Denitriﬁcation-Potential Evaluation and Nitrate-Removal-Pathway Analysis of Aerobic Denitriﬁer Strain Marinobacter hydrocarbonoclasticus RAD-2

: An aerobic denitriﬁer was isolated from a long-term poly (3-hydroxybutyrate- co -3-hydroxyvalerate) PHBV-supported denitrification reactor that operated under alternate aerobic/anoxic conditions. The strain was identiﬁed as Marinobacter hydrocarbonoclasticus RAD-2 based on 16S rRNA-sequence phylogenetic analysis. Morphology was observed by scanning electron microscopy (SEM), and phylogenetic characteristics were analyzed with the API 20NE test. Strain RAD-2 showed efﬁcient aerobic denitriﬁcation ability when using NO 3 − -N or NO 2 − -N as its only nitrogen source, while heterotrophic nitriﬁcation was not detected. The average NO 3 − -N and NO 2 − -N removal rates were 6.47 mg/(L · h)and 6.32 mg/(L · h), respectively. Single-factor experiments indicated that a 5:10 C/N ratio, 25–40 ◦ C temperature, and 100–150 rpm rotation speed were the optimal conditions for aerobic denitriﬁcation. Furthermore, the denitrifying gene napA had the highest expression on a transcriptional level, followed by the denitrifying genes nirS and nosZ . The norB gene was found to have signiﬁcantly low expression during the experiment. Overall, great aerobic denitriﬁcation ability makes the RAD-2 strain a potential alternative in enhancing nitrate management for marine recirculating aquaculture system (RAS) practices.


Introduction
Recirculating aquaculture systems (RAS) are a potential alternative to traditional aquaculture systems due to their intensive production and environmental sustainability [1]. In practice, RAS mainly use biological filters to oxidize ammonium to nitrate through nitrification, with nitrite as the intermediate product since ammonium and nitrite have direct toxicity to most fish species [2]. Nitrate concentration accumulates and reaches high concentrations during intensive fish farming. Therefore, nitrate management is very important due to its explicit long-term stress effect on cultured species [3], as well as its contribution to environmental eutrophication [4,5]. In various nitrate-removal 1.0 g of K 2 HPO 4 , and 10 mL of a trace-element solution in 1 L of distilled water. The heterotrophic nitrification medium (HNM) was prepared by dissolving 2.0 g sodium acetate, 0.3 g of NH 4 Cl, 0.2 g of MgSO 4 ·7H 2 O, 6.7 g of Na 2 HPO 4 , 1.0 g of KH 2 PO 4 , and 10 mL of a trace-element solution in 1 L of distilled water. The composition of the trace-element solution was 50.0 g of EDTA, 2.2 g of ZnSO 4 , 5.5 g of CaCl 2 , 5.06 g of MnCl 2 ·4H 2 O, 5.0 g of FeSO 4 ·7H 2 O, 1.1 g of (NH 4 ) 6 Mo 7 O 2 ·4H 2 O, 1.57 g of CuSO 4 ·5H 2 O, and 1.61 g of CoCl 2 ·6H 2 O in 1 L of distilled water. The Luria-Bertani (LB) medium was prepared by dissolving 5.0 g yeast extract, 10.0 g peptone, and 25.0 g NaCl in 1 L of distilled water and 1.5% (w/v) agar. The initial pH of all media was set to 7.2, and all media were autoclaved for 20 min at 121 • C.

Bacteria Isolation, Screening, and Identification
Strain RAD-2 was isolated from the biofilms of a long-term aerobic/anoxic denitrifying reactor using PHBV as simultaneous carbon source and carrier. The reactor setup and operation conditions were according to our previous study [18]. The reactor was placed in a dark artificial-climate room to retain the temperature at 26 ± 2 • C. The influent NO 3 − -N concentration was set at 70 mg/L and HRT (hydraulic retention time) was 4 h. In detail, 20 g of matured PHBV samples and 10 mL solution samples were aseptically transferred to a flask with 100 mL sterile water and 10 small glass balls.
To suspend the biofilms attached to the PHBV granules, the flask was shaken on a rotary shaker at 200 rpm for 30 min. The homogenized suspensions were serially diluted and plated using a DM, and then incubated at 28 • C for 72 h. A single colony with a white circle was purified by streaking onto an LB medium plate, which was then incubated for three days at 28 • C. Several colonies were obtained after strict investigation of their purity. Among the isolates, a colony that was white, irregular circle-shaped with opaque, wet, and smooth surfaces, 1-2 mm in diameter was distinguished as RAD-2.
The purified isolate was stored in a 30% glycerol solution at −80 • C. The genomic DNA of the RAD-2 strain was isolated using a DNA extraction kit (TaKaRa Biotechnology Co. Ltd, Beijing, China). The 16S rRNA gene was PCR-amplified using bacterial universal primers F27 (5 -AGAGTTTGATCMTGGCTCAG-3 ) and R1492 (5 -TTGGYTCCTTGT TACGACT-3 ), under the following conditions: 2 min at 95 • C, 25 cycles of 20 s at 95 • C, 20 s at 55 • C, 30 s at 72 • C, and a final step of 10 min at 72 • C. PCR products were detected on 1% agarose gel electrophoresis and ethidium bromide staining. The amplified products were purified and sequenced by the Zhejiang Institute of Microbiology (Hangzhou, Zhejiang, China). The sequence was submitted to the NCBI database (accession numbers MH725589) and compared with other available 16S rRNA gene sequences in Genbank by BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi). A phylogenetic tree was constructed using Molecular Evolutionary Genetics Analysis software (MEGA) version X by the neighbor-joining method with 1000 bootstrap replicates.

Nitrogen-Removal Performance Evaluation
To evaluate aerobic denitrification capacity, a sole nitrogen source of NO 3 − -N (around 300 mg/L) or NO 2 − -N (around 300 mg/L) was tested in DM containing KNO 3 or NaNO 2 , respectively. Afterward, a 3 mL seed suspension was inoculated in 250 mL Erlenmeyer flasks and cultured for 48 h with aeration at 27 • C and 150 rpm. To evaluate the capacity for heterotrophic nitrification, a similar operation was carried out that only replaced the HNM with a sole nitrogen source of NH 4 Cl (around TAN 90 mg/L). Cell-growth and inorganic-nitrogen changes were measured every 4 h. The nitrogen-removal rate was calculated as below: where R N = nitrogen removal rate, mg/(L·h); C I = initial NO 3 − -N or NO 2 − -N concentration, mg/L; C F = final NO 3 − -N or NO 2 − -N concentration, mg/L; V = volume, mL; T = incubated time, h.
Single-factor experiments were also carried out to evaluate the effect of various conditions on the aerobic denitrification performance of strain RAD-2. The operation conditions for DM were as follows: NO 3 − -N concentration of around 300 mg/L, C/N ratio 10, NaCl 25%, temperature 25 • C, rotation 150 rpm, and 1.2% inoculation (v/v). For temperature experiments, the temperature was set to 5 • C, 10 • C, 15 • C, 25 • C, and 40 • C. For C/N ratio experiments, the C/N ratios were set to 2, 5, 10, 15, and 20. For dissolved oxygen (DO) experiments, the rotations were set to 0, 50, 100, 150, and 200 rpm. Cell growth and indexes (nitrate, nitrite, DOC, and pH) were measured during the experimental period. All tests were conducted in triplicate and none-seeded samples were used as blank control.

RT-qPCR Analysis
To quantitatively analyze the potential aerobic denitrification pathways of strain RAD-2, real-time PCR was conducted to amplify the denitrifying genes napA, nirS, norB, nosZ, and 16S rRNA (housekeeping gene) with RNA samples in 48 h experiments. Total RNA extraction and cDNA synthesis were performed by using an RNAprep Bacteria Kit and FastQuant RT Kit (Tian Gen Biotech Co. Ltd, Beijing, China), respectively. Primers are listed in Table S1. PCR amplification was performed with the following protocol: an initial denaturation step of 10 min at 95 • C, followed by 40 cycles of denaturation at 95 • C for 10 s, annealing at 60 • C (napA, nirS, and nosZ) or 56 • C (16S V3 region and norB) for 30 s, and a final extension at 72 • C for 30 s [30]. All quantitative amplifications were conducted in triplicate using the SYBR Green Real-Time PCR Kit (Novland, Shanghai, China) and respective primers on an Mx3000P qPCR System (Agilent Technologies Co. Ltd., Beijing, China) [19].

Analytical Methods
The solution samples were collected and filtered through a 0.45 µm filter membrane before water-quality analysis. TAN, NO 2 − -N, and NO 3 − -N concentrations were analyzed according to standard methods [31]. Cell growth (OD 600 ) was measured by using a spectrophotometer at 600 nm. DOC was measured using a TOC analyzer (Multi N/C 2100, Analytik Jena, Jena, Germany). DO was measured using a DO meter (SG9-FK2, Mettler Toledo, Zurich, Switzerland). Morphological analysis was performed by scanning electron microscopy (SEM) (SU8010; HITACHI, Tokyo, Japan). Fresh colonies grown on LB agar for 2 days were fixed in 1% glutaraldehyde (prepared in cacodylate buffer, pH 7.4) at 4 • C overnight, and then completely dehydrated in ethanol. Cells were coated with gold-palladium and observed with a HITACHI 8010 scanning electron microscope (HITACHI, Tokyo, Japan). Physiological and biochemical characteristics were tested using API 20NE kits (BioMérieux Shanghai Co. Limited, Shanghai, China). API 20NE test strips was checked after incubation for 24 h [29].

Characteristics and Identification
In this study, more than six pure isolates were obtained from solid DM and tested for aerobic denitrification performance by monitoring changes in nitrite and nitrate concentration in the liquid DM. A particular isolate, namely, RAD-2, exhibited the highest efficiency in nitrate and nitrite removal and was subject to further investigation. Strain RAD-2 was slightly halophilic and able to grow under aerobic conditions. The colonies of RAD-2 were yellow, small, circular in shape, semitransparent, slabby, and presented a wet surface on the LB medium. The cells were Gram-negative, bacilliform, with a size of 0.3-0.4 µm in diameter and 1.0-2.0 µm in length ( Figure S1).
Analysis of 16S rRNA gene sequencing showed that strain RAD-2 belongs to the species Marinobacter hydrocarbonoclasticus, having 98% similarity with Marinobacter hydrocarbonoclasticus strain ATCC 49840 and Marinobacter hydrocarbonoclasticus strain VT8. Phylogenetic analyses of the 16S rRNA gene sequencing showed that strain RAD-2 formed a distinct clade with strain ATCC 49840 and strain VT8, and this clade clustered with the nearest clade containing Marinobacter sp. NN5, Marinobacter sp. U1369-101122-SW163, and Marinobacter hydrocarbonoclasticus strain NY-4 ( Figure 1). The phylogenetic position of this strain indicated that it presented a subspecies of the species Marinobacter hydrocarbonoclasticus. Table 1. Characteristics of strain RAD-2 determined by API 20 NE tests.

API 20 NE Results
Strain RAD-2 The phylogenetic position of this strain indicated that it presented a subspecies of the species Marinobacter hydrocarbonoclasticus.

Aerobic Nitrogen-Removal Ability of Strain RAD-2
The aerobic denitrification ability of strain RAD-2 under varied nitrogen sources is shown in Figure 2. After 48 h of incubation, NO 3 − -N concentration decreased from the initial 310.94 mg/L to the final 5.17 mg/L, which indicated 98.34% removal efficiency (Panel A). The obvious lag phase was observed between 0 and 24 h, while the logarithmic growth phase was observed between 24 and 36 h. Nitrite accumulation occurred between 20 and 36 h, while peak concentration of 5.05 mg/L was observed at 32 h. In addition, a slight ammonium concentration of 2.96 mg/L was also found in the final concentration. The biomass growth of OD 600 reached 1.34. Additionally, when nitrite was used as the sole nitrogen source (Panel B), NO 2 − -N concentration also decreased from 303.69 mg/L to 0.52 mg/L, which was 99.83% removal efficiency. However, the backward lag phase was found in 0-44 h, with a final biomass of 0.71, which indicated that strain RAD-2 might be more adaptable under a nitrate condition. However, the removal rates of 6.47 mg/(L·h) and 6.32 mg/(L·h) were detected for strain RAD-2 when nitrate or nitrite was used as the sole nitrogen source, respectively. It should be noted that the maximum nitrite-removal rate of strain RAD-2 that could be achieved was 56.20 mg/(L·h) at 44-48 h of the logarithmic growth phase (Panel B).
Water 2018, 10, x FOR PEER REVIEW 6 of 13

Aerobic Nitrogen-Removal Ability of Strain RAD-2
The aerobic denitrification ability of strain RAD-2 under varied nitrogen sources is shown in Figure 2. After 48 h of incubation, NO3 − -N concentration decreased from the initial 310.94 mg/L to the final 5.17 mg/L, which indicated 98.34% removal efficiency (Panel A). The obvious lag phase was observed between 0 and 24 h, while the logarithmic growth phase was observed between 24 and 36 h. Nitrite accumulation occurred between 20 and 36 h, while peak concentration of 5.05 mg/L was observed at 32 h. In addition, a slight ammonium concentration of 2.96 mg/L was also found in the final concentration. The biomass growth of OD600 reached 1.34. Additionally, when nitrite was used as the sole nitrogen source (Panel B), NO2 − -N concentration also decreased from 303.69 mg/L to 0.52 mg/L, which was 99.83% removal efficiency. However, the backward lag phase was found in 0-44 h, with a final biomass of 0.71, which indicated that strain RAD-2 might be more adaptable under a nitrate condition. However, the removal rates of 6.47 mg/(L·h) and 6.32 mg/(L·h) were detected for strain RAD-2 when nitrate or nitrite was used as the sole nitrogen source, respectively. It should be noted that the maximum nitrite-removal rate of strain RAD-2 that could be achieved was 56.20 mg/(L·h) at 44-48 h of the logarithmic growth phase (Panel B).  The heterotrophic nitrification performance of the strain RAD-2 is illustrated in Figure 3. After 48 h of incubation, TAN concentration decreased slightly from the initial 89.64 mg/L to the final 80.73 mg/L, which indicated only 9.94% removal efficiency. No nitrite accumulation was found at any period, while around 0.70 mg/L nitrate was produced. Biomass built up to 0.16 after incubation, which indicated poor growth performance. Therefore, the strain RAD-2 was found to have no heterotrophic nitrification ability under current conditions. The heterotrophic nitrification performance of the strain RAD-2 is illustrated in Figure 3. After 48 h of incubation, TAN concentration decreased slightly from the initial 89.64 mg/L to the final 80.73 mg/L, which indicated only 9.94% removal efficiency. No nitrite accumulation was found at any period, while around 0.70 mg/L nitrate was produced. Biomass built up to 0.16 after incubation, which indicated poor growth performance. Therefore, the strain RAD-2 was found to have no heterotrophic nitrification ability under current conditions.

Single-Factor Experiments of Strain RAD-2
The effects of several environmental factors on the aerobic denitrification performance of strain RAD-2 are shown in Table 2. Aerobic denitrification efficiency relied on the amount of the carbon source, which served as electron donor and energy source. In this study, C/N ratio 5:10 was found optimal for strain RAD-2, having more than 95% nitrate-removal efficiency. A low C/N ratio of 2 lowered nitrate-removal efficiency to 33.59% and had inadequate cell growth, with a final OD600 of 0.32 after 48 h incubation. It should be noted that excess C/N ratio also led to a decrease in denitrification performance. On a C/N ratio of 20, only 29.80% nitrate-removal efficiency was achieved, with a final cell growth value of 0.42.
In general, denitrification performance is typically sensitive to temperature variations due to the differences in bacteria species. In this study, strain RAD-2 presents a mesophilic characteristic in the aerobic denitrification process. When the temperature range was 2-15 °C , notably low nitrateremoval efficiency of less than 10% was obtained. Increased temperature could significantly improve denitrification performance, as near as 100% nitrate-removal efficiency, which was gained in the 25-40 °C range.
The different rotation speeds that presented the DO effects on denitrification efficiency were also tested. In this study, strain RAD-2 gained ideal nitrate-removal efficiency in rotation speeds of 100 and 150, which is the equivalent of DO centration at 5.55 and 6.23 mg/L, respectively. Otherwise, only 54.41% nitrate-removal efficiency was obtained under no rotations (DO 0.82 mg/L). In addition, 200 rpm (DO 7.2 mg/L) slightly decreased nitrate-removal efficiency to 89.90%.

Single-Factor Experiments of Strain RAD-2
The effects of several environmental factors on the aerobic denitrification performance of strain RAD-2 are shown in Table 2. Aerobic denitrification efficiency relied on the amount of the carbon source, which served as electron donor and energy source. In this study, C/N ratio 5:10 was found optimal for strain RAD-2, having more than 95% nitrate-removal efficiency. A low C/N ratio of 2 lowered nitrate-removal efficiency to 33.59% and had inadequate cell growth, with a final OD 600 of 0.32 after 48 h incubation. It should be noted that excess C/N ratio also led to a decrease in denitrification performance. On a C/N ratio of 20, only 29.80% nitrate-removal efficiency was achieved, with a final cell growth value of 0.42.
In general, denitrification performance is typically sensitive to temperature variations due to the differences in bacteria species. In this study, strain RAD-2 presents a mesophilic characteristic in the aerobic denitrification process. When the temperature range was 2-15 • C, notably low nitrate-removal efficiency of less than 10% was obtained. Increased temperature could significantly improve denitrification performance, as near as 100% nitrate-removal efficiency, which was gained in the 25-40 • C range.
The different rotation speeds that presented the DO effects on denitrification efficiency were also tested. In this study, strain RAD-2 gained ideal nitrate-removal efficiency in rotation speeds of 100 and 150, which is the equivalent of DO centration at 5.55 and 6.23 mg/L, respectively. Otherwise, only 54.41% nitrate-removal efficiency was obtained under no rotations (DO 0.82 mg/L). In addition, 200 rpm (DO 7.2 mg/L) slightly decreased nitrate-removal efficiency to 89.90%.

Expression of Denitrifying Genes by RT-qPCR Analysis
The expression of key denitrifying genes in the aerobic denitrification of strain RAD-2 is shown in Figure 4. On a transcriptional level, the napA gene showed the highest expression level in this study, which indicated the aerobic denitrification characteristic of strain RAD-2. The nirS and nosZ genes had similar expression intensity, which was one order of magnitude lower than that of the napA gene. Moreover, the norB gene was found to have significantly low expression during the whole period, and its intensity was negligible when compared with other genes (napA, nirS, and nosZ). All genes showed a decrease or low expression intensity during 0-24 h. Then, notable synergetic expressions of napA, nirS, and nosZ genes showed an increase in the range 24-36 h. It should be noted that though maximum expression intensity was found at 0 h, this time point should reflect the transcriptional state of strain RAD-2 in LB media, as we obtained the samples immediately after the inoculation.

Characteristics and Identification
In this study, Marinobacter hydrocarbonoclasticus strain RAD-2 was isolated from a denitrifying reactor using PHBV as the carbon source and biofilm carrier. In general, Marinobacter hydrocarbonoclasticus is the species of the genus Marinobacter, which belongs to the class

Characteristics and Identification
In this study, Marinobacter hydrocarbonoclasticus strain RAD-2 was isolated from a denitrifying reactor using PHBV as the carbon source and biofilm carrier. In general, Marinobacter hydrocarbonoclasticus is the species of the genus Marinobacter, which belongs to the class Gammaproteobacteria. Species of this genus are Gram-staining-negative, rod-shaped, and motile [31]. A notable feature of Marinobacter hydrocarbonoclasticus is the utilization of various hydrocarbons as sole carbon and energy sources [32]. For example, using waste frying oil as the inducer carbon source, the produced biosurfactant of the strain Marinobacter hydrocarbonoclasticus SdK644 could be applied to improve crude-oil solubilization in a marine environment [33]. Therefore, strain RAD-2 might have the ability to use biodegradable polymers (PHBV etc.) for denitrification.
Based on the 16S rRNA gene sequences, strain RAD-2 formed a distinct branch with strain ATCC49840 and strain VT8, and this clade was close to the groups containing Marinobacter sp. NN5, Marinobacter sp. U1369-101122-SW163, and Marinobacter hydrocarbonoclasticus strain NY-4. However, the genus Marinobacter was reported to have many different phenotypic characteristics in the denitrification process. For example, strain RAD-2, Marinobacter sp. NN5, and Marinobacter sp. F6 were found to have efficient aerobic denitrification ability [20,25], while Marinobacter hydrocarbonoclasticus strain NY-4 only had anaerobic denitrification ability [34].

Nitrogen-Removal Performance Evaluation
In this study, strain RAD-2 presented efficient aerobic denitrification performance. An average removal rate of 6.47 mg/(L·h) and 6.32 mg/(L·h) was found in strain RAD-2 when nitrate or nitrite was used as the sole nitrogen source, respectively ( Figure 2). This was much faster than several other Marinobacter strains. For example, Marinobacter sp. NN5 and Marinobacter sp. F6 were reported to have a 4.7 mg/(L·h) and 1.46 mg/(L·h) NO 3 − -N removal rate, respectively [20,25]. In other genera, Bacillus methylotrophicus L7 was found to have a 5.81 mg/(L·h) NO 2 − -N removal rate [35].
Otherwise, strain RAD-2 cannot perform heterotrophic nitrification, which was consistent with Marinobacter sp. NN5 [20]. Only Marinobacter sp. F6 was reported to have the simultaneous ability of heterotrophic nitrification and aerobic denitrification in the Marinobacter genus [25]. Based on several single-factor experiments, strain RAD-2 showed good ecological width in marine-aquaculture conditions. A 5:10 C/N ratio, 25-40 • C temperature, and 100-150 rpm rotation speed were the optimal conditions for aerobic denitrification (Table 2). It is reported that Marinobacter sp. NN5 has 35 • C temperature, 6:8 C/N ratio, and 150 rpm rotation speed as optimal conditions [20]. Therefore, strain RAD-2 could adapt to a lower temperature of 25 • C, which might increase its application in marine aquaculture, as temperatures 25-35 • C are the optimal environmental conditions for most cultured species. However, to better use the strain in practice, toxicology research should also be performed for strain RAD-2 in the future [38,39].

Aerobic Denitrification Pathways Analysis
The expression of key denitrifying genes in the aerobic denitrification of strain RAD-2 is shown in Figure 4. In general, aerobic denitrification has two different electron-transfer pathways [22]. The expression of the napA gene can guarantee that the aerobic denitrification strain still has electron-transfer capacity under aerobic conditions [22]. In anoxic denitrification, electron transfer to nitrate can be blocked as encoding gene narG is sensitive to oxygen [13]. In this study, the napA gene had a maximum expression level, which was responsible for the efficient aerobic nitrate removal performance. The synergetic expressions of the napA, nirS, and nosZ genes increased during 24-36 h, which resulted in strain growth and nitrate elimination. (Figure 2A). It should also be noted that the norB gene showed very low expression (Figure 4). The norB gene was in charge of nitric oxide reductase production [22], which transfers NO to N 2 O. N 2 O emission has recently been attracting more attention due to its environmental impact [40]. In the Marinobacter genus, many strains were reported as having zero N 2 O emissions. For example, Marinobacter sp. NN5 has total N 2 production without N 2 O in aerobic conditions, while Marinobacter hydrocarbonoclasticus strain NY-4 was reported to produce no N 2 O in anaerobic conditions [20,34]. Conventionally, the high activity of nitrous oxide reductase, which was encoded by the nosZ gene, was charged with the efficient transfer of N 2 O to N 2 . Here, we also give molecular evidence that the Marinobacter hydrocarbonoclasticus RAD-2 strain has little expression intensity, which might be another reason for its zero N 2 O emissions. Since N 2 O is an important greenhouse gas, and aquaculture systems are considered an important anthropogenic source of N 2 O emission [41], strain RAD-2 might have great potential for aerobic denitrification in marine RAS applications.

Conclusions
An aerobic denitrifier strain was isolated from a long-term PHBV-supported denitrification reactor that was operated under alternate aerobic/anoxic conditions. The strain was identified as Marinobacter hydrocarbonoclasticus RAD-2 based on 16S rRNA-sequence phylogenetic analysis. Strain RAD-2 showed high efficiency for aerobic denitrification when using NO3 − -N or NO 2 − -N as the sole nitrogen source, while almost being unable to perform heterotrophic nitrification. The average NO 3 − -N and NO 2 − -N removal rates were 6.47 mg/(L·h) and 6.32 mg/(L·h), respectively. Single-factor experiments indicated that a 5:10 C/N ratio, 25-40 • C temperature, and 100-150 rpm rotation speed were the optimal conditions for aerobic denitrification. Furthermore, the denitrifying gene napA had maximum expression intensity on a transcriptional level, followed by nirS and nosZ. The norB gene was found to have significantly low expression during the whole period. Therefore, the denitrifying pathways showed its aerobic denitrification characteristic and potentially fewer N 2 O emissions. Overall, the efficient aerobic denitrification performance of strain RAD-2 makes it a potential candidate for bioaugmentation to improve the effluent treatment of marine RAS.