Enhanced Alcaligenes faecalis Denitrification Rate with Electrodes as the Electron Donor

Abstract

The utilization by Alcaligenes faecalis of electrodes as the electron donor for denitrification was investigated in this study. The denitrification rate of A. faecalis with a poised potential was greatly enhanced compared with that of the controls without poised potentials. For nitrate reduction, although A. faecalis could not reduce nitrate, at three poised potentials of +0.06, −0.06, and −0.15 V (versus normal hydrogen electrode [NHE]), the nitrate was partially reduced with −0.15- and −0.06-V potentials at rates of 17.3 and 28.5 mg/liter/day, respectively. The percentages of reduction for −0.15 and −0.06 V were 52.4 and 30.4%, respectively. Meanwhile, for nitrite reduction, the poised potentials greatly enhanced the nitrite reduction. The nitrite reduction rates for three poised potentials (−0.06, −0.15, and −0.30 V) were 1.98, 4.37, and 3.91 mg/liter/h, respectively. When the potentials were cut off, the nitrite reduction rate was maintained for 1.5 h (from 2.3 to 2.25 mg/liter/h) and then greatly decreased, and the reduction rate (0.38 mg/liter/h) was about 1/6 compared with the rate (2.3 mg/liter/h) when potential was on. Then the potentials resumed, but the reduction rate did not resume and was only 2 times higher than the rate when the potential was off.

INTRODUCTION

Alcaligenes faecalis is a Gram-negative heterotrophic bacterium that is common in soil (1). A. faecalis was reported to aerobically produce nitrite (NO₂⁻), nitrate, nitric oxide (NO), and nitrous oxide in both peptone-meat extract and defined media with ammonium and citrate as the sole nitrogen and carbon sources (2). Nevertheless, A. faecalis also had denitrification ability as the bacterium had a copper-containing nitrite reductase that catalyzes the reduction of nitrite (NO₂⁻) to nitric oxide (NO) (3). Recent research by Lu et al. revealed that A. faecalis might be able to use extracellular electrons from semiconducting mineral photocatalysis for metabolism (4). In their study, with photoelectrons A. faecalis gradually became the dominant species in the soil microbial community, indicating this bacterium might have a potential to utilize photoelectrons (4).

The interaction between microbes and solid minerals is becoming a worldwide hot spot in the field of geomicrobiology, an interdisciplinary field involving geology and microbiology. Especially, the electron flows between microbes and minerals (or electrodes) are a major research interest in this field. Currently, the models of microbes donating electrons to minerals (or electrodes) are categorized into three types: direct contact, electron shuttle, and microbial extracellular appendages (5). Direct contact between cells and the electrode surface will no doubt facilitate the electron exchanges (6). Under some circumstances, the environmental chemicals or microbially secreted chemicals could act as an electron shuttle, transferring electrons via redox reactions (7, 8). Besides secreting chemical compounds, some bacteria have evolved extracellular structures, the conductive flagella called “nanowires,” to harvest electrons from solid electrodes or minerals (9).

Most research on electron transfer between microbes and minerals (or electrodes) focuses on the electron transfer from microbes to minerals or electrodes, such as in microbial fuel cell (MFC) applications and microbial iron reduction in clay minerals (10, 11). Few studies have focused on microbes accepting electrons from electrodes or minerals. Strychaz et al. (12) and Aulenta et al. (13) revealed that a graphite electrode could be the sole electron donor for Geobacter lovleyi to dechlorinate the tetrachlorethene. However, although a series of papers on microbes utilizing electrons from electrodes have been published, many of them mainly focused on the microbes' dechlorination (14) and carbon dioxide transformation under anaerobic conditions (15, 16).

Recently, many researchers have focused on microbial denitrification in a bioelectrochemical system (BES). Park et al. used electrodes as electron donors for nitrate reduction, but this was performed with high current (∼200 mA) and under anaerobic conditions (17). Sakakibara and Kuroda applied a wide range of voltages (0 to 37 V) to cathodic biofilm in a dual-chamber device and found that the current and nitrogen gas had a linear relationship (18). In Sakakibara's system, hydrogen was produced in the cathode as the potential for hydrogen at pH 7 was −410 mV (versus a normal hydrogen electrode [NHE]) (19). Doan et al. investigated the effect of different current densities (0.2, 1, 5, 10, and 20 mA/cm² corresponding to a voltage of 1 to 5 V) on bioelectrochemical denitrification in organic carbon-free wastewater (20). In Tong and He's study, nitrate was reduced in an anode in microbial fuel cells by heterotrophic bacteria (21). Su et al. reported that Pseudomonas alcaliphila strain MBR reduced nitrate and nitrite with an electrode as the sole electron donor (22).

In this study, a strain of Alcaligenes faecalis (ATCC 8750 = DSM 30030 = NCIB 8156) was cultured aerobically in a dual-chamber microbial fuel cell (MFC)-like device. The A. faecalis ATCC 8750 strain was fed electrons from a graphite electrode and reduced nitrate and nitrite. A. faecalis ATCC 8750 was reported previously to be unable to reduce nitrate (23, 24, 25). Our studies revealed that A. faecalis could reduce nitrate and nitrite under aerobic conditions with electrodes as the electron donor, and the current densities were low (<0.02 mA/cm²).Our previous study showed that a chemoautotrophic bacterium, Acidithiobacillus ferrooxidans, utilized the cathode as an electron donor for growth (4), and A. faecalis had a good response to semiconducting mineral photoelectrons (4). The potential ability of A. faecalis to directly use external electrons from semiconducting minerals will help us to understand the synergism between microbes and semiconducting minerals in natural environments.

MATERIALS AND METHODS

Bacterial cultivation.

Alcaligenes faecalis ATCC 8750 was provided by China General Microbiological Culture Collection Center (CGMCC). The bacterium was first activated and then enriched with Luria-Bertani (LB) medium containing 10 g/liter peptone, 5 g/liter yeast extract, and 10 g/liter sodium chloride.

For nitrite and nitrate reduction in batch cultivation aerobically and anaerobically, the medium was a simple basic medium containing the following: K₂HPO₄, 9.28 g/liter; KH₂PO₄, 1.81 g/liter; (NH₄)₂SO₄, 661 mg/liter; sodium citrate, 1.18 g/liter; Na₂MoO₄·2H₂O, 24.2 mg/liter; FeSO₄·7H₂O, 5.6 mg/liter; MnCl₂·4H₂O, 0.99 mg/liter; CaCl₂·2H₂O, 51.5 mg/liter; and MgSO₄·7H₂O, 200 mg/liter. For nitrite and nitrate treatments, the nitrite and nitrate concentrations were both 4 mmol/liter. The anaerobic experiments were performed in 150-ml serum bottles in a glove box (Plas-Labs, Lansing, MI). The aerobic experiments were performed with flasks in an HZQ-X100 shaking incubator (Hualong, Jintan, China) at 35°C.

In the poised potential experiment, the simple basic mineral medium described above was applied but without sodium citrate. The microbial electronic reduction of nitrate and nitrite was conducted in a dual-chamber device. The culture medium was simple mineral medium without organics in both anodic and cathodic chambers. The nitrite and nitrate were added to the cathodic chamber, and the final concentrations were about 1 and 4 mmol/liter, respectively.

Equipment setup.

The bioelectrochemical system (BES) was a dual-chamber glass device (Fig. 1). The two chambers were a counter chamber with a counter electrode (anode) and a bacterial chamber with a working electrode (cathode) and a reference electrode. The two chambers were separated by a proton exchange membrane (PEM) (Dupont). The counter and working electrodes were brush-shaped graphite electrodes (Shanghai Hongfeng Industrial Co., Ltd., Shanghai, China). The reference electrode was a 232-type saturated calomel electrode (SCE) (potential at 0.2412 V versus NHE). Three electrodes were connected to a CHI1000B potentiostat (ChenHua Instruments, Shanghai, China). The bacterial chamber was open to the air but had a 0.22-μm-pore nitrocellulose membrane to avoid contamination.

FIG 1.

Schematic diagram of the dual-chamber device used in this study. PEM, proton exchange membrane.

Biofilm preparation.

The A. faecalis cells were deposited on the surface of the electrode in the bacterial chamber.

(i) Step 1.

A. faecalis was first enriched in LB medium, and when it reached the logarithmic phase, the culture medium was transferred to the bacterial chamber.

(ii) Step 2.

When the system current was stable, the medium in the bacterial chamber was replaced with a fresh simple mineral medium with 1.18 g/liter sodium citrate as the carbon source, and the system was run until the system current was stable. This step was repeated three times to remove residual LB medium.

(iii) Step 3.

The whole medium in the bacterial chamber was then replaced with fresh simple mineral medium without organics. The working electrode with bacterium film was gently washed three times with sterilized 0.9% NaCl solution to remove the remaining organics on the working electrode.

The biofilm on the surface of the working electrode remained in the environment without any carbon sources. The system was run at poised potentials for 24 h. Then, the sodium nitrite or sodium nitrate was injected, and the nitrite and nitrate concentrations were measured by UV-visible (UV-Vis) spectrophotometry at standard intervals.

Analytical methods and procedures.

The protein concentration on the electrode surface indirectly reflected the bacterial density. A small piece of electrode was cut and immersed in 1 ml of 0.2 mol/liter NaOH in a 1.5-ml Eppendorf tube. The tube was heated in boiling water bath for 10 min to release protein from cells on the electrode. Then the protein solution was analyzed with a modified Bradford protein assay (26), measuring Coomassie brilliant blue at 450 nm and 590 nm.

The bacterium optical density at 600 nm (OD₆₀₀) was measured using an Agilent Cary 8453 UV-Vis spectrophotometer (Agilent).

Citrate was determined by ion chromatography (ICS-1100; Thermo) with an IonPac AG11 Guard column and AS11 analytical column. The eluent was 35 mM potassium hydroxide. The injection volume was 30 μl, and the eluent flow rate was 1.0 ml/min. The column temperature was 30°C. Before sample injection, the sample was pretreated with an OnGuard II RP pretreatment column.

The concentrations of nitrite and nitrate were routinely measured during the experiment. The sample was first centrifuged at 10,000 rpm (approximately 9,400 × g) with an Eppendorf 5424 centrifuge. Then the supernatant was collected for determination of the nitrate and nitrite concentrations by the sodium salicylate and sulfanilamide methods, respectively (27).

The electrodes for scanning electron microscopy (SEM) observation were prepared by the critical point drying method. The bacterial cells on the electrode were first fixed with 2% paraformaldehyde and 2.5% glutaraldehyde in phosphate-buffered saline (PBS [pH 6.8 to ∼7.2]) for 50 min. Then the electrode was immersed in PBS for 10 to ∼15 min, and this was repeated four times, followed by sequential dehydration with a graded ethanol series. The electrode was dried with the Emithech K850 critical point dryer (Quorum, United Kingdom). The electrode, coated with gold, was observed with an FEI Quanta650 scanning electron microscope using a 20-keV accelerating voltage and an 8.5- to ∼9.6-mm working distance.

RESULTS

Nitrite and nitrate reduction under anaerobic and aerobic conditions.

It was obvious that the reduction of nitrite by A. faecalis happened under aerobic conditions within 72 h (from 3.8 mmol/liter to 2.9 mmol/liter) (Fig. 2A), but nitrite did not decrease under anaerobic conditions even after 144 h. Neither under anaerobic conditions nor under aerobic conditions could the nitrate be reduced. The cell densities (OD₆₀₀) were stable under both conditions (Fig. 2), except for the drastic decrease under anaerobic conditions during first 12 h. Compared with the citrate concentration under anaerobic conditions, more citrate was consumed under aerobic conditions. Given the organics as carbon sources in the culture medium, A. faecalis could only reduce nitrite aerobically and could not reduce nitrate aerobically or anaerobically.

FIG 2.

Aerobic and anaerobic growth of A. faecalis and nitrite/nitrate reduction by A. faecalis in the medium with citrate. (A and B) Aerobic cultivation with nitrite and nitrate. (C and D) Anaerobic cultivation with nitrite and nitrate. solid triangles, nitrite concentration in panels A and C and nitrate concentrations in panels B and D; gray circles, citrate concentration; gray squares, OD₆₀₀.

Bacterial-electrochemical nitrite and nitrate reduction. (i) Biofilm observation.

The detailed biofilm cultivation method is described in the section “Biofilm preparation.” In step 3, two small pieces of graphite electrode from each experiment set were cut off. One was processed for scanning electron microscopy observation, and the other was used for protein determination to estimate the relative bacterial density on the electrode. The surface area with bacterium cells was approximately 2.3 cm². The protein densities on electrodes with +0.06, −0.06, and −0.15 V were 7.89, 16.6, and 19.9 mg/cm², respectively. This result was in accordance with the SEM observation (Fig. 3). It was obvious that the electrode with the −0.15-V poised potential (Fig. 3C) had more bacterial cells on the surface, and as the poised potential decreased, the cell densities on the electrode increased.

FIG 3.

SEM images of A. faecalis on graphite electrodes with +0.06-V (A), −0.06-V (B), and −0.15-V (C) poised potentials. Scale bars, 5 μm.

(ii) Bacterial-electrochemical nitrate reduction.

In the batch culture, nitrate could not be reduced by A. faecalis either aerobically or anaerobically. However in the BES, A. faecalis behaved differently in terms of nitrate reduction. The potentials poised in this experiment were +0.06, −0.06, and −0.15 V (versus NHE). Each experiment set had a corresponding sterile control. A blank control without poised potentials and bacterium (open circuit in Fig. 4) was used.

FIG 4.

Nitrate reduction by A. faecalis with poised potentials (A) and the current densities (B). Poised potentials: ▲ and △, +0.06 V; ● and ○, −0.06 V; ■ and □, −0.15 V. Solid symbols indicate experiment sets with poised potentials and A. faecalis inoculation; open symbols indicate experiment sets with poised potentials but without A. faecalis inoculation.

The results summarized in Table 1 indicated that in the BES, nitrate reduction rates were higher than those of sterile controls. In the blank control (open circuit), the nitrate concentration varied between 257.9 and 255.0 mg/liter. In sterile controls, the nitrate concentration slightly decreased, and the reduction rates for +0.06, −0.06, and −0.15 V were 2.74, 4.42, and 3.48 mg/liter/day, respectively. In the experiment with the +0.06-V poised potential and A. faecalis, the nitrate concentration slowly decreased, from 263.9 mg/liter to 250.3 mg/liter in 10 days. With the −0.06-V poised potential, the nitrate concentration decreased from 256.4 mg/liter to 178.3 mg/liter, corresponding to 30.4% nitrate, and the reduction rate was 17.3 mg/liter/day (from day 4 to day 10), while in the experiment with the −0.15-V poised potential, the percentage of nitrate reduced was 52.4%, from 252.7 mg/liter to 120.3 mg/liter, and the reduction rate was 28.5 mg/liter/day (from day 4 to day 10). Meanwhile, the current densities at −0.06 and −0.15 V were higher than the current density at +0.06 V (Fig. 4B). Generally, as the poised potentials decreased, the reduction rates increased.

TABLE 1.

Nitrate reduction with cathodes under conditions with and without A. faecalis

Expt set	Concn of nitrite reduced (mg/liter)	% of reduction	Reduction rate (mg/liter/day)
Without A. faecalis
+0.06 V	27.4	10.6	2.74
−0.06 V	44.2	17.2	4.42
−0.15 V	34.8	14.0	3.48
With A. faecalis
+0.06 V	13.6	5.15	1.36
−0.06 V	78.1	30.4	17.3a
−0.15 V	132.4	52.4	28.5a

^aCalculated from day 4 to day 10.

(iii) Bacterial-electrochemical nitrite reduction.

In the BES, the A. faecalis biofilm was previously prepared for nitrite reduction. Ten experiments were performed that were divided into the four groups described as follows. In group 1, there were three different potentials (−0.06, −0.15, and −0.30 V versus NHE) with A. faecalis. In group 2, there were three sterilized controls. In group 3, there were three controls only, with A. faecalis in flasks with organic-free simple medium under aerobic conditions. The bacterium inoculation amount was the same as those for the three sets in group 1. The bacterium amount was estimated by cell proteins, and two sets had the same proteins because two sets in group 1 had a similar amount of bacterium (∼1,661 mg protein on electrodes). Group 4 consisted of a blank control only with nitrite.

The current densities in group 1 were all negative (Fig. 5B), and according to the potentiostat software setup, the negative current meant that the electron flow was from the potentiostat to the cathode biofilm. The current density (absolute value) of the −0.06-V (versus NHE) experiment set was lower than those in the −0.15-V (versus NHE) and −0.30-V (versus NHE) experiment sets.

FIG 5.

Nitrite reduction by A. faecalis with poised potentials (A) and current densities of experiment sets with poised potentials and A. faecalis (B). Poised potentials: solid squares, −0.06 V; gray circles, −0.15 V; and gray triangles, −0.30 V. Solid lines without symbols indicate the sterile control, and dashed lines without symbols indicate the controls with A. faecalis but without potentials (A).

In the control groups, the nitrite concentration did not decrease, ranging from 46 mg/liter to 59.8 mg/liter. In group 2, the absence of nitrite reduction indicated that the electrodes with the potentials set in this study could not reduce nitrite. Although A. faecalis could reduce nitrite in the presence of sodium citrate in simple mineral medium under aerobic conditions (see “Nitrate and nitrate reduction under anaerobic and aerobic conditions”), in group 3, the bacterium could not reduce nitrite in the medium lacking sodium citrate. In group 4, no obvious nitrite oxidation by oxygen was observed.

The nitrite could only be reduced by the synergism of the bacterium and electrodes in the medium without carbon sources (Fig. 5A). The nitrite reductions behaved as the zero order kinetic reactions, according to Table 2. The slopes of three equations reflected the reaction rates, which were 1.98, 4.37, and 3.91 mg/liter/h for −0.06, −0.15, and −0.30 V, respectively. While in the aerobic batch culture described in the section “Nitrate and nitrate reduction under anaerobic and aerobic conditions,” the nitrite reduction rate was 0.50 mg/liter/h, which was much lower than the rates in the experiment sets with −0.06-, −0.15-, and −0.30-V poised potentials.

TABLE 2.

Nitrite reduction rates of three experiments in group 1

Voltage (V)a	Nitrite reduction rate (mmol/liter/h)
−0.06	1.98
−0.15	4.37
−0.30	3.91

^aThe result from −0.06 V was calculated with data from 0 to 30 h, and the results from −0.15 and −0.30 V were calculated with data from 0 to 12 h.

It would be better to directly detect and measure nitrite reduction products, such as nitric oxide, nitrous oxide, nitrogen gases, and ammonium (equations 1 to 4). However, in order to maintain an aerobic environment, the experiment system was an open system, and thus the gaseous phase could not be determined. As the currents of three experiment sets were recorded, it was possible to estimate the quantities of electrons flowing from the potentiostat to biofilm on the cathode and then reduction products. The theoretic stoichiometric ratios of nitrite reduced and electrons transferred are shown by the following equations:

$${{\text{NO}}_2}^{-}+{3\mathrm{H}}^{+}+{\mathrm{e}}^{-}\to \text{NO}+{\mathrm{H}}_2\mathrm{O}$$ (1)

$${{\text{NO}}_2}^{-}\hspace{0.17em}+{3\mathrm{H}}^{+}\hspace{0.17em}+\hspace{0.17em}{2\text{e}}^{-}\to 1/2\hspace{0.17em}{\mathrm{N}}_2\mathrm{O}+3/2\hspace{0.17em}{\mathrm{H}}_2\mathrm{O}$$ (2)

$${{\text{NO}}_2}^{-}+{4\mathrm{H}}^{+}+{3\mathrm{e}}^{-}\to 1/2\hspace{0.17em}{\mathrm{N}}_2+{2\mathrm{H}}_2\mathrm{O}$$ (3)

$${{\text{NO}}_2}^{-}+{8\mathrm{H}}^{+}+{6\mathrm{e}}^{-}\to {{\text{NH}}_4}^{+}+{2\mathrm{H}}_2\mathrm{O}$$ (4)

In this study, the reduction of 1 molecule of nitrite needed 1.63 (−0.06 V), 1.69 (−0.15 V), or 2.54 (−0.30 V) electrons at the maximum (Table 3). Thus, from the nitrite reduction/electron transfer ratio, the reduction products might be nitric oxide (NO [1 electron per nitrite molecule]), nitrous oxide (N₂O [2 electrons per nitrite molecule]), nitrogen gas (N₂ [3 electrons per nitrite]), or a mixture of three gases.

TABLE 3.

Amounts of reduced nitrite and electrons from the cathode with A. faecalis

Voltage (V)	Time range (h)	Amt (mmol) of:		Ratio of electrons/nitrite reduced
		Electrons	Nitrite reduced
−0.06	0–30	0.226	0.139	1.63
−0.15	0–15	0.238	0.141	1.69
−0.30	0–15	0.330	0.130	2.54

Nitrite reduction with electricity cutoff.

Further study focused on the issue that how external electrons influenced nitrite reduction in a microbial-electrochemical system. Here, a poised potential at −0.15 V versus NHE was applied due to the fastest nitrite reduction rate in the section “Bacterial-electrochemical nitrite reduction.” Three controls were set: the first was the sterilized control with only a −0.15-V poised potential, the second was the control with only the same amount of bacterium cells as the experiment set in simple medium without a carbon source, and the third was a blank control with neither bacterium nor poised potential.

The first 12 h was the fast nitrite reduction period, and the nitrite concentration decreased from 54.7 mg/liter to 23.92 mg/liter (Fig. 6) at a reduction rate of 2.3 mg/liter/hour. This rate was lower than that of the same potential experiment set in the section “Bacterial-electrochemical nitrite reduction,” and this might be caused by the different biofilm activities in different experiments.

FIG 6.

Potentiostat cutoff and turn-on and the influence on nitrite reduction by A. faecalis with poised potential (−0.15 V). Solid symbols indicate the experiment set with −0.15 V (versus NHE) potential in the presence of A. faecalis, and open symbols indicate controls with only potential (open squares) or only A. faecalis (open circles) and the blank control without potential and A. faecalis (open triangles).

At the time point of 12 h, the potentiostat was cut off to stop poising the potential on the cathode. The nitrite reduction continued for 1.5 h, from 23.92 mg/liter to 20.7 mg/liter. The reduction rate during this 1.5 h was estimated as 2.25 mg/liter/h, which was close to the reduction rate during the first 12 h, indicating that although the bacterium lacked external electrons, the nitrite reduction ability did not change. From 13.5 h to 24 h, the nitrite concentration slightly decreased from 20.75 mg/liter to 20.06 mg/liter, corresponding to a reduction rate of 0.063 mg/liter/h. At the 24-h time point, the potentiostat was turned on to continue powering the cathode. The nitrite reduction rate did not resume to the rate in the first 12-h period, and the reduction rate from 24 to 51 h was only 0.38 mg/liter/h, which was only 1/6 of the reduction rate in the first 12 h. In the other three controls, the nitrite concentrations were around 46 mg/liter.

DISCUSSION

Nitrate and nitrite reduction enhanced by cathode in BES.

The microbial denitrification in BES has been widely studied recently. In microbial fuel cells, denitrification could happen in the anode (21) and cathode (28). In some research, a series of voltages were applied between the anode and cathode (18, 20), and in the cathode, hydrogen was always produced. Hydrogen could be the energy source for microbial growth, such as that in the autohydrogenotrophic bacteria (29). Hydrogen generation in the cathode at least needed a potential of −410 mV (versus NHE) (19), and considering the overpotential in solutions, a potential of −440 mV was needed (19, 30). In this study, the potentials on the cathode were fixed at +0.06, −0.06, −0.15, and −0.30 V (versus NHE), and as a result, no hydrogen was produced in the cathode. Thus, the only energy source for A. faecalis was the cathode.

A. faecalis was not a nitrate reducer under either aerobic or anaerobic conditions, which was proved in the section “Nitrite and nitrate reduction under anaerobic and aerobic conditions” and in previous papers (23, 24). Nevertheless, in BES with cathodes at −0.06- and −0.15-V potentials (versus NHE), nitrate was partially reduced in 10 days. The nitrate reduction rates were 17.3 and 28.5 mg/liter/day at −0.06 and −0.15 V, respectively. These reduction rates were not the highest ever reported. Puig et al. reported a reduction rate of 75.7 ± 12.4 g N/m³/day (corresponding to 335 ± 55 mg NO₃/liter/day) in an MFC cathode with a chemolithoautotrophic bacterium, Oligotropha carboxidovorans strain OM5, which was capable of using hydrogen as an energy source (31). Su et al. reported that Pseudomonas alcaliphila, a nitrate reducer (32), reduced nitrate with electrodes as the solo electron donor (22). The nitrate reduction rate in Su's research was 0.16 ± 0.04 mol N/liter/day/m² (corresponding to 15.9 ± 4 mg/liter/day), which was slower than the nitrate reduction in the −0.15-V experiment set. Although denitrification in BES has been widely studied over the past several years, our study is the first one that used a non-nitrate reducer to reduce nitrate with an electrode as the sole electron donor, and the reduction rate was higher than the rates by the nitrate reducer Pseudomonas alcaliphila, with more negative potential (−303 mV versus NHE) (22).

A. faecalis could aerobically reduce nitrite, and the nitrite reduction ability originated from the copper-containing nitrite reductase coded for by nirK (33, 34). In the BES with an electrode as the sole electron donor, the reduction rates increased to 1.98, 4.37, and 3.91 mg/liter/h for −0.06, −0.15, and −0.30 V, respectively. Also in Su's study, the nitrite concentration of nearly 1.4 mmol/liter was not completely reduced in 5 days by Pseudomonas alcaliphila with an electrode as the sole electron donor. In this study, all nitrite (initial concentration of 1 mmol/liter) was depleted in less than 30 h (Fig. 5). Although the difference in nitrite reduction rates between this study and Su's study might be caused by the different cell densities on electrodes, from another point of view, this might indicate that A. faecalis had a higher affinity to electrode electrons than Pseudomonas alcaliphila.

The possible denitrification mechanism.

The mechanisms that microbial cells use with electrodes or mineral electrons have been widely discussed. A. faecalis cells were able to live on the electrodes (Fig. 3), as shown by the SEM images. Clearly, A. faecalis did not evolve a conductive “nanowire” to absorb electrons from the electrode. A. faecalis might gain electrons from the electrode by direct contact. Two hypotheses were proposed. A. faecalis might absorb electrons by outer membrane proteins, and thus the unknown outer membrane protein was reduced. One pathway was that the reduced protein immediately reduced nitrate or nitrite. In this situation, no intercellular reactions happened, and the denitrification site was on the cell surface. Another pathway was that the reduced protein transferred the electrons into the cell and might generate reductive compounds like FADH₂ or NADH via a biological pathway. The stored energy could then participate in the cellular denitrification pathway.

Considering the electricity cutoff experiment in the section “Nitrite reduction with electricity cutoff,” immediately after the cutoff of electricity, the nitrite reduction rate by A. faecalis was almost maintained (2.25 mg/liter/h versus 2.3 mg/liter/h before electricity cutoff). This indicated that the bacterium might have an “electron sink” to store electrons for reduction. After 1.5 h, the reduction rate drastically decreased to 0.063 mg/liter/h, which was almost 1/36 of the previous reduction rate. When the potentiostat turned on again, the nitrite reduction rate recovered to 0.38 mg/liter/h, which was only 1/6 of the rate before the cutoff. It was proposed that RubisCo might be an electron sink for a photoferrotrophic bacterium, Rhodopseudomonas palustris TIE-1, during cultivation with 100-mV poised potentials in the dark (35). The “electron sink” proposed in this study might be the redox protein on the cell surface or the intercellular reductive compounds.

Outlook.

A. faecalis was believed to have high affinity to photoelectrons derived from semiconducting mineral photocatalysis (4). In Lu's previous study, this bacterium was a dominant species in soil when cultivated with photoelectrons (4). This study confirmed that the simulated photoelectrons could enhance the A. faecalis denitrification process. As the semiconducting minerals and microbes are ubiquitous on the surface of the Earth (4), the photoelectrons generated by semiconducting minerals under solar irradiation could be utilized by microbes to support microbial metabolism. This process extends the knowledge of mineral and microbe interactions (including microbial dissolution of minerals and microbial formation of minerals) (36) and helps us to better understand how minerals influence the microbial world.