Nitrate-Dependent Ferrous Iron Oxidation by Anaerobic Ammonium Oxidation (Anammox) Bacteria

Abstract

We examined nitrate-dependent Fe²⁺ oxidation mediated by anaerobic ammonium oxidation (anammox) bacteria. Enrichment cultures of “Candidatus Brocadia sinica” anaerobically oxidized Fe²⁺ and reduced NO₃⁻ to nitrogen gas at rates of 3.7 ± 0.2 and 1.3 ± 0.1 (mean ± standard deviation [SD]) nmol mg protein⁻¹ min⁻¹, respectively (37°C and pH 7.3). This nitrate reduction rate is an order of magnitude lower than the anammox activity of “Ca. Brocadia sinica” (10 to 75 nmol NH₄⁺ mg protein⁻¹ min⁻¹). A ¹⁵N tracer experiment demonstrated that coupling of nitrate-dependent Fe²⁺ oxidation and the anammox reaction was responsible for producing nitrogen gas from NO₃⁻ by “Ca. Brocadia sinica.” The activities of nitrate-dependent Fe²⁺ oxidation were dependent on temperature and pH, and the highest activities were seen at temperatures of 30 to 45°C and pHs ranging from 5.9 to 9.8. The mean half-saturation constant for NO₃⁻ ± SD of “Ca. Brocadia sinica” was determined to be 51 ± 21 μM. Nitrate-dependent Fe²⁺ oxidation was further demonstrated by another anammox bacterium, “Candidatus Scalindua sp.,” whose rates of Fe²⁺ oxidation and NO₃⁻ reduction were 4.7 ± 0.59 and 1.45 ± 0.05 nmol mg protein⁻¹ min⁻¹, respectively (20°C and pH 7.3). Co-occurrence of nitrate-dependent Fe²⁺ oxidation and the anammox reaction decreased the molar ratios of consumed NO₂⁻ to consumed NH₄⁺ (ΔNO₂⁻/ΔNH₄⁺) and produced NO₃⁻ to consumed NH₄⁺ (ΔNO₃⁻/ΔNH₄⁺). These reactions are preferable to the application of anammox processes for wastewater treatment.

INTRODUCTION

Iron is a prevalent redox-active metal element present in the Earth and occurs in two oxidation states (Fe²⁺ and Fe³⁺) in nature. The state of iron depends on various physicochemical parameters, particularly pH, temperature, and redox potential (1, 2). Some microorganisms drive the oxidation of Fe²⁺ to gain energy for growth, with molecular oxygen or NO₃⁻ as the electron acceptor (3). There are four types of biological Fe²⁺ oxidation, (i) aerobic Fe²⁺ oxidation under acidophilic conditions; (ii) neutrophilic Fe²⁺ oxidation under microaerobic conditions; (iii) neutrophilic, light-dependent Fe²⁺ oxidation under anoxic conditions; and (iv) neutrophilic, nitrate-dependent Fe²⁺ oxidation under anoxic conditions (4, 5, 6, 7). Because the redox potential of Fe³⁺/Fe²⁺ is +0.77 V at pH 2, molecular oxygen is used for Fe²⁺ oxidation under acidophilic conditions. Nitrate can also be used for Fe²⁺ oxidation under neutrophilic conditions since the redox potentials of iron composites are sufficiently low to reduce nitrate (+0.42 V for NO₃⁻/NO₂⁻), e.g., −0.236 V for Fe(OH)₃/Fe²⁺, +0.2 V for Fe(OH)₃ + HCO₃⁻/FeCO₃, and from −0.1 to +0.1 V for ferrihydrite/Fe²⁺ at pH 7 (5, 8). Nitrate-dependent Fe²⁺ oxidation involves the oxidation of Fe²⁺ coupled with denitrification or dissimilatory nitrate reduction to ammonium (DNRA); therefore, it influences both N and Fe cycles in the environment. Nitrate-dependent Fe²⁺ oxidation has been observed in various environments, including freshwater springs (9), paddy soils (10), blackish-water and marine sediments (11), freshwater sediments (12, 13, 14), groundwater (15, 16), and activated sludge (17).

Anammox bacteria are obligate anaerobic and chemoautotrophic bacteria affiliated with a monophyletic group in the bacterial order Brocadiales in the phylum Planctomycetes (18). Anammox bacteria anaerobically oxidize ammonium with nitrite to produce nitrogen gas (N₂) through a hydrazine intermediate; the stoichiometry is shown in the following equation (19, 20): NH₄⁺ + 1.32NO₂⁻ + 0.066HCO₃⁻ + 0.13H⁺ → 1.02N₂ + 0.26NO₃⁻ + 2.03H₂O + 0.066CH₂O_0.5N_0.15.

In addition to the anammox reaction, these bacteria are capable of performing dissimilatory nitrate (21, 22), Fe³⁺, and Mn(IV) (23, 24) reduction by using organic matter as an electron donor. Interestingly, the ability to oxidize Fe²⁺ with NO₃⁻ as an electron acceptor has also been described for an anammox bacterium, “Candidatus Kuenina stuttgartiensis.” When an enrichment culture of “Ca. Kuenina stuttgartiensis” was anaerobically incubated with added Fe²⁺ and NO₃⁻, Fe²⁺ was oxidized at a rate of 4.8 nmol mg protein⁻¹ min⁻¹ (23). However, the details of nitrate-dependent Fe²⁺ oxidation by anammox bacteria, including the stoichiometry of the reaction, affinity for NO₃⁻, and impact on the anammox reaction, are poorly understood.

Our objectives were (i) to examine nitrate-dependent Fe²⁺ oxidation by anammox bacteria other than “Ca. Kuenina stuttgartiensis,” (ii) to characterize the physiological traits of nitrate-dependent Fe²⁺ oxidation, including temperature and pH dependence, and the mean half-saturation constant (K_s) for NO₃⁻, and (iii) to examine the influence of nitrate-dependent Fe²⁺ oxidation on the stoichiometry of the anammox reaction. Experiments were carried out primarily by using an anammox bacterium, “Candidatus Brocadia sinica,” that was enriched from activated sludge from a municipal wastewater treatment plant in Japan (25, 26). Activity of nitrate-dependent Fe²⁺ oxidation was also examined in another anammox bacterium, “Candidatus Scalindua sp.,” that was originally obtained from marine sediment in Hiroshima Bay, Japan (27, 28).

MATERIALS AND METHODS

Bacteria strains.

Planktonic “Ca. Brocadia sinica” cells were obtained from a membrane bioreactor (MBR) as previously described (29). A hollow-fiber membrane unit composed of 300 polyethylene tubes (pore size, 0.1 μm; tube diameter, 1 mm; length, 70 mm) was installed in a 3-liter jar fermentor (Eyela, Tokyo, Japan) to sustain the “Ca. Brocadia sinica” cells in the MBR. This MBR has been operated at 37 ± 1°C for more than 1 year with a continuous supply of inorganic nutrient medium containing NH₄⁺ and NO₂⁻ (each at 40 to 60 mM). Inorganic nutrient medium contained the following (mM): MgSO₄ · 7H₂O (0.24), CaCl₂ (0.47), KH₂PO₄ (0.18), KHCO₃ (1.0), yeast extract at 1.0 mg liter⁻¹ (Becton, Dickinson Co., Franklin Lakes, NJ), and 0.5 ml of trace element solutions I and II (30). Fluorescence in situ hybridization (FISH) analysis with the oligonucleotide probe AMX820 (31) showed that anammox bacteria accounted for more than 94% of the total biomass. The dominance of “Ca. Brocadia sinica” in a monospecies culture was confirmed by determining the 16S rRNA gene sequence of cells in the enrichment culture. To obtain a highly enriched “Ca. Brocadia sinica” culture, Percoll separation was performed as described previously (32). A red-colored band in the bottom layer of a centrifugation tube was recovered with a plastic syringe and washed twice with inorganic nutrient medium. “Ca. Brocadia sinica” accounted for more than 99.9% of the total biomass in this highly enriched culture.

The small floccular biomass (<100 μm) of “Ca. Scalindua sp.” was obtained from another MBR. An enrichment culture of “Ca. Scalindua sp.” previously described by Kindaichi et al. (27) was inoculated into the MBR, and it was operated for more than half of a year at 20°C with a continuous supply of inorganic nutrient medium containing artificial sea salt (2.5% [wt/vol]; Marine Tech, Tokyo, Japan). Anammox bacteria accounted for more than 90% of the total biomass in the culture, as determined by FISH analysis with the oligonucleotide probe AMX820. Partial 16S rRNA gene nucleotide sequences (Escherichia coli 16S rRNA position, bp 225 to 645) retrieved from the culture showed 99% sequence similarity to the clone husup-a7 (accession number AB573103) (27).

Activity tests.

Standard anaerobic techniques (26, 33) were used with an anaerobic chamber (Coy Laboratory Products, Grass Lake Charter Township, MI). Biomass taken from the MBRs was centrifuged (10,000 × g, 10 min, 20°C), and the biomass pellets were washed twice and resuspended at a biomass concentration of 0.1 to 1.0 mg protein ml⁻¹ in an inorganic nutrient medium (pH 7.3 to 7.5) supplemented with 10 to 15 mM Fe²⁺ (in the form of FeSO₄) and 2.5 mM NO₃⁻. Inorganic nutrient medium was prepared by purging with N₂ gas (99.999%) for 30 min, and this medium was left for 1 week in an anaerobic chamber. Prior to activity testing, KHCO₃ and FeSO₄ · 7H₂O were weighed in the anaerobic chamber and dissolved in medium.

Biomass suspension (1 to 10 ml) was dispensed into 5- or 25-ml serum glass vials sealed with butyl rubber stoppers and aluminum caps. The headspace was replaced with He gas (>99.99995%), and the vials were incubated at 37°C in the dark for 10 to 98 h. To determine the extent of chemical Fe²⁺ oxidation or NO₃⁻ reduction, negative-control vials without biomass were incubated in the same manner.

Fe²⁺ oxidation and NO₃⁻ reduction activities were measured under different pH and temperature conditions. The influence of pH was examined over a pH range of 2.7 to 11.2. The pH was adjusted by adding 2 N H₂SO₄ or 1 N NaOH. The influence of temperature was examined over a range of 4 to 75°C.

The value of K_s for NO₃⁻ was determined under NO₃⁻-limiting conditions. The initial concentrations of Fe²⁺ and NO₃⁻ were set at 10 and 0.7 mM, respectively. The value of K_s was evaluated on the basis of the Hanes-Wool plot.

The stoichiometry of nitrate-dependent Fe²⁺ oxidation by “Ca. Brocadia sinica” was determined by incubating the biomass with 2.5 mM ¹⁵N-labeled NO₃⁻ (Cambridge Isotope Laboratories, Andover, MA) and 15 mM Fe²⁺. As controls, vials (i) with autoclaved biomass, (ii) without added Fe²⁺, and (iii) without added NO₃⁻ were incubated in parallel. Additionally, “Ca. Brocadia sinica” cells were incubated with 500 mg liter⁻¹ penicillin G (Sigma-Aldrich, St. Louis, MO) and 200 mg liter⁻¹ chloramphenicol (Wako Pure Chemicals, Osaka, Japan).

Incorporation of CO₂.

Incorporation of CO₂ was examined by incubating the biomass with NaH¹⁴CO₃ (PerkinElmer, Waltham, MA) at a final concentration of 20 μCi mg volatile suspended solids⁻¹ as previously described (34). Biomass was collected after 4 days of incubation, washed three times with phosphate-buffered saline (PBS), and mixed with scintillation cocktail Ultima Gold XR (PerkinElmer). Radioactivity was measured with an LSC-6100 liquid scintillation counter (Hitachi-Aloka Medicals, Tokyo, Japan). Additionally, cold runs were performed in parallel to determine the rates of NO₃⁻ reduction and Fe²⁺ oxidation. Autoclaved biomass was incubated in the same manner to examine nonbiological CO₂ incorporation.

Co-occurrence of Fe²⁺ oxidation and anammox.

Concurrent Fe²⁺ oxidation and anammox were examined by incubating “Ca. Brocadia sinica” cells with ¹⁵N-labeled NO₃⁻, unlabeled NH₄⁺ (each 2.5 mM), and Fe²⁺ (10 mM). The influence of nitrate-dependent ferrous iron oxidation on the stoichiometry of the anammox reaction was further examined by incubating the biomass with NH₄⁺, NO₂⁻ (each at 2 mM), and Fe²⁺ (1 mM). Controls included vials (i) without Fe²⁺ and (ii) with 1 mM Fe³⁺ instead of Fe²⁺, which were incubated in parallel.

Chemical analysis.

To determine Fe²⁺, NH₄⁺, NO₂⁻, and NO₃⁻ concentrations, 1 ml of biomass suspension was mixed with 50 μl of 4 M HCl and incubated for 30 min at ambient temperature. A portion (∼10 μl) of this mixture was for Fe²⁺ concentration measurement. The remaining biomass was centrifuged at 18,200 × g for 10 min, and NH₄⁺, NO₂⁻, and NO₃⁻ were measured in the supernatant. These procedures were conducted in an anaerobic chamber to prevent chemical oxidation of Fe²⁺. Headspace gas was collected from the vials with a gas-tight glass syringe (VICI AG International, Schenkon, Switzerland) and immediately analyzed by gas chromatography-mass spectrometry (GC-MS) as described below.

Concentrations of NH₄⁺ and NO₃⁻ were determined with an IC-2010 ion chromatograph equipped with a TSKgel SuperIC-Anion HS or a TSKgel SuperIC-Cation HS column (Tosoh, Tokyo, Japan) for anion or cation analysis, respectively. When the value of K_s for NO₃⁻ was determined, the concentration of NO₃⁻ was measured via the colorimetric brucine method (35). For colorimetric determination of NO₃⁻, removal of Fe²⁺ and Fe³⁺ from the bulk solution was essential because they interfere with brucine color development. Liquid samples were passed over a TOYOPAK IC-SP cation-exchange column (Tosoh, Tokyo, Japan). This filtration does not decrease or increase the NO₃⁻ concentration. Filtered samples were mixed with 80% (vol/vol) sulfuric acid and brucine sulfanilic acid solution (10 mM brucine and 5.8 mM sulfanilic acid) and heated in boiling water for 20 min. After cooling, the absorbance at 405 nm was determined with an ARBO MX microplate reader (PerkinElmer).

The NO₂⁻ concentration was determined by the naphthylethylenediamine method (36). The sample was mixed with 4.9 mM naphthylethylenediamine solution, and the absorbance at 540 nm was measured.

The Fe²⁺ concentration was determined by the 1,10-phenanthroline method (36). Briefly, the sample was mixed with 5.5 mM 1,10-phenanthroline solution and 1 M acetate buffer (pH 4.6) in an anaerobic chamber and then the absorbance at 405 nm was measured.

¹⁵N-labeled gaseous compound concentrations (²⁹N₂, ³⁰N₂, ³¹NO, and ⁴⁶N₂O) were determined by GC-MS analysis as previously described (37, 38). Fifty microliters of headspace gas was injected at a split ratio of 1:50 into a GCMS-QP2010 SE gas chromatograph (Shimadzu, Kyoto, Japan) equipped with a CP-Pora BOND Q fused silica capillary column (Agilent Technologies, Santa Clara, CA). A standard curve for ²⁹N₂ and ³⁰N₂ gas quantification was prepared with ³⁰N₂ gas (>98% purity) (Cambridge Isotope Laboratories).

Biomass concentration.

The culture fluid was centrifuged at 18,200 × g for 10 min, and the biomass pellets were resuspended in 10% (wt/vol) sodium dodecyl sulfate solution. After boiling for 10 min, the supernatant was obtained after centrifugation at 18,200 × g for 10 min, followed by determination of the protein concentration. The protein concentration was determined with a DC-protein assay kit (Bio-Rad, Hercules, CA) by following to the manufacturer's instructions. Bovine serum albumin (Wako Pure Chemicals) was used as the protein standard.

Microscopy.

Epifluorescence microscopy was performed after staining the biomass with 1:1,000 SYBR gold (Life Technologies, Carlsbad, CA) for 10 min. Specimens on the glass slide (Matsunami Glass, Osaka, Japan) were examined with an AxioVert 200 microscope equipped with an Axio Cam MRC charge-coupled device camera (Carl Zeiss, Jena, Germany).

Scanning electron microscopy (SEM) was performed as previously described (39). Rusty precipitates were washed two times with PBS and fixed with 2.5% glutaraldehyde in PBS overnight at 4°C. The sample was washed three times with PBS for 10 min. The fixed samples were dehydrated in a sequential acetone series (50, 70, 80, 90, and 95% for 15 min each and 100% three times for 15 min) and substituted with isoamyl acetate. Samples were dried with a critical-point drier with liquid CO₂ and coated with platinum and palladium for 2 min. The coated samples were examined by SEM with a model S-4000 microscope (Hitachi).

RESULTS

Nitrate-dependent Fe²⁺ oxidation by anammox bacteria.

Nitrate-dependent Fe²⁺ oxidation was demonstrated by incubating “Ca. Brocadia sinica” cells with Fe²⁺ and ¹⁵NO₃⁻. As shown in Fig. 1, a simultaneous decrease in Fe²⁺ and NO₃⁻ concentrations was observed until NO₃⁻ was depleted (after 50 h of incubation). The ³⁰N₂ gas accumulated in the headspace after 30 h of incubation, and nearly all of the ¹⁵N-labeled nitrogen supplemented as ¹⁵NO₃⁻ was recovered as ³⁰N₂ gas at 98 h. Accumulation of NH₄⁺ and NO₂⁻ in the liquid phase and the production of ²⁹N₂, ³¹NO, and ⁴⁶N₂O gas in the headspace were not detected throughout incubation. These results clearly indicate that “Ca. Brocadia sinica” oxidized Fe²⁺ coupled with the reduction of NO₃⁻ to N₂.

Fig 1.

Oxidation of Fe²⁺ (gray bars, left y axis) with reduction of ¹⁵NO₃⁻ (closed circles, right y axis) and production of ³⁰N₂ gas (empty circles, right y axis). “Ca. Brocadia sinica” cells (2.5-ml sample in a 5-ml glass vial) were incubated at 37°C with ¹⁵NO₃⁻ (2.5 mM) and Fe²⁺ (15 mM). Error bars indicate the range of SDs derived from three replicated vials.

The contribution of “Ca. Brocadia sinica” to nitrate-dependent Fe²⁺ oxidation was further verified by using a highly enriched culture (>99.9% of the total biomass). This highly enriched “Ca. Brocadia sinica” culture also oxidized Fe²⁺ and reduced NO₃⁻ at rates of 3.7 ± 0.2 and 1.3 ± 0.1 nmol mg protein⁻¹ min⁻¹ (mean ± SD), respectively (Fig. 2). Therefore, the calculated molar ratio of reduced NO₃⁻ (ΔNO₃⁻) to oxidized Fe²⁺ (ΔFe²⁺) was 0.351. Fe²⁺ oxidation and NO₃⁻ reduction activities were not strongly affected by the addition of penicillin G or chloramphenicol (Fig. 2). In contrast, Fe²⁺ oxidation was greatly (>95%) reduced when the cells were incubated without NO₃⁻. Additionally, no activity of Fe²⁺ oxidation or NO₃⁻ reduction was observed in vials containing autoclaved biomass.

Fig 2.

Specific nitrate-dependent Fe²⁺ oxidation activities of a highly enriched culture of “Ca. Brocadia sinica” supplemented with (i) Fe²⁺ and NO₃⁻; (ii) Fe²⁺, NO₃⁻, and penicillin G; or (iii) Fe²⁺, NO₃⁻, and chloramphenicol. Penicillin G and chloramphenicol inhibit the activities of most bacteria but not anammox bacteria. Specific activities were also examined during incubation with added Fe²⁺ only. Additionally, the autoclaved biomass was incubated with Fe²⁺ and NO₃⁻ in the same manner as the control. Error bars indicate the range of SDs derived from three replicated vials. NA, not applicable; ND, not detected.

“Ca. Brocadia sinica” cells were planktonic and colored red (Fig. 3A); however, they formed rusty precipitates (diameters of 20 to 800 μm) (Fig. 3B) within 1 day of incubation. Epifluorescence microscopy after staining with SYBR gold revealed that rusty precipitates were aggregates of microbial cells (Fig. 3C and D). SEM revealed that the cells in aggregates were covered with small amorphous particles (Fig. 3E), which were likely oxidized iron-containing minerals since no amorphous particles were observed prior to incubation (Fig. 3F).

Fig 3.

Microscopy of rusty precipitates developed during nitrate-dependent Fe²⁺ oxidation by “Ca. Brocadia sinica.” Cell suspensions of “Ca. Brocadia sinica” were anaerobically incubated with NO₃⁻ and Fe²⁺ at 37°C (A). Rusty precipitates exhibiting a reddish color appeared at the bottoms of the vials after 1 day of incubation (B). Precipitates were stained with SYBR gold and subsequently examined by epifluorescence microscopy. Precipitates gave a signal for SYBR gold. Panels C and D are bright-field and fluorescence microscopy images, respectively, of SYBR gold. The precipitates were further examined by SEM (E). Cells were covered with small amorphous particles and were attached via the particles. As a control, “Ca. Brocadia sinica” cells were examined prior to incubation (F). Amorphous particles were not observed in image F. Scale bars are 6 cm (A and B), 100 μm (C and D), and 1 μm (E and F).

Fe²⁺ oxidation and NO₃⁻ reduction activities were also determined in the enrichment culture of “Ca. Scalindua sp.,” with the cells exhibiting Fe²⁺ oxidation and NO₃⁻ reduction activities at rates of 4.7 ± 0.59 and 1.45 ± 0.05 nmol mg protein⁻¹ min⁻¹, respectively. The molar ratio of ΔNO₃⁻ to ΔFe²⁺ was 0.309 for “Ca. Scalindua sp.,” which is similar to that for “Ca. Brocadia sinica” (i.e., 0.351).

Temperature and pH dependence and K_s for NO₃⁻.

Higher activities of Fe²⁺ oxidation and NO₃⁻ reduction were seen in the range of 30 to 45°C (Fig. 4a), which corresponds to the physiological anammox temperature of “Ca. Brocadia sinica” (25 to 45°C) (26). The pH range was determined to be 5.9 to 9.8 (Fig. 4b). Fe²⁺ oxidation and NO₃⁻ reduction activities were not observed below pH 5.0, indicating that “Ca. Brocadia sinica” is a neutrophilic Fe²⁺-oxidizing bacterium. The molar ratio of ΔNO₃⁻ to ΔFe²⁺ increased from 0.26 at 20°C to 0.42 at 76°C. These molar ratios ranged from 0.39 to 1.55 over a pH range of 5.9 to 9.8. The mean (± SD) K_s value for NO₃⁻ was determined to be 51 ± 21 μM on the basis of three independent batch experiments (see Fig. S1 in the supplemental material).

Fig 4.

Temperature (a) and pH (b) dependence of Fe²⁺ oxidation (filled squares) and NO₃⁻ reduction (gray bars) activities. “Ca. Brocadia sinica” cells were (a) incubated at pH 7.8 over a temperature range of 4 to 75°C and (b) incubated at 37°C over a pH range of 2.7 to 11.2. Error bars indicate the range of SDs derived from three replicated vials. ND, not detected for the NO₃⁻ reduction rate.

Autotrophic growth by nitrate-dependent Fe²⁺ oxidation.

Incorporation of ¹⁴C-labeled CO₂ was below the detection limit (<0.12 mM), although “Ca. Brocadia sinica” consumed 2.76 mM Fe²⁺ and 0.868 mM NO₃⁻ over 4 days of incubation. This indicates that the molar ratio of oxidized Fe²⁺ to incorporated CO₂ was greater than 23, which agrees with the high molar ratio obtained for the aerobic Fe²⁺ oxidizer Acidothiobacillus ferrooxidans (i.e., 71) (40).

Co-occurrence of Fe²⁺ oxidation and anammox.

When “Ca. Brocadia sinica” cells were incubated with Fe²⁺, ¹⁵N-labeled NO₃⁻, and unlabeled NH₄⁺, ²⁹N₂ gas was produced immediately after incubation started (Fig. 5). In addition to the production of ²⁹N₂ gas, accumulation of ³⁰N₂ gas was observed in the headspace, but its production was lower than that of ²⁹N₂ gas.

Fig 5.

Accumulation of ²⁹N₂ (circles) and ³⁰N₂ gas (triangles). “Ca. Brocadia sinica” cells (10-ml samples in 25-ml glass vials) were incubated at 37°C with ¹⁵NO₃⁻, ¹⁴NH₄⁺ (each at 2.5 mM), and Fe²⁺ (10 mM).

The presence of Fe²⁺ influenced the stoichiometry of the anammox reaction. When “Ca. Brocadia sinica” was cultivated in the presence of Fe²⁺, NH₄⁺, and NO₂⁻, the molar ratios of ΔNO₂⁻/ΔNH₄⁺ and ΔNO₃⁻/ΔNH₄⁺ were significantly lower than those found when Fe²⁺ was absent and when Fe³⁺ rather than Fe²⁺ was present (Fig. 6). The presence of Fe²⁺ and Fe³⁺ (up to 1 mM) did not result in an inhibitory effect on the activity of anaerobic ammonium oxidation.

Fig 6.

Stoichiometric ΔNO₂⁻/NH₄⁺ (gray bars) and ΔNO₃⁻/NH₄⁺ (white bars) ratios in “Ca. Brocadia sinica” cultures incubated in the presence of (i) NH₄⁺, NO₂⁻, and NO₃⁻ (labeled as w/o); (ii) NH₄⁺, NO₂⁻, NO₃⁻, and Fe²⁺ (labeled as Fe²⁺); and (iii) NH₄⁺, NO₂⁻, NO₃⁻, and Fe³⁺ (labeled as Fe³⁺). Error bars indicate the range of SDs derived from three replicated vials. P values were calculated with Student's t test. *, P < 0.01.

DISCUSSION

Nitrate-dependent Fe²⁺ oxidation has been observed in various bacterial and archaeal species (3, 41). We showed that anammox bacteria “Ca. Brocadia sinica” and “Ca. Scalindua sp.” are capable of carrying out nitrate-dependent iron oxidation. Their contributions to nitrate-dependent Fe²⁺ oxidation, particularly that of “Ca. Brocadia sinica,” were carefully evaluated by using highly enriched cultures and supplementation with antibiotics (penicillin G and chloramphenicol) that are not active against anammox bacteria but inhibit the activity of most heterotrophs (42, 43). Moreover, control experiments without the addition of biomass were performed to examine the extent of abiotic Fe²⁺ oxidation [e.g., Fe²⁺ oxidation by Mn(III) and Mn(IV) oxide, copper, green rust, NO₂⁻, and N₂O] (44, 45, 46, 47) and NO₃⁻ reduction. Therefore, the activities of Fe²⁺ oxidation coupled with NO₃⁻ reduction described in this report represent microbial activities.

Several metabolic pathways, including nitrate reduction to nitrite (NO₃⁻ → NO₂⁻), denitrification (NO₃⁻ → NO₂⁻ → NO → N₂O → N₂), and DNRA (NO₃⁻ → NO₂⁻ → NH₄⁺), can be coupled with Fe²⁺ oxidation (47, 48, 49). We confirmed the production of ²⁹N₂ gas from ¹⁵N-labeled NO₃⁻ and unlabeled NH₄⁺ (Fig. 5), indicating that NO₃⁻ was reduced to NO₂⁻ and the ¹⁵N-labeled NO₂⁻ produced was subsequently utilized for the anammox reaction. This is because (i) ²⁹N₂ gas is produced solely in the anammox reaction (50) and (ii) NO₃⁻ is not a direct substrate of the anammox reaction (19, 20). Moreover, accumulation of ³⁰N₂ gas shows that ¹⁵N-labeled NO₂⁻ was further reduced to NH₄⁺ and then consumed in the anammox reaction. DNRA is the likely pathway for NO₃⁻ reduction to NH₄⁺ via NO₂⁻, as previously demonstrated by anammox bacteria other than “Ca. Brocadia sinica” (21, 22, 51). Indeed, we found that “Ca. Brocadia sinica” carried a gene essential for DNRA (nrfA-encoding ammonia-forming nitrite reductase) and expressed this protein at moderate levels. On the other hand, “Ca. Brocadia sinica” does not contain a gene for denitrification (nosZ, which encodes nitrous oxide reductase) (M. Oshiki et al., unpublished data). However, the contribution of NrfA to DNRA by “Ca. Brocadia sinica” remains unknown and requires additional analysis. Notably, the production of ³⁰N₂ gas lagged behind NO₃⁻ reduction (Fig. 1) and the production of ²⁹N₂ gas (Fig. 5), indicating that reduction of the intermediate produced from NO₃⁻ was the rate-limiting step in the anammox reaction. This observation agrees with a previous study stating that NO₂⁻ reduction to NH₄⁺ is the rate-limiting step in the DNRA reaction of “Ca. Kuenina stuttgartiensis” (22).

The two mechanisms postulated for microbial Fe²⁺ oxidation are cell surface and intracellular Fe²⁺ oxidation (52). In the former case, electrons derived from Fe²⁺ oxidation must be transported into cells and finally passed to nitrate reductase to reduce nitrite. In the latter case, membrane proteins are essential for Fe²⁺ uptake and further for Fe³⁺ export out of cells after Fe²⁺ oxidation (53). Since anammox bacterial cells contain three separate compartments (paryphoplasm, riboplasm, and anammoxosome) (54), Fe²⁺ or electrons must pass through these compartments. The biochemistry of nitrite-dependent Fe²⁺ oxidation by anammox bacteria is of great interest and requires further investigation.

The molar ratios of ΔNO₃⁻ to ΔFe²⁺ (0.351 in “Ca. Brocadia sinica” and 0.309 in “Ca. Scalindua sp.” at 37°C and pH 7.3 to 7.5) were higher than the theoretical ratio of Fe²⁺ oxidation coupled to NO₃⁻ reduction to the equimolar NO₂⁻ and NH₄⁺ (i.e., 0.20). Moreover, the theoretical ratio further decreases when electrons derived from Fe²⁺ oxidation are utilized in the subsequent anammox reaction; i.e., 1 and 3 mol of electrons are required for reduction of nitrate to nitric oxide and synthesis of hydrazine from equimolar levels of NO and NH₄⁺ (19, 20). High molar ratios of ΔNO₃⁻ to ΔFe²⁺ were previously observed in cultures of freshwater sediment microorganisms (55, 56). These studies examined the microbial activities of Fe²⁺ oxidation coupled with DNRA, and the observed molar ratios of ΔNO₃⁻ to ΔFe²⁺, respectively, were 0.191 and 0.25. Those values were nearly twice as high as the excepted molar ratio of 0.125. A possible explanation is that Fe³⁺ may have been reduced by endogenous respiration and/or the oxidation of dead cells and extracellular polymeric substances as electron donors. The metabolic capability of Fe³⁺ reduction by anammox bacteria has been demonstrated in “Ca. Kuenina stuttgartiensis” and “Candidatus Scalindua profunda” with exogenous formate as an electron donor (23, 24). An alternative explanation is that the production of unidentified N species occurred during NO₃⁻ reduction (47). For instance, the occurrence of incomplete denitrification under unfavorable pH conditions has been previously demonstrated in a denitrifying culture (57) and in soil samples (58).

The Fe²⁺ oxidation activities of three phylogenetically distinct anammox bacteria, “Ca. Kuenina stuttgartiensis” (23), “Ca. Brocadia sinica,” and “Ca. Scalindua sp.,” are in the range of 3.7 to 8.7 nmol Fe²⁺ mg protein⁻¹ min⁻¹. These values correspond to 3.8 to 8.9 mmol Fe²⁺ mmol C⁻¹ h⁻¹ when a conversion factor of 0.0585 mol C g protein⁻¹ (26) is used, indicating that the three anammox bacteria show Fe²⁺ oxidation activity lower than that of acidophilic Fe²⁺-oxidizing bacteria (6.8 to 20 mmol Fe²⁺ mmol C⁻¹ h⁻¹) (59, 60). “Ca. Brocadia sinica” and “Ca. Scalindua sp.” showed NO₃⁻ reduction activities that were an order of magnitude lower than their anammox activities, which are in the range of 10 to 75 nmol NH₄⁺ mg protein⁻¹ min⁻¹ (26, 28). On the basis of the K_s value for NO₃⁻ of “Ca. Brocadia sinica,” 51 ± 21 μM, the nitrate-dependent Fe²⁺ oxidation reaction by “Ca. Brocadia sinica” can be anticipated in circumneutral environments where NO₃⁻ and Fe²⁺ are available at several micromolar concentrations, such as in anoxic groundwater sediments (15, 61).

The molar ratios of ΔNO₃⁻/ΔNH₄⁺ and ΔNO₂⁻/ΔNH₄⁺ in the anammox reaction decreased significantly upon the addition of Fe²⁺ to the culture medium (Fig. 6). NO₃⁻ produced during the anammox reaction was utilized for DNRA (NO₃⁻ → NO₂⁻ → NH₄⁺) coupled with Fe²⁺ oxidation; the intermediate of DNRA, NO₂⁻, can also be utilized for the anammox reaction. Thus, the molar ratios of ΔNO₃⁻/ΔNH₄⁺ and ΔNO₂⁻/ΔNH₄⁺ were lower than those without Fe²⁺ addition. In contrast, there are two other possible explanations for decreased molar ratios of ΔNO₃⁻/ΔNH₄⁺ and ΔNO₂⁻/ΔNH₄⁺. The reducing power of CO₂ fixation may be obtained not only from NO₂⁻ oxidation but also from Fe²⁺ oxidation, resulting in lower production of NO₃⁻, and maintenance energy is increased in the presence of Fe²⁺, leading to a decreased growth yield of anammox bacteria. These hypotheses must be examined in future studies. Decreased ΔNO₃⁻/ΔNH₄⁺ and ΔNO₂⁻/ΔNH₄⁺ ratios are preferable in the application of anammox for wastewater treatment because they require lower levels of NO₂⁻ in the influents and there is a low discharge of NO₃⁻ (62, 63). We recommend the further study of nitrogen removal performance in anammox processes wherein ferrous iron is supplemented with wastewater.