The deep-sea samples were collected by the RV KEXUE from a typical cold seep at the Formosa Ridge (Site F) in the South China Sea (E119°17’07.322”, N22°06’58.598”) as described by Cao et al. (2021). The relevant chemical parameters of the cold seep are shown in Supplementary Table 1 . The temperature, salinity, and underwater depth of sampling sites were recorded in real time by an SBE 25plus Sealogger CTD (SBE, USA). Concentrations of CO₂ and CH₄ in surface sediments were measured in situ with CONTROSHydroCO₂ (CONTROS, Norway) and HydroCH₄ (CONTROS, Norway) sensors, respectively. All sensors were mounted on the Discovery ROV (Cao et al., 2021). Dissolved oxygen was measured above the vent site and the measurement method of dissolved oxygen can refer to the article of Cao et al. (2021) and more details can be found in Materials and Methods. Specific site information is shown in Supplementary Table 2 . To enrich the polysaccharide degrading bacteria, the sediment needs to be suspended with seawater sterilized three times to ensure that the sediment is not contaminated. A temperature gradient is then set before the culture is carried out, with the aim of slowly reviving the microbes in the sediment, which is very important for the whole isolation and culture process. Later, the sediment samples were cultured at 28°C for 6 months in an anaerobic inorganic medium supplemented with 1.0 g/L of mixed seaweed polysaccharides, namely, porphyra polysaccharide (PP), agar, carrageenan, laminarin, fucoidan, starch, alginate, ulvan, and enteromorpha polysaccharide (EP). The compositions of inorganic medium are 1.0 g NH₄Cl, 1.0 g NaHCO₃, 1.0 g CH₃COONa, 0.5 g KH₂PO₄, 0.2 g MgSO₄, 0.7 g cysteine hydrochloride, 500 µl 0.1% (w/v) resazurin in 1 L filtered seawater, pH 7.0. This medium was prepared anaerobically as previously described (Fardeau et al., 1997). After the 6-month enrichment, 50 µl dilution portions were spread on Hungate tubes covered by a modified organic medium (1.0 g NH₄Cl, 1.0 g NaHCO₃, 1.0 g CH₃COONa, 0.5 g KH₂PO₄, 0.2 g MgSO₄, 1.0 g yeast extract, 1.0 g peptone, 0.7 g cysteine hydrochloride, 500 µl 0.1% (w/v) resazurin in 1 L filtered seawater, pH 7.0) supplemented with 15.0 g/L agar. The Hungate tubes were anaerobically incubated at 28°C for 7 days. Individual colonies with distinct morphology were selected using sterilized bamboo sticks and then cultured in the anaerobic liquid 2216E medium (5.0 g/L yeast extract, 1.0 g/L peptone, 0.7 g cysteine hydrochloride, 500 µl 0.1% (w/v) resazurin, 1 L filtered seawater, pH 7.0). Strain zth2 was isolated and purified with 2216E medium by repeated use of the Hungate roll-tube methods for several rounds until it was considered to be pure. The purity of strain zth2 was conventionally confirmed repeated partial sequencing of the 16S rRNA gene. Strain zth2 is preserved at −80°C in 2216E medium supplemented with 20% (v/v) glycerol.

To observe the morphological characteristics of strain zth2, its cell pellets were washed with 10 mM phosphate buffer solution (PBS, pH 7.4) and then centrifuged at 5,000 g for 5 min, and taken by immersing copper grids coated with a carbon film for 20 min, washed for 10 min in PBS and dried for 20 min at room temperature (Buchan et al., 2014). Ultrathin-section electron microscopic observation was performed as described previously (Sekiguchi et al., 2003; Graham and Orenstein, 2007). Briefly, the samples were firstly preserved in 2.5% (v/v) glutaraldehyde overnight at 4°C and then washed three times using PBS in the next day. Later, samples were dehydrated in ethanol solutions of 30, 50, 70, 90, and 100% for 10 min each time, and then the dehydrated samples were embedded in a plastic resin. Ultrathin sections (50–70 nm) of cells were prepared with an ultramicrotome (Leica EM UC7, Germany), stained with uranyl acetate and lead citrate. All samples were examined using TEM (HT7700, Hitachi, Japan) with a JEOL JEM 12000 EX (equipped with a field emission gun) at 100 kV.

The temperature, pH, and NaCl concentration ranges for the growth of strain zth2 were determined in the anaerobic liquid 2216E medium incubated for 5 days. Growth assays were performed at different temperatures (4, 28, 30, 37, and 60°C). The pH range for growth was tested from pH 2.0 to 10.0 with increments of 1 pH unit. Salt tolerance was tested with 0–10% (w/v) NaCl (1.0% intervals) in 2216E medium prepared with sterile water instead of sea water. Substrates utilization of strain zth2 was tested in the medium consisting of (L⁻¹): 1.0 g NH₄Cl, 1.0 g NaHCO₃, 1.0 g CH₃COONa, 0.5 g KH₂PO₄, 0.2 g MgSO₄, 0.7 g cysteine hydrochloride, 500 µl 0.1% (w/v) resazurin in 1 L filtered seawater, pH 7.0. Single substrate (namely, glucose, mannitol, laminarin, fructose, dextrin, fucoidan, maltose, mannose, glycine, pullulan, lactose, polyethylene glycol, butyrate, sucrose, acetate, formate, starch, trehalose, galactose, cellulose, xylose, lactate, ethanol, D-mannose, glycerin, phosphate, and sorbitol) was added from sterile filtered stock solutions to the final concentration at 1.0 g/L, respectively. Cell culture containing basic medium without adding any other substrates was used as a control. For each substrate, three biological replicates were performed.

According to the instructions of the manufacturer, genomic DNA of strain zth2 was extracted using a QIAGEN Genomic-tip (QIAGEN Biotech Co., Ltd., Germany). For genome sequencing, 15 ml strain zth2 culture was inoculated in a 2 L anaerobic bottle containing 1.5 L anaerobic 2216E medium. After cell growth reached the logarithmic phase, the culture solution was centrifuged at 8,000 g for 10 min to obtain cell pellets.

Its cell pellets were washed with 10 mM phosphate buffer solution (PBS, pH 7.4) and then centrifuged at 8,000 g for 5 min. Then the supernatant was poured out and the cell pellets were preserved at −80°C for sequencing. Quality inspection of DNA purity, concentration, and integrity was performed by NanoDrop (Thermo Scientific, USA), and 0.35% agarose gel electrophoresis, respectively. Then, a genomic DNA library for Nanopore sequencing was constructed by using a ligation sequencing kit (Oxford Nanopore Technologies Ltd., UK). The library was sequenced using a Promethion sequencer 48. Canu v1.5 (Koren et al., 2017) was used to assemble the filtered subreads and Pilon v1.22 (Walker et al., 2014) was used to correct the assembly sequence with higher accuracy. Sequencing work was conducted by Biomarker Technologies Co., Ltd. (Beijing, China).

Genome related values are calculated by several methods: Average Nucleotide Identity (ANI) based on the MUMMER ultra-rapid aligning tool (ANIm), ANI based on the BLASTN algorithm (ANIb), the tetra-nucleotide signatures (Tetra), and in silico DNA–DNA similarity. ANIm, ANIb, and Tetra frequencies were calculated using JSpecies WS (http://jspecies.ribohost.com/jspeciesws/). The recommended species criterion cut-offs for the ANIb and ANIm were respectively used as 95% and 0.99 for the Tetra signature. The calculation of in silico DNA–DNA similarity values was performed by the Genome-to-Genome Distance Calculator (GGDC) (http://ggdc.dsmz.de/) (Meier-Kolthoff et al., 2013). The isDDH results were calculated based on the recommended formula 2.

The 16S rRNA gene based tree was constructed with the full-length 16S rRNA sequences by the maximum likelihood method. In short, the full-length 16S rRNA (1,361 bp) gene sequence of strain zth2 was obtained from its whole genome (accession number MW729759), and other 16S rRNA sequences of related taxa used for phylogenetic analysis were obtained from NCBI (www.ncbi.nlm.nih.gov/). The multiple sequence alignment of sequences was performed by MAFFT ver.7 (https://mafft.cbrc.jp/alignment/server/). The method of Gappyout was used to trim sequences. Phylogenetic trees were constructed using W-IQ-TREE web server (http://iqtree.cibiv.univie.ac.at/) (Trifinopoulos et al., 2016) with GTR + F + I + G4 model. The edition of the phylogenetic tree was performed by the online tool Interactive Tree of Life (iTOL v4) (Letunic and Bork, 2019).

The genome tree was constructed from a concatenated alignment of 37 protein-coding genes (Wu et al., 2013) that extracted from each genome by Phylosift (v1.0.1) (Darling et al., 2014) with automated setting. The genomes used to construct the genome tree included both draft and finished genomes from NCBI. Aligned sequences were trimmed using TrimAl (version 1.2) with gappyout function. The phylogenetic trees were constructed using W-IQ-TREE web server (http://iqtree.cibiv.univie.ac.at) (Trifinopoulos et al., 2016) with GTR + F + I + G4 model, and the online tool Interactive Tree of Life (iTOL v6) was used for the tree edition (Letunic and Bork, 2019).

Growth assays of strain zth2 were performed at atmospheric pressure. To detect the effect of polysaccharides on the growth of strain zth2, the modified 1/5 2216E medium (containing 1 g/L yeast extract, 0.2 g/L peptone, 0.7 g cysteine hydrochloride, 500 µl 0.1% (w/v) resazurin, 1 L seawater, pH 7.0) supplemented either with or without 1 g/L of different polysaccharides, namely, PP (porphyra polysaccharide), carrageenan, laminarin, starch or ulvan. For all the growth assays, 100 µl strain zth2 culture was inoculated in 15 ml Hungate tubes containing 10 ml medium. All of the Hungate tubes were anaerobically incubated at 28°C. Bacterial growth status was monitored by measuring the OD₆₀₀ value via a microplate reader (Infinite M1000 Pro; Tecan, Mannedorf, Switzerland) every day until cell growth reached the stationary phase.

To explore the effects of phosphate and iron on bacterial growth, strain zth2 was grown in basic medium (containing 10 g/L glucose, 0.3 g/L NaCl, 0.5 g/L (NH₄)₂SO₄, 0.3 g/L MgSO₄, 0.2 g/L KCl, 0.03 g/L MnSO₄, 0.2 g/L peptone, 0.7 g cysteine hydrochloride, 500 µl 0.1% (w/v) resazurin, 1 L seawater, pH 6.0–6.5) supplemented with Na₃PO₄ (2.6 mM), or FeSO₄ (0.03 g/L), or both FeSO₄ (0.03 g/L) and Ca₃(PO₄)₂(1.0 g/L), or both FeSO₄ (0.03 g/L) and Na₃PO₄ (2.6 mM). Similarly, the growth assays were performed as described above.

Transcriptomic analysis was performed by Novogene (Tianjin, China). For the transcriptomic analysis of the effects of polysaccharides on the growth of strain zth2, strain zth2 was cultured in the above modified 1/5 2216E medium supplemented either with or without 1 g/L of polysaccharides (laminarin or starch) in 2 L anaerobic bottles for 4 days. For the transcriptomic analysis of the effects of phosphate and iron on the growth of strain zth2, cells suspension of strain zth2 was cultured in the above basic medium supplemented with Na₃PO₄ (2.6 mM), or FeSO₄ (0.03 g/L), or both FeSO₄ (0.03 g/L) and Na₃PO₄ (2.6 mM) in 2 L anaerobic bottles for 5 days. After incubation, the cells were collected by centrifugation at 8,000 g, 4°C, 10 min. For transcriptomic analyses, total RNAs of strain zth2 were extracted using TRIzol reagent (Invitrogen, USA) and DNA contamination was removed using the MEGA clear™Kit (Life technologies, USA). Detailed protocols of the following procedures, namely, library preparation, clustering and sequencing and data analyses were described in the Supplementary Information .

To better understand the ecological role of strain zth2, we mimicked the ocean environment in the anaerobic bottles. Briefly, dialysis bags (containing 50 ml sterilized seawater, named SW; 50 ml sterilized seawater with strain zth2 cells, named SWL; 50 ml sterilized seawater with 1 g/L of laminarin, named SWP; 50 ml sterilized seawater with 1 g/L of laminarin and strain zth2 cells, named SWLP) were respectively put into four anaerobic bottles filled with filtered sea water using medium speed double circle qualitative filter paper from the First Sea Bathing Beach, Qingdao, China. After a month of cultivation, bacterial cells grown outside of the dialysis bags were respectively collected and used for OTUs sequencing that was performed by Novogene (Tianjin, China). Total DNAs from the above four samples were extracted by the CTAB/SDS method (Murray and Thompson, 1980) and were diluted to 1 ng/µl with sterile water and used for PCR template. 16S rRNA genes of distinct regions (16S V3/V4) were amplified using specific primers (341F: 5′-CCTAYGGGRBGCASCAG and 806R: 5′-GGACTACNNGGGTATCTAAT) with the barcode. The PCR products were purified with a Gel Extraction Kit (Qiagen, Germany) and prepared to construct libraries. Sequencing libraries were generated using a TruSeq^® DNA PCR-Free Sample Preparation Kit (Illumina, USA) following the instructions of the manufacturer. The library quality was assessed on the Qubit@ 2.0 Fluorometer (Thermo Scientific, USA) and Agilent Bioanalyzer 2100 system (Agilent, USA). The library was sequenced on an Illumina NovaSeq platform. Paired-end reads (250 bp) were merged using FLASH (V1.2.7, http://ccb.jhu.edu/software/FLASH/) (Magoc and Salzberg, 2011). Quality filtering on the raw tags was performed according to the QIIME (V1.9.1, http://qiime.org/scripts/split_libraries_fastq.html) quality controlled process (Bokulich et al., 2013).

The tags were compared with the reference database (Silva database, release 111-2012_07, https://www.arbsilva.de/) using UCHIME algorithm (UCHIME Algorithm, http://www.drive5.com/usearch/manual/uchime_algo.html) (Edgar et al., 2011) to detect chimera sequences, and then the chimera sequences were removed (Haas et al., 2011). Sequence analyses were performed by Uparse software (Uparse v7.0.1001, http://drive5.com/uparse/) (Edgar, 2013). Sequences with ≥97% similarity were assigned to the same OTUs. The representative sequence for each OTU was screened for further annotation. For each representative sequence, the Silva Database (release 111-2012_07, http://www.arb-silva.de/) (Quast et al., 2013) was used based on Mothur algorithm to annotate taxonomic information

In our previous work, we successfully isolated a novel Bacteroidetes species through a polysaccharide degradation-driven strategy from the deep-sea cold seep (Zheng et al., 2021a). Beyond our expectation, we found there was a small amount of Lentisphaerae bacteria within the enrichment based on the OTUs analysis. Given that members of Lentisphaerae are typical difficult-to-culture microbes (Choi et al., 2015), we spent much effort for a couple of rounds of purification and 16S rRNA sequencing confirmation and eventually obtained the pure culture of a Lentisphaerae bacterium (strain zth2) ( Figure 1A ). Under the TEM observation, strain zth2 possessed an irregular coccoid cell (~400–500 nm in size, no flagellum) with the extracellular slime layer ( Figure 1B ), which was very similar to the morphologies of other Lentisphaerae members reported previously (Zoetendal et al., 2003; Cho et al., 2004).

To gain more insights into strain zth2, the whole genome was sequenced and analyzed. The genome size of strain zth2 is 3,198,720 bp with a DNA G+C content of 43.77% and it is complete ( Supplementary Figure S1A and Supplementary Table 3 ). Annotation of the genome of strain zth2 revealed 2,683 predicted genes that included 66 RNA genes (6 rRNA genes, 52 tRNA genes and 8 ncRNAs). The genome contains 146 CAZymes that may be related to its capability of polysaccharides degradation ( Supplementary Figure S1B ). The genome relatedness values were calculated by the average nucleotide identity (ANI), in silico DNA–DNA similarity (isDDH) and the tetranucleotide signatures (Tetra), against two available genomes of type strains of Lentisphaerae bacteria: V. vadensis ATCC BAA-548^T (Zoetendal et al., 2003) and L. araneosa HTCC2155^T (Cho et al., 2004) ( Supplementary Table 3 ). The average nucleotide identities (ANIb) of strain zth2 with BAA-548^T and HTCC2155^T were 65.32 and 65.45%, respectively. The average nucleotide identities (ANIm) of strain zth2 with BAA-548^T and HTCC2155^T were 89.31 and 83.99%, respectively. The tetra values of strain zth2 with BAA-548^T and HTCC2155^T were 0.69176 and 0.48857, respectively. Based on digital DNA–DNA hybridization employing the Genome-to-Genome Distance Calculator GGDC, the in silico DDH estimates for strain zth2 with BAA-548^T and HTCC2155^T were 26.1 and 25.5%, respectively. These results together demonstrated the genome of strain zth2 to be clearly below established ‘cut-off’ values (ANIb: 95%, ANIm: 95%, AAI: 95%, isDDH: 70%, Tetra: 0.99) (Goris et al., 2007; Meier-Kolthoff et al., 2013) for defining bacterial species ( Supplementary Table 3 ), suggesting it represents a novel taxon of Lentisphaerae.

To clarify the taxonomic status of strain zth2, we further performed the phylogenetic analysis with 16S rRNA genes of some cultured and uncultured representatives of PVC superphylum. The maximum likelihood tree of 16S rRNA placed strain zth2 as a distinct cluster separating from other three identified families, namely, Victivallaceae (from order Victivallales), Lentisphaeraceae (from order Lentisphaerales) and Oligosphaeraceae (from order Oligosphaerales) ( Figure 1C ). Consistently, a genome-based tree ( Supplementary Figure S2 ) further confirmed the position of strain zth2, located between the branches of families of Victivallaceae and Lentisphaeraceae. Moreover, based on the 16S rRNA gene sequence of strain zth2, a sequence similarity calculation using the NCBI server indicated that the closest relatives of strain zth2 were V. vadensis ATCC BAA-548^T (88.77%, family Victivallaceae), L. araneosa HTCC2155^T (86.09%, Lentisphaeraceae), L. marina IMCC11369^T (85.73%, Lentisphaeraceae), L. profundi SAORIC-696^T (85.72%, Lentisphaeraceae), and O. ethanolica 8KG-4^T (83.87%, Oligosphaeraceae). Recently, the proposed median and minimum sequence identity values of 16S rRNA sequence for a new family have been revealed as 92.25 and 87.65% (Yarza et al., 2014), respectively. Therefore, combining the results of 16s rRNA phylogenetic tree, a genome-based tree and ANI analysis, we propose that strain zth2 is classified as the type strain of a novel family. Given its isolated site, we proposed the name Candidatus Coldseepensis marina gen. nov., sp. nov. for strain zth2. In addition, we also proposed the associated genus and family as Candidatus Coldseepensis gen. nov. and Candidatus Anaerobicaceae fam. nov., respectively.

As previously reported, members of the PVC superphylum have very complex and specific cytoarchitectural features (Rivas-Marin and Devos, 2018). Ultrathin TEM sections showed that strain zth2 possessed distinctive membrane and compartmentalization systems ( Figures 2A, B ). In particular, its outer membrane possessed a lot of twists and turns ( Figure 2A ), which effectively increased the surface area of the cell membrane. Notably, strain zth2 divided by budding ( Figure 2A ) but not the usual binary fission manner that adopted by most bacteria. In the division process, daughter cells always originated from one site of the mother cell ( Figure 2A , panel I), and then the bud became bigger and bigger along with the cellular content distributing into the daughter cell ( Figure 2A , panels II to VI). Once the content and size of daughter cell were equal to those of mother cell, the daughter cell would separate from the mother cell ( Figure 2A , panels VII to VIII), and a cell division cycle was finished.

Another dramatic feature of strain zth2 was that there were several narrow protrusions around the cells ( Figure 2B , panels I–IV, the red arrow points). These protruded 80–114 nm from the cell and were 40–80 nm in diameter measured at the base of the structure preserved slender shape. They seemed to be coincided with the definition of prosthecae as cellular appendages or extensions of the cell containing cytoplasm (Staley, 1968).

Based on our enrichment and isolation strategy of strain zth2 ( Figure 1A ), this strain was proposed to possess a prominent capability of degrading polysaccharides. To test the potential polysaccharide substrate spectra of strain zth2, we checked the growth status of strain zth2 in the medium supplemented with different kinds of polysaccharides at a final concentration of 1 g/L. The results showed the supplementation of porphyra polysaccharide (PP), carrageenan and ulvan could only modestly promote the growth of strain zth2 ( Figure 3A ). In contrast, the supplement of starch or laminarin could significantly facilitate the growth of strain zth2, resulting in an almost 4-fold increased growth compared to the control group ( Figure 3A ).

To further validate the mechanism of polysaccharides degradation mediated by strain zth2, we performed transcriptomic analysis of strain zth2 grown in the medium supplemented with starch or laminarin. Indeed, the expression of many genes encoding glycan degradation was evidently upregulated when 1 g/L of starch was added in the medium. These included genes encoding amylase, glycosyl hydrolases (GH families 3, 9, 18, and 32), glycosyl transferases (GT1, GT2, and GT41), and glucanases ( Figure 3B ). Meanwhile, many proteins associated with ATP production (namely, ATP corrinoid adenosyltransferase, NAD/NADPH binding domain proteins, electron transfer flavoprotein, etc.) were also significantly upregulated in the presence of starch ( Figure 3C ), which might contribute to the growth promotion of strain zth2. Out of our expectation, the supplementation of starch could also dramatically upregulate the expression of many proteins responsible for phosphorus and iron metabolisms ( Figure 3D ). These included phosphoadenosine phosphosulfate reductase, PhoU, Fe–S binding protein and so on. Similarly, the supplement of 1 g/L of laminarin in the medium also markedly upregulated the expression of proteins responsible for glycan degradation ( Figure 3E ), ATP production ( Figure 3F ) and phosphorus and iron metabolisms ( Figure 3G ). Significantly, GH9 family proteins were annotated as laminarinase in CAZy database, which may play the key role in the degradation of laminaria.

Based on the above transcriptomic results of polysaccharide utilization, we speculate that the utilization of polysaccharide could promote phosphorus and iron metabolisms ( Figure 3D ). Therefore, we next sought to explore the effects of the supplementation of phosphorus and iron on the growth of strain zth2. First, we analyzed the growth status of strain zth2 in a basic medium supplemented with Na₃PO₄ (2.6 mM), or FeSO₄ (0.03 g/L), or both FeSO₄ (0.03 g/L) and Ca₃(PO₄)₂ (1 g/L), or both FeSO₄ (0.03 g/L) and Na₃PO₄ (2.6 mM). The growth curves showed that strain zth2 could only grow to an OD₆₀₀ value about 0.06 in the basic medium amended with FeSO₄ after two weeks incubation ( Figure 4A ). In contrast, in the presence of FeSO₄, the growth of strain zth2 could be significantly promoted to OD₆₀₀ values about 0.20 and 0.68 (about 11-fold increase) when insoluble Ca₃(PO₄)₂ or soluble Na₃PO₄ was added into the basic medium ( Figure 4A ). These results indicated the soluble phosphate was necessary for the growth of strain zth2 and this bacterium could transform insoluble Ca₃(PO₄)₂ to soluble phosphate for utilization to some extent. On the other hand, in the presence of Na₃PO₄, the absence of FeSO₄ in the medium could decrease about a quarter of the growth rate (from OD₆₀₀ value 0.68 to 0.53) ( Figure 4A ), suggesting that iron is also a key factor for the growth of strain zth2.

We next analyzed the effects of Na₃PO₄ or FeSO₄ on the growth of strain zth2 through a transcriptomic approach. The results showed that, compared to the medium only amended with FeSO₄, the supplement of both Na₃PO₄ and FeSO₄ could upregulate the expression of proteins responsible for energy production, such as citrate synthase, malate dehydrogenase, GTPase, ATP synthase, FAD/NADH dependent enzymes and electron transfer flavoprotein ( Figure 4B ). This result was well consistent with the growth cure shown in Figure 4A . Strikingly, many proteins associated with iron metabolisms ( Figure 4C ) and polysaccharide degradation ( Figure 4D ) were also evidently upregulated when Na₃PO₄ was supplemented in the medium beside FeSO₄, suggesting the presence of soluble phosphate could facilitate iron metabolisms and polysaccharide degradation. On the other hand, the absence of FeSO₄ in the medium indeed weakened the metabolisms of phosphorus, iron, and polysaccharide and the eventual energy generation even though Na₃PO₄ was present in the medium ( Figure 4E ). This result is a good explanation that the absence of FeSO₄ in the medium could decrease a half of growth rate as shown in Figure 4A . In the combined growth assay and transcriptomics results, we conclude that the utilization of polysaccharide, phosphate, and iron by strain zth2 is correlative and mutually promoted, and these three nutrient sources are essential factors supporting the best growth of strain zth2.

Given its prominent capability of polysaccharide degradation and acquiring phosphorus and iron, we next sought to ask whether strain zth2 could provide extra nutrient to other bacteria and thereby promote microbial diversity as shown in the previous report of the ecological function of Bacteroides (Zheng et al., 2021a). We therefore mimicked the natural ecological environment for bacterial growth and survival in the laboratory ( Figure 5A ). Correspondingly, four anaerobic dialysis bags (namely, SW, SWL, SWP, and SWLP) were respectively incubated in different anaerobic bottles filled with filtered but non-sterilized sea water using medium speed double circle qualitative filter paper for one month ( Figure 5A ). After 30-day incubation, the cells growing in the non-sterilized sea water were collected and used for OTUs analysis to determine the variety and abundance of bacteria in each sample. The OTUs analysis results revealed that the bacterial diversity of samples SW and SWP was very low: at the phylum level, only 3–4 phyla including Proteobacteria, Bacteroides and Firmicutes (or Gracilibacteria) were identified in the sea water ( Figure 5B ). In contrast, the bacterial diversity of samples SWL and SWPL was significantly increased ( Figure 5B ). Especially, the bacterial variety was increased to 8–9 phyla in sample SWL, and many bacteria belonging to unidentified bacteria and different known phyla (namely, Tenericutes, Fusobacteria, Actinobacteria, and Verrucomicrobia) were enriched, indicating that strain zth2 is indeed able to increase the environmental bacteria diversity. Similarly, at the family level, the abundance and variety of novel members belonging to other phyla rather than Proteobacteria were obviously enhanced in both SWL and SWLP samples, further confirming the proposal that strain zth2 contributes to the increase and maintainance of microbial diversity regardless of the presence of polysaccharide.

In the combined data on the utilization of polysaccharide, phosphate and iron, and also the contribution to the microbial diversity conducted by strain zth2, we proposed a model representing the central metabolism and ecological function of strain zth2 ( Figure 6 ). In this model, different polysaccharides were degraded and then transported into the cells by special transporters; then the oligosaccharide could be further transformed into glucose (or other simple sugar) and enters into the Embden–Meyerhof–Parnas (EMP) glycolysis pathway, which thereby joining to the tricarboxylic acid cycle (TCA pathway) and producing energy; the degradation products can also promote the growth of other microorganisms; on the other hand, different kinds of phosphate could be effectively obtained and utilized as a nutrient source for phosphorus cycling and ATP production; meanwhile, iron could be acquired and transported into the cells to activate essential iron-associated enzymes (e.g., Fe–Fe hydrogenase and Fe–S protein) for various metabolic pathways and thereby generating energy. As mentioned above, the utilization of polysaccharide, phosphate, and iron by strain zth2 is correlative and jointly contributes to carbon, phosphorus, and iron metabolisms and cycling, which thereby supports bacterial growth and promotes the environmental microbial diversity.

Candidatus Anaerobicaceae (An.ae’ro.bica.ce.ae. N.L. fem. adj. Candidatus Coldseepensis type genus of the family; suff. -aceae, ending to denote a family; N.L. fem. pl. adj. Candidatus Anaerobicaceae the family of the genus Candidatus Coldseepensis.) The type genus is Candidatus Coldseepensis.

For the genus name Candidatus Coldseepensis, cold.seep’ensis. N.L. fem. n. Given its isolated site, we proposed this name. The type and only species of the genus is Candidatus Coldseepensis marina.

Candidatus Coldseepensis marina (ma.ri’na. L. fem. adj. marina of the sea, marine). Cells are coccoid, ~400–700 nm in size, and have no flagellum; strictly anaerobic; the temperature range for growth is 28–37°C with an optimum at 28°C ( Supplementary Figure S3A ); growing at pH values of 6.0–8.0 (optimum, pH 7.0) ( Supplementary Figure S3B ); growth occurs at NaCl concentrations between 2.0 and 4.0% with an optimum growth at 3% NaCl ( Supplementary Figure S3C ); growth is stimulated by using glucose, maltose, fructose, sucrose, mannose, starch, cellulose, laminarin, dextrin, lactose, galactose, xylose, mannitol or sorbitol as a sole carbon source ( Supplementary Table 4 ). The type strain, zth2^T, was isolated from the sediment of deep-sea cold seep, P.R. China.