Dirammox Is Widely Distributed and Dependently Evolved in Alcaligenes and Is Important to Nitrogen Cycle.

Nitrogen cycle is an essential process for environmental health. Dirammox (direct ammonia oxidation), encoded by the dnfT1RT2ABCD cluster, was a novel pathway for microbial N2 production defined in Alcaligenes ammonioxydans HO-1. Here, a copy of the cluster dnfT1RT2ABCD as a whole was proved to have existed and very conserved in all Alcaligenes genomes. Phylogenetic analyses based on 16S rRNA gene sequences and amino acid sequences of DnfAs, together with G + C content data, revealed that dnf cluster was evolved associated with the members of the genus Alcaligenes. Under 20% O2 conditions, 14 of 16 Alcaligenes strains showed Dirammox activity, which seemed likely taxon-related. However, the in vitro activities of DnfAs catalyzing the direct oxidation of hydroxylamine to N2 were not taxon-related but depended on the contents of Fe and Mn ions. The results indicated that DnfA is necessary but not sufficient for Dirammox activity. The fact that members of the genus Alcaligenes are widely distributed in various environments, including soil, water bodies (both freshwater and seawater), sediments, activated sludge, and animal-plant-associated environments, strongly suggests that Dirammox is important to the nitrogen cycle. In addition, Alcaligenes species are also commonly found in wastewater treatment plants, suggesting that they might be valuable resources for wastewater treatment.

Introduction

The nitrogen cycle is important for nitrogen balance in the biosphere and global environmental health (Gruber and Galloway, 2008). Biological nitrogen removal of fixed nitrogen is not only an essential part of the natural nitrogen cycle (Gruber and Galloway, 2008), but also the basic principle in the development of nitrogen removal technologies in wastewater treatment (Kuypers et al., 2018). Up to date, only two pathways have been identified and recognized as biological N2 generation pathways, namely, nitrification/denitrification (Van Cleemput and Samater, 1995; Sigman and Fripiat, 2019) and anaerobic ammonia oxidation (anammox) (Mulder et al., 1995; Strous et al., 1999). In nitrification/denitrification pathway, ammonia oxidation (NH4+ → NH2OH → NO2– → NO3–) is the energy source of autotrophic nitrifers, ammonia can be oxidized aerobically via hydroxylamine to nitrite, nitrate, or nitrogenous gas by a single-strain Nitrospira species (comammox, complete ammonia oxidation), or by a combination of ammonia-oxidizing bacteria or ammonia-oxidizing archaea, and nitrite-oxidizing bacteria (Abell et al., 2010; Pester et al., 2012; Baptista et al., 2014). The enzyme ammonia monooxygenase catalyzes the oxidation of ammonia to hydroxylamine (Ensign et al., 1993; Gilch et al., 2009), which is then converted to NO by hydroxylamine oxidoreductase (Maalcke et al., 2014; Caranto and Lancaster, 2017). In denitrification pathway, nitrate or nitrite is sequentially reduced to N2 (NO3– → NO2– → NO → N2O → N2) when O2 is limited (Skiba, 2008). In anammox pathway, N2 is produced from oxidation of ammonia, with nitrite produced by nitrifiers as the electron acceptor (NH4+ + NO2– → N2 or NH4 + + 1.32 NO2– + 0.066 HCO3– + 0.13 H+ → 1.02 N2 + 0.26 NO3– + 0.066 CH2O0.5N0.15 + 2.03 H2O) under anaerobic conditions by anammox bacteria (phylum Planctomycetes) (Mulder et al., 1995; van de Graaf et al., 1996; Strous et al., 1999).

We previously isolated and characterized Alcaligenes ammonioxydans strain HO-1 (Wu et al., 2021) from a SHARON bioreactor treating ammonium-rich piggery wastewater (Du et al., 2016). Strain HO-1 could be tolerant to high ammonia load up to 3,400 mg/L and directly convert ammonia via hydroxylamine to N2 under aerobic conditions (Wu et al., 2021). The high ammonia tolerance property makes it a potential strain in wastewater treatment plants (WWTPs). Furthermore, it also demonstrated its ability to directly convert ammonia to N2 under aerobic conditions as a result of a novel pathway, termed Dirammox (direct ammonia oxidation) (Wu et al., 2021). It was also demonstrated that the production of N2 from ammonia in HO-1 was the activity of Dirammox, but not that of nitrite denitrification (Wu et al., 2021). This result was further confirmed in strain JQ135 (Xu et al., 2022), another Alcaligenes strain with the same ability of Dirammox as HO-1. The knockout of denitrification pathway in strain JQ135 did not change its N2 yield under aerobic conditions (Xu et al., 2022). This novel pathway is conducted by enzymes encoded by the gene cluster dnfT1RT2ABCD defined in the genome of HO-1; and just dnfABC could enable recombinant Escherichia coli strain to generate hydroxylamine and N2 from ammonia (Wu et al., 2021). These seven genes were annotated as 3-phosphoserine/phosphohydroxythreonine transaminase, PLP-dependent aminotransferase family protein, serine hydroxymethyltransferase, diiron N-oxygenase, 2Fe–2S iron–sulfur cluster binding domain-containing protein, glutamine amidotransferase, and pyridoxine/pyridoxal/pyridoxamine kinase, respectively (Wu et al., 2021). Among the seven genes, only dnfA encoded an N-oxygenase catalyzing the oxidation of hydroxylamine to N2, and its deletion resulted in the total loss of Dirammox activity in strain JQ135 (Xu et al., 2022). Therefore, dnfA was selected as a marker gene of Dirammox pathway for the following investigations.

Dirammox is quite different from the two known N2 production pathways mentioned above, namely, (1) N2 generation in Dirammox is quite efficient under physiological O2 levels by a single heterotrophic bacterium, while denitrification and anammox occur anaerobically; and (2) Dirammox is the simplest N loss process, of which N2 is directly generated from ammonia NH3+ (-NH2) via hydroxylamine by a single bacterium such as A. ammonioxydans HO-1. However, N2 produced by nitrification/denitrification is the result of the subsequent cooperation of multiple strains as mentioned above. Although an anammox bacterium could oxidize ammonia directly to N2 with nitrite as the final electron acceptor, it needs the cooperation of nitrifiers for providing nitrite (Kuypers et al., 2018).

Currently, there are five species in the genus, namely, Alcaligenes faecalis, A. aquatilis (Van Trappen et al., 2005), A. pakistanensis (Abbas et al., 2015), A. endophyticus (Lu et al., 2017), and A. ammonioxydans HO-1 (Wu et al., 2021). Members of the genus Alcaligenes are widely distributed in natural environments such as soil (Liu X. et al., 2016), water (Regar et al., 2016), as well as human and other vertebrates (Simmons et al., 1981; Kaliaperumal et al., 2006; Laham et al., 2017). With the disclosed Dirammox in Alcaligenes strain HO-1 (Wu et al., 2021), for which the N2 produced aerobically is not related to aerobic denitrification, whether the nitrogen removal of Alcaligenes strains comes from heterotrophic nitrification and aerobic denitrification (HNAD, the process: NH4+ → NH2OH → NO2– → NO3– → NO2– → NO → N2O → N2) (Wehrfritz et al., 1993; Richardson et al., 1998, Stein, 2011) or the novel pathway remained to be clarified.

However, it is still not clear (1) whether Dirammox is widely distributed in Alcaligenes or not and (2) whether N2 generation is exclusively from Dirammox or the combination of Dirammox and HNAD under physiological conditions in Alcaligenes strains? In this study, the distribution of dnfT1RT2ABCD in Alcaligenes genomes was analyzed widely by GenBank. Meanwhile, the sequences of this gene cluster defined in the genomes were further compared and phylogenetically studied. Furthermore, 16 Alcaligenes strains were obtained and investigated for their Dirammox and denitrification activities using stable isotope incubations. In addition, the representatives of each phylogenetic type of DnfA were investigated for their enzymatic activity catalyzing oxidation of hydroxylamine to N2. The results provided insights into the distribution and evolution of Dirammox in the genus Alcaligenes, activity of Dirammox among Alcaligenes strains, as well as its mechanism, and thereby the possible importance of Dirammox to nature.

Materials and Methods

Bioinformatic Analysis

To investigate the distribution of gene cluster dnfT1RT2ABCD in Alcaligenes genomes, DnfT1, DnfR, DnfT2, DnfA, DnfB, DnfC, and DnfD of strain HO-1 were used as a query individually against the non-redundant protein sequences database (NR) restricted to Alcaligenes genomes to perform BlastP program in NCBI blast web (Johnson et al., 2008), respectively. The algorithm parameter “Max target sequences” was adjusted to 5,000. The posted back results were filtered with conditions, namely, expect threshold ≤0.001, query coverage ≥70%, identity ≥30%, and extract the accession numbers of proteins. Get identical protein groups by the above accession numbers using the NCBI Batch Entrez tool. Then, detailed information about each homologous protein was obtained, including nucleotide accession number, start and stop site in nucleotide, strand, accession number of protein, protein name, organism, strain, and assembly. Finally, only the genomes containing all the seven genes of the dnf cluster arranged in one DNA fragment of the strains of a bacterial species were screened as the genomes containing gene cluster dnfT1RT2ABCD.

Phylogenetic trees based on 16S rRNA gene sequences or amino acid sequences of DnfAs were reconstructed using the neighbor-joining method (Saitou and Nei, 1987) performed in MEGA X software (Kumar et al., 2018) with 1,000 bootstrap resamplings. Clustal Omega (Sievers and Higgins, 2013) or Clustal W (Thompson et al., 1994) was used to align multiple sequences. ESPript 3.0[1] (Robert and Gouet, 2014) with default setting was used to render sequence similarities.

Bacterial Strains, Plasmid, Media, and Cultivation Conditions

Bacterial strains and plasmids used in this study are listed in Table 1. Luria–Bertani medium (LB) was prepared according to Liu Y. et al. (2016). The basal medium was prepared by dissolving KH2PO4 0.50 g, Na2HPO4⋅12H2O 1.25 g, MgSO4⋅7H2O 0.2 g, NaCl 10 g, and trace element solution 2 ml in 1 L of distilled water, pH 7.0–7.5. The trace element solution was prepared according to Joo et al. (2005). A total of 4.72 g of succinate was added as carbon source in the basal medium per liter. Ammonia or nitrite was added as nitrogen source as needed listed in Table 2. Alcaligenes strains were cultivated at 30°C, 160 rpm in synthetic media (Table 2). E. coli strains were cultivated at 37°C in LB media. Ampicillin (100 mg/L) was supplemented with media as necessary. For agar plates, 1.5% (w/v) of agar was added to the media.

TABLE 1

Bacterial strains and plasmids used in this study*.

Strain or plasmid	Description	Source/References
Alcaligenes strain
A. ammonioxydans HO-1	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	SHARON bioreactor (Wu et al., 2021)
A. aquatilis CGMCC 1.0767	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	Rubber
A. aquatilis GL12	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	Farming wastewater; this study
A. aquatilis CGMCC 1.9053	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	Activated sludge
A. faecalis subsp.faecalis PC01	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	Farming wastewater; this study
A. faecalis subsp. faecalis CGMCC 1.1799	Aerobically converting 15NO2– to 15N2O	Faecalis (Málek et al., 1963)
A. faecalis subsp. faecalis CGMCC 1.2006T	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	(Papen et al., 1989)
A. faecalis subsp. faecalis CGMCC 1.2908	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	ND
A. faecalis subsp. faecalis CGMCC 1.2388	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	ND
A. faecalis subsp. faecalis CGMCC 1.3937	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	ND
A. faecalis subsp. faecalis CGMCC 1.1837	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	Tomur Peak, Xinjiang
A. faecalis subsp. phenolicus DSM 16503T	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	Graywater bioprocessor (Rehfuss and Urban, 2005)
A. faecalis subsp. phenolicus RL12	Aerobically converting 15NO2– to 15N2O and 15N2	Farming wastewater; this study
A. faecalis subsp. parafaecalis DSM 13975T	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	Eutrophic garden pond (Schroll et al., 2001)
A. faecalis JQ135	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	Municipal wastewater (Zhang et al., 2018)
A. sp. CGMCC 1.10269	Aerobically converting 15NH4+ to 15N2 with hydroxylamine accumulation and 15NO2– to 15N2O	Dyeing wastewater, this study
Escherichia coli strain
Transetta(DE3)	Protein expression host	Transgen, China
Transetta(DE3)/pET21a-WP_003803202.1	E. coli Transetta(DE3) carrying plasmid pET21a-WP_003803202.1	This study
Transetta(DE3)/pET21a-WP_042487153.1	E. coli Transetta(DE3) carrying plasmid pET21a-WP_042487153.1	This study
Transetta(DE3)/pET21a-WP_045930341.1	E. coli Transetta(DE3) carrying plasmid pET21a-WP_045930341.1	This study
Transetta(DE3)/pET21a-WP_009459326.1	E. coli Transetta(DE3) carrying plasmid pET21a-WP_009459326.1	This study
Transetta(DE3)/pET21a-WP_094197465.1	E. coli Transetta(DE3) carrying plasmid pET21a-WP_094197465.1	This study
Transetta(DE3)/pET21a-WP_035272004.1	E. coli Transetta(DE3) carrying plasmid pET21a-WP_035272004.1	This study
Transetta(DE3)/pET21a-WP_137431195.1	E. coli Transetta(DE3) carrying plasmid pET21a-WP_137431195.1	This study
Transetta(DE3)/pET21a-HCA15598.1	E. coli Transetta(DE3) carrying plasmid pET21a-HCA15598.1	This study
Plasmid
pET21a(+)	Ampr, Expression vactor	Lab stock
pET21a-WP_003803202.1	pET21a(+) haboring dnfA of A. ammonioxydans HO-1.	This study
pET21a-WP_042487153.1	pET21a(+) haboring dnfA of A. faecalis subsp. faecalis CGMCC 1.2006T.	This study
pET21a-WP_045930341.1	pET21a(+) haboring dnfA of A. faecalis subsp. phenolicus DSM 16503T.	This study
pET21a-WP_009459326.1	pET21a(+) haboring dnfA (N879_RS10420) of Alcaligenes sp. EGD-AK7 (GCA_000465875.3); whole gene synthesis.	This study
pET21a-WP_094197465.1	pET21a(+) haboring dnfA of A. aquatilis CGMCC 1.9053	This study
pET21a-WP_035272004.1	pET21a(+) haboring dnfA of A. faecalis subsp. faecalis CGMCC 1.3937	This study
pET21a-WP_137431195.1	pET21a(+) haboring dnfA (D0C27_RS13980) of A. faecalis AU14 (GCA_005311025.1); whole gene synthesis.	This study
pET21a-HCA15598.1	pET21a(+) haboring dnfA (DEO64_00430) of A. faecalis UBA11281 (GCA_003521065.1); whole gene synthesis.	This study

*CGMCC and DSM denote China General Microbiological Culture Collection Center and German Collection of Microorganisms and Cell Cultures GmbH, respectively. All Alcaligenes strains were tested in a sealed bottle with the V

TABLE 2

Nitrogen source used for different media.

Nitrogen source	Medium A	Medium B	Medium C	Medium D	Medium E
5 mM (NH4)2SO4	+	−	−	−	+
5 mM (15NH4)2SO4	−	+	+	−	−
3 mM NaNO2	−	−	+	+	−
3 mM Na15NO2	−	−	−	−	+

Amplification of 16S rRNA Genes and dnfAs of Tested Strains

Primers used in this study are listed in Table 3. Primer pair dnfA-upF/dnfA-downR targeting complete dnfA were designed based on upstream and downstream sequences of dnfAs of Alcaligenes fetched from GenBank. High-fidelity DNA polymerase TransStart® FastPfu (Transgen, China) was used. The PCR products were sequenced by a commercial company. Finally, the start and end codons were determined by aligning with dnfA sequence of strain HO-1. EMBOSS Transeq[2] was applied to translate nucleotides into proteins. These sequence data are available in the GenBank[3] under accession number OM293488-511.

TABLE 3

Primers used in this study.

Primer	Sequence (5′→3′)	Description	Source/Reference
27F	AGAGTTTGATCCTGGCTCAG	Universal primers amplifying 16S rRNA gene	Wilson et al., 1990
1492R	GGTTACCTTGTTACGACTT
dnfA-upF	CAAATCCTTTTAAGCCTGCG	Universal primers amplifying complete dnfA of Alcaligenes.	This study
dnfA-downR	ACTTTGACGYTTCCGGCCT
dnfA-F	CGCGGATCATGACWATCAAAAGCTACGAAAC	Universal former primer with BamHI restriction site targeting dnfA of Alcaligenes for clones.	This study
dnfA-R1	CCCAAGCTTTTGCAGCGCCTCCTGTTGTTCG	Reverse primer with HindIII restriction site targeting dnfA of A. faecalis subsp. faecalis CGMCC 1.3937 for cloning pET21a-WP_035272004.1.	This study
dnfA-R2	CCCAAGCTTTTGCAGCGCCTCCTGTTGCTCG	Reverse primer with HindIII restriction site targeting dnfA of A. ammonioxydans HO-1T and A. aquatilis CGMCC 1.9053, for cloning pET21a-WP_003803202.1 and pET21a-WP_094197465.1, respectively.	This study
dnfA-R3	CCCAAGCTTTTGCAGCGCCTCCTGCTGTTCG	Reverse primer with HindIII restriction site targeting dnfA of A. faecalis subsp. faecalis CGMCC 1.2006T and A. faecalis subsp. phenolicus DSM 16503T, for cloning pET21a-WP_042487153.1 and pET21a-WP_045930341.1, respectively.	This study

Growth of Ammonia or Nitrite as the Sole Nitrogen Source

Growth (OD600) and hydroxylamine were monitored over time when ammonia (medium A, Table 2) was employed as the sole nitrogen source in Erlenmeyer flasks. The ability to grow on nitrite as the sole nitrogen source (medium D, Table 2) was determined over time in sealed bottles as previously described (Wu et al., 2021) but without displacement of the gas of the headspace, in which the seeds were prepared in agar plates, resuspended, and injected into the bottles.

Nitrogen Removal Characteristics of Dirammox and Aerobic Denitrification

The ability to convert ammonia to N2 was tested with 15N2 production from (15NH4)2SO4 (medium B, Table 2) or in the presence of nitrite (medium C, Table 2). For detection of gaseous nitrogen products (N2 and N2O) of Dirammox and aerobic denitrification, strains were cultivated in media B or E (Table 2) in air-tight bottles prepared according to the previous description (Wu et al., 2021). All samples were assayed when ammonia and formed hydoxylamine were completely consumed.

Chemicals

15N-labeled hydroxylamine (15NH2OH), Na15NO2, and (15NH4)2SO4 were all 99 atom% 15N and got from Cambridge Isotope Laboratories Inc., United States.

Plasmid Construction, Expression, and Purification of DnfA and the Homologs

In this study, the gene sequences of WP_009459326.1, WP_137431195.1, and HCA15598.1 were obtained from GenBank and synthesized by a commercial company, then the obtained genes were cloned into pET-21a(+) (Table 1). For construction of other plasmids listed in Table 1, PCR products amplified by TransStart® FastPfu were purified and digested with BamHI (NEB) and HindIII (NEB). The digested amplicons and BamHI/HindIII digested pET-21a(+) plasmid were ligated by T4 ligase (Takara) at 4°C overnight. For expression of DnfA and its s, E. coli Transetta (DE3) containing pET-21a ligating dnfA or dnfAs from other Alcaligenes strains was cultured in LB supplemented with 100 μg/ml ampicillin at 37°C on a rotary shaker (160 rpm) to OD600 = 0.3–0.6, then induced with 0.5 mM isopropyl-β-D -thiogalactopyranoside on a rotary shaker (160 rpm) for another 20 h at 16°C. DnfA or its s was purified individually using Ni-NTA resin (QIAGEN) with the following procedures, namely, cells were harvested by centrifugation (5,000 × g, 30 min, 4°C), resuspended in buffer A (100 mM Tris-HCl, 100 mM NaCl, 10 mM imidazole, pH 8.0), lysed by ultrasonication, and centrifuged (14,000 × g, 30 min, 4°C). Supernatant was applied to Ni-NTA resin (QIAGEN) column pre-equilibrated with buffer A. Ni-NTA matrix was washed with buffer A and added with 10–50 mM imidazole to remove impurities. 6xHis-tagged proteins were eluted with buffer B (100 mM Tris-HCl, 100 mM NaCl, 250 mM imidazole, pH 8.0), desalted with centrifugal filter devices (MilliporeSigma), and excess imidazole was removed with PD-10 desalting column (GE Healthcare).

To analyze the distribution of the genus Alcaligenes in nature, BlastN was performed using the 16S rRNA gene sequence (1542 nt) of strain HO-1 as a query against the nucleotide database (NT). Only the sequences with coverage >70% and similarity >95% were included. The isolation source or sample source was extracted and used to analyze the environmental distribution of the genus Alcaligenes.

Hydroxylamine Oxygenase Activity Assays

In vitro enzymatic assays, DnfA, and its homologs catalyzing the conversion of hydroxylamine to N2 were carried out in 300 μl reaction mixtures in 20 mM buffer Tris-HCl (pH 8.5) containing 330 μM DnfA or its homologs, 10 mM 15NH2OH, 10 mM NADH, and 20 μM FAD, as described in our preprint[4] (Wu et al., 2020). Mixtures lacking DnfA were used as the control. The reaction was started with the addition of 15NH2OH, and the mixture was directly injected into 10 ml gastight tubes whose top air had been completely replaced by 20% O2 and 80% He mixed gas. The reactions were incubated at 30°C without agitation in the dark for 100 min. Enzymatic assays were performed in triplicate.

Analytical Methods and Statistical Analysis

Bacterial growth and concentrations of ammonium (NH4+) of nitrite (NO2–), nitrate (NO3–), hydroxylamine, 15N2, and 15N2O were determined as described previously (Wu et al., 2021), except nitrite determine in enzymatic activity system, which was monitored by ion chromatography. For determination of gas products, strain cultivation and DnfA enzymatic reaction were performed in 250 ml sealed bottles containing 20 ml media or 10 ml gastight tube with 300 μl reaction system, respectively. The air was replaced completely by 80% He and 20% O2 mixed gas. PerkinElmer Optima 5300EV ICP-OES was used for the determination of atoms, including Fe, Mn, Ca, Cu, K, Mg, Ni, P, Zn, Si, and S. Results are shown as mean ± SD. The correlation relationship was demonstrated by Pearson’s correlation coefficient.

Results

Gene Cluster dnfT1RT2ABCD Present in the Known Genomes of Alcaligenes

All the available Alcaligenes genomes in GenBank (updated to May 2015) were analyzed for the presence of the gene cluster dnfT1RT2ABCD. As shown in Figure 1, a copy (and only one copy) of the gene cluster dnfT1RT2ABCD as a whole was defined in the 36 well-assembled Alcaligenes genomes (totally 49 genomes of unique Alcaligenes strain), including all 19 complete genomes, with the same arrangement and position as those in HO-1 (Figure 1). Among them, A. faecalis was the most dominant, with 29 genomes containing dnfT1RT2ABCD, followed by unnamed Alcaligenes strains (four genomes) and A. aquatilis (three genomes) (Figure 1B). In addition, parts of the dnf cluster were also defined in the rest of the 15 low-level assembled genomes of the total 49 Alcaligenes genomes of unique strain deposited in GenBank. Furthermore, the dnf clusters from different Alcaligenes genomes exhibited very high sequence similarity (>97% for dnfABC), suggesting that the dnf cluster was very conserved in the genus Alcaligenes. Based on these results, it was deduced that the conserved dnf cluster and so the Dirammox were commonly distributed in the genus Alcaligenes.

FIGURE 1

Arrangement (A) and statistics on distribution (B) of dnfT1RT2ABCD among the genomes of the genus Alcaligenes. Here, A. faecalis includes all three subspecies, namely, A. faecalis subsp. faecalis, A. faecalis subsp. Parafaecalis, and A. faecalis subsp. phenolicus. There are 49 genomes of unique Alcaligenes strains in GenBank updated to May 2021, and strain HO-1 was recognized as one of Alcaligenes sp.

Aerobic N2 Release in Alcaligenes Depends on the Activity of Dirammox

A total of 16 Alcaligenes strains (Table 1) were obtained for verifying the existence of Dirammox pathway and nitrogen removal ability. As shown in Figure 2A, ammonia could be used as the sole nitrogen source for aerobic growth of all Alcaligenes strains. In contrast, no growth for all tested strains was observed on nitrite as the sole nitrogen source (Figure 2A), suggesting that Alcaligenes spp. have no the ability to assimilate nitrite. Analysis of the gaseous nitrogen products indicated that most of the tested Alcaligenes strains (except of strains RL12 and CGMCC 1.1799) could aerobically convert 15NH4+ to 15N2 when 15NH4+ was used as the sole nitrogen source, in spite of their difference in N2-producing ability (Figure 2B). The results in Figure 2B also indicate that 15N2 amount released was not affected by the addition of nitrite. Figure 2C displays the gaseous nitrogen products released from aerobic denitrification of nitrite by the strains tested. The results showed that no tested strain could produce 15N2 from nitrite denitrification, and most of them could release 15N2O from 15NO2– (medium E, in Table 2) under aerobic conditions. All these results were in good agreement with those of HO-1 (Wu et al., 2021). These results strongly suggested that the production of N2 resulted from the direct conversion of ammonia in Alcaligenes spp. Therefore, it was concluded that the activity of Dirammox was the only way to produce N2 in Alcaligenes spp. under aerobic conditions.

FIGURE 2

Growth and Dirammox (direct ammonia oxidation) activity of Alcaligenes strains in different medium. (A) Aerobic growth in medium A and medium D, and OD600 shown here were data of stationary growth phase. (B) 15N2 release from (15NH4)2SO4 (initial amount 100 μmol) in the absence or presence of NaNO2 (medium B and medium C, respectively). (C) Gaseous products of aerobic denitrification of Alcaligenes strain in medium E (60 μmol of initial amount of Na15NO2).

Dirammox Evolves Associated With Members of the Genus Alcaligenes

Figure 3 depicts the results of phylogenetic analyses of the genus Alcaligenes. It is noticed that the topology of the tree based on 16S rRNA gene sequences (Figure 3A, Table 4) was almost the same as that of the tree based on the amino acid sequences of DnfAs (Figure 3B), and both trees can be divided into two clades, namely, clade I and clade II (Figures 3A,B). Dirammox activities of Alcaligenes spp. seemed likely taxon-related, as indicated by the right column of (A) showing the heatmap of conversion (%) of consumed 15NH4+ to 15N2 in medium B, the members in clade I exhibited lower Dirammox activities than those in clade II with the exception of strains CGMCC 1.10269 and CGMCC 1.1799 (Figure 3), both strains had very low or no Dirammox activities. Furthermore, the G + C contents of dnf clusters of Alcaligenes spp. were all around 56.3%, a value very close to that (∼57.2%) of the genomes of Alcaligenes spp. These results, together with the very high sequence similarity of the dnf clusters among Alcaligenes species, significantly referred that the dnf cluster was evolved associated with the members of the genus Alcaligenes. This means that Alcaligenes species obtain dnf cluster and so the Dirammox genetic potential ability from heredity, but not from horizontal gene transfer from other microbes.

FIGURE 3

Phylogenetic analyses of the genus Alcaligenes based on 16S rRNA gene sequences (A) and amino acid sequences of DnfAs (B). The protein accession numbers shown in parentheses represent the best hit of DnfA against NR database (the same in Table 4). The columns on the right of (A) represent the heatmap showing the conversion (%) of consumed 15NH4+ to 15N2 in medium B.

TABLE 4

BlastP results of DnfAs against NR database.

Source strain	Best hit accession number	Identity (%)
A. faecalis subsp. phenolicus DSM 16503T	WP_045930341.1	100
A. faecalis subsp. phenolicus RL12	WP_042487153.1	100
A. faecalis subsp. faecalis CGMCC 1.2006T	WP_042487153.1	100
A. faecalis subsp. faecalis PC01	WP_042487153.1	100
A. faecalis subsp. faecalis CGMCC 1.1837	WP_042487153.1	99.68
A. faecalis subsp. faecalis CGMCC 1.2908	WP_035272004.1	99.37
A. faecalis subsp. faecalis CGMCC 1.3937	WP_035272004.1	100
A. faecalis subsp. parafaecalis DSM 13975T	WP_207875549.1	100
A. faecalis subsp. faecalis CGMCC 1.1799	WP_003803202.1	100
A. sp. CGMCC 1.10269	WP_003803202.1	100
A. ammonioxydans HO-1	WP_003803202.1	100
A. aquatilis GL12	WP_094197465.1	100
A. aquatilis CGMCC 1.9053	WP_094197465.1	100
A. faecalis JQ135	WP_094197465.1	100
A. aquatilis CGMCC 1.0767	WP_125277662.1	100
A. faecalis subsp. faecalis CGMCC 1.2388	WP_125277662.1	100

In vitro Activities of DnfA Homologs Catalyzing the Direct Oxidation of Hydroxylamine to N2

Results of multiple sequence alignment of DnfA and its homologs (DnfAs) from Alcaligenes species showed only a few amino acid differences (Figure 4A), suggesting that the Alcaligenes DnfAs share the same function. In addition, Figure 4A also shows that Alcaligenes DnfAs can be divided into 14 subtypes. Among them, DnfA with accession number WP_042487153.1 was the most abundant one that existed in 18 genomes, followed by DnfA with accession number WP_035272004.1, which existed in five genomes (Figure 4B). To verify the functions of DnfAs, some representatives were selected and their genes were obtained by PCR amplification or gene synthesis for their enzymatic activity assays.

FIGURE 4

Multiple-sequence alignment (A) and statistics (B) of Alcaligenes DnfAs. Accession numbers of the proteins (DnfAs, or DnfA homologs) are shown. All genomes (49) in Figure 1 were analyzed, and five of them with incomplete DnfA sequence were excluded.

It was found that DnfAs could catalyze the direct oxidation of hydroxylamine to N2 with a molecular ratio of 2:1 in the presence of molecular O2, FAD, and NADH. Therefore, in vitro reconstitution enzymatic activity assays using FAD as the chemical electron mediator were performed for DnfAs as shown in Figure 5, according to previous method (Choi et al., 2008). The results (Figure 5B) indicated that all the enzyme reaction samples could release 15N2 from 15NH2OH, in spite of the different amounts of 15N2 released, suggesting that all tested DnfAs could oxidize hydroxylamine directly to N2. These results also showed that the different original DnfAs exhibited different enzyme activity. DnfAs with accession numbers WP_042487153.1 and WP_045930341.1 released 0.63 ± 0.01 and 0.43 ± 0.04 μmol 15N2-N, respectively, showing the lowest activity (Figure 5B). The results also suggested that the in vitro enzymatic activities of the tested DnfAs seemed likely none phylogenetically related, a phenomenon quite different from the Dirammox activities of Alcaligenes strains (Figure 3). Furthermore, when combined with the results presented in Figure 3, an interesting fact was noticed. The Alcaligenes strains with low in vitro DnfA activity also exhibited low Dirammox activity, such as strains RL12, CGMCC 1.2006, PC01, and DSM 16503. In contrast, the strains with low Dirammox activity may harbor DnfA of high in vitro activity, such as strain CGMCC 1.3937.

FIGURE 5

Hydroxylamine oxidation activities of DnfAs from Alcaligenes strains. Phylogenetic tree of the representatives of all DnfA homologs defined in the genomes of all Alcaligenes strains deposited in GenBank (A); 15N2 release of the representatives of DnfA homologs from 15NH2OH under 20% O2 and 80% He in vitro (B). The content of Fe and Mn atoms of DnfAs (C). The data are shown as the mole content of metal atoms per mole of DnfA.

Based on homology alignment against the public database, DnfA was annotated as a diiron oxygenase and related to the AurF N-oxygenase (Choi et al., 2008; Leipoldt et al., 2017) and CmlIs (Knoot et al., 2016; Komor et al., 2018). Similarly, DnfAs are predicted to contain a diiron motif. In this case, atoms of DnfAs were determined. Only Fe and Mn existed in DnfA, and the results (Figure 5C) showed that all the tested DnfAs contained both Fe and Mn atoms with contents ranging from 0.11 ± 0.00 to 1.60 ± 0.01 and from 0.03 ± 0.00 to 0.48 ± 0.00 mmol per mmol of protein, respectively (Figure 5C). Notably, a positive correlationship was defined between the in vitro activities (release amount of 15N2) of DnfAs and the contents of Fe, Mn, or Fe + Mn with correlation coefficient r = 0.92, 0.86, and 0.93, respectively.

Distribution of the Genus Alcaligenes in Environments

The above results indicated that Dirammox pathway was commonly distributed in the genus Alcaligenes, in spite of a few Alcaligenes species exhibiting low or even no ability of transforming ammonia to N2 under 20% oxygen. In this case, the distribution of the genus Alcaligenes in environments was investigated first based on 16S rRNA gene sequences.

For concise purpose, samples were briefly divided into five categories, namely, soil, water, sediment, sludge, and animal–plant-associated, according to their sources. Each category includes samples from both natural environments and human-made surroundings. Clinical samples and those directly derived from animals or plants were classified as animal/plant-associated samples. A total of 824 different Alcaligenes 16S rRNA gene sequences with definite environment/source information were defined in NT database of the GenBank. The results (Figure 6) showed that there were more Alcaligenes 16S rRNA gene sequences (∼35%, 288/824) from animal/plant-associated samples than other samples, followed by soil samples (27.3%, 225/824). Also, there were 11.7, 6.90, and 4.85% (96, 57, and 40 of 824, respectively) of the Alcaligenes 16S RNA gene sequences from samples of water, sludge, and sediment, respectively. These results indicated that members of the genus Alcaligenes are widely distributed in various environments with quite high biodiversity, in good agreement with the results described in the literature as mentioned in the “Introduction” section 1. Therefore, the genus Alcaligenes is an important conductor of elemental biogeochemical cycles, including the nitrogen cycle, and so the Dirammox pathway is probably important to the nitrogen cycle since most members of the genus Alcaligenes could exhibit Dirammox activity (Figure 2B). In addition, approximately 31% (256/824) Alcaligenes strains were derived from WWTPs or various industrial effluents and agricultural wastewater. This means that Alcaligenes might also be efficient candidates for wastewater treatment.

FIGURE 6

Distribution of the genus Alcaligenes in nature. Categories of samples and numbers of Alcaligenes 16S rRNA gene sequences are indicated. The results were based on the analysis of 16S rRNA gene sequences defined in GenBank of NCBI database.

Discussion

In our previous work, a novel microbial N2 production pathway, termed Dirammox, was defined in Alcaligenes strain HO-1 and encoded by the gene cluster dnfT1RT2ABCD (Wu et al., 2021). Recently, dnfABC was further proved to be the genetic basis for Dirammox and regulated by dnfR in another Alcaligenes strain, JQ135 (Xu et al., 2022). Here, it was found that high homogeneity of Dirammox was among the members of the genus Alcaligenes. Our results showed that the production of N2 from ammonia by an Alcaligenes strain must be the results of Dirammox pathway, but not the results of aerobic denitrification that could only emit N2O (Figure 2) under aerobic conditions. The results are quite different from previous studies on nitrogen removal of Alcaligenes strains (van Niel et al., 1992; Joo et al., 2006; Liu et al., 2015; Shoda and Ishikawa, 2016; Chen et al., 2021), which all attributed to N2 producing or nitrogen removal to HNAD. Although it was also speculated that N2 generation in Alcaligenes strains was through hydroxylamine (NH4+ → NH2OH → N2O → N2) (Joo et al., 2006; Zhao et al., 2012), the pathway and molecular basis remained unclear. Here, we proved the use of Dirammox, a process of direct oxidization of ammonia to N2 via hydroxylamine in Alcaligenes strains.

In this study, this gene cluster dnfT1RT2ABCD as a whole was investigated for its distribution in members of the genus Alcaligenes, and it was found that dnf cluster not only existed in all members of the genus Alcaligenes (Figure 1), but also was highly conserved among Alcaligenes species. Phylogenetic data revealed that DnfA was taxon-related (Figure 3), referring to the Alcaligenes species acquiring Dirammox pathway by heredity but not horizontal gene transfer from other microbes, which in turn determines different Dirammox activities among Alcaligenes members in different clades. The function and activity of DnfA are of course important, but other conditions, such as low or even no transcription or expression of dnf genes and the activity of DnfB or DnfC, will also influence the Dirammox activity. So, we can see strains CGMCC 1.10269 and CGMCC 1.1799 breaking the rule, whose Dirammox activities differ from their very closely related neighbors HO-1 (Figure 3).

DnfA, annotated as a diiron oxygenase, contained both Fe and Mn atoms whose content greatly affected the in vitro activities of DnfA catalyzing the direct oxidation of hydroxylamine to N2. It was reported that Mn and Fe have similar atomic numbers and Fe/Fe and Mn/Mn are common active states in many proteins. Although DnfAs were predicted to contain a diiron site, Mn atoms might take up the site in place of Fe atoms. This means that the contents of Fe or Mn or Fe + Mn were a very important factor determining the in vitro activities or even physiological activities of DnfAs, and this needs further investigation in the future. The differences of Fe and Mn atoms among Alcaligenes spp. might result from the sequences of amino acid of DnfA, and the change of some amino acid residues might decrease the combination of protein and Fe and Mn. The reasons for this might be quite complex and need to be further revealed in the future.

With general N2-producing ability of Alcaligenes members and the wide distribution of the genus Alcaligenes in nature as shown above, it was speculated that Dirammox was also important to the nitrogen cycle. Certainly, the Dirammox activity in situ and the assessment of N-fluxes contributed by Dirammox conducted by Alcaligenes or even other bacteria should be investigated urgently in the near future.

Conclusion

In this study, the cluster dnfT1RT2ABCD encoding Dirammox proved to be universally distributed, conserved, and dependently evolved in the genus Alcaligenes. Most of the Alcaligenes strains exhibited Dirammox activities, which seemed likely taxon-related, in spite of their different performance. In contrast, the in vitro activities of DnfAs catalyzing the direct oxidation of hydroxylamine to N2 were not taxon-related but depended on the contents of Fe and Mn ions. Dirammox might be important to the nitrogen cycle, as well as the environment, since the genus Alcaligenes was widely distributed in environment.

Data Availability Statement

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/supplementary material.

Author Contributions

Z-PL and S-JL contributed to conception. L-LM and T-TH designed the study, analyzed data, and wrote the first draft of the manuscript. T-TH and LM contributed to the data collection. Z-PL reviewed and edited the manuscript. J-SP, QH, YL, and M-RW provided resources for the study. G-MA built the N2 and N2O detection methods. All authors contributed to manuscript revision, read and approved the submitted version.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.