INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

Nitrogen present in freshwaters is derived from three key sources: (i) agriculture, (ii) domestic sewage and industrial effluents, and (iii) rainfall and dry deposition (11). In Northern Ireland, associated with large inputs of agriculturally derived N substrates into the aquatic environments, there have been reports of elevated nitrite (NO₂) concentrations (31) which are believed to be toxic to aquatic biota (12). Concentrations of NO₂, an intermediate in oxidative and reductive pathways mediated by bacteria (Fig. 1), regularly exceed the European Union guideline of 0.003 mg of N liter¹ for rivers supporting salmonid fish (9). The two main substrates for NO₂ production are NH₄⁺, through nitrification, and NO₃, via a number of NO₃-reductive pathways. Previous experiments in our laboratory have shown that NH₄⁺ oxidation via nitrification, occurring only in the oxygen diffusion zone, is mainly responsible for the elevated NO₂ levels observed in fast-flowing aerobic small streams (30). However, the high concentrations of NO₂ found in larger rivers in summer under warm, slow-flowing conditions have recently been attributed to anaerobic NO₃-reducing processes (13). Our laboratory incubation experiments showed that maximal concentrations of NO₂ were found in anaerobic sediments deeper than 6 cm and were associated with a high concentration of added metabolizable C where dissimilatory NO₃ reduction to NH₄⁺ (DNRA) was the predominant NO₃-reducing pathway.

View larger version (28K): [in this window] [in a new window]   FIG. 1.   Nitrogen transformation processes that involve NO2 production and consumption. Pathway 1, denitrification; pathway 2, DNRA; pathway 3, nitrification. (Reproduced with permission from Kelso et al. [13]).

The partitioning of NO₃ between denitrification and fermentative DNRA is dependent on the activities of two distinct bacterial populations (21). The complexity of the functioning of these multispecies microbial communities is only beginning to be understood (5), but it is believed that the availability of organic matter is probably of prime importance in regulating the relative rates of DNRA and denitrification. Organic matter has the potential to mediate between DNRA and denitrification directly by providing a direct electron donor (C substrate) and indirectly by taking up oxygen, thus creating anoxic conditions (35). The ratio between available C, which acts as an electron donor, and NO₃, an electron acceptor, is important not only in influencing which NO₃-reducing pathway is followed (36) but also in determining what end products are produced (28). Denitrification is the dominant process in NO₃-rich sediments with a poor C supply; conversely, DNRA is the dominant process in environments rich in C which are preferentially colonized by fermentative bacteria (36). To ensure their survival, microbial communities must be very versatile, and this versatility is reflected in their ability to be able to metabolize a large range of C substrates which are available in the absence of oxygen (3). However, the chemical structure of the C source may have diverse effects on the biochemical reduction rates of NO₃ and NO₂ (18, 27).

The aim of this study was to investigate the effects of a range of C compounds commonly found in agricultural pollutants on the biochemical pathways of NO₂ accumulation. Our previous anaerobic study (13) focused on NO₂ accumulation in rivers where sediment, taken from a range of depths, was supplemented with C in the form of glucose under a high concentration of NO₃ (13 mg of N liter¹) and negligible NH₄⁺ concentrations. In the present study differentially ¹⁵N-labeled NH₄NO₃ and a range of C supplements were provided at environmental concentrations for sediment incubations in order to deduce which N pathway supported NO₂ accumulation.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Experimental design. Approximately 20 liters of surface water was collected from the midstream section of a continuously monitored "unpolluted" station on the Upper Bann River, stored in polyethylene bottles, and returned immediately to the laboratory. The water was filtered through GF/C filters (pore size, 0.45 µm; Whatman International Ltd., Kent, United Kingdom) and refrigerated at 4°C. Sediment was collected in October 1997 from a sampling station on the Upper Bann River (Irish Grid reference, J192 429, described previously in the report of Kelso et al. [13]), which drains an agricultural catchment in the counties of Armagh and Down, Northern Ireland. Eleven sediment cores (5 cm [diameter] by 15 cm [depth]) were collected by pushing Plexiglas tubes into areas of sediment accumulation. The top end of the column was closed with a rubber stopper, and the column was then pulled out with the intact core. The stopper was removed, and a plunger was inserted into the bottom end to push the core out. The sediments from the cores were pooled, sieved (sieve pore size, <8 mm), thoroughly mixed, and left at room temperature for 48 h. One-ninth of a core (approximately one-third of the core size used in our previous study [13]) was added to approximately 180 ml of river water in 90 Kilner jars. Then one of two N treatment samples in 10-ml aliquots was added to the jars to give a N concentration of 6 mg of N liter¹, either as ¹⁵NH₄NO₃, where the NH₄⁺ moiety was labeled at 40 atom% excess, or as NH₄¹⁵NO₃, where the NO₃ moiety was labeled at 40 atom% excess. In addition to the two differentially labeled substrates, five C treatment solutions (a distilled-water control, glucose, glycine, acetate, and formate) were added to obtain a final C concentration of 1.0 g of C liter¹. Acetate and formate were prepared by neutralizing the respective acids to pH 7 with KOH (22). During incubation there was approximately 5 cm of water above the sediment. Immediately after the addition of all substrates, a nylon lid with a gas-sampling septum was fitted to each jar with an O-ring to form a gastight seal. Each treatment was replicated three times, with replicates randomly distributed in incubators maintained at a temperature of 23°C. Analyses were made at time zero, with subsequent destructive sampling carried out at 6, 24, and 48 h.

Gas analyses. At each sampling time prior to destructive sampling, two 12-ml gas samples were taken through the gas-sampling septum by using a 20-ml gastight syringe with a push-button valve into evacuated (<100 Pa) septum-capped Exetainers (Europa Scientific, Crewe, United Kingdom). The gas was analyzed to determine the ¹⁵N contents of N₂O and N₂ and the concentration of N₂O by continuous-flow isotope-ratio mass spectrometry. Analyses were performed with a Europa Scientific model 20-20 stable-isotope analyzer interfaced to a Europa Scientific trace gas preparation system with a Gilson auto-sampler. The valve switching was automated so that the ¹⁵N contents of N₂ and N₂O could be determined from the same sample. The ion currents (I) at m/z 44, 45, and 46 enabled ⁴⁵R(⁴⁵I/⁴⁴I) and ⁴⁶R(⁴⁶I/⁴⁴I), where R is the ratio of I and the superscript number indicates the mass/charge ratio, to be calculated for N₂O. The ¹⁵N content of the N₂O was calculated from either the ⁴⁵R, with equations 5 and 7, or the ⁴⁶R, with equations 6 and 7 in the work of Stevens et al. (33). For N₂, the ion currents at m/z 28, 29, and 30 were measured by isotope-ratio mass spectrometry (33). The differences between ²⁹R(²⁹I/²⁸I) and ³⁰R(³⁰I/²⁸I) for enriched and normal atmospheres enabled the flux of N₂ to be calculated (19). The concentration of N₂O was calculated as described by Stevens et al. (33) from measurements of ⁴⁴I, ⁴⁵I, and ⁴⁶I. The flux of N₂O was calculated from the change in N₂O concentration with time. It was assumed that the concentration of N₂O at the start of each flux measurement was 310 ppb. A further two 5-ml gas samples were transferred to helium-filled 10-ml crimp-capped septum vials to determine the concentrations of CO₂ and CH₄ in the headspace by using a Varian Genesis headspace auto-sampler interfaced to a Varian model 3800 chromatograph fitted with a 5-m by 2-mm Porapak gas QS column (80-100 mesh). Carbon dioxide was measured with a thermal conductivity detector, and CH₄ was measured with a flame ionization detector.

Analysis of N fractions. The pH of the liquid fraction was measured with the aid of an Orion expandable ion analyzer, model EA940. For determining the ¹⁵N contents of NH₄⁺, NO₂, and NO₃, a sufficient amount of KCl was added to the solution to produce a 2 M KCl solution, which was subsequently filtered sequentially through GF/C and GF/F filters (Whatman International Ltd.). Concentrations of NO₂ were determined spectrophotometrically (Hitachi spectrophotometer model V-2000) by the sulfanilamide-naphthylene ethylene diamine reaction (24). Concentrations of NO₃ and NH₄⁺ in the KCl solutions were determined by segmented-flow analysis (Technicon Random Access Automated Chemistry System 800+) (4). Particulate organic N (PON) was calculated by the methodology of Koike and Hattori (14) as follows: PON = (total ¹⁵N added originally)  (¹⁵N in NH₄⁺ + NO₂ + NO₃ + N₂O + N₂). The ¹⁵N contents of NO₂, NO₃, and NH₄⁺ were analyzed by methods based on their conversion to N₂O (16, 32).

Calculation of simultaneous nitrification and NO₃ reduction rates. The equations employed were those developed by Koike and Hattori (15):

(1)

(2)

where j is time (6, 24, or 48 h); N_j is the NO₃ plus NO₂ concentration at time j (in milligrams of N per liter); X_j is the ¹⁵N content of NO₃ plus NO₂ in excess of the level in nature (natural abundance) at time j (atom percent excess); is the average ¹⁵N content of NO₃ plus NO₂ in excess of natural abundance between time j and time j plus 1 day (atoms percent excess); _a is the average ¹⁵N content of NH₄⁺ in excess of natural abundance between time j and time j plus 1 day (atoms percent excess); Z is the rate of nitrification per time interval; and Y is the rate of NO₃ reduction per time interval. Equation 1 states that the change in NO₂ and NO₃ concentrations per time interval is equal to the amounts of NO₂ and NO₃ produced from nitrification minus the reduction in NO₂ and NO₃ brought about by NO₃-reducing processes in that same time interval, i.e., the nitrification rate minus the NO₃ reduction rate per time period. Equation 2 states that the change in the ¹⁵N content of NO₂ and NO₃ per time interval is equal to the average ¹⁵N content of NH₄⁺ nitrified minus the average of the combined ¹⁵N contents of NO₂ and NO₃ reduced, all within that time period. Natural abundance is assumed to be 0.37%, and the rates of nitrification and NO₃ reduction are assumed to be constant during each time interval.

Statistical analyses. To determine the significance of the effects of the different C treatments, the logarithms of concentration and enrichment data from the sediment studies were subjected to analysis of variance with Genstat software (10).

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

To deduce and quantify the mechanisms by which NO₂ is produced, we employed the paired-incubation technique where the NH₄⁺ and NO₃ pools are differentially labeled with ¹⁵N and used the results in ¹⁵N pool dilution calculations (15). The rationale for the paired-incubation technique is that if NO₃-reductive pathways, e.g., denitrification and DNRA, are solely responsible for NO₂ accumulation (i.e., all the accumulated NO₂ came from the NO₃ pool), then the ¹⁵N enrichment of the NO₂ pool derived from NH₄¹⁵NO₃ (Table 1) would be expected to match that of the NO₃ pool once isotopic equilibrium had been established. In contrast, if nitrification was the only source of NO₂ during ¹⁵NH₄NO₃ supplementation (Table 2) then the levels of enrichment of the NO₂ and NH₄⁺ pools would be similar.

Control treatment. Because organic C is normally present in freshwater sediment, additional organic C supplementation is not essential as an energy source to stimulate NO₃-reductive pathways (34). This was observed in the control treatment, where although no additional organic C was provided, 44.5% of the initial ¹⁵NO₃ label was consumed during a 48-h period (Fig. 2) at an average rate of 1.1 mg of N liter¹ day¹ (Table 3). This consumption was reciprocated by an increase in NO₂ concentrations, which showed a peak at 24 h with a concentration of 82 µg of N liter¹ (Fig. 3). The average ¹⁵N enrichment of the NO₂ produced under NH₄¹⁵NO₃-enriched conditions (29.4 atom% excess) was closer to that of NO₃ (36.7 atom% excess) than to that of NH₄⁺ (0.35 atom% excess) (Table 1), suggesting that the NO₂ was predominantly of NO₃ origin. Although the average rate of NO₃ production via nitrification was substantially lower than the rate of NO₃ utilization (Table 3), the 4 atom% excess enrichment of the NO₂ pool detected under ¹⁵NH₄NO₃-supplemented conditions (Table 2) occurring in conjunction with dilution of the NO₂ pool arising from NH₄¹⁵NO₃ incubations (Table 1) is evidence that nitrification of NH₄⁺ does contribute to the NO₂ pool. Nitrous oxide can be produced by both NO₃ reduction and NH₄⁺ oxidation (23, 25); however, the absence of a statistical difference between the ¹⁵N atom percent excesses of NO₃ and N₂O produced from NH₄¹⁵NO₃ (Table 1) indicates that only NO₃-reductive pathways were responsible for the 146 µg of N₂O-N produced. Nitrogen gas, the terminal product of denitrification, was produced in concentrations similar to those of N₂O, acting as a sink for 10% of the original ¹⁵NO₃ label (Fig. 2). There is evidence for the DNRA pathway via NH₄⁺ production, since 1% of NO₃ was detected in NH₄⁺ fractions (Fig. 2). Further, CH₄ concentrations, a possible indicator of fermentative activity, increased by 40% from a base concentration of 3 ppm (base concentrations were similar in all treatments), while CO₂ concentrations, indicative of fermentative and mineralization activity, doubled from a typical initial concentration of 2,000 ppm. It was not possible to directly determine whether the fate of NH₄⁺ originating from the DRNA pathway was assimilated into PON. After 48 h the unexplained fraction, which we assume to be NH₄⁺ assimilated into PON, represented as much as 22% of the original ¹⁵NO₃ label (Fig. 2).

View larger version (30K): [in this window] [in a new window]   FIG. 2.   Fate of NO3 label during anaerobic incubation with NH415NO3 under the control conditions and with four carbon treatments. The remainder of the label was still present in the NO3 pool.

View this table: [in this window] [in a new window]   TABLE 3.   Rates of nitrification and NO3 reduction in sediment enriched with differentially 15N-labeled NH4NO3a

View larger version (11K): [in this window] [in a new window]   FIG. 3.   Concentrations of NO2 produced under the control conditions and with four carbon treatments. Error bars indicate the standard errors of the means (n = 6).

Glycine treatment. The value of glycine as an additional energy source appeared to be limited, as the rate of NO₃ reduction was similar to that of the control (Table 3). However, glycine was an important source of additional N via deamination and provided an NH₄⁺ influx of 60 mg of N liter¹. A maximal NO₂ accumulation of 74 µg of N liter¹ was reported at 24 h (Fig. 3). Under NH₄¹⁵NO₃-supplemented conditions, the ¹⁵N enrichment of the NO₂ pool (Table 1) indicated that there was an additional source of NO₂ other than that formed from NO₃ reduction. However a nitrification rate of 0.04 mg of N liter¹ day¹ (Table 3) in addition to a NO₂ pool enrichment of 0.25 to 0.51 atom% excess reported after ¹⁵NH₄NO₃ treatments (Table 2) is inadequate to represent another large source of NO₂. The mole fraction of N₂O [i.e., N₂O/(N₂O + N₂)] remained constant at almost 50%, at concentrations resembling that of the control (Fig. 2). A substantial proportion (7.67%) of the initial ¹⁵NO₃ label was present in the NH₄⁺ pool (Fig. 2) and, together with almost a trebling in CO₂ concentrations and a 20% increase in CH₄, was strong evidence that DNRA was active. After 48 h, the unexplained fraction was equivalent to 22.9% of the initial ¹⁵NO₃ label.

Acetate treatment. The addition of acetate, for potential usage as an energy source, significantly retarded (P < 0.001) rates of NO₃ utilization (Table 3), with 72% of the initial ¹⁵NO₃ label still unprocessed after 48 h (Fig. 2). In contrast to what occurred with other treatments, NO₂ accumulated at a constant rate to approximately 80 µg of N liter¹ at the end of the 48-h monitoring period (Fig. 3). Although nitrifying bacteria were inactive as evidenced by the low ¹⁵N enrichment of NO₃ produced under ¹⁵NH₄NO₃-enriched conditions (0.08 atom% excess at most [Table 2]) and the almost negligible nitrification rates (Table 3) under NH₄¹⁵NO₃-enriched conditions, the NO₂ pool was not entirely derived from NO₃, with the ¹⁵N enrichment of the NO₂ 18.8 atom% excess initially increasing to 32.9 atom% excess at the termination of the investigation. The concentration of nitrous oxide produced was significantly lower than that of the control (P < 0.001), with only 56 µg of N₂O-N detected in a 48-h period. None of the N₂O was reduced further to N₂. Only 0.35% of the initial ¹⁵NO₃ label accumulated as NH₄⁺, and 21.76% was calculated by difference to be present in the PON pool.

Formate treatment. Apart from a significant accumulation of NO₂, concentrations of terminal products detected in formate-amended sediments were not significantly different from concentrations detected after the acetate treatment. However, NO₂ accumulated in a linear fashion to a concentration of 1,343 µg of N liter¹ (Fig. 3), which contrasted with the small N₂O concentrations (72 µg of N₂O-N) arising from NO₂ reduction and zero N₂ production (Fig. 2). After 48 h, 11.8% of the initial ¹⁵NO₃ label was detected in the PON pool. Although CO₂ concentrations were twice the original levels, only 0.3% of the initial ¹⁵NO₃ label was processed by DNRA into the NH₄⁺ pool (Fig. 2). Nitrification rates were shown to be negligible (Table 3).

Glucose treatment. In glucose-enriched sediments, NO₃ reduction pathways were markedly stimulated to the extent that 5.3 mg of N liter¹ (90.6% of the ¹⁵NO₃ label) was metabolized within 48 h at an average rate of 2.6 mg of N liter¹ day¹ (Table 3). At 24 h, NO₂ concentrations showed a peak at 163 µg of N liter¹; thereafter, the NO₂ concentration decreased to 53 µg of N liter¹ (Fig. 3). Throughout the investigation, the ¹⁵N enrichment of the NO₂ pool under NH₄¹⁵NO₃ conditions was approximately 36 atom% excess, which is very similar to the initial 37 atom% excess of the NO₃ pool (Table 1), indicating that NO₃ reduction accounted for NO₂ production. The rapid NO₂ depletion occurred in association with a trebling in CH₄ efflux and a 10-fold increase in CO₂ concentrations. Of the initial NO₃ pool, 48% was transformed to N₂ via complete denitrification. Fermentative DNRA activity as indicated by CO₂, CH₄, and NH₄⁺ measurements was also significant, with 2.01% of the ¹⁵N label being detected in the NH₄⁺ pool. Assimilation of NO₃ into PON was 50% greater than that of the other treatments, incorporating 34.8% of the ¹⁵NO₃ label. Nitrification was responsible for 3.5% of NO₂ produced (Table 3). However, activity was retarded after 24 h of activity, when the enrichment of the NO₂ pool derived from ¹⁵NH₄NO₃ declined from 1.24 to 0.59 atom% excess (Table 2), probably as a consequence of increasing anoxia.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

Potential of carbon substrates to influence NO₂ formation pathways. Nitrite is a common intermediate in at least three different biochemical pathways that occur in freshwater sediments: nitrification, denitrification, and DNRA (Fig. 1). The relative contribution of these processes to the accumulation of NO₂ is governed by the species compositions of the bacterial populations resident in the sediments, available C and N substrates, and the surrounding environmental conditions. In this study, NO₃-reducing processes deemed responsible for large NO₂ concentrations were predominantly controlled by the C substrate present. Two dissimilar patterns of NO₂ accumulation reflecting different formation pathways were observed (Fig. 3). The most common pattern, observed in the glucose, glycine, and control treatments, exhibited a NO₂ peak at 24 h. The second pattern, observed with the acetate and formate treatments, exhibited continuous NO₂ production that resulted in a high concentration of NO₂ (1,300 µg of N liter¹ [approximate]) in the formate treatment and a low concentration of NO₂ (<100 µg of N liter¹) in the acetate treatment.

Glucose, glycine, and control treatments. Nitrite accumulation patterns detected in the glucose, glycine, and control treatments were indicative of the multistage DNRA and denitrification processes occurring simultaneously. Not all denitrifiers have the capability of completely reducing NO₃ to N₂, with the enzyme NO₂ reductase commonly being absent (35). Rather than being limited by the genetic capability of the organism, reduction beyond the NO₂ step is restricted more by the environmental conditions (29). For NO₂ to accumulate in the environment, activity of the NO₃ reductase enzyme, which reduces NO₃ to NO₂, must function at a higher rate than the corresponding NO₂ reductase. Often it is the environmental O₂ concentrations that regulate the reduction of NO₂ in denitrification, through stimulation of the enzyme NO₂ reductase when suitable conditions are induced. This finding is in contrast to what occurs with the corresponding NO₃ reductase, which is ever present and functional in natural environments (6, 8) and may elevate NO₂ concentrations before sufficient synthesis of the NO₂ reductase has occurred. The accumulation of NO₂ is not entirely restricted to denitrification, since it is also believed to be a common trait of DNRA (7) due to either inhibitory effects of NO₃ on the fermentative NO₂ reductase enzyme (28) or repression of this enzyme (21).

Acetate and formate treatments. Acetate and formate are capable of reducing NO₃ only through the denitrification pathway (3). In Escherichia coli, it has been shown that formate-dependent NO₂ reductase is active only when NO₃ is scarce and NO₂ is available (6). This can lead to substantial concentrations of NO₂, which under acidic conditions accumulates predominantly as the protonated species, nitrous acid (HNO₂) (1, 2, 20). It has been suggested that toxic effects on the cell may be exerted by HNO₂, which is capable of increasing the proton permeability of the cell membrane by shuttling protons between the two sides. To expedite the release of protons from a cell, a large proportion of energy (26), which rations the C essential to promote N-reducing processes, is required. Because acetate is a terminal product of C metabolism with a low redox potential (37), the low NO₃ utilization rates resulting in reduced NO₂ accumulations observed in the present study may have arisen from the inability of acetate to support further N transformations.

Environmental implications. In Northern Ireland, agriculture has a major impact on the environment and has the potential to significantly elevate NO₂ concentrations in freshwaters, either directly through leaching or indirectly by providing C and N substrates for sediment transformations. The application of slurry to grassland is a common agricultural practice. Acetic acid is the largest organic acid constituent of slurry (4.6 g liter¹) and has the potential to stimulate NO₂ accumulation by denitrifying populations in freshwaters. On the other hand, silage effluent, the most common pollutant of rivers (17), is capable of supplying readily available C-rich substrates, e.g., glycine and glucose, which may support NO₂ accumulation by both denitrifying and DNRA bacteria.