1 Introduction

Denitrification is an important microbial process in which nitrate (NO3 −) undergoes dissimilatory reduction to nitrogen gas (N2) in the following sequential reactions: NO3 − → NO2 − → NO → N2O → N2 (Xu and Enfors 1996). Most denitrifying microorganisms are aerobic heterotrophic organisms that transfer redox equivalents from the oxidation of a carbon source to an N oxide under anoxic conditions (Zumft 1997). As the oxidation of ferrous iron (Fe(II)) is coupled to the denitrification process (Straub et al. 1996), this phenomenon has been discovered in various habitats, such as freshwater sediments, costal marine sediments, and paddy soils (Muehe et al. 2009; Laufer et al. 2015; Li et al. 2016). The interaction between denitrification and Fe(II) oxidation not only affects the biogeochemistry transformation of nitrogen and iron (Mejia et al. 2016; Wang et al. 2016) but also influences the fate of other elements, such as carbon and heavy metals (Yu et al. 2017).

Several bacteria have been reported to couple nitrate reduction and Fe(II) oxidation in pure cultures, such as Pseudogulbenkiania sp. (Weber et al. 2006) and Acidovorax sp. (Kappler et al. 2005; Byrne-bailey et al. 2010; Chakraborty et al. 2011). There is evidence that the capacity for nitrate-reducing Fe(II) oxidation is widespread and likely innate to the denitrifying bacteria, such as Paracoccus denitrificans and Pseudomonas stutzeri (Muehe et al. 2009; Klueglein et al. 2014). The intermediate nitrite (NO2 −) in the denitrification pathway can chemically oxidize Fe(II) (chemodenitrification), producing nitric oxide (NO) and nitrous oxide (N2O) gases along with Fe(III) (hydr)oxides as byproducts (Carlson et al. 2013; Melton et al. 2014). However, the contribution of the microbial and chemical reactions in the observed nitrate-reducing Fe(II) oxidation remains unclear. As the number of denitrifying bacteria in the environment is significantly higher than that of nitrate-reducing Fe(II) oxidizers (Straub and Buchholz-cleven 1998; Muehe et al. 2009), it is important to investigate how microbial and chemical reactions are involved in nitrate reduction and Fe(II) oxidation under denitrifying conditions.

Stable isotope analysis of N2O (particularly the 15N–isotopomer) can be useful to disentangle the different pathways leading to N2O formation in natural environments and pure cultures (Toyoda et al. 2005; Sutka et al. 2006; Heil et al. 2014). Regarding the microbial pathways, the different isotopic compositions of N2O in the denitrification and nitrification processes have been used to evaluate their relative contribution in N2O production (Sutka et al. 2006; Wunderlin et al. 2012). For chemical pathways, the characteristic isotopic composition of N2O involved in chemodenitrification has been recently reported (Jones et al. 2015). The δ15N values of N2O were utilized to distinguish between chemical and biological reduction of NO2 − by an iron-reducing bacterium in which the N2O produced during NO3 − reduction in the presence of goethite was primarily of abiotic origin (Cooper et al. 2003). In this study, the δ15N in N2O was characterized to determine the N2O origin during Fe(II) oxidation and nitrate reduction by denitrifying bacteria.

A variety of Fe(III) (hydr)oxide minerals can be formed during nitrate-reducing Fe(II) oxidation, including ferrihydrite (Lack et al. 2002), goethite (Senko et al. 2005), lepidocrocite (Larese-Casanova et al. 2010), magnetite (Chaudhuri et al. 2001), green rust (Chaudhuri et al. 2001), and vivianite (Miot et al. 2009). The Fe(III) mineral type is strongly dependent on the geochemical solution conditions, such as carbonate, phosphate, pH, and humic acids (Larese-Casanova et al. 2010). Formation of green rust as an intermediate mineral (followed by goethite formation as final Fe(III) product) and cell encrustation have been characterized due to abiotic Fe(II) oxidation by NO2 − from heterotrophic denitrification (Nordhoff et al. 2017). Chemodenitrification can also produce ferrihydrite, lepidocrocite, and goethite which, in combination with the characteristic isotopic fractionation, is proposed to be useful to distinguish qualitatively between microbially and chemically emitted N2O (Jones et al. 2015).

In this study, a denitrifying bacterium, Pseudomonas stutzeri LS-2, isolated from red paddy soil in southern China was used to investigate the kinetics of nitrate reduction and Fe(II) oxidation under denitrifying conditions. Stable isotope fractionation of δ15N-N2O, as well as Fe(III) mineral formation, was characterized to clarify how chemical and biological reactions were involved in the NO3 − reduction and Fe(II) oxidation mediated by P. stutzeri LS-2.

2 Materials and methods

2.1 Soil sampling and isolation of the strain

P. stutzeri strain LS-2 was isolated from paddy soil in Shantou City, Guangdong Province, China (23° 38′ 30.4″ N, 116° 50′ 4.7″ E). The geochemical properties of the paddy soil are as follows: pH 7.3, organic matter: 13.1 g kg−1, total As: 0.301 g kg−1, total Fe: 30.5 g kg−1, DCB-extractable (dithionite-citrate-bicarbonate) Fe: 17.3 g kg−1, amorphous Fe: 0.403 g kg−1. The soil sample was collected 10–20 cm below the soil surface and stored in an anaerobic sterile-sealed container. Upon arrival to the laboratory, the soil sample was cultured in 100-mL sterilized sealed bottles with butyl rubber stoppers containing sterile denitrifier basal medium (DBM) with a ratio of soil to DBM at 1:10 for enrichment. The bottle headspaces were filled with N2 and the enrichment cultures were incubated in an anaerobic workstation at 30 ± 1 °C in the dark. By regular subculturing at 7-day intervals, 10% (v/v) inoculum was transferred to fresh DBM medium. The enriched populations were serially diluted and streaked on the DBM agar plates for isolation. Distinct colonies were picked and tested in the DBM medium to confirm their ability to reduce NO3 −. The DBM medium contained 30 mM PIPES [piperazine-N, N`-bis (2-ethanesulfonic acid)] buffer (pH 7.0 ± 0.2), 10 mM NaNO3, 5 mM sodium acetate, and 1 mL L−1 each of the trace element solution and vitamin solution (Bruce et al. 1999). The trace element and vitamin solutions were filtered using a 0.22 μm filter and other medium was autoclaved at 120 °C for 25 min before use.

The genomic DNA of the strain was extracted using the EZNA™ Bacterial DNA Kit (Omega Bio-Tek, Inc., Norcross, GA, USA) after the manufacturer’s instructions. The 16S rRNA gene was amplified by PCR using universal bacterial primers (27F, 5-AGAGTTTGATCCTGGCTCAG-3 and 1492R, 5-GGTTACCTTGTTACGACTT-3) as described by Delong (1992). The amplified PCR product was tested by gel electrophoresis and purified using a DNA Gel Extraction Kit (Omega Bio-Tek, Inc., Norcross, GA, USA). The purified PCR product was cloned using the pGEM-T Vector System (Promega, Madison, WI) and transformed into E. coli JM109 competent cells. Positive clones were selected randomly and sequenced by Sangon Biotech (Shanghai, China). The sequences were searched against the GenBank database (http://blast.ncbi.nlm.nih.gov/) to determine the closest matches using the BLAST program. A phylogenetic tree was constructed by using the neighbor-joining (NJ) algorithm in the MEGA 6.0 program.

2.2 Fe(II) oxidation and nitrate/nitrite reduction experiments

The isolated strain was cultured in LB medium for 12 h in a shaker at 180 rpm and 30 °C. The LB medium contained (g L−1): tryptone, 10; yeast extract, 5; NaCl, 10; pH 7.0 ± 0.2. The strain was harvested by centrifugation at 10,000 rpm for 10 min at 4 °C following being washed and re-suspended with sterile and anoxic PIPES buffer (30 mM, pH 7.0 ± 0.2) for three times. A total of seven treatments were conducted, and the details are presented in Table 1. The initial concentrations of FeCl2·4H2O, NaNO3, NaNO2, and acetate were 5, 10, 4, and 5 mM, respectively. The cell density was approximately 3.8 × 108 cells mL−1. PIPES buffer (30 mM) was added to maintain the pH at 7.0 ± 0.2. The serum bottles were purged with N2 for 30 min, sealed with butyl rubber stoppers and aluminum crimp seals, and incubated at 30 ± 1 °C in the anaerobic workstation in the dark. NO3 −, NO2 −, N2O, and Fe(II) concentrations were measured at intervals. The kinetics constants of Fe(II) oxidation and NO3 −/NO2 − reduction were calculated based on the first-order reaction formulas k t = ln(C0/Ct).

Table 1 Experimental setups of the treatments
Full size table

2.3 Measurements of nitrate, nitrite, N2O, Fe(II)

To determine NO3 − and NO2 − concentrations, a sample aliquot was exposed to O2 to oxidize Fe(II) rapidly followed by filtration using a syringe filter containing a 0.22 μm mixed cellulose ester membrane. NO3 − and NO2 − were quantified by ion chromatography with conductivity detection (Dionex ICS-90 fitted with an AS-4A column; Dionex Corp., Sunnyvale, CA). A mobile phase containing Na2CO3 (8 mM) and NaHCO3 (1 mM) was operated at a flow rate of 1 mL min−1. N2O was quantified by gas chromatography (GC7900, Techcomp, Shanghai) on a packed column [1 m (L) × 3 mm (OD) × 2 mm (ID); 80/100 mesh Porapak] with an electron capture detector. Ferrous iron was quantified photometrically with ferrozine (Lovley and Phillips 1987).

2.4 Stable isotope methods

Headspace N2O produced from NO3 − or NO2 − reduction was dispensed into duplicate gastight serum vials for isotope analysis. 15N/14N was quantified by gas chromatography combustion isotope ratio mass spectrometry (GC-IRMS, MAT-253, Finnigan, Bremen, Germany). Experiments were conducted under a He atmosphere in specially fabricated Pyrex media bottles (500 ml) equipped with a vacuum stopcock and crimp-sealed sampling port. N2O after pre-concentration on line by a PreCon instrument (ThermoFisher Scientific, USA) were processed further in a modified PreCon unit, where H2O and CO2 were removed through a long glass tube packed with MgClO4 and NaOH. Then, purified N2O was introduced into the mass spectrometer for isotope ratio monitoring. The isotopic composition of nitrogenous material is commonly expressed as a delta value:

$$ {\updelta}^{15}\mathrm{N}=\left({\mathrm{R}}_{\mathrm{sample}}/{\mathrm{R}}_{\mathrm{standard}}\hbox{-} 1\right)\times 1000 $$
(1)

where Rsample and Rstandard are the 15N/14N of the sample and standard, respectively. The standard used for reporting 15N/14N ratios is atmospheric N2 (15Rair N2 = 0.0036765) (Jones et al. 2015).

δ15N of NO3 − and NO2 − nitrogen isotopes were determined using azide method (Mcilvin and Altabet 2005). Prior to NO3 − nitrogen isotope analyses, NO3 − was reduced to NO2 − by activated cadmium powder. The NO2 − was dispensed into duplicate gastight serum vial and purged with N2 gas, removing any N2O in the sample produced during the reaction prior to the addition of sodium azide. Finally, the N2O produced from NO2 − during the azide step was then analyzed by GC-IRMS.

2.5 Electron microscopy

After 6 days of incubation, the cell-mineral precipitates formed after Fe(II) oxidation were collected for electron microscopy imaging. For scanning electron microscopy (SEM), the precipitates were washed with phosphate buffer (10 mM, pH 7.5) and were fixed with 2.5% glutaraldehyde at 4 °C overnight. The samples were later dehydrated using a series of ethanol dilutions (30, 70, 95, and 100% [twice]) and fixed on cover slips (dried on a molecular sieve) (Klueglein et al. 2014). The cover slips were mounted onto SEM stubs via clear double-sided sticky tape and gold-coated. The samples were observed with SEM/EDS (Merlin, Zeiss, Germany) using 5 kV accelerating voltage and at a working distance of 8.7 mm. The EDS spectra provided a primary mineral composition (Zhao et al. 2013).

For transmission electron microscopy (TEM) analyses, the precipitates were fixed for 2 h in 2.5% glutaraldehyde at 4 °C, centrifuged (5000 rpm, 10 min), and rinsed three times in phosphate buffer (10 mM, pH 7.5) for 12 h at 4 °C. The samples were post-fixed for 90 min in 1% OsO4 in the same buffer, rinsed three times in distilled water, dehydrated in graded ethanol (30, 70, 95, and 100% [once]) and propylene oxide-1,2 and progressively embedded in epoxy resin (Epoxy, Sigma) (Klueglein et al. 2014). Ultrathin sections (40-nm thick) were cut with an ultramicrotome (UCT, Leica, Germany). The ultrathin sections were fixed on copper grids and were observed with TEM (Tecnai 12, FEI, Netherlands) using 80 kV accelerating voltage.

2.6 X-ray diffraction (XRD) analysis

After 6 days of incubation, the Fe(III) precipitates were centrifuged for 15 min at 10,000 rpm and subsequently freeze-dried for 48 h at − 60 °C. Minerals were identified with a Bruker D8 Advance X-ray diffraction instrument (Bruker AXS GmbH, Germany) equipped with a Co KX-ray tube and operating at 40 kV and 40 mA. The MDI Jade 7 software was used for mineral phase identification.

3 Results

3.1 Identification of the strain

The strain designated LS-2 was a gram-negative, facultative anaerobic bacterium. Cells were rod-shaped with a length of 1–2 μm and had monopolar flagella (Fig. 1a). A BLAST analysis of the 16S rRNA gene revealed that strain LS-2 belonged to the gamma subclass of Proteobacteria and showed the closest relationship with the denitrifying bacterium Pseudomonas stutzeri ATCC 17588T (AF094748), with a 99.2% similarity (Fig. 1b). Therefore, the strain was identified as Pseudomonas stutzeri LS-2 and has been deposited in the China General Microbiological Culture Collection Center (CGMCC 11556). The 16S rRNA sequences of the strain LS-2 were deposited in GenBank under the accession number KY274147.

Fig. 1
figure 1

a TEM image of cell of strain Pseudomonas stutzeri LS-2 grown anaerobically with 5 mM acetate and 10 mM nitrate (scale bar 500 nm). b A neighbor-joining tree based on 16S rRNA sequences shows the phylogenetic affiliation of P. stutzeri strain LS-2. Bootstrap values were determined based on 1000 replicates. The numbers in parentheses are the accession numbers. The scale bar represents a 2% sequence difference

Full size image

3.2 Kinetics of Fe(II) oxidation during nitrate reduction

To investigate microbial NO3 − reduction by P. stutzeri LS-2 in the presence of Fe(II), four treatments were carried out (i.e., Fe(II) + NO3 −, Cell + Fe(II) + NO3 −, Cell + NO3 −, and Cell + Fe(II)) (Table 1). Fe(II) was completely oxidized within 3 days in the Cell + Fe(II) + NO3 − treatment (Fig. 2a) with a rate of Fe(II) oxidation of 1.2 d−1 (Table 2). Neither Fe(II) oxidation nor NO3 − reduction was observed in the treatment of Cell + Fe(II) or Fe(II) + NO3 −. NO3 − was completely reduced within 2 days in the Cell + NO3 − treatment, while NO3 − reduction was slowed down in the Cell + Fe(II) + NO3 − treatment with only 21% of the NO3 − reduced at the end of incubation (day 6) (Fig. 2b). The rates of NO3 − reduction were 3.5 and 0.026 d−1 in the treatments of Cell + NO3 − and Cell + Fe(II) + NO3 −, respectively (Table 2). Most of the NO3 − consumed was transferred to NO2 − in which the NO2 − at the end of incubation was accumulated up to 7.2 mM and 1.9 mM in the treatments of Cell + NO3 − (72% of the NO3 − consumed) and Cell + Fe(II) + NO3 − (85% the NO3 − consumed), respectively (Fig. 2c). The N2O concentration increased over time and reached a maximum value of 0.02 mM and 0.04 mM in the Cell + NO3 − and Cell + Fe(II) + NO3 − treatments on day 4, respectively (Fig. 2d). The above results suggested that Fe(II) oxidation only occurred in the presence of both bacterial cells and nitrate. NO3 − could be reduced quickly and completely by P. stutzeri LS-2 and the microbial reduction was inhibited by the presence of Fe(II). Fe(II) also facilitated the N2O production.

Fig. 2
figure 2

Concentrations of a dissolved Fe(II), b NO3 −, c NO2 −, and d N2O with time dependence in the treatments of Fe(II) + NO3 −, Cell + Fe(II) + NO3 −, Cell + NO3 −, and Cell + Fe(II). Data were presented as the means ± standard deviations (SD) of triplicate

Full size image
Table 2 The first-order rate constants of Fe(II) oxidation and NO3 −/NO2 − reduction in different treatments
Full size table

3.3 Kinetics of Fe(II) oxidation during nitrite reduction

Three treatments were set up (i.e., Fe(II) + NO2 −, Cell + Fe(II) + NO2 −, and Cell + NO2 −) to investigate the Fe(II) oxidation and NO2 − reduction by P. stutzeri LS-2. Similar rates of Fe(II) oxidation were observed in both Fe(II) + NO2 − (0.32 d−1) and Cell + Fe(II) + NO2 − (0.31 d−1) treatments (Fig. 3a and Table 2). At the end of incubation, 87−90% of the Fe(II) was oxidized in these two treatments. In the Cell + NO2 − treatment, NO2 − was completely reduced at a rate of 1.78 d−1 within 2 days (Fig. 3b and Table 2). However, the NO2 − reduction rate was only 0.13 and 0.15 d−1 in the Fe(II) + NO2 − and Cell + Fe(II) + NO2 − treatments, respectively (Table 2), in which 53–58% of the NO2 − were reduced (Fig. 3b). In the Cell + NO2 − treatment, the concentration of N2O was 0.03 mM on day 1, and then decreased to 0 mM during days 2–6. The N2O concentration increased over time in a similar trend in the Fe(II) + NO2 − and Cell + Fe(II) + NO2 − treatments, both of which produced 0.15 mM at the end of incubation (Fig. 3c). The above results indicated that the presence of Fe(II) inhibited the NO2 − reduction by P. stutzeri LS-2, but the Fe(II) oxidation and NO2 − reduction were not affected by the presence of P. stutzeri LS-2.

Fig. 3
figure 3

Concentrations of a dissolved Fe(II), b NO2 −, and c N2O with time dependence in the treatments of Fe(II) + NO2 −, Cell + Fe(II) + NO2 −, and Cell + NO2 −. Data were presented as the means ± standard deviations (SD) of triplicate

Full size image

3.4 Stable isotope fractionation of δ15N-N2O

The results in Fig. 4 display the changes of δ15N-N2O with time dependence in the treatments with N2O production during NO3 − or NO2 − reduction. The δ15N values of NO3 − and NO2 − supplied in this study are − 0.2 and − 7.3‰, respectively. The values of δ15N-N2O decreased from + 7.8 to − 9.9‰ over time in the microbial treatments of Cell + NO2 − and Cell + NO3 −. Those values were constant at − 16 ± 0.64 ‰ in the chemical treatment of Fe(II) + NO2 − during the time course of experiment, which was clearly different from those of the microbial treatments. The δ15N-N2O values in the Cell + Fe(II) + NO2 − treatment increased gradually from − 23‰ on day 1 to − 18‰ on day 6, which was close to those values of the Fe(II) + NO2 − treatment. This finding suggested that the N2O produced from the Cell + Fe(II) + NO2 − treatment was primarily of chemical origin. In the Cell + Fe(II) + NO3 − treatment, the δ15N-N2O values increased from − 37‰ on day 1 to − 25‰ on day 6, which was different from those of microbial or chemical treatments.

Fig. 4
figure 4

Evolution of isotopic fractionation δ15N in N2O during NO3 − or NO2 − reduction in the treatments of Cell + NO2 −, Cell + NO3 −, Fe(II) + NO2 −, Cell + Fe(II) + NO2 −, and Cell + Fe(II) + NO3 −. The δ15N values of NO3 − and NO2 − supplied in this study are − 0.2 and − 7.3‰, respectively (dash lines). Data are presented as the means ± standard deviations (SD) of triplicate reactions

Full size image

3.5 Fe(III) mineral formation and cell encrustation

Mineral precipitates were only formed in the three treatments with Fe(II) oxidation (i.e., Cell + Fe(II) + NO3 −, Cell + Fe(II) + NO2 −, and Fe(II) + NO2 −), all of which showed regular lath-like structures based on the SEM images in Fig. 5a, b, c. The EDS data indicated the presence of iron and oxygen in the minerals (Table S1, Electronic Supplementary Material). The XRD patterns showed that a classical lepidocrocite (γ-FeOOH) phase was formed in the three treatments (Fig. 5f). The intensities of the characteristic diffraction peaks of γ-FeOOH in the chemical treatment of Fe(II) + NO2 − were significantly lower than those of the Cell + Fe(II) + NO3 − and Cell + Fe(II) + NO2 − treatments, suggesting that the γ-FeOOH formed in the Fe(II) + NO2 − treatment had a lower level of crystallinity compared with the other two treatments (Fig. 5f). In the Cell + Fe(II) + NO3 − and Cell + Fe(II) + NO2 − treatments, the bacterial cells were fully covered by the Fe(III) minerals as indicated by the arrows in Fig. 5a, b. The TEM images of cell-mineral precipitates showed that the cell surface was encrusted with Fe(III) minerals and there was a mineral layer within the periplasm, as well (arrow pointing in Fig. 5d, e).

Fig. 5
figure 5

Scan electron microscopic (SEM) images (a–c), transmission electron microscopic (TEM) images (d and e), and X-ray diffraction (XRD) spectra of the Fe mineral precipitates (f) from the treatments of Cell + Fe(II) + NO3 −, Cell + Fe(II) + NO2 −, and Fe(II) + NO2 − after 6 days of inoculation. Yellow arrows indicate LS-2 cells were surrounded by minerals, and blue arrows indicate mineral precipitations were observed in the periplasm

Full size image

4 Discussion

4.1 Effect of Fe(II) oxidation on nitrate reduction by P. stutzeri LS-2

P. stutzeri is among the most active denitrifying heterotrophic bacteria and possesses all the enzymes involved in the four successive denitrification pathway steps (Lalucat et al. 2006). After complete NO3 − reduction by P. stutzeri LS-2 in the Cell + NO3 − treatment, 72 and 0.2% of the NO3 − consumed was determined to be NO2 − and N2O, respectively. NO2 − accumulation is a widespread phenomenon observed during the denitrification process mediated by P. stutzeri, which is considered to be the result of an unbalanced NO3 − and NO2 − reduction (Xu and Enfors 1996). In previous studies regarding Fe(II) oxidation coupled to NO3 − reduction by denitrifying bacteria, complete NO3 − reduction was observed by P. denitrificans strain ATCC19367 or strain Pd 1222 (Klueglein et al. 2014), while only 17–21% of the added NO3 − was consumed by P. stutzeri ATCC 17588 or P. denitrificans ATCC 17741 (Muehe et al. 2009). In this study, the presence of Fe(II) resulted in an inhibition on NO3 − reduction by P. stutzeri LS-2 (Fig. 2b), indicating that Fe(II) oxidation may restrain nitrate reductase activity or block NO3 − reduction. Accumulation of NO2 − was up to 85% in the Cell + Fe(II) + NO3 − treatment (Fig. 2c), which was suggested to be a result of encrustation of the nitrite reductase in the periplasm (Klueglein and Kappler 2013). However, there was no NO2 − accumulation during the nitrate reduction and Fe(II) oxidation by P. stutzeri ATCC 17588 or P. denitrificans ATCC 17741 (Muehe et al. 2009). The effect of Fe(II) oxidation on NO3 − reduction and NO2 − accumulation by denitrifying bacteria appears to vary on a case by case basis.

N2O is commonly produced during chemodenitrification as well as Fe(II) oxidation and nitrate reduction by denitrifying bacteria (Muehe et al. 2009; Jones et al. 2015). The Fe(II) oxidation strongly affects the N2O emission via donating electrons to denitrification in paddy soils (Wang et al. 2016). The δ15N-N2O values on the first day varied among different treatments (Fig. 4), which may be due to the difference in N transformation. In Cell + NO2 − treatment, 85% of the NO2 − added was consumed on day 1, with only 0.03 mM of N2O formed. Based on N mass balance considerations, about 3.4 mM of the reduced NO2 − was transformed as NO and/or N2. Generally, 14N reacts more quickly relative to 15N (Barford et al. 1999). Hence, it is speculated that N2 was the dominant product from NO2 − reduction in Cell + NO2 − treatment, resulting in relatively high δ15N accumulation in N2O. Meanwhile, in Cell + NO3 − treatment, only 6% of the NO3 − added was consumed and 90% of the consumed NO3 − was present as NO2 −, with no formation of N2O and only 0.1 mM of NO/N2. Its δ15N-N2O value is close to that of δ15N in NO3 −, which is probably due to the fact that most of the N is present as NO3 − and the yield of N intermediates was limited. The concentration of N2O was less than 0.02 mM on day 1 in Fe(II) + NO2 −, Cell + Fe(II) + NO2 −, and Cell + Fe(II) + NO3 − treatments, resulting in 0.3 mM, 1.4 mM, and 0.4 mM transferred as NO/N2, respectively. It is reported that NO is a heavy pool of N accumulation and 14N is preferentially reacted to N2O (Toyoda et al. 2005; Jones et al. 2015). Accordingly, the relatively low δ15N accumulation in N2O in these three treatments suggests that NO is likely the dominant product at this time interval.

The δ15N-N2O in the two microbial treatments (i.e., Cell + NO3 − and Cell + NO2 −) decreased since day 1–2, which is possibly attributed to that the NO3 −/NO2 − added was completely reduced and further N transformation was ceased, resulting in a balanced δ15N-N2O value close to that of δ15N in NO2 − during the rest of incubation. The δ15N-N2O in our chemical treatment of Fe(II) + NO2 − (− 16 ± 0.64 ‰) is within the wide range of δ15N-N2O during the chemodenitrification (− 16.8 to 0.2‰) as reported previously (Jones et al. 2015). The analogous reaction rates of NO2 − reduction and Fe(II) oxidation (Table 2), in combination with the similar values of δ15N-N2O (Fig. 4) between the Fe(II) + NO2 − and Cell + Fe(II) + NO2 − treatments, indicated that the N2O produced during NO2 − reduction and Fe(II) oxidation in the presence of cells was mainly due to the chemodenitrification process between NO2 − and Fe(II). Similarly, the N2O produced from the NO3 − reduction by Shewanella putrefaciens in the presence of iron oxide is primarily of chemical origin per the stable isotope studies of δ15N in N2O (Cooper et al. 2003). The slight increase in δ15N-N2O of Cell + Fe(II) + NO2 − treatment may be caused by a further reduction from N2O to N2 in which 14N is preferentially reacted (Barford et al. 1999).

In the Cell + Fe(II) + NO3 − treatment, the presence of Fe(II) facilitated N2O production during denitrification by P. stutzeri LS-2 (Fig. 2d), and its δ15N-N2O values were different from those of microbial or chemical treatments (Fig. 4). Previous study blocks the N2O reduction process using an acetylene inhibition technique and observes negative δ15N-N2O values (from − 37 to − 29‰) during denitrification (Toyoda et al. 2005). Fe(III) precipitation occurred in the periplasm (Fig. 5d), by which both periplasimc NO2 − and N2O reductases can be easily coated and inactivated (Kappler et al. 2005; Carlson et al. 2013), resulting in accumulation of NO2 − and N2O (Fig. 2c, d). As such, the increase in N2O in the Cell + Fe(II) + NO3 − treatment should be from NO reduction which can be mediated via the cytoplasmic-membrane bound NO reductase and chemodenitrification. If this is the case that N2O reduction is blocked by the Fe(III) precipitation, similar negative δ15N-N2O could be obtained as well, and further N2O formation from NO with highly accumulated 15N could also lead to an increase in δ15N-N2O in some extent.

4.2 Effect of nitrate reduction on Fe(II) oxidation by P. stutzeri LS-2

The added Fe(II) was completely oxidized within 3 days during the denitrification by P. stutzeri LS-2 (Fig. 2a). Similarly, complete Fe(II) oxidation has also been reported during NO3 − reduction by the denitrifier Paracoccus denitrificans (Klueglein et al. 2014), while incomplete Fe(II) oxidation was observed in the nitrate-reducing Fe(II) oxidation culture of other denitrifiers such as Pseudomonas stutzeri, Nocardioides sp., and Rhodanobacter sp. (Muehe et al. 2009; Nordhoff et al. 2017). Such a discrepancy may be associated with a difference in bacterial species, cell metabolic states, Fe(II) and NO3 − initial concentrations, and incubation medium compositions used in different studies, which also leads to a difference in Fe(III) mineralization and cell encrustation (Larese-Casanova et al. 2010; Nordhoff et al. 2017).

Several studies have reported nitrate-reducing Fe(II) oxidation in pure Pseudomonas sp. culture (Muehe et al. 2009; Su et al. 2015; Zhang et al. 2015); however, few studies characterize the iron mineralization after Fe(II) oxidation. Lepidocrocite formation appears to be favorable in a buffer without any iron-complexing anions such as MOPS buffer (Larese-Casanova et al. 2010), while goethite is the primary Fe(III) phase formed in a carbonate buffer even with the same bacterial strain (Klueglein et al. 2014; Senko et al. 2005). This result can explain the formation of lepidocrocite in this study since all the experiments were carried out in the PIPES buffer without any iron-complexing anions. It has been demonstrated that the Fe(III) mineralogy is mainly controlled by the geochemical condition of the solution but not the mode of Fe(II) oxidation (biotic or abiotic) (Larese-Casanova et al. 2010). These results support our finding that lepidocrocite was formed in all the treatments with Fe(II) oxidation occurring in the same medium (Fig. 5f). Treatments of Cell + Fe(II) + NO3 − and Cell + Fe(II) + NO2 − showed relatively higher crystallinity of lepidocrocite than Fe(II) + NO2 −, indicated that the presence of P. stutzeri LS-2 probably favors the formation of highly crystalline Fe(III) minerals.

The chemodenitrification reactions between Fe(II) and NO2 − in a PIPES buffer can generate a variety of Fe(III) (hydr)oxide minerals including ferrihydrite, lepidocrocite and goethite (Jones et al. 2015). The experiment with equimolar concentrations of Fe(II) and NO2 − results in a nearly equal proportion of the three Fe(III) minerals after 2 days of reaction (Jones et al. 2015). Only lepidocrocite was detected after a 6-day incubation in our chemical treatment of Fe(II) + NO2 − with initial concentrations of Fe(II) and NO2 − of 5 and 4 mM, respectively (Table 1). This result may be caused by the difference in the composition of reaction solution and the reaction time.

Fe(II) oxidation in the presence of denitrifying bacteria is considered to be driven by the intermediates (mainly NO2 −) produced during denitrification (Muehe et al. 2009; Klueglein et al. 2014; Nordhoff et al. 2017); this is because accumulation of NO2 − in these cultures is commonly observed (Fig. 2) and the chemodenitrification reaction is rapid even without catalysts (Fig. 3) (Klueglein and Kappler 2013; Jones et al. 2015). The chemical oxidation of Fe(II) by NO2 − in denitrifying bacterial culture does not allow control over the location of Fe(III) precipitation and even occurs in the periplasm where the NO2 − was formed after NO3 − reduction (Lalucat et al. 2006; Miot et al. 2009; Nordhoff et al. 2017). The cell encrustation on the cell surfaces and in the periplasm in the Cell + Fe(II) + NO3 − treatment (Fig. 5d, e) suggested that chemodenitrification contributed to the Fe(II) oxidation. Because of encrustation, the mineral coatings on the cell surface may hinder NO3 − transport and nutrient uptake into the cell and may eventually result in inhibition of cell growth (Klueglein et al. 2014). This can be supported by the substantially lower concentrations of cell protein observed in Cell + Fe(II) + NO3 − and Cell + Fe(II) + NO2 − treatments, relative to those of Cell + NO3 − and Cell + NO2 − treatments (Fig. S1, Electronic Supplementary Material).

The presence of a mineral surface has a catalytic effect on Fe(II) oxidation by NO3 −/NO2 − (Sørensen and Thorling 1991; Tai and Dempsey 2009). Supplementary experiments were conducted with Fe(II) and NO3 −/NO2 − in the absence and presence of lepidocrocite. The kinetics results indicated that the presence of lepidocrocite did accelerate the reactions between Fe(II) and NO3 −/NO2 − (Fig. 6, and Table S2 in the Electronic Supplementary Material), which is consistent with previous study showing the accelerated effect of goethite on Fe(II) oxidation by NO2 − (Klueglein and Kappler 2013). Hence, the formation of lepidocrocite in this study can contribute to the Fe(II) oxidation during the denitrification by P. stutzeri LS-2. In summary, the Fe(II) oxidation and nitrate reduction mediated by P. stutzeri LS-2 included both microbial and chemical processes, as presented in the Fig. 7.

Fig. 6
figure 6

Concentrations of a dissolved Fe(II), and b NO3 − or NO2 − with time dependence in the treatments of Fe(II) + NO3 −, γ-FeOOH + Fe(II) + NO3 −, Fe(II) + NO2 −, and γ-FeOOH + Fe(II) + NO2 −. The initial concentration of γ-FeOOH was 5 mM. Data are presented as the means ± standard deviations (SD) of triplicate reactions

Full size image
Fig. 7
figure 7

The coupled interaction between denitrification and Fe(II) oxidation mediated by Pseudomonas stutzeri LS-2 included microbial and chemical processes

Full size image

5 Conclusions

The present study observed that the denitrifying bacterium P. stutzeri LS-2, isolated from paddy soil in southern China, could rapidly reduce NO3 − or NO2 − in the absence of Fe(II). The addition of Fe(II) slowed the microbial NO3 − or NO2 − reduction, which was probably due to encrustation on the cell surface and in the periplasm. The NO2 − produced during NO3 − reduction by P. stutzeri LS-2 primarily reacted chemically with Fe(II). Fe(II) oxidation during denitrification by P. stutzeri LS-2 resulted in the formation of lepidocrocite, which further accelerated the chemical reactions between Fe(II) and NO3 −/NO2 −. Our findings suggested that stable isotope fractionations of δ15N-N2O in combination with the transformation kinetics of iron and nitrogen are helpful approaches to distinguish the chemical and biological reactions involved in Fe(II) oxidation and nitrate reduction by denitrifying bacteria.