cytidine deaminase; H1: extracellular nuclease; I1: guanine/adenine/adenosin deaminase; J1: nucleobase transport; K1: nucleoside phosphorylase; L1: nucleoside transport; M1: pentose phosphate pathway; N1: phospholipid transport system; O1: fatty acid transport; P1: fadD; Q1: acd and bcd; R1: crt and paaF; S1: fadN and paaH; T1: fadA and atoB. Pathway identification was performed with the software METABOLIC. All annotations are provided in Supplementary Data 3. Source data are provided as a Source Data file. Subsequently, regarding polysaccharide degradation, most bacterial MAGs encoded diverse extracellular CAZymes capable of cleaving a wide range of carbohydrates, coupled with numerous sugar transport systems for the uptake of oligo- and monomers. These breakdown products can subsequently enter the ubiquitous glycolysis pathway (Fig. 2). In contrast, the vast majority of archaeal MAGs lacked genes for extracellular CAZymes (Supplementary Data 4), suggesting a limited capacity for direct polysaccharide hydrolysis. Nonetheless, nearly all archaeal and bacterial MAGs harbored glycolysis genes, indicating a conserved potential for fermentative metabolism of simpler sugars—potentially derived from the partial hydrolysis of polysaccharides by co-occurring bacteria, which was consistent with previous studies33. While nucleic acids are recognized as a fundamental trophic resource and energy source in the deep sea35, their degradation potential across the microbial biosphere has remained largely an enigma. Prior investigations have primarily focused on specific bacterial isolates, such as our previous demonstration of extracellular DNA utilization by deep-sea Xianfuyuplasma coldseepsis zrk1336. Our results showed that the genomic potential for nucleic acid degradation was distributed across both bacterial and archaeal lineages (Fig. 2 and Supplementary Data 5). These MAGs encoded a suite of genes facilitating efficient heterotrophic nutrition. Key capabilities included extracellular nucleases and phosphohydrolases for liberating nucleosides and nucleobases, complemented by transporters for their uptake and dedicated pathways for purine and pyrimidine degradation. The resultant ribose could be channeled into the pentose phosphate pathway. Furthermore, the prevalence of genes for lipid catabolism was observed, encompassing phospholipid and fatty acid transporters along with complete beta-oxidation machineries for efficient energy generation (Fig. 2 and Supplementary Data 6). However, certain archaeal lineages, such as Nanobdellota, Aenigmatarchaeota, Altiarchaeota, Halobacteriota, Hydrothermarchaeota, and Iainarchaeota, lacked the genetic machinery for fatty acid degradation. Collectively, these findings indicate that the majority of deep-sea microorganisms possess a substantial capacity to degrade COM. We acknowledge, however, that the limited number of bacterial and archaeal MAGs recovered in this study constrains the generalizability of these observations. While our MAG-centric analysis provides a framework for assessing genomic potential, we acknowledge that the prevalence estimates presented here are inherently influenced by bin completeness and contamination—factors that may lead to under- or overestimation of gene frequencies, particularly among medium-quality bins. Future work integrating complementary read-based profiling and quantitative activity measurements will be necessary to fully resolve the distribution and expression of these catabolic pathways in situ. Ubiquitous polysaccharide degradation potential in deep-sea microbes In marine ecosystems, polysaccharides primarily derived from the cell wall constituents of phytoplankton and macroalgae, represent a substantial carbon input into the deep sea upon sedimentation37,38. These recalcitrant molecules are considered a major carbon source for benthic microbial communities. However, current understanding of polysaccharide degradation in deep-sea environments remains notably limited—particularly regarding the largely overlooked PVC superphylum—largely due to the prevailing paradigm that primarily associates such capabilities with specific bacterial lineages like the Bacteroidota phylum38,39. This dominant focus has largely overlooked the broader metabolic repertoire of other deep-sea microorganisms. To elucidate the metabolic potential of deep-sea microorganisms for polysaccharide degradation and utilization, metagenomic and metatranscriptomic analyses were conducted on four deep-sea sediment samples collected in 2022, with a focus on CAZyme repertoires. Our screen for CAZymes revealed their widespread distribution across the generated MAGs, encompassing major families including GHs, GTs, PLs, CEs, AAs, and CBMs (Fig. 3 and Supplementary Data 7). To assess the functional relevance of these genes in deep-sea in situ environments, initial metatranscriptomic analysis showed widespread transcription of diverse CAZyme families—predominantly GHs and GTs—across most MAGs via the normalized transcripts per million (TPM), underscoring active polysaccharide degradation (Supplementary Fig. 3, Supplementary Data 8–13). To move beyond simple transcript abundance and address statistical support, we performed a targeted differential expression analysis by normalizing GH and GT transcript levels against 10 conserved marker genes40 (fold change analysis, Fig. 3 and Supplementary Data 14–15). These upregulated GH and GT families were primarily associated with dominant lineages, including Thermodesulfobacteriota, Bacteroidota, Alphaproteobacteria, Planctomycetota, and Omnitrophota, with predominant enrichment in Site F cold seep sediments. The expression profiles of PVC bacteria revealed a clear niche adaptation: GH-encoding genes in three Planctomycetota and one Omnitrophota MAGs were predominantly upregulated in cold seep sediments, particularly Site F, suggesting a potential role for these lineages in polysaccharide degradation within cold seep habitats (Supplementary Fig. 4). Together, these results demonstrate that polysaccharide degradation in deep-sea sediments is mediated by distinct microbial lineages exhibiting habitat-specific patterns of CAZyme expression. Notably, Planctomycetota and Omnitrophota belong to the PVC superphylum, a group previously reported for its potential to degrade polysaccharides in other environments22,41. While we did not directly measure extracellular enzyme activities or polysaccharide hydrolysis rates—and thus mRNA abundance alone does not directly quantify functional turnover—the convergence of genomic potential and in situ transcriptional activity, supported by replicate-level statistical inference, provides evidence that these lineages may contribute to polysaccharide degradation in deep-sea environments. These findings reveal that this understudied PVC superphylum bacteria may contribute to polysaccharide degradation in deep-sea environments alongside other lineages, and provide an empirical framework for the targeted cultivation of previously recalcitrant PVC bacteria.Fig. 3Relative abundance of potential carbohydrate-active enzyme genes across deep-sea microbial communities.The left panel shows the abundance of key carbohydrate-active enzyme (CAZyme) families (GH, glycoside hydrolase; GT, glycosyltransferase; PL, polysaccharide lyase; CE, carbohydrate esterase; AA, auxiliary activity; CBM, carbohydrate-binding module) within deep-sea MAGs. MAGs are organized into distinct phylogenetic clusters from four deep-sea sites: 2022 Haima cold seep (HMZ), 2022 Zhenbei seamount (HSX), 2022 Site F cold seep 1 (TPK), and 2022 Site F cold seep 2 (TPKF). The right panel displays the relative abundance of genes encoding GH and GT within these MAGs from metatranscriptomic data (three replicates per site). The relative abundance of each gene was normalized against the average abundance of the 10 selected marker genes and quantified as log2-transformed fold change (log2 fold change) based on metatranscriptomic data (three replicates per site). Sites are denoted as follows: I, HMZ; II, HSX; III, TPK; IV, TPKF. Source data are provided as a Source Data file. Relative abundance of potential carbohydrate-active enzyme genes across deep-sea microbial communities. The left panel shows the abundance of key carbohydrate-active enzyme (CAZyme) families (GH, glycoside hydrolase; GT, glycosyltransferase; PL, polysaccharide lyase; CE, carbohydrate esterase; AA, auxiliary activity; CBM, carbohydrate-binding module) within deep-sea MAGs. MAGs are organized into distinct phylogenetic clusters from four deep-sea sites: 2022 Haima cold seep (HMZ), 2022 Zhenbei seamount (HSX), 2022 Site F cold seep 1 (TPK), and 2022 Site F cold seep 2 (TPKF). The right panel displays the relative abundance of genes encoding GH and GT within these MAGs from metatranscriptomic data (three replicates per site). The relative abundance of each gene was normalized against the average abundance of the 10 selected marker genes and quantified as log2-transformed fold change (log2 fold change) based on metatranscriptomic data (three replicates per site). Sites are denoted as follows: I, HMZ; II, HSX; III, TPK; IV, TPKF. Source data are provided as a Source Data file. Strict laminarin utilization by deep-sea PVC superphylum bacteria To functionally link the observed polysaccharide degradation potential in deep-sea PVC superphylum members to isolates, we employed an optimized, selective enrichment strategy utilizing laminarin as the sole carbon and energy source. This cultivation yielded two robustly growing strains (WC338 and WC36) belonging to the Planctomycetota and Lentisphaerota (PVC superphylum) phyla. Initial phenotypic assessment via transmission electron microscopy (TEM) revealed distinct morphologies: strain WC338 displayed a spherical shape (∼1.0 µm) with a single polar flagellum (Fig. 4a), while strain WC36 was oval-shaped (∼1.0 µm) and appeared aflagellate (Fig. 4b). A maximum-likelihood phylogenetic tree based on genomes revealed that strain WC338 clusters robustly within the genus Poriferisphaera (Planctomycetota), forming a distinct, monophyletic clade separate from its closest described relatives, such as Poriferisphaera corsica KS4 (98.06%, based on full-length 16S rRNA gene identity) and Poriferisphaera heterotrophicis ZRK32 (97.19%) (Fig. 4c). This sequence divergence falls below established species demarcation thresholds (98.65%)42, thus suggesting that strain WC338 represents a novel species within Poriferisphaera. In parallel, strain WC36 was taxonomically placed within the order Victivallales (Lentisphaerota). Strain WC36 forms a deeply branching lineage, with sequence identity to its nearest neighbors, such as Victivallis lenta BBE-744-WT-12 (85.95%) and Victivallis vadensis Cello (85.14%), falling below the recognized threshold for family delineation (86.5%)43. This significant phylogenetic distance suggests that strain WC36 may represent a novel family within the order Victivallales, substantially expanding the known taxonomic breadth of this understudied order. To further corroborate their taxonomic status, we calculated average nucleotide identities (ANIb, based on BLAST + ; ANIm, based on MUMmer) and tetranucleotide signatures (Tetra) for both strains. The results showed that the genomes of strains WC338 (ANIb, 68.8%; ANIm, 86.13%; Tetra, 0.93; Supplementary Figs. 5–7) and WC36 (ANIb, 65.01%; ANIm, 89.1%; Tetra, 0.69; Supplementary Figs. 8–10) fell well below the established cut-off values (ANIb: 95%, ANIm: 95%, Tetra: 0.99) for species classification, strongly suggesting that these strains represent novel taxa as currently defined.Fig. 4Isolation, morphology, and phylogenetic analysis of cultured deep-sea PVC superphylum bacteria.a, b Transmission electron microscopy (TEM) micrographs of Poriferisphaera sp. WC338 (a) and Lentisphaerota sp. WC36 (b). These experiments were independently repeated more than three times with similar results. Representative images are shown. c Maximum likelihood phylogenetic tree of deep-sea Poriferisphaera sp. WC338 and Lentisphaerota sp. WC36 was reconstructed using the Genome Taxonomy Database Toolkit (https://github.com/Ecogenomics/GTDBTk). The tree is inferred and reconstructed under the maximum likelihood criterion and nodes with greater than 80% bootstrap support are labeled with a gray circle (expressed as percentages of 1000 replications). Bar, 0.1 substitutions per nucleotide position. Source data are provided as a Source Data file. Isolation, morphology, and phylogenetic analysis of cultured deep-sea PVC superphylum bacteria. a, b Transmission electron microscopy (TEM) micrographs of Poriferisphaera sp. WC338 (a) and Lentisphaerota sp. WC36 (b). These experiments were independently repeated more than three times with similar results. Representative images are shown. c Maximum likelihood phylogenetic tree of deep-sea Poriferisphaera sp. WC338 and Lentisphaerota sp. WC36 was reconstructed using the Genome Taxonomy Database Toolkit (https://github.com/Ecogenomics/GTDBTk). The tree is inferred and reconstructed under the maximum likelihood criterion and nodes with greater than 80% bootstrap support are labeled with a gray circle (expressed as percentages of 1000 replications). Bar, 0.1 substitutions per nucleotide position. Source data are provided as a Source Data file. Metabolic analysis confirmed the direct utilization of laminarin by these two PVC superphylum strains. When cultured in basal medium where laminarin was supplied as the sole carbon source, both strains reached the stationary phase within 5–6 days (OD600 ~ 0.16) (Fig. 5a, b). This growth correlated directly with laminarin consumption, evidenced by a decrease in laminarin relative content from 100% down to 55.7% (strain WC338) and 66.5% (strain WC36). These results demonstrate that these PVC superphylum bacteria can effectively catabolize laminarin to sustain growth and maintain viability. Furthermore, the growth of both strains exhibited strict nutritional dependence on laminarin. Supplementation of the basal medium with common energy sources, including glucose, N-acetylglucosamine, yeast extract, or peptone, failed to support measurable growth (Fig. 5c, d). Intriguingly, while N-acetylglucosamine is a frequently utilized substrate by Planctomycetota bacteria23, strain WC338 could not utilize it. Laminarin, a key component of the marine carbon cycle, constitutes approximately 26% of the marine particulate organic carbon pool, underscoring its substantial contribution4,5,7. Therefore, this strict dependence on laminarin coupled with the inability to assimilate simpler carbon sources like N-acetylglucosamine, suggesting that these deep-sea PVC superphylum bacteria possess unique, highly specialized metabolic pathways adapted specifically for laminarin assimilation, potentially reflecting their distinct niche in the deep-sea carbon cycle.Fig. 5Strict laminarin dependency drives growth of deep-sea PVC superphylum strains WC338 and WC36.a Growth of strain WC338 is coupled with laminarin consumption. Growth curves (Purple line, OD600) and relative laminarin concentrations (Orange-yellow line, relative percentage) of strain WC338. b Growth of strain WC36 is coupled with laminarin consumption. Growth curves (Purple line, OD600) and relative laminarin concentrations (Orange-yellow line, relative percentage) of strain WC36. c, d Growth of strains WC338 (c) and WC36 (d) on various carbon sources. Bacterial growth was assessed in basal medium alone or supplemented with 3.0 g/L of either laminarin, glucose, N-acetylglucosamine, yeast extract, or peptone as the sole carbon source. The asterisks indicate statistically significant differences between basal medium group and laminarin group for strains WC338 and WC36. Data are presented as mean values ± SD (n = 5, five biologically independent replicates). For strain WC338, ***P < 0.001 (exact P = 1.79 ×; 10−4); for strain WC36, ****P < 0.0001 (exact P = 5.32 × 10−5) by one‑way ANOVA with Dunnett’s multiple-comparisons test. Source data are provided as a Source Data file. Strict laminarin dependency drives growth of deep-sea PVC superphylum strains WC338 and WC36. a Growth of strain WC338 is coupled with laminarin consumption. Growth curves (Purple line, OD600) and relative laminarin concentrations (Orange-yellow line, relative percentage) of strain WC338. b Growth of strain WC36 is coupled with laminarin consumption. Growth curves (Purple line, OD600) and relative laminarin concentrations (Orange-yellow line, relative percentage) of strain WC36. c, d Growth of strains WC338 (c) and WC36 (d) on various carbon sources. Bacterial growth was assessed in basal medium alone or supplemented with 3.0 g/L of either laminarin, glucose, N-acetylglucosamine, yeast extract, or peptone as the sole carbon source. The asterisks indicate statistically significant differences between basal medium group and laminarin group for strains WC338 and WC36. Data are presented as mean values ± SD (n = 5, five biologically independent replicates). For strain WC338, ***P < 0.001 (exact P = 1.79 ×; 10−4); for strain WC36, ****P < 0.0001 (exact P = 5.32 × 10−5) by one‑way ANOVA with Dunnett’s multiple-comparisons test. Source data are provided as a Source Data file. The strategy for laminarin catabolism in deep-sea PVC superphylum bacteria To decipher the mechanisms underlying laminarin degradation and utilization by strains WC338 and WC36, we first performed a systematic analysis of their CAZymes. Strain WC338 exhibited a substantially larger complement of glycoside hydrolases (GHs) and glycosyltransferases (GTs) compared to strain WC36 (Fig. 6a, b). Notably, WC338 possessed canonical laminarin-degrading enzymes, specifically GH16 and GH134, which were absent in strain WC36. To further probe the metabolic characteristics of strains WC338 and WC36, we performed transcriptomic analyses under laminarin induction. A volcano plot analysis revealed significant transcriptional upregulation of 255 genes in strain WC338 (Fig. 6c), with genes encoding GHs, sugar transporters, and components involved in flagellar motility. Consistent with its genomic potential and laminarin degradation phenotype, strain WC338 showed upregulation of genes encoding GH16 and GH13 (Fig. 6d). Additionally, genes encoding other GH families (GH1, GH2, GH4, GH20) and numerous sugar transporter proteins (Fig. 6e), important for importing breakdown products, were also upregulated. Furthermore, a substantial cluster of genes related to flagellar motility and chemotaxis exhibited elevated expression (Fig. 6f). Given that marine bacteria are known to exhibit chemotaxis towards laminarin44, this upregulation potentially facilitates rapid capture of laminarin, a role further suggested by the single polar flagellum in WC338. In parallel, under laminarin induction, strain WC36 displayed the upregulation of 97 genes (Fig. 6g). Intriguingly, although lacking laminarin-degrading enzymes like GH16 and GH13 found in WC338, strain WC36 showed the upregulation of genes encoding GH2 (Fig. 6h). GH2 enzymes are known to hydrolyze β-1,4-glycosidic bonds45, a prevalent linkage in laminarin46, suggesting that WC36 may employ a complex GH2-associated pathway for laminarin breakdown—a mechanism not previously documented for this polysaccharide47. While we cannot completely rule out that the observed activity is partly due to minor impurities in the commercial laminarin preparations, the specific upregulation of GH2 genes may suggest a potential role in degrading the polysaccharide itself. Consistent with this, genes encoding sugar transporters were also upregulated in WC36 (Fig. 6i), indicating efficient uptake of breakdown products. Interestingly, while strain WC36 appears to lack flagellar structures, genes associated with flagellar protein synthesis were also upregulated (Fig. 6j). Although flagellar genes can play roles beyond motility in some bacteria, the function of this upregulation in WC36 remains to be determined. Collectively, these findings reveal differential strategies employed by closely related deep-sea PVC superphylum bacteria for laminarin catabolism, highlighting the metabolic plasticity and specialized adaptations of these lineages in the deep-sea environment.Fig. 6CAZymes composition and transcriptomic analyses of Poriferisphaera sp. WC338 and Lentisphaerota sp.WC36. Statistical analysis of the types and numbers of CAZymes in Poriferisphaerasp.WC338 (a) and Lentisphaerotasp.WC36 (b). c Volcano plot showing the number of significantly differentially expressed genes in strain WC338 cultured with 3.0 g/L laminarin. Differential expression analysis was performed using edgeR with a two-sided test. Differentially expressed genes were defined as P ≤ 0.05 and |log2 (fold change)| ≥ 1.0. No adjustment for multiple comparisons was applied. Red dots indicate significantly upregulated genes (n = 255), blue dots indicate significantly downregulated genes (n = 287), and gray dots indicate non-significant genes (n = 2885). Transcriptomic analysis of strain WC338 cultured with 3.0 g/L laminarin reveals the expression of genes encoding GHs (d), sugar transporters (e), and components involved in flagellar motility (f). g Volcano plot showing the number of significantly differentially expressed genes in strain WC36 cultured with 3.0 g/L laminarin. Two-sided edgeR test; P ≤ 0.05 and |log2 (fold change)| ≥ 1.0 as significance threshold. No multiple-comparison correction. Up: 97 genes, down: 305 genes. Transcriptomic analysis of strain WC36 cultured with 3.0 g/L laminarin reveals the expression of genes encoding GHs (h), sugar transporters (i), and components involved in flagellar motility (j). “Lam” indicates strains WC338 and WC36 grown in the basal medium with 3.0 g/L laminarin; “Control” indicates strains WC338 and WC36 grown in the basal medium with 0.05 g/L laminarin. The values shown above the bars in panels D-F and H-J represent the fold change of gene expression (by using the log2 value). Source data are provided as a Source Data file. CAZymes composition and transcriptomic analyses of Poriferisphaera sp. WC338 and Lentisphaerota sp. WC36. Statistical analysis of the types and numbers of CAZymes in Poriferisphaerasp.WC338 (a) and Lentisphaerotasp.WC36 (b). c Volcano plot showing the number of significantly differentially expressed genes in strain WC338 cultured with 3.0 g/L laminarin. Differential expression analysis was performed using edgeR with a two-sided test. Differentially expressed genes were defined as P ≤ 0.05 and |log2 (fold change)| ≥ 1.0. No adjustment for multiple comparisons was applied. Red dots indicate significantly upregulated genes (n = 255), blue dots indicate significantly downregulated genes (n = 287), and gray dots indicate non-significant genes (n = 2885). Transcriptomic analysis of strain WC338 cultured with 3.0 g/L laminarin reveals the expression of genes encoding GHs (d), sugar transporters (e), and components involved in flagellar motility (f). g Volcano plot showing the number of significantly differentially expressed genes in strain WC36 cultured with 3.0 g/L laminarin. Two-sided edgeR test; P ≤ 0.05 and |log2 (fold change)| ≥ 1.0 as significance threshold. No multiple-comparison correction. Up: 97 genes, down: 305 genes. Transcriptomic analysis of strain WC36 cultured with 3.0 g/L laminarin reveals the expression of genes encoding GHs (h), sugar transporters (i), and components involved in flagellar motility (j). “Lam” indicates strains WC338 and WC36 grown in the basal medium with 3.0 g/L laminarin; “Control” indicates strains WC338 and WC36 grown in the basal medium with 0.05 g/L laminarin. The values shown above the bars in panels D-F and H-J represent the fold change of gene expression (by using the log2 value). Source data are provided as a Source Data file. Ubiquitous laminarin catabolism by PVC superphylum bacteria across diverse environments Extending the investigation to the broader PVC superphylum reveals widespread laminarin-degrading potential. Building upon the discovery that strains WC338 and WC36 exhibit efficient laminarin degradation, we hypothesized that this capability might be a common trait across the PVC superphylum. To test this, we conducted a comprehensive genomic survey of laminarin-relevant glycoside hydrolase (GH) genes, specifically focusing on GH16 and GH2, as well as other GH families (GH3, GH17, GH55, GH64) found in related taxa. GH16 is a well-known laminarinase (β-1,3-glucanase), while GH2 targets β-1,4 linkages that may be present in minor amounts in laminarin or its degradation products. Our analysis encompassed 171 available genomes in the NCBI database of the PVC superphylum bacteria, including members from the phyla Planctomycetota, Verrucomicrobiota, Lentisphaerota, and Kiritimatiellota. These isolates were derived from diverse environmental niches worldwide, including marine sediments, the water column, soils, and algal samples (Supplementary Data 16), underscoring the broad ecological distribution of this bacterial group (Supplementary Fig. 11a). Our genomic analysis revealed that genes encoding GH16 and GH2 are widely distributed within the PVC superphylum bacteria (Supplementary Fig. 11b and Supplementary Data 17), suggesting a prevalent capacity for laminarin utilization. In contrast, genes for GH3, GH17, GH55, and GH64 were found in a smaller subset of these bacteria, indicating more specialized or context-dependent roles. The high prevalence of GH16 and GH2, particularly in isolates from diverse marine environments, strongly supports the hypothesis that laminarin serves as a widespread carbon source in deep-sea and other marine ecosystems7, and that PVC superphylum bacteria have evolved enzymatic machinery to exploit it. This broad distribution of laminarin-degrading genes highlights the previously underappreciated role of PVC superphylum bacteria in the marine carbon cycle and underscores their potential importance in the breakdown of algal biomass, a major component of oceanic organic matter6. Moreover, laminarin degradation in deep-sea sediments is likely mediated by a diverse consortium, with PVC bacteria contributing alongside other specialized lineages such as Bacteroidota48. The relative contribution of PVC bacteria compared with these other lineages remains to be quantified. Laminarin facilitates the isolation of deep-sea PVC superphylum bacteria The ability of laminarin to selectively enrich Planctomycetota strain WC338 and Lentisphaerota strain WC36 from deep-sea cold seep sediments suggests that laminarin-based cultivation strategies might be broadly applicable to the isolation of other recalcitrant bacterial taxa across diverse deep-sea environments. To test this hypothesis, we enriched sediment samples from cold seeps, hydrothermal vents, and seamounts for one month in basal medium supplemented with laminarin. Amplicon sequencing revealed that laminarin amendment led to an enrichment of the PVC superphylum, specifically Lentisphaerota, Planctomycetota, and Verrucomicrobiota. In parallel, we observed a substantial decline in the relative abundance of Pseudomonadota and Actinomycetota, suggesting a niche displacement under laminarin-replete conditions (Fig. 7a). Furthermore, Fusobacteriota—a lineage recently recognized for its role in the degradation of complex organic matter49—were enriched in cold seep and seamount samples, suggesting that they may contribute indirectly through the utilization of metabolic intermediates generated by primary laminarin degraders, although a direct role in laminarin breakdown cannot be excluded. To date, only two Planctomycetota strains (Gimesia benthica E7 and Poriferisphaera heterotrophicis ZRK32)24,25 and one Lentisphaerota strain (Lentisphaera profundi SAORIC-696)50 have been successfully isolated and characterized from deep-sea habitats, highlighting the challenges associated with accessing and studying these deep-sea PVC superphylum bacteria. Given the limited number of cultured PVC bacteria from deep-sea environments, we focused on analyzing the number of PVC bacteria that were enriched. We found that many PVC bacteria were enriched from sediments using laminarin, with the most from cold seep samples, followed by hydrothermal vent and seamount (Fig. 7b), indicating that laminarin is a very effective material for enriching and isolating deep-sea PVC bacteria. Laminarin-supplemented enrichment cultures enabled the isolation of additional PVC strains from other deep-sea environments: Planctomycetota strain WC47 from a hydrothermal vent sediment and Lentisphaerota strain WC29 from a seamount sediment (Fig. 7c), complementing the strains WC338 and WC36 previously obtained from cold seeps. The successful isolation of diverse PVC superphylum members via laminarin enrichment supports previous metagenomic and metatranscriptomic evidence suggesting their metabolic capability to degrade laminarin. Together with the established role of Bacteroidota in laminarin degradation47,51,52, our findings suggest that laminarin turnover in deep-sea sediments is mediated by a diverse microbial consortium, with PVC bacteria contributing alongside other functionally specialized lineages. However, in the absence of quantitative rate measurements, their ecological contribution relative to other taxa, such as Bacteroidota, cannot be determined from the current data. Future studies using stable isotope probing or lineage-resolved enzyme assays will be needed to partition laminarin degradation rates among co-occurring microbial groups.Fig. 7Laminarin facilitates the isolation and cultivation of PVC superphylum bacteria from deep-sea sediments.a Relative abundance of bacterial phyla in deep-sea sediment communities after 30 days of incubation in basal medium (Control) or basal medium supplemented with 3.0 g/L laminarin (Laminarin). Data represent mean ± SD (n = 3, three biologically independent replicates) for cold seep, hydrothermal vent, and seamount samples. b Absolute abundance of PVC superphylum bacteria enriched from these sediments with laminarin as the primary carbon source. Data are presented as box plots showing the median (centre line), bounds of box (first and third quartiles), whiskers (minimum to maximum, excluding outliers), and all individual data points (n = 3). *P < 0.05, **P < 0.01, with one‑way ANOVA and Dunnett’s multiple-comparisons test. c Maximum-likelihood phylogenetic tree of Planctomycetota strain WC47 and Lentisphaerota strain WC29. Scale bar, 0.1 substitutions per nucleotide position. Source data are provided as a Source Data file. Laminarin facilitates the isolation and cultivation of PVC superphylum bacteria from deep-sea sediments. a Relative abundance of bacterial phyla in deep-sea sediment communities after 30 days of incubation in basal medium (Control) or basal medium supplemented with 3.0 g/L laminarin (Laminarin). Data represent mean ± SD (n = 3, three biologically independent replicates) for cold seep, hydrothermal vent, and seamount samples. b Absolute abundance of PVC superphylum bacteria enriched from these sediments with laminarin as the primary carbon source. Data are presented as box plots showing the median (centre line), bounds of box (first and third quartiles), whiskers (minimum to maximum, excluding outliers), and all individual data points (n = 3). *P < 0.05, **P < 0.01, with one‑way ANOVA and Dunnett’s multiple-comparisons test. c Maximum-likelihood phylogenetic tree of Planctomycetota strain WC47 and Lentisphaerota strain WC29. Scale bar, 0.1 substitutions per nucleotide position. Source data are provided as a Source Data file. Methods Sample collection Sediment samples were collected from seven deep-sea sites representing three distinct habitat types—cold seeps (Site F and Haima), hydrothermal vents (Sites H1 and H2), and a seamount (Zhenbei)—during two expeditions in 2018 and 2022 by RV KEXUE. These habitats were selected to capture variation in geochemical regimes and organic matter inputs: cold seeps are influenced by hydrocarbon-rich fluids, hydrothermal vents by high-temperature sulfur-rich fluids, and seamounts by enhanced currents and organic matter deposition. Specifically, deep-sea sediment samples were collected using the remotely operated vehicle (ROV) Faxian. Surface sediments were retrieved with push corers deployed by the ROV, with careful attention to preserving the integrity of the uppermost sediment layer. Only the top 1 cm of each sediment core was retained for downstream analyses. To minimize disturbance and prevent contamination from deeper strata, the push corers were inserted to a controlled depth and monitored to avoid core overfilling. This approach ensured the recovery of pristine surface sediment material suitable for subsequent geochemical and microbiological investigations. These sediment samples were immediately stored at -80 °C following rapid collection. Detailed information for each sample, including geographic coordinates, depth, and habitat type, is provided in Supplementary Data 1. Amplicon sequencing To understand the microbial community structure of these seven sediment samples, 16S rRNA gene sequencing analysis was performed as previously described53. Briefly, the total DNA was extracted from these four sediment samples using the E.Z.N.A.® soil DNA Kit (Omega Bio-tek, USA) according to the manufacturer’s instructions. The quality and concentration of DNA were determined by 1.0% agarose gel electrophoresis and a NanoDrop2000 spectrophotometer (Thermo Scientific, USA) and kept at -80 °C prior to further use. The V3-V4 region of the 16S rRNA gene was amplified from total DNA using the forward primer 338 F (5’-ACTCCTACGGGAGGCAGCAG-3’) and the reverse primer 806 R (5’-GGACTACHVGGGTWTCTAAT-3’). The library was prepared using the TruSeqTM DNA Sample Prep Kit (Illumina, USA) and sequenced on the Illumina Nextseq2000 platform (Illumina, USA) according to the standard protocols by Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China). After demultiplexing, the resulting sequences were quality filtered with FASTQ v0.19.654 and merged with FLASH v1.2.1155. Then the high-quality sequences were de-noised using the DADA2 plugin56 in the QIIME2 v2024.2 pipeline57 with recommended parameters, which obtains single-nucleotide resolution based on error profiles within samples. DADA2 denoised sequences are usually called amplicon sequence variants (ASVs). Taxonomic assignment of ASVs was performed using the naive Bayes consensus taxonomy classifier implemented in Qiime2 and the SILVA database v138.1. Metagenomic sequencing, assembly, and binning Total microbial DNA was extracted from deep-sea sediment samples using the E.Z.N.A.® Stool DNA Kit (Omega Bio-tek, Norcross, GA, USA) following the manufacturer’s instructions. DNA integrity was assessed by 1% (w/v) agarose gel electrophoresis, and concentration was determined using a Qubit® 3.0 fluorometer (Invitrogen, Carlsbad, CA, USA). For each sample, 1.0 μg of genomic DNA was sheared using a Covaris S220 focused-ultrasonicator (Woburn, MA, USA) to generate fragment sizes of approximately 450 bp. Sequencing libraries were constructed using the NEB Next® Ultra™ DNA Library Prep Kit for Illumina (New England BioLabs, Ipswich, MA, USA) according to the manufacturer’s protocol. Paired-end sequencing (2 × 150 bp) was performed on the Illumina HiSeq X platform (Illumina, San Diego, CA, USA). To ensure consistency in data processing, all raw sequencing reads were analyzed using a standardized bioinformatics pipeline established by Shanghai Biozeron Biological Technology Co., Ltd. (Shanghai, China). Raw sequence reads underwent quality trimming was performed by SOAPnuke v1.5.6 (parameters: -l 20 -q 0.2 -n 0.05 -Q 2 -d -c 0 -5 0 -7 1)58 and the clean data were assembled using MEGAHIT v1.2.9 (parameters:--min-count 2 --k-min 33 --k-max 83 --k-step 10)59. Genome binning was performed using three complementary algorithms. Coverage information for each sample was generated by mapping clean reads to the assembled contigs using Bowtie2 v2.3.5 with default settings. The resulting coverage profiles were used as input for binning with MaxBin2 v2.2.7 (-markerset 40)60, MetaBAT2 v2.12.1 (-m 1500 and --unbinned parameters)61 and Concoct v0.4.0 (default parameters)62. The initial bin sets were refined and consolidated using MetaWRAP v1.3.263 with default settings to generate the final metagenome-assembled genomes (MAGs). The completeness and contamination of each MAG were assessed using CheckM v1.0.1864 with the lineage-specific workflow. Only bins with estimated completeness >50% and contamination <10% were retained for downstream analysis. Taxonomic assignments of MAGs Genome taxonomy of MAGs was assigned using GTDB-Tk v2.1.065. The concatenated alignments of 120 bacterial or 122 archaeal marker genes from 238 MAGs were generated via GTDB-Tk’s identify and align workflow. IQ-TREE v2.2.066 was used to construct a maximum-likelihood tree. The resultant tree was visualized with the Interactive Tree of Life (iTOL) v5.0 online tool67. Functional annotations To characterize the functional potential of all MAGs from deep-sea sediment samples, we performed functional annotation using METABOLIC v4.068. Predicted coding sequences were pooled and clustered at 95% nucleotide sequence similarity using CD-HIT v4.8.169 (parameters: -c 0.95 -T 0 -M 0 -G 0 -aS 0.9 -g 1 -r 1 -d 0). Gene abundance was quantified by mapping metagenomic reads to this catalog using Salmon v1.5.070 in the mapping-based mode (parameters: -validateMappings -meta), with values expressed as genes per million (GPM). We also employed DRAM71 with default parameters to annotate MAGs against the KOfam, MEROPS, and dbCAN databases, identifying CAZymes, peptidases, lipases, nucleases, transporters, and other proteins. Metatranscriptomic analysis To investigate the in situ transcriptional activity of carbohydrate-active enzyme (CAZyme)-encoding genes within MAGs, we performed metatranscriptomic sequencing on four sediment samples (2022 Site F cold seep 1, 2022 Site F cold seep 2, 2022 Haima cold seep, and 2022 Zhenbei seamount). For each site, three biological replicate samples were collected and processed independently. Metatranscriptomic library preparation and sequencing were conducted at Shanghai Biozeron Biological Technology Co., Ltd. (Shanghai, China). Total RNA was extracted from sediment samples using TRIzol® Reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol, followed by DNase I digestion (TaKaRa, Kusatsu, Japan) to remove residual genomic DNA. RNA integrity and concentration were assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and a NanoDrop ND-2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), respectively. Only high-quality RNA samples meeting the following criteria were used for library construction: OD260/280 ratio between 1.8 and 2.2, OD260/230 ratio ≥ 2.0, RNA integrity number (RIN) ≥ 6.5, 28S:18S rRNA ratio ≥ 1.0, and total RNA yield > 10 μg. For each sample, 5 μg of total RNA was used as input for library preparation with the TruSeq™ Stranded Total RNA Sample Preparation Kit (Illumina, San Diego, CA, USA) following the manufacturer’s instructions. Ribosomal RNA was depleted using the Ribo-Zero™ rRNA Removal Kit (Illumina, San Diego, CA, USA). After fragmentation, first- and second-strand cDNA synthesis was performed, followed by end repair, A-tailing, and ligation of Illumina-indexed adapters. cDNA fragments of 200–300 bp were size-selected on 2% Low Range Ultra Agarose gels and subsequently amplified by 15 cycles of PCR using Phusion DNA Polymerase (New England BioLabs, Ipswich, MA, USA). The final libraries were quantified and sequenced on the Illumina HiSeq 2500 platform (Illumina, San Diego, CA, USA) in paired-end mode (2 × 150 bp) with an average insert size of approximately 450 bp. Raw metatranscriptomic reads were processed using the same quality control pipeline applied to the metagenomic datasets. Adapter trimming and quality filtering were performed using fastp v0.19.6 with default parameters. High-quality clean reads were then quantified against the predicted protein-coding gene sequences derived from all MAGs using Salmon v1.5.070 in mapping-based mode with the -validateMappings and -meta flags enabled. Transcript levels of functional genes encoding CAZymes were normalized to transcripts per million (TPM). Detailed information on the 10 selected marker genes is provided in Supplementary Data 14. Enrichment and cultivation of deep-sea PVC bacteria To obtain PVC bacterial strains from deep-sea sediments, a sample collected from the Site F cold seep was added to 15 mL anaerobic tubes containing 10 mL of basal medium supplemented with 3.0 g/L laminarin. The basal medium consisted of 1.0 g/L NH4Cl, 1.0 g/L CH3COONa, 1.0 g/L NaHCO3, 0.5 g/L MgSO4, 500 µL/L of 0.1% (w/v) resazurin, 0.7 g/L cysteine, 1.0 L seawater, and adjusted to pH 7.0. The medium was prepared under an N2 gas phase and sterilized by autoclaving at 115 °C for 30 minutes. Sediment inoculum was introduced into 15-mL anaerobic Hungate tubes containing 10 mL of the laminarin-supplemented medium. Cultures were incubated at 28 °C for one month under anoxic conditions. To promote colony formation and isolation, the medium was solidified by the addition of 15 g/L agar. The agar medium was evenly distributed as a thin layer on the inner surface of Hungate tubes to prepare roll tubes. After the initial enrichment period, 100 µL of the enriched culture was transferred anaerobically into roll tubes, spread onto the agar surface, and incubated at 28 °C for five days. Individual colonies were picked under anaerobic conditions using sterile bamboo sticks and transferred to fresh liquid medium (10 mL in 15-mL Hungate tubes) for further propagation. This single-colony isolation procedure was repeated multiple times until pure cultures were obtained. Putative PVC bacterial strains were identified by PCR amplification and sequencing of the 16S rRNA gene. Culture purity was routinely monitored by transmission electron microscopy (TEM) and confirmed by repeated partial 16S rRNA gene sequencing. TEM observation For morphological observation of strains WC338 and WC36, cell pellets were obtained by centrifuging cultures grown in basal medium with 3.0 g/L laminarin (5000 × g, 10 min). Cell suspensions were adsorbed onto copper grids for 20 min, washed with 10 mM PBS (pH 7.4), and air-dried. Transmission electron micrographs were acquired using a Hitachi HT7700 with the JEOL JEM 12000 EX at 100 kV. Phylogenetic analysis Complete genome sequences for strains WC338 and WC36 were determined in this study, while comparative genomes were retrieved from the Genome Taxonomy Database (GTDB). A genome-based phylogenetic tree was reconstructed using the GTDB Toolkit (GTDB v214, https://github.com/Ecogenomics/GTDBTk). Full-length 16S rRNA gene sequences from strains WC338, WC36, WC47, WC29 (obtained from their genomes), and related taxa (from NCBI) were aligned using the secondary-structure-aware aligner (SINA, v1.2.12)72. Phylogenetic trees were constructed using the maximum likelihood method via the W-IQ-TREE web server (http://iqtree.cibiv.univie.ac.at)73 with automatic model selection. The K3P + I + G4 model was selected as the best fit based on the Bayesian Information Criterion (BIC). All the trees were edited and visualized with the Interactive Tree of Life (iTOL) v5.0 online tool67. Growth assay To assess laminarin’s effect on strains WC338 and WC36, 100 µL of fresh cultures were inoculated into 10 mL basal medium with or without 3.0 g/L laminarin in triplicate Hungate tubes. Cultures were incubated anaerobically at 28 °C. Growth was monitored daily by measuring OD600 using a microplate reader (Tecan Infinite M1000 Pro, Switzerland) until the stationary phase was reached. Determination of laminarin content Total laminarin content was determined using a modified phenol-sulfuric acid colorimetric assay74. Briefly, a standard curve was generated using known concentrations of glucose (0-100 µg/mL) as the standard. Subsequently, aliquots (2 mL) of culture supernatant were collected daily from the initial inoculation up to day 9 of incubation. Each experimental condition was performed with three biological replicates. For each sample (both glucose standards and culture supernatants), a 2 mL sample was combined with 1 mL of 5% phenol, and 5 mL of concentrated sulfuric acid was rapidly introduced. The mixture was vortexed for 5 s, incubated at 37 °C for 30 min, and its absorbance was measured at 490 nm after cooling to room temperature. The laminarin content was calculated based on the absorbance values and the established standard curve. Genome sequencing, assembly, and annotation Deep-sea strains WC338 and WC36 were cultivated anaerobically at 28 °C for 5 d in basal medium supplemented with 3.0 g/L laminarin. Cells were harvested by centrifugation, and genomic DNA was extracted using the PowerSoil DNA Isolation Kit (Mo Bio Laboratories, Carlsbad, CA, USA) according to the manufacturer’s protocol. DNA purity, integrity, and concentration were assessed using three complementary approaches: (1) agarose gel electrophoresis (1% w/v) to visualize degradation or contamination; (2) spectrophotometric analysis (NanoDrop, Thermo Fisher Scientific, Waltham, MA, USA) to evaluate OD260/280 ratios; and (3) fluorometric quantification using the Qubit® dsDNA Assay Kit with a Qubit® 3.0 fluorometer (Invitrogen, Carlsbad, USA). Whole-genome sequencing was performed at Novogene Bioinformatics Technology Co., Ltd. (Beijing, China) using a hybrid strategy combining short- and long-read technologies. Short-insert paired-end libraries (350 bp insert size) were constructed using the NEB Next® Ultra™ DNA Library Prep Kit for Illumina (New England BioLabs, Ipswich, MA, USA) and sequenced on the Illumina NovaSeq 6000 platform (PE150). For long-read sequencing, native genomic DNA was processed without amplification and sequenced on the Oxford Nanopore PromethION platform. Base calling of raw nanopore signals (fast5) was performed using Guppy v3.1.5, and resulting fastq files were demultiplexed by barcode. Read quality metrics were evaluated using NanoPlot v1.29.1. High-quality short and long reads were co-assembled using Unicycler v0.4.8 with a hybrid assembly approach. For each strain, the assembly converged to a single circular contig, which was manually validated and circularized by trimming terminal overlaps. Gene prediction and functional annotation were carried out by aligning predicted protein sequences against multiple public databases using BLAST (E-value ≤ 1×10⁻⁵, minimum alignment coverage ≥ 40%). The following databases were used: Pfam (release 33.1), GO (release 2022-01)75, KEGG (release 104.1)76, COG (release 2021-11)77, NCBI NR (release 2022-03), TCDB (release 2021-01), and Swiss-Prot (release 2021-05)78. Genomic characteristics The genomic relatedness between strains WC338 and WC36 was assessed using multiple complementary metrics, including Average Nucleotide Identity based on MUMMER alignment (ANIm), Average Nucleotide Identity based on BLASTN (ANIb), and tetranucleotide (Tetra) frequency analysis. All calculations were performed using the JSpecies Web Services platform (http://jspecies.ribohost.com/jspeciesws/). Species boundaries were defined using established thresholds: ≥95% for both ANIb and ANIm, and ≥0.99 for Tetra signatures. Transcriptomic analysis Strains WC338 and WC36 were cultured in basal medium supplemented with either 0.05 g/L or 3.0 g/L laminarin for 5 days to evaluate transcriptional effects. Following incubation, biological triplicates from each condition were immediately harvested and submitted to Novogene (Tianjin, China) for RNA sequencing, employing protocols referenced previously29,36. The sequencing utilized an Illumina NovaSeq 6000 platform, generating paired-end reads of 150 bp length (PE150). Initial processing involved adapter removal and stringent quality filtering of raw reads using Trimmomatic79. Subsequent steps focused on rRNA and tRNA removal using SortMeRNA80 at default settings, ensuring only mRNA sequences were retained. Quantification was performed by mapping the filtered reads to the reference genome using the Burrows-Wheeler Aligner (BWA, v.0.7.17-r1188)81. Resulting alignments were converted to BAM format via SAMtools (v.1.13), and read distribution across genomic features was determined using BEDTools (v.2.30.0)82. Gene expression levels were quantified as Fragments Per Kilobase Million (FPKM)83, a standard metric normalizing counts by both gene length and total sequencing depth. Differential expression analysis across the three experimental conditions (with three biological replicates each) was conducted using the DESeq R package (v.1.20.0)84. This statistical framework employs a negative binomial distribution model suitable for digital gene expression data. To control the false discovery rate, P-values were adjusted using Benjamini and Hochberg’s method. Significant differential expression was defined by an adjusted P-value < 0.05 and a log2 (Fold change) threshold of 1. Amplicon analysis of enrichment cultures supplemented with laminarin To enrich for deep-sea PVC bacteria, we established enrichment cultures using a basal medium as a control and the basal medium supplemented with 3.0 g/L laminarin. Specifically, we selected surface sediment samples from cold seep, hydrothermal vent, and seamount environments. These samples were cultured for one month at 28 °C in these two distinct enrichment media. Each enrichment condition was performed in triplicate biological replicates. To elucidate the bacterial community structure within these enrichment cultures, 2 mL of each culture was harvested and subjected to amplicon sequencing analysis by Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China). The detailed procedures were described above. Statistical analysis All data were presented as means ± SD. Statistical analysis was performed using OriginPro 2025. For comparing each experimental group to the control group, one-way ANOVA was performed, followed by Dunnett’s multiple-comparisons test. Differences of P < 0.05 were considered statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). Differences of P > 0.05 were considered not significant (ns). Supplementary information Supplementary information Description of Additional Supplementary Files Supplementary Data 1- 17 Transparent Peer Review file Supplementary information Description of Additional Supplementary Files Supplementary Data 1- 17 Transparent Peer Review file Source data Source Data Source Data