For preliminary isolation of the aerobic denitrifying bacteria, a denitrification screening medium with low N (DM-L) was employed. The composition of DM-L medium included the following components per liter: sodium succinate 0.25 g, sodium citrate dihydrate 0.25 g, Na₂HPO₄ 1.6 g, KH₂PO₄ 1.0 g, NaCl 0.5 g, NaNO₂ 0.07 g, KNO₃ 0.1 g, (NH₄)₂SO₄ 0.066 g, MgSO₄·7H₂O 0.2 g, 0.2% (v/v) trace element solution (TES), and 0.1% (v/v) mixed carbon source solution (CSS). The pH was adjusted to pH 7.2. Solid plates were prepared by supplementing DM-L with 15.0 g/L agar. Before use, the medium (excluding MgSO₄·7H₂O, TES, and CSS) was sterilized by autoclaving at 121 °C (0.11 MPa) for 30 min. Filter-sterilized MgSO₄·7H₂O, TES, and CSS were aseptically supplemented to the autoclaved medium. The formulations of TES and CSS followed protocols established in our previous work (Zhang et al., 2022).

Samples collected from Penaeus vannamei aquaculture water in Jiangmen City, Guangdong Province, China, were diluted (10⁻¹ to 10⁻⁴), and 0.1 mL from 10⁻² to 10⁻⁴ dilutions was spread on DM-L agar, incubated aerobically at 30 °C for 2–7 days. Colonies with distinct morphology were purified and stored at −80 °C in sterile physiological saline supplemented with 25.0% (v/v) glycerol. For denitrifying bacteria isolation, selected strains were grown statically at 30 °C in DM-L medium with 5.0 mg/L NO₂⁻-N, and residual NO₂⁻-N was measured using the Griess reaction.

Strain JM12B12 was cultured on LB agar at 30 °C for 48 h, with colonies observed macroscopically and cells via electron microscopy (H7650, Hitachi). Bacterial genomic DNA was extracted using the HiPure Bacterial DNA Kit (Magen Biotech., China). The 16S rRNA genes were PCR-amplified with primers 27F and 1492R, sequenced by GENEWIZ (Suzhou, China), and aligned against the EzBioCloud database (Chalita et al., 2024). Strain JM12B12 showed the highest NO₂⁻-N removal efficiency and was further analyzed for morphology, phylogeny, and genome. The 16S rRNA sequences were aligned using MAFFT v7.526 under the L-INS-i iterative refinement (Rozewicki et al., 2019). Maximum-likelihood (ML) phylogenetic tree was reconstructed with IQ-TREE v2.1.2 with integrated ModelFinder for evolutionary model selection, selecting the best model via Bayesian Information Criterion and assessing node support with 1,000 bootstrap replicates (Hoang et al., 2018; Minh et al., 2020). The phylogenetic tree was visualized using MEGA 11 (Tamura et al., 2021).

Our results showed that JM12B12 exhibited more effective capability in the removal of NO₂⁻-N compared to other isolates. Therefore, the effects of various conditions on the N removal efficiency of JM12B12 were investigated. Each variable was adjusted independently, and the optimal conditions were applied in subsequent experiments.

A single colony of JM12B12 was grown in LB medium for 12 h, washed, and resuspended to OD₆₀₀ of 1.0. The suspension was inoculated into media to study N removal, testing sodium acetate, citrate, succinate, lactate, glucose, and sucrose as carbon sources at C/N ratios of 1 to 20. The basal medium (BM) without carbon and N sources was formulated as (per liter): Na₂HPO₄ 1.6 g, KH₂PO₄ 1.0 g, NaCl 0.5 g, MgCl₂·6H₂O 0.1 g, D-biotin 5.0 mg, cobalamin 5.0 mg. Furthermore, the effects of various culture conditions on the denitrification of JM12B12 were investigated, including initial pH (5–11), concentrations of NaCl (0–3.0%), temperature (20–45 °C), and shaking speeds (0–200 rpm). The BM was supplemented with NO₃⁻-N and sodium lactate as sole N and carbon sources, and the C/N ratio was adjusted to 10. After 48 h of incubation, samples were analyzed for OD_600, and the relevant culture supernatants were used for the measurement of N concentrations (NO₂⁻-N, NO₃⁻-N, and NH₄⁺-N). Uninoculated media were used as controls.

To evaluate JM12B12’s denitrification process and N balance, sodium lactate-supplemented BM with either NO₂⁻-N (BM1) or NO₃⁻-N (BM2) was used. During incubation, samples were taken to measure OD_600, and the relevant supernatants were used for detecting concentrations of NO₂⁻-N, NO₃⁻-N, and NH₄⁺-N. Additionally, total N (TN-N) and intracellular N (CN-N) concentrations were measured at 0 h and 48 h. Uninoculated media were used as controls.

Strain JM12B12 was cultured in the BM with NO₃⁻-N and sodium lactate as the sole N and carbon source, respectively. After 48 h of incubation, the bacterial suspension (BS) was centrifuged at 12,000 g for 10 min to obtain cell-free supernatant (CFS) and cells (CE). The flocculation capabilities of BS, CFS, and CE were determined by the kaolin suspension method (Chen L. et al., 2024). Briefly, the experimental flocculation system comprised 2.0 mL sample (cells were resuspended with the physiological saline solution), 3.0 mL 1.0% CaCl₂ (w/v), and 95.0 mL 4.0 g/L kaolin suspension, with pH maintained at 7.5 via HCl and/or NaOH adjustment. The mixture was stirred at 200 rpm for 2 min, then stirred at 50 rpm for 5 min, and then allowed to settle for 5 min. The 3 mL of the supernatant was aspirated from a consistent depth of 1 cm below the air-liquid interface. The optical density measurement was conducted immediately post-collection using a calibrated spectrophotometer at 550 nm wavelength. The flocculating efficiency (FloE) was calculated according to the equation: FloE (%) = (A-B)/A × 100, where A represented the absorbance values of samples (BS, CFS, and CE) at 550 nm, and B represented the reference absorbance values (sterile culture medium for BS and CFS; physiological saline for CE) at 550 nm.

The complete genome of strain JM12B12 was sequenced at Shanghai Majorbio Bio-Pharm Technology Co., Ltd. (China) using a hybrid approach combining Nanopore PromethION (Oxford Nanopore, Oxford, UK) and Illumina HiSeq 2,500 (Illumina, Inc., San Diego, CA, USA). For Illumina sequencing, genomic DNA was fragmented to 400–500 bp using Covaris M220, and libraries were prepared with the NEXTFLEX Rapid DNA-Seq kit. For Nanopore sequencing, DNA fragments were repaired, purified, and ligated with sequencing adapters from the SQK-LSK kit before library preparation and sequencing. Raw Illumina reads were filtered using fastp v0.23.0, while Nanopore reads were processed (basecalling, demultiplexing, trimming) with a minimum Q score of 7. Hybrid assembly was performed using Unicycler v0.4.8, and Pilon v1.22 was used for error correction. The complete genome was reconstructed by integrating data from both platforms.

GeneMarkS v4.3, tRNA-scan-SE v2.0.12, and barrnap v0.9 were used to predict coding sequences (CDS), tRNA, and rRNA, respectively. The predicted CDS were annotated using COG (202006) and KEGG (202209) databases via sequence alignment tools. Core genome circular maps, COG, and KEGG analyses were conducted on Majorbio’s Cloud platform (Han et al., 2024). The genome was also annotated with RAST v2.0 under the Classic RAST scheme. Digital DNA–DNA hybridization (dDDH) and average nucleotide identity (ANI) values were calculated using Genome-to-Genome Distance Calculator 3.0 (formula 2) and FastANI (Jain et al., 2018; Meier-Kolthoff et al., 2013). The phylogenomic tree was constructed using Up-to-date Bacterial Core Gene sets (UBCGs) (Riesco and Trujillo, 2024).

Genomic DNA of strain JM12B12 was extracted using the method described in Section 2.2. Primer sets napA_5F/ napA_3R, nirS1_5F/ nirS1_3R, nirS2_5F/ nirS2_3R, norB_5F/ norB_3R, and nosZ_5F/ nosZ_3R were utilized to amplify the denitrification genes from JM12B12 (Supplementary Table S2). PCR amplification was performed using 2 × Phanta Max Master Mix (Vazyme Biotech, Nanjing, China) in a total reaction volume of 20 μL, comprising the following components: 10 μL of 2 × Phanta Max Master Mix, 0.4 μL of primers (10 μM each), 1 μL of DNA template (50 ng/μL), and 8.2 μL of ddH₂O. PCR reaction program was as follows: 95 °C for 5 min; 30 cycles consisting of 95 °C for 30 s, 58 °C for 30 s, and 72 °C for 2 min 30 s; followed by 72 °C for 10 min. The PCR products were verified by comparing with Trans2K® Plus II DNA Marker (TransGen Biotech, Beijing, China) electrophoretically using a 1.5% agarose gel.

NO₂⁻-N, NO₃⁻-N, and NH₄⁺-N concentrations were quantified using N-(1-naphthyl) ethylenediamine dihydrochloride, ultraviolet spectrophotometry, and Nessler’s reagent photometry at the wavelength of 420 nm, respectively (Baird et al., 2017). TN-N and CN-N concentrations were quantified using alkaline potassium persulfate digestion followed by spectrophotometric detection at 220 nm and 275 nm, respectively (Baird et al., 2017). Standard curves for the determination of N concentrations were shown in Supplementary Figure S1. OD₆₀₀ was determined at a wavelength of 600 nm. N removal efficiency was calculated by the following formula: N removal efficiency (%) = ([N_x]ⁱ - [T3N]^f)/[N_x]ⁱ × 100, where [N_x]ⁱ is the initial NO₃⁻-N, NO₂⁻-N, and NH₄⁺-N concentration, respectively; [T3N]^f is the sum of NO₃⁻-N, NO₂⁻-N, and NH₄⁺-N concentrations. All experiments were carried out in quadruplicate, and results were expressed as the mean of four replicates ± standard deviation (mean ± SD).

There were 10 isolates obtained, and strain JM12B12 exhibited excellent performance in removing NO₂⁻-N (Supplementary Table S1). The colonies of strain JM12B12 appeared off-white, opaque, with a smooth surface, regularly circular, and neat-edged, measuring approximately 1 mm in diameter (Figure 1a). Cells of this strain were observed to be rod-shaped (0.7–0.9 × 1.9–2.5 μm) with a single polar flagellum (Figure 1b). The 16S rRNA gene sequence of JM12B12 obtained through PCR amplification (1,400 bp) showed complete identity with the corresponding sequences extracted from its genomic DNA (1,537 bp). The complete 16S rRNA gene sequence was deposited in the NCBI GenBank database under accession number PP716610. The sequence comparison showed that JM12B12 shared the highest similarity with T. chlorobenzoica 3CB-1 (99.2%), followed by strains T. selenatis ATCC 55363 (99.2%), T. aminoaromatica S2 (99.1%), T. phenylacetica B4P (99.0%), and T. mechernichensis TL1 (98.9%). Phylogenetic reconstruction based on the 16S rRNA sequences revealed that JM12B12 formed a distinct clade within the genus Thauera cluster, supported by 73% bootstrap values (Figure 1c). Therefore, JM12B12 was taxonomically assigned to the genus Thauera.

Denitrifying bacteria require an array of reductases to accomplish the denitrification pathway. According to RAST annotation of genome, 36 genes were associated with denitrification and nitrite/nitrate assimilation pathways (Figure 5a). Dissimilatory nitrate reduction to nitrite was catalyzed by two types of enzymes: the membrane-bound nitrate reductase (Nar) and the periplasmic nitrate reductase (Nap), and Nar and Nap reductases were often associated with anaerobic and aerobic denitrification, respectively (Hu et al., 2023). A napFDAGHBC gene cluster (gene_3698/3697/3696/3695/3694/3693/3692) encoding NapA and relevant enzymes was identified in JM12B12, which was consistent with Shewanella oneidensis MR-1 (Liu et al., 2021). Dissimilatory nitrite reduction to nitric oxide was catalyzed by two structurally distinct enzymes: a cytochrome cd1-dependent nitrite reductase (encoded by nirS) and a copper-containing nitrite reductase (encoded by nirK). Interestingly, two nirS gene (gene_1573 and gene_1693) sequences with a similarity of 66.2% were found in the genome of JM12B12, which was inconsistent with Bradyrhizobium diazoefficiens (Pacheco et al., 2022). Previous studies have demonstrated that nirS-type denitrifying bacteria possessed robust metabolic systems for energy conservation, facilitating their survival under environmental stresses (Ming et al., 2024). The genome of JM12B12 harbored a functionally enriched cluster of energy metabolism-associated genes, demonstrating evolutionary adaptations critical for environmental persistence. The key functional genes for reducing NO to nitrous oxide (N₂O) (norB, gene_1714) and reducing N₂O to N₂ (nosZ, gene_0523) were also identified.

Furthermore, the key genes involved in denitrification were amplified via PCR from the genomic DNA of JM12B12. As shown in Figure 5b, the napA, nirS1, nirS2, norB, and nosZ genes were determined via agarose gel electrophoresis. The results were consistent with the predicted sizes of the relevant functional genes within the JM12B12 genome assembly. Overall, the strain JM12B12 harbored all the denitrification genes, which indicated its potential for complete N removal through the denitrification pathway: NO₃⁻-N → NO₂⁻-N → NO→N₂O → N₂.

Bioflocculants produced by different bacteria were various in their distribution. Some were secreted into the extracellular environment, and others were tightly adhered to the cellular surface. To investigate the flocculation characterization of JM12B12, the BS, CFS, and CE were obtained after it grew in the denitrification medium. As shown in Figure 6, the flocculation efficiencies of BS, CFS, and CE were 83.7, 91.4, and 21.8%, respectively. The results indicated that the flocculants produced by JM12B12 were found mainly in the cell-free supernatant, consistent with that of strains Stenotrophomonas pavanii GXUN74707, Pseudomonas sp. XD-3, and Providencia huaxiensis OR794369.1 (Chen L. et al., 2024; Qin et al., 2024; Selepe and Maliehe, 2024). Bioflocculants produced by bacteria were found to be EPS, which were mainly composed of polysaccharides, protein polymers, glycoproteins, etc. (Selepe and Maliehe, 2024). Therefore, JM12B12 was a novel denitrifying bacterium capable of producing bioflocculants, suggesting that it possesses important features in practical applications of bioremediation and wastewater treatment.

Our results indicated that JM12B12 was capable of producing extracellular bioflocculants. In general, bacterial bioflocculants were composed of EPS, which included polysaccharides, proteins, extracellular nucleic acids, and lipids, and extracellular polysaccharides (exopolysaccharides) constituted the dominant component of EPS (Flemming et al., 2025; Vandana and Das, 2023). Therefore, the genes associated with exopolysaccharide biosynthesis and protein secretion systems were identified by analyzing the genome annotations of JM12B12. Exopolysaccharides production was a multistage process where polysaccharides were intracellularly synthesized and exported outside (Stephens et al., 2023). According to analyses of annotations, there were 31 genes related to polysaccharide biosynthetic, polysaccharide export, and biopolymer transport processes, such as genes alg, rfb, udg, wbp, eps, exb, and tol, etc. (Supplementary Table S3). These genes might be essential for exopolysaccharide production of JM12B12. Extracellular proteins, one of the key amphiphilic macromolecules within EPS, played a crucial role in governing flocculating ability (He et al., 2024). Secretion systems were protein export machines that enable bacteria to exploit their environment through the release of protein effectors (Lauber et al., 2024). There were 49 genes identified as components of the bacterial protein secretion machineries, encompassing the Sec (general secretory pathway) and Tat (twin-arginine translocation) systems, along with Type I, II, III, IV, and VI secretion systems (Supplementary Table S4). Overall, the genes involved in exopolysaccharides production and protein secretion systems might be the potential functional genes associated with the flocculating activity of JM12B12. However, the mechanism of bioflocculant synthesis in this bacterium remains to be fully clarified in the future.

The circus map of genome characteristics of JM12B12 was comprehensively displayed in Figure 5c. The genomic analysis of JM12B12 revealed a single circular chromosome spanning 4,171,389 bp, characterized by a GC content of 67.9%. Notably, no plasmid was identified in this bacterium. The absence of plasmid reduced the probability of genes (such as genes related to antimicrobial resistance) transfer within strain JM12B12, potentially resulting in enhanced safety compared to the plasmid-harboring bacteria. Genome data has been deposited in the NCBI GenBank database under the accession number CP154859. Compared with related strains in the genus Thauera, JM12B12 had ANI values of 84.2–87.8% and dDDH values of 26.3–31.4% (Figures 1d,e), which were all below the classical standard species delineation threshold values of 95.0 and 70%, respectively (Goris et al., 2007; Richter and Rosselló-Móra, 2009). The phylogenomic tree constructed using UBCGs indicated that JM12B12 fell in a large clade with members of the genus Thauera, which was consistent with the phylogenetic relationships inferred from 16S rRNA gene sequence analysis (Supplementary Figure S2). Moreover, JM12B12 together with T. aminoaromatica S2 and T. phenylacetica B4P formed a clade supported by the bootstrap value of 100%. Phylogenomic analyses incorporating dDDH values, ANI values, and phylogenetic reconstruction conclusively confirmed that strain JM12B12 was a novel species in the genus Thauera (Riesco and Trujillo, 2024).

Aggregate annotation results indicated that a total of 3,771 genes, including 3,707 protein-coding sequences (CDS), 55 tRNA genes, and 9 rRNA genes, were identified. The metabolic pathways were analyzed based on genome annotations. The RAST annotation showed that there were 363, 301, 300, and 277 genes associated with ‘Amino acids and derivatives’, ‘Cofactors, vitamins, prosthetic groups, pigments’, and ‘Carbohydrates’, respectively (Supplementary Figure S3). The KEGG annotation showed that the top four pathways with the most enriched number of genes were ‘Global and overview maps’, ‘Energy metabolism’, and ‘Amino acid metabolism’, respectively (Supplementary Figure S4). The COG annotation revealed that 287, 282, 282, and 259 genes were categorized under the functional classes ‘Energy production and conversion’, ‘Amino acid transport and metabolism’, and ‘Signal transduction mechanisms’, respectively (Supplementary Figure S5). The reconstruction of this bacterium’s carbon metabolic pathways based on genome annotation revealed that the tricarboxylic acid cycle (TCA), glyoxylate cycle (GAC), glycolysis pathway (Embden-Meyerhof-Parnas, EMP), and gluconeogenesis pathway (GNG) were all complete. These carbon metabolic pathways could supply electrons for denitrification and provide precursors for the synthesis of extracellular polysaccharides and proteins.

Our research studied Thauera sp. JM12B12, an aerobic denitrifying bacterium that uniquely combines the capability to produce bioflocculants with efficient denitrification, even under low C/N ratio conditions. Consequently, we proposed that JM12B12 held significant promise as a candidate for wastewater treatment in RASs. However, it is important to acknowledge the limitations of our study, which can be addressed and supplemented in the future. Firstly, we will elucidate the chemical composition and mechanisms of EPS produced by JM12B12. To advance the practical application of Thauera sp. JM12B12 in N removal from aquaculture water, future research could focus on assessing its biosafety, validating its denitrification efficiency in real RASs aquaculture water. Additionally, employ the molecular biology, genetic biology, transcriptomics, and metabolomics techniques to elucidate the underlying mechanisms behind JM12B12’s flocculation and its exceptional denitrification performance at low C/N ratios.