The samples were collected using a TV grab sampler at several sites at a depth of 1,453–2,653 m in hydrothermal fields of the southwest Indian Ocean in February 2020 during the cruise of DaYang Hao (Table S1). The sediments, sulfides, and oxides were fractionated on a clean bench and subsample was further used to inoculate (1/10, w/v) a sterile liquid enrichment media at 25°C on board. The sulfides were composed primarily of hydrothermal chimney fragments, predominantly consisting of pyrite (FeS₂), chalcopyrite (CuFeS₂), and sphalerite (ZnS), among others. The oxide samples were dominated by oxides/hydroxides of Fe, Mn, and Si. Furthermore, the sediments were composed of calcareous ooze, with traces of sulfides, exhibiting a significant proportion of calcium carbonate, reaching up to 50% in volume.

Enrichment of copper-resistant bacteria was conducted in 100 mL of marine broth (MB) with 6 mM Cu (II) as copper sulfate (CuSO₄•5H₂O) in 250 mL Erlenmeyer flasks. Per 1,000 mL of MB (Becton, Dickinson) contained 5.0 g peptone, 1.0 g yeast extract, 0.1 g ferric citrate, 19.45 g NaCl, 5.9 g MgCl₂, 1.8 g CaCl₂, 0.55 g KCl, 0.16 g NaHCO₃, 34.0 mg SrCl₂, 22.0 mg H₃BO₃, 0.08 g KBr, 4.0 mg Na₂SiO₃, 2.4 mg NaF, 1.6 mg NH₄NO₃, 8.0 mg Na₂HPO₄, 3.24 g MgSO₄, and the solution was adjusted to pH 7.0. The enrichment culture was incubated at 28°C for 3–5 days on board and stored with glycerol at −20°C. In the lab, the subculture was incubated at 28°C for 24 h, with shaking (150 rpm). Cu (II)-resistant bacterial strains were purified by repeatedly streaking them on marine broth agar (MA) plates containing 6 mM Cu (II). These isolates were deposited in the Marine Culture Collection of China, MCCC¹ (Table 1).

DNA of each isolate was extracted by the Genomic DNA extraction kit (SBS Genetech Co., Ltd., Shanghai). Two universal primers corresponding to E. coli positions 27F (5’-AGATTTGATCMTG GCTCAG-3′) and 1492R (5’-TACGGYTACCTTGTTACGACTT-3′) were used for PCR amplification of the 16S ribosomal RNA of the isolates, and sequences information were obtained through Sanger sequencing. EzBioCloud² was used for homology searches on 16 s rDNA genes (Yoon et al., 2017a). Phylogenetic analysis was performed using MEGA X (64-bit) (Larkin et al., 2007) after multiple alignments of the sequence data with Clustal W. Distances were calculated using distance options according to Kimura’s two-parameter model and clustering was performed using the neighbor-joining (NJ) methods. Sequences of related taxa were obtained from the GenBank database and supported with bootstrap values based on 1,000 replications.

To identify potential EPS production by bacterial isolates, the Congo red agar (CRA) plate test was utilized, which contained 5.0 g/L peptone, 1.0 g/L yeast extract, 30.0 g/L NaCl, and 15.0 g/L agar, with pH adjusted to 7.0. After autoclaving (121°C, 15 min), 10.0 g/L of glucose and 0.8 g/L of Congo red indicator were added (Maalej et al., 2015). The purified strains were inoculated onto a CRA plate and incubated at 28°C for 48 h. Positive colonies were observed to turn red to black.

The selected strains for EPS production were inoculated into MB medium containing 1 mM copper (250 mg/L) and incubated at 28°C for 72 h in a shaker at 150 rpm. The MB medium without copper served as a control. The cells were centrifuged at 8000 rpm for 10 min, and the supernatant was mixed with two volumes of cold 95% ethanol, followed by overnight incubation at 4°C. Most EPS was precipitated by centrifugation at 8000 rpm for 20 min at 4°C. It was then washed twice with cold 95% ethanol for purification dissolved in 10 mL of ddH₂O and filtered through a 0.22-μm sterile filter (Whatman Uniflo, United States). The polysaccharide in EPS extract was determined using the phenol–sulfuric acid method with glucose as the standard (Dubois et al., 1951). The proteins in EPS extract were detected using the BCA Protein Assay Kit (Thermo Scientific, USA) with bovine serum albumin as the standard (Smith et al., 1985). All the optical density measurements were detected using a multimode microplate reader (Thermo Scientific, Model: Varioskan LUX, United States).

The resistance of the isolates to Cu was tested by determining their growth on solid and liquid MB medium with various concentrations of CuSO4 (0, 0.4, 0.8, 1.2, 1.6, 2, 4, 6, 7, 8, 9, 10 mM) at 28°C for 5 days in triplicate. Use the formula (R_i = ∣(V₀-V_i)/V₀∣) to determine the effect of copper concentration on growth rates. In this formula, R_i represents the relative change in growth rate at an i mM copper concentration. V denotes the growth rate of the strains during the exponential growth phase, which is observed from 8 to 16 h in this study. V₀ is the exponential growth rate at 0 mM copper concentration, serving as a baseline control. V_i is the exponential growth rate at the i mM copper concentration. Use statistical software SPSS to perform an analysis of variance (ANOVA) to assess the statistical significance of the growth rate differences under different copper concentrations among strains.

The tolerance of the bacteria against various concentrations of heavy metals was tested. To evaluate the sorption capability, isolates were cultured in triplicate in 100 mL of MB medium containing 1 mM Cu (250 mg/L) in a 250 mL flask at 150 rpm for 24 h at 28°C. The supernatants of cultures were sampled every 24 h, filtered, and then diluted with 10% HNO₃ for the estimation of copper in the medium. To estimate the biosorption capacity of EPS, the extracted EPS was dissolved to a final concentration of 1 g/L, followed by the addition of different concentrations of CuSO₄ solution ranging from 0 to 400 mg/L and after 1-h incubation then filter diluted with 10% HNO_3. Copper ion concentrations were estimated using an atomic adsorption spectrometer (Srivastava and Majumder, 2008) and measured using an inductively coupled plasma optical emission spectrometer (ICP-OES, PerkinElmer, Model: Optima 7,300 V, United States). Using an uninoculated medium without CuSO₄ standard solution served as a control.

The GenBank/EMBL/DDBJ accession numbers for the 16S rRNA gene sequence of isolates are shown in Table 1. The whole-genome project for M. metalliresistant CuT 6 has been deposited in GenBank under the accession number ASM3809864v1 and is submitted in the NCBI under BioSample number SAMN29498721.

Twelve copper-resistant bacteria were isolated from hydrothermal fields, of which 11 were assigned to the γ-Proteobacteria and one to the α-Proteobacteria as determined by their 16S rDNA sequences. They were affiliated with the Pseudoalteromonas (4), Marinobacter (3), Halomonas (2), Psychrobacter (1), Pseudomonas (1) genus in the γ-Proteobacteria group, and the Sphingomonas (1) genera in the α-Proteobacteria group (Figure 1; Table 1). All the isolates had growth temperatures between 4 and 37°C, with the optimal growth temperature being higher than 15°C, specifically approximately 28°C. As all the strains had growth at 4°C, they could be assigned as psychrotrophs. All 12 strains were Gram-negative, neutrophilic, and aerobic.

Among these genera, Pseudomonas and Halomonas strains have often been found heavy metal resistant ability and used to sequester heavy metals such as Cd (II), Cu (II), Cr (VI), and Ni (II) from aqueous solution (Gabr et al., 2008; Fang et al., 2011; Ozturk et al., 2012; Govarthanan et al., 2015; Mădălina et al., 2016; Chug et al., 2021). Strain CuT 1–1 closest to Pseudomonas stutzeri ATCC 17588 was isolated from sediments and tolerated with 9 mM copper. Two Halomonas strains CuT 3–1 and CuT 4–2 showed extra high copper-resistant capacity of up to 10 mM, which are close to extremely halotolerant bacteria Halomonas meridiana DSM 5425 isolated from Antarctic saline lake (Kaye and Baross, 2000; Jonathan et al., 2011).

Pseudoalteromonas, Marinobacter, and Psychrobacter are broadly distributed in marine environments, including hydrothermal vents, subseafloor, and deep-sea sediments (Holmstrom and Kjelleberg, 1999). However, little reported their metal-resistant ability. For example, the EPS-producing Pseudoalteromonas sp. MER144 from Antarctic seawater showed metal adsorption capacity (Caruso et al., 2017). In this study, three strains showed 99.57–99.93% similarity with marine Pseudoalteromonas agarivorans DSM 14585 and had high copper-resistant ability 9–10 mM. Another Pseudoalteromonas sp. strain CuT 3–2 showed 99.01% 16 s rDNA similarity with Pseudoalteromonas carrageenovora IAM12662. Three Marinobacter strains CuT 1–2, CuT 4–1, and CuT 6 were close to M. vinifirmus FB1^T (99.93% 16 s rDNA similarity) and M. guineae M3B^T (99.21, 96.88% 16 s rDNA similarity). Strain CuT 6 was identified as a novel species and further characterized as follows. Strain CuT 10 had 98.89% 16 s rDNA similarity with Psychrobacter oceani 4k5^T, which was isolated from marine sediment.

Sphingomonas genus in the α-Proteobacteria group was frequently isolated from soil and encoded heavy metal resistance-related genes such as multicopper oxidase and the czc efflux system (Altimira et al., 2012; Kim et al., 2020). Those strains possess multifaceted functions ranging from remediation of environmental contaminations to promoting plant growth (Asaf et al., 2020). In this study, strain CuT 7 closest to Sphingomonas dokdonensis DS-4 was isolated. Sphingomonas dokdonensis DS-4 was isolated from soil, but no metal-resistant ability was found previously (Yoon et al., 2006).

Diverse and adaptable heterotrophic α- and γ-proteobacteria have consistently been detected and repeatedly isolated from different venting areas throughout a wide range of environmental conditions in hydrothermal plumes, sulfide structures, and associated sediments, demonstrating their prevalence and dominance in the deep-sea hydrothermal field (Edwards et al., 2003; Kato et al., 2013; Meier et al., 2016; He et al., 2022). In the sample area (Southwest Indian Ocean Ridge), metagenomic analysis also showed that α- and γ-proteobacteria were abundant in two sulfides samples, accounting for 7.79–24.54% and 20.6–61.0% (Zhong et al., 2022). Our findings implied that their adaptability might encompass their remarkable resistance ability to heavy metals.

In this study, all 12 isolates showed highly copper-resistant activities up to 6 to 10 mM in the liquid MB medium with various concentrations of CuSO₄ (Table 1; Supplementary Figure S1). Four strains, including Halomonas sp. CuT 3–1 and CuT 4–2, Pseudoalteromonas sp. CuT 4–3 and CuT 5, could tolerate high concentrations of Cu (II) up to 10 mM. Among the three isolated Marinobacter strains (CuT 1–2, CuT 4–1, and CuT 6), strain CuT 6 exhibited an exceptional capacity to withstand copper concentrations reaching as high as 8 mM. These three distinct genera, including Halomonas sp. CuT 3–1, Pseudoalteromonas sp. CuT 4–3, and Marinobacter metalliresistant CuT 6, were chosen as representatives for the analysis of their copper-resistant abilities, as they exhibited high levels of copper resistance.

The effects of copper with treatment concentrations ranging from 0 to 10 mM on the bacterial growth curve and the maximum bacterial growth number are shown in Supplementary Figure S1 and Figure 2. The growth of Halomonas sp. CuT 3–1 was inhibited in the presence of higher than 2 mM copper (Supplementary Figure S1A). Furthermore, it could have growth with OD600 (optical density at 600 nm) of 0.09 in the 18 h culture with 10 mM Cu (II) (Supplementary Figure S1A). There was almost no effect on the highest cell growth number of Pseudoalteromonas sp. CuT 4–3 under 2.0 mM Cu (II) and M. metalliresistant CuT 6 under 1.6 mM Cu (II) (Figure 2). Pseudoalteromonas sp. CuT 4–3 could grow up to OD600 = 0.3 at 6 mM Cu (II) (Supplementary Figure S1B). M. metalliresistant CuT 6 showed the highest growth in MB with 0.4–0.8 mM Cu (II) and reached an optical density (OD600) of 0.25 with 8 mM Cu (II). Previous reports showed that most copper-resistant bacteria tolerate 0.5–3.5 mM of Cu (II) (Altimira et al., 2012). Bacillus sp. showed the highest copper-resistant ability up to 5.5 mM (Kunito et al., 1997). Some copper-resistant bacteria could decrease copper (<1.5 mM) for the remediation of metal-contaminated waters or soils (Kugler et al., 2022). The isolates in this study displayed high resistance to Cu (II) up to 6–10 mM, which indicated their highly heavy metal resistant ability might contribute to their adaptation in the hydrothermal fields. Besides, the genus Marinobacter is very broadly distributed in marine and terrestrial environments (Cooper et al., 2022). However, only one report demonstrated that Marinobacter adhaerens HP15 had zinc resistance by two CzcCBA efflux pumps and helped it to colonize aggregates (Stahl et al., 2015; Will et al., 2020). In this study, typical heavy metals, including Cu (II), Cr (VI), Cd (II), Co (II), Zn (II), and Hg (II), were further determined to study the heavy metal-resistant abilities of 30 Marinobacter species (28 type species and 2 new isolates from hydrothermal vents) (Supplementary Figure S2). Most of them exhibited certain copper resistance, with a minimum tolerance of 2 mM and a maximum tolerance of 6 mM Cu (II). Among them, strain CuT 6 showed the highest copper resistance capacity up to 8 mM and could tolerate 2 mM Co (II), 3 mM Zn (II), and 0.3 mM Hg (II) (Figure 3; Supplementary Figure S2). Among their phylogenetically closest relatives, M. guineae strain M3B^T and M. profundi strain PSW21^T demonstrated the highest copper resistance at a concentration of only 2 mM, which is different from the growth rate changes observed in CuT 6 when CuT 6 was exposed to copper. The differences between strains were especially obvious at concentrations of 4, 6, and 8 mM. (p < 0.05, Table. S6). Besides, each of Marinobacter spp. displayed resistance to at least one type of heavy metal. Among them, Marinobacter fuscus NH169-3^T showed metal-resistant ability to all the tested heavy metals.

Except for strain CuT 3–2, 11 out of 12 isolates exhibited positive EPS production ability, forming red or black colonies on the CRA plates compared with colonies on MB plates (Supplementary Figure S3). The EPS was subsequently purified from a 24-h liquid culture of Halomonas sp. CuT 3–1, Pseudoalteromonas sp. CuT 4–3, and M. metalliresistant CuT 6 as described methods (Maalej et al., 2015).

Compared to the absence of Cu (II) in the medium, the addition of 1 mM Cu (II) significantly induced the production of EPS by these three isolates, which were mainly composed of proteins and sugars (Supplementary Figure S4). The increase in the concentration of EPS from 379.65 mg/L to 1013.00 mg/L in the culture of Halomonas sp. strain CuT 3–1 with copper stimulation. The polysaccharide and the protein contents in the EPS of CuT 3–1 increased by 3.64-fold [from 21.64 mg/L (5.70%) to 100.49 mg/L (9.92%)] and 2.55-fold [from 358.04 mg/L (94.30%) to 912.06 mg/L (90.08%)] with copper stimulation. Contents of EPS of Pseudoalteromonas sp. strain CuT 4–3 consistently increased from 366.70 mg/L to 1057.92 mg/L as Cu increased. The polysaccharide content of strain CuT 4–3 increased from 33.37 mg/L (9.10%) to 66.12 mg/L (6.25%), and the protein part increased from 333.44 mg/L (90.10%) to 992.00 mg/L (93.75%). The total EPS content of strain CuT 6 was lower than those of strain CuT 3–1 and strain CuT 4–3. The EPS content of strain CuT 6 increased a little from 214.35 mg/L to 447.17 mg/L with copper stimulation. The copper ion addition also induced an increase in polysaccharide content [from 50.48 mg/L (23.55%) to 82.19 mg/L (18.38%)] and protein content [from 214.31 mg/L (76.45%) to 364.96 mg/L (81.62%)] of strain CuT 6. In addition, we examined the composition and contents of EPS produced by strain CuT 6 and its closely related type strains under conditions with and without copper. The results indicated that strain CuT 6 produced different components and contents of EPS, and it responded more sensitively to copper stimulation compared to the other five Marinobacter strains (Supplementary Figure S10).

The copper adsorption capacity of the culture medium of copper-tolerant bacteria during growth at 250 mg/L of CuSO₄ was monitored (Supplementary Figure S5). The binding ability of EPS produced by these bacteria at different concentrations of 0–400 mg/L of CuSO₄ was also determined, as shown in Figure 4. With the addition of CuSO₄, the copper ion content in the mediums was reduced by copper-resistant bacterial strains. After 24 h, strain CuT 3–1 reduced 75.38 mg/L of CuSO₄. After 48-h cultivation, strain CuT 4–3 and strain CuT 6 reduced 74.60 mg/L and 79.21 mg/L of CuSO₄, respectively. The binding of metal by EPS increased linearly with the content of Cu (II). The EPS of M. metalliresistant CuT 6 was found to have the highest binding ability up to 52.24 mg CuSO₄/g EPS. The maximum amount of Cu (II) bound to the EPS of strain CuT 4–3 and CuT 3–1 was 37.78 and 35.92 mg CuSO₄/g EPS. These three copper-tolerant bacterial strains all exhibited abilities to reduce CuSO₄ concentrations in the medium. However, their EPS adsorption capacities varied significantly. The EPS produced by strain CuT 6 had a higher capacity to adsorb CuSO₄ than those produced by the other two strains. The results indicated that the copper ions could be adsorbed by the cells themselves and by the secreted EPS. The findings revealed that upon exposure to copper, the production of EPSs by copper-resistant bacteria increased, along with a fluctuating ratio of polysaccharides and proteins within the EPS matrix. Notably, the EPS surrounding the cells displayed remarkable adsorption capabilities for heavy metals, which might contribute to bacterial adaptation in the harsh environments of deep-sea hydrothermal vents.