Denitrification and Nitrate-Dependent Fe(II) Oxidation in Various Pseudogulbenkiania Strains

Pseudogulbenkiania is a relatively recently characterized genus within the order Neisseriales, class Betaproteobacteria. This genus contains several strains that are capable of anaerobic, nitrate-dependent Fe(II) oxidation (NDFO), a geochemically important reaction for nitrogen and iron cycles. In the present study, we examined denitrification functional gene diversities within this genus, and clarified whether other Pseudogulbenkiania sp. strains perform denitrification and NDFO. Seventy strains were analyzed, including two type strains, a well-characterized NDFO strain, and 67 denitrifying strains isolated from various rice paddy fields and rice-soybean rotation fields in Japan. We also attempted to identify the genes responsible for NDFO by mutagenesis. Our comprehensive analysis showed that all Pseudogulbenkiania strains tested performed denitrification and NDFO; however, we were unable to obtain NDFO-deficient denitrifying mutants in our mutagenesis experiment. This result suggests that Fe(II) oxidation in these strains is not enzymatic, but is caused by reactive N-species that are formed during nitrate reduction. Based on the results of the comparative genome analysis among Pseudogulbenkiania sp. strains, we identified low sequence similarity within the nos gene as well as different gene arrangements within the nos gene cluster, suggesting that nos genes were horizontally transferred. Since Pseudogulbenkiania sp. strains have been isolated from various locations around the world, their denitrification and NDFO abilities may contribute significantly to nitrogen and iron biogeochemical cycles.

The genus Pseudogulbenkiania, which belongs to the family Chromobacteriaceae (1), order Neisseriales, class Betaproteobacteria, was first proposed by Lin et al. (27). Two species have since been reported: Pseudogulbenkiania subflava isolated from a cold spring in Taiwan (27) and Pseudogulbenkiania gefcensis isolated from soil in South Korea (26).
The genus Pseudogulbenkiania also contains the anaerobic, nitrate-dependent Fe(II)-oxidizing bacterium strain 2002 (also called "Pseudogulbenkiania ferrooxidans"), isolated from a freshwater lake in Illinois, USA (42,44), and strain MAI-1, isolated from a freshwater lake in Indonesia (23). Anaerobic, nitrate-dependent Fe(II) oxidation (NDFO) is a process in which ferrous iron (Fe [II]) is oxidized to ferric iron (Fe[II]) coupled with the reduction of nitrate under anoxic, circumneutral conditions (5,9,38,43,45). This reaction is ecologically important (28,29) and has great potential for biotechnological applications such as the bioremediation of toxic metals (11,19,24,32,33). However, Fe(II) oxidoreductase has not yet been identified among NDFO microbes (6), which limits our understanding of this geochemically important biological reaction. Since some Pseudogulbenkiania strains have the potential to be genetically engineered (23), their use is advantageous for the study of NDFO.
We previously isolated 67 Pseudogulbenkiania strains by using a functional single-cell isolation method (3) from rice paddy fields and rice-soybean rotation fields in Kumamoto, Niigata, and Yamagata in Japan (39). These strains showed strong denitrification and N 2 O reduction activities. The findings of a culture-independent RNA-based analysis also suggest that Pseudogulbenkiania spp. strongly contribute to denitrification and N 2 O reduction in rice paddy soils (46). Rice paddy fields are abundant in nitrate and Fe(II) (12,16,35) and NDFO activity has also been detected (12,16,35); therefore, Pseudogulbenkiania spp. may be involved in NDFO in the environment. However, the NDFO ability of Pseudogulbenkiania denitrifiers isolated from rice paddy soils has not yet been examined. Furthermore, relatedness among the Pseudogulbenkiania denitrifiers, NDFO strains 2002 and MAI-1, and other Pseudogulbenkiania species has not been analyzed to date. We targeted denitrification functional genes (nitrite reductase gene [nirS] and nitrous oxide reductase gene [nosZ]) in order to analyze diversity among strains.
The objectives of the present study are (i) to examine the NDFO abilities of Pseudogulbenkiania strains, (ii) to identify the genes responsible for NDFO, and (iii) to analyze denitrification functional gene diversities within the genus Pseudogulbenkiania.

Iron oxidation assays
Standard anaerobic culturing techniques with anoxic grove box (Coy laboratories) with an N 2 :CO 2 :H 2 (80:10:10) atmosphere were used for this experiment. Cells grown on R2A agar were suspended in anoxic basal medium supplemented with 5 mM nitrate in test tubes with butyl rubber stoppers. Basal medium contained the following chemicals (L -1 ): 0.25 g of NH 4 Cl, 0.6 g of NaH 2 PO 4 , 0.1 g of KCl, 2.52 g of NaHCO 3 , and 10 mL each of the trace metal solution and vitamin solution (45). The cells were then incubated under anoxic conditions (N 2 :CO 2 :H 2 = 80:10:10) at 30°C. This incubation was performed in order to deplete carbon in the cells. After a 3-d incubation, cells were pelleted and washed three times with anoxic PIPES (piperazine-N,N'-bis[2-ethanesulfonic acid]) buffer (10 mM, pH 7.0). Washed cells (final conc. 5 × 10 7 cells mL -1 as established by microscopy and OD 600 measurements [15]) were inoculated into anoxic basal medium (20 mL) supplemented with 10 mM FeCl 2 and 5 mM NaNO 3 in test tubes with butyl rubber stoppers. After replacing the air phase of the test tubes with H 2 -free anoxic gas (N 2 :CO 2 = 80:20), cells were incubated at 30°C. Samples (2 mL) were periodically collected using syringe needles. Aliquots (100 μL) of the samples were immediately mixed with 700 μL of 1 M HCl to stabilize Fe(II) and dissolve Fe(II) minerals. Twenty microliters was kept frozen at -20°C for a quantitative PCR (qPCR) analysis (see below). The remainders of the samples (ca. 1.9 mL) were filtered through 0.20-μm pore membrane filters and stored at -20°C until used for ion chromatography.
Fe(II) concentrations were measured spectrophotometrically using the ferrozine method as described by Hegler et al. (10). In order to measure total Fe concentrations (Fe[II] + Fe[III]), Fe(III) was reduced by 50% (w/v in 1M HCl) hydroxylamine hydrochloride prior to measurements by the ferrozine method. Nitrate and nitrite concentrations were measured using an ion chromatograph IC-2010 equipped with the TSKgel SuperIC-Anion HS column (Tosoh).
A high-throughput iron oxidation assay was performed using 96-well plates. In brief, cells grown on R2A agar were suspended in anoxic basal medium (150 μL) supplemented with 5 mM nitrate. After a 3-d anoxic incubation, cells (15 μL) were transferred to anoxic basal medium (135 μL) supplemented with 10 mM FeCl 2 and 5 mM NaNO 3 , and incubated under anoxic conditions (N 2 :CO 2 :H 2 = 80:10:10) at 30°C. Before and after a 1-week incubation, 25 μL of the cell culture was mixed with 175 μL of 1 M HCl solution in order to measure Fe(II) and total Fe concentrations, as described above. Pseudogulbenkiania sp. strain 2002 was used as a positive control.

Testing for autotrophy by NDFO
In an attempt to clarify whether cells have the ability to grow under NDFO conditions, we performed quantitative PCR by using strain-specific primers (IAC_23F and IAC_92R primers; Table S1) (15) and the KOD Sybr qPCR Mix (Toyobo). Culture medium (1 μL) in the NDFO assay was directly used as a template for qPCR. Standard DNA was prepared by diluting the linearized plasmid (15) with basal medium containing the same concentrations of potential PCR inhibitors (i.e., Fe 2+ ) as the samples. The amplification efficiency (E) of the assay was 76% with a linear dynamic range from 20 -2 × 10 6 copies μL -1 .
We also performed stable isotope probing using 13 C-labeled bicarbonate. Cells were prepared as described above, and incubated in anoxic basal medium (20 mL) supplemented with 10 mM FeCl 2 and 5 mM NaNO 3 . All bicarbonate in this medium was labeled with 13 C (Cambridge isotope laboratories). As a control, cells were incubated in medium with 12 C bicarbonate. DNA was extracted after a 9-d incubation by using the PowerSoil DNA Isolation Kit (MoBio Laboratories). Cesium chloride density gradient ultracentrifugation and DNA collection were performed as previously described (13,30). DNA concentrations in each of the density-resolved fractions were quantified using the Quant-iT PicoGreen ds DNA Assay Kit (ThermoFisher Scientific).

Transposon mutagenesis
Conjugation was performed as described by Kim et al. (20). In brief, E. coli pRL27 donor strain WM3064 and the recipient Pseudogulbenkiania sp. strain NH8B were mixed at a 1:4 ratio, and spotted on R2A agar supplemented with DAP (300 μg mL -1 ). After an overnight incubation, cells were suspended in R2A broth and spread onto R2A agar supplemented with kanamycin (100 μg mL -1 ). Colonies grown on this agar were selected and tested for their iron-oxidizing ability, as described above.

DNA extraction, PCR, and Sequencing
DNA was extracted by heating cells in 100 μL of 0.05 M NaOH at 95°C for 15 min (3). After centrifugation, the supernatant was diluted 10-fold and used for PCR as described below.
The PCR reaction mixture (50 μL) contained 1 × Ex Taq buffer (Takara Bio), 0.2 μM of each primer, 0.2 mM of each dNTP, 1 U of Ex Taq DNA polymerase (Takara Bio), and 1 μL of a DNA template. PCR was performed using a Veriti 96-well thermal cycler (Applied Biosystems) with the following conditions: initial annealing at 96°C for 5 min, followed by 35 cycles at 95°C for 30 s, the annealing temperature shown in Table S1 for 30 s, and 72°C for the extension time shown in Table S1. After the final extension at 72°C for 7 min, the PCR mixtures were stored at 4°C. The sizes of the PCR products were verified by agarose gel electrophoresis.
PCR products were purified using the FastGene Gel/PCR Purification Kit (Nippon Genetics) and sequenced using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) and ABI 3730xl capillary sequencer (Applied Biosystems) with the primers shown in Table S1 and Fig. S1.

Transcription analysis
Pseudogulbenkiania sp. strain 2002 and strain NH8B were grown in basal medium (5 mL) supplemented with 10 mM acetate under oxic or anoxic conditions. When grown under anoxic conditions, 5 mM nitrate was added to this medium as an electron acceptor. Total RNA was extracted using the RiboPure Bacteria kit (Ambion). After the DNase treatment, complementary DNA (cDNA) was synthesized using random hexamers and a PrimeScript RT reagent kit (Perfect Real Time) (Takara Bio). The resulting cDNA was used to detect gene transcription by PCR (Table S2). While genomic DNA was used as a positive control, RNA samples were used as negative controls in order to verify the absence of contamination by genomic DNA. Experiments were performed in triplicate (i.e., three test tubes per condition for each strain).

Genomic and phylogenetic analyses
Multiple reads obtained by sequencing reactions were assembled using the Phred-Phrap program (7). The nucleotide sequences of multiple clones were aligned using ClustalW, and used to generate phylogenetic trees. Phylogenetic trees were constructed based on the maximum likelihood method by using MEGA ver. 6 (40).
The genome of Pseudogulbenkiania sp. strain NH8B was previously sequenced (14). The genome sequences of Pseudogulbenkiania sp. strain 2002 (4) and Chromobacterium violaceum ATCC 12472 were retrieved from the GenBank database. These genomes were compared using GenomeMatcher (31).

Anaerobic nitrate-dependent Fe(II) oxidation
Pseudogulbenkiania sp. strain NH8B exhibited NDFO activity: Fe(II) was rapidly oxidized to Fe(III) along with the reduction of nitrate to nitrite (Fig. 1), similar to previous findings obtained with Pseudogulbenkiania sp. strain 2002 (42,44). In contrast, the oxidation of Fe(II) was not observed when no cells or heat-killed cells were added to the test tubes (data not shown). These results indicate that Fe(II) was oxidized as a result of biological activity, either by a direct or indirect reaction.
The amount of nitrate consumed (ΔNO 3 -= -1.89 ± 0.08 mM; Table S4) in the 7-d incubation was larger than the amount of nitrite produced (ΔNO 2 -= 0.76 ± 0.08 mM), indicating that nitrite was further reduced to gaseous nitrogen oxides and dinitrogen. Based on the electron balance of the NDFO reaction (eq.1 and 2), two and five moles of Fe(II) are required to reduce one mole of nitrate to nitrite and to N 2 gas, respectively. The reactions shown in eq. 3-7 proceed abiotically, particularly under acidic conditions (e.g., during the HCl extraction of Fe minerals) (21). Furthermore, NO may be oxidized back to NO 2 under oxic conditions (eq. 8). Although our HCl extraction was performed in the anoxic grove box, the ferrozine assay was conducted under a 20% O 2 atmosphere. Thus, some Fe(II) may have been abiotically oxidized during sample processing. This may partly explain the discrepancy between the rates of Fe(II) oxidation and nitrate reduction: i.e., most Fe(II) was oxidized within the first 2 d, while nitrate was reduced for a longer time period (ca. 4-7 d) in Fig. 1. The accumulation of nitrite may be due to the depletion of electron donors (e.g., Fe[II] and organic substrates stored in cells). Cell growth was not observed during the incubation period (Fig. S2). In addition, 13 C-bicarbonate was not incorporated into nucleic acids based on the SIP analysis (Fig. S3), suggesting that NDFO does not support the autotrophic growth of strain NH8B. Thus, NDFO-dependent growth may differ according to the strains (33,38,44) or test conditions. The oxidation of Fe(II) was also observed in all other Pseudogulbenkiania sp. strains based on our high-throughput iron oxidation assay (Table S3), suggesting that NDFO is a common characteristic among Pseudogulbenkiania sp. strains.
In order to identify the genes associated with Fe(II) oxidoreductase, we performed transposon mutagenesis using strain NH8B. A total of 1,440 mutants were obtained and screened based on NDFO activity. As expected, mutants unable to grow under anoxic conditions on R2A agar supplemented with nitrate did not exhibit NDFO activity. We were unable to obtain mutants that did not oxidize iron, but had the ability to grow by anaerobic nitrate reduction, indicating that mutants disrupted with Fe(II) oxidoreductase-associated genes were not obtained in this study. In addition, the genome of strain NH8B did not contain homologs to mtrAB and pioAB, which are involved in respiratory iron reduction and phototrophic iron oxidation, respectively (17,34). The homologs to mtrAB and pioAB are candidates for a potential Fe(II) oxidoreductase gene.
Similar to the present study, a previous attempt to identify specific Fe(II) oxidoreductase in the NDFO strain, Acidovorax ebreus strain TPSY, using a proteomic approach was not successful (6). These authors showed that respiratory complexes such as nitrate reductase (Nar) directly mediated NDFO, indicating that a specific Fe(II) oxidoreductase is not necessarily required for NDFO (6,24). In addition, biotically or abiotically formed NO 2 or NO may also react with Fe(II) to form Fe(III) oxides (6,21,22). These results suggest that NDFO exhibits an intrinsic ability in all nitrate-reducing bacteria, and may explain the failure to obtain the mutants of interest (i.e., NDFO-incapable, nitrate-reducing mutants). Strong NDFO abilities among Pseudogulbenkiania spp. may depend on their toxicity tolerance of Fe(II) and nitrogen oxides (e.g., NO 2and NO), the extent of the accumulation of nitrogen oxides, and/or the nature of their respiratory proteins (6).
Complete sets of denitrification functional genes were identified on the genomes of the strains NH8B and 2002. Strains NH8B and 2002 both possessed gene clusters for membrane-bound Nar, cytochrome cd 1 -containing nitrite reductase (Nir), nitric oxide reductase (Nor), and nitrous oxide reductase (Nos). While nar-nir-nor operons were located next to each other, the nos operon was distantly located on the genome in strains NH8B and 2002 ( Fig. 2A).
The gene arrangement within the nar-nir-nor cluster was similar between strains NH8B and 2002 (Fig. 2B). In addition, gene sequences were highly conserved between the two strains. For example, the nirS sequence was 98.6% similar between strain NH8B and strain 2002. In contrast, the gene arrangement within the nos cluster was different between the two strains, and nos gene sequence similarities were relatively low (Fig. 2C). These results suggest that one of the two nos clusters was horizontally acquired. Further studies (e.g., comparative genome analysis with many other strains) may be necessary in order to obtain concrete evidence of the horizontal gene transfer event.
The nirS sequences from all Pseudogulbenkiania sp. strains were relatively highly conserved among the 70 strains, with >90.4% similarity (367 bp). Pseudogulbenkiania sp. strains isolated from different locations were clustered together based on their nirS sequences (Fig. S6).
All 15 NH8B-type strains had a nosZ-like pseudogene similar to strain NH8B, based on NH8B_3641-specific qPCR (15) and an analysis of the sequences downstream of nosZ.
The intergenic region between nosZ and the nosZ-like pseudogene was relatively well conserved among the 15 strains (Fig. S7). The intergenic region is known to be generally more variable than protein-coding sequences (8). The highly conserved intergenic regions among the 15 strains indicate that these strains may have diversified relatively recently from a common ancestor, and were then distributed to different locations.
In this region, we detected a FNR-binding motif (FNR box; TTGAT----ATCAA), which is recognized by FNR-like transcription regulators and reported to be essential for the transcription of nosZ in Pseudomonas aeruginosa (2). We detected the transcription of nosZ and nosZ-like pseudogenes in the cells of Pseudogulbenkiania sp. strain NH8B under oxic and anoxic conditions ( Fig. S8A and S8B). In contrast, the transcription of nosZ was only detected in the cells of Pseudogulbenkiania sp. strain 2002 (Fig. S8C) grown under anoxic conditions, suggesting that transcription regulation may differ between strains NH8B and 2002. The role of FNR-like transcription regulators in the transcription of nosZ and related proteins is currently unknown. Since Pseudogulbenkiania sp. strain NH8B reduces exogenous N 2 O (14), a deeper understanding of the transcription mechanism of nosZ will be useful for mitigating N 2 O emissions.

Conclusions
Our comprehensive analysis showed that diverse strains within the genus Pseudogulbenkiania perform denitrification and NDFO. The failure to detect NDFO-deficient denitrifying mutants in our mutagenesis experiment suggests that NDFO is linked to nitrate-reducing activity (i.e., abiotic Fe[II] oxidation by reactive nitrogen species formed during nitrate reduction) without the involvement of specific Fe(II) oxidoreductase. The NDFO abilities of Pseudogulbenkiania spp. may depend on their toxicity tolerance of Fe(II) and nitrogen oxides (e.g., NO 2and NO), the extent of the accumulation of nitrogen oxides, and/or the nature of their respiratory proteins (6). Different gene arrangements within the nos cluster and low sequence similarity within nos gene sequences among Pseudogulbenkiania sp. strains indicate the occurrence of horizontal gene transfer. Since Pseudogulbenkiania sp. strains have been isolated from various locations (e.g., agricultural soils and freshwater lakes) around the world, their denitrification and NDFO abilities may contribute significantly to nitrogen and iron biogeochemical cycles.