Anaerobic Ammonium Oxidation Measured in Sediments along the Thames Estuary, United Kingdom

Abstract

Until recently, denitrification was thought to be the only significant pathway for N₂ formation and, in turn, the removal of nitrogen in aquatic sediments. The discovery of anaerobic ammonium oxidation in the laboratory suggested that alternative metabolisms might be present in the environment. By using a combination of ¹⁵N-labeled NH₄⁺, NO₃⁻, and NO₂⁻ (and ¹⁴N analogues), production of ²⁹N₂ and ³⁰N₂ was measured in anaerobic sediment slurries from six sites along the Thames estuary. The production of ²⁹N₂ in the presence of ¹⁵NH₄⁺ and either ¹⁴NO₃⁻ or ¹⁴NO₂⁻ confirmed the presence of anaerobic ammonium oxidation, with the stoichiometry of the reaction indicating that the oxidation was coupled to the reduction of NO₂⁻. Anaerobic ammonium oxidation proceeded at equal rates via either the direct reduction of NO₂⁻ or indirect reduction, following the initial reduction of NO₃⁻. Whether NO₂⁻ was directly present at 800 μM or it accumulated at 3 to 20 μM (from the reduction of NO₃⁻), the rate of ²⁹N₂ formation was not affected, which suggested that anaerobic ammonium oxidation was saturated at low concentrations of NO₂⁻. We observed a shift in the significance of anaerobic ammonium oxidation to N₂ formation relative to denitrification, from 8% near the head of the estuary to less than 1% at the coast. The relative importance of anaerobic ammonium oxidation was positively correlated (P < 0.05) with sediment organic content. This report of anaerobic ammonium oxidation in organically enriched estuarine sediments, though in contrast to a recent report on continental shelf sediments, confirms the presence of this novel metabolism in another aquatic sediment system.

Since the 1970s, substantial research has focused on the ability of estuarine sediments to attenuate riverine nitrogen (N) loads before they affect coastal seas (4, 18, 19, 31, 33). Estuarine sediments are essentially anaerobic below a few surface millimeters, and the mineralization of organic matter proceeds via alternate electron acceptors such as NO₃⁻ and SO₄²⁻ (20, 27). In turn, the reduction of NO₃⁻ removes NO₃⁻ from the overlying waters. Until recently, it was largely thought that NO₃⁻ could be either reduced to N₂ gas via denitrification (a facultative metabolism mediated by a variety of bacteria) and lost from the system or reduced to ammonium (NH₄⁺) by fermentative metabolisms and hence conserved within the sediments (dissimilatory nitrate reduction to ammonium [DNRA]) (8, 23). It had been demonstrated that changes in sediment organic loadings and estuarine NO₃⁻ concentrations may affect the partitioning between these two end products of NO₃⁻ reduction (11, 12). The discovery within the laboratory (17) of anaerobic ammonium oxidation revealed a novel metabolism that could short circuit the N cycle, bypassing what was previously thought to be a critical aerobic nitrification phase and potentially providing an alternative pathway for N₂ gas formation in the environment (Fig. 1).

FIG. 1.

Simplified N cycle showing how anaerobic ammonium oxidation (Anammox) (broken arrows) bypasses the classic coupling of aerobic nitrification to denitrification, with the oxidation of NH₄⁺ and N₂ formation being driven by the reduction of NO₂⁻. Note the process of DNRA and that NO₂⁻ is an intermediate of both DNRA and nitrification.

Originally, it was thought (17) that anaerobic ammonium oxidation coupled the oxidation of NH₄⁺ to the reduction of NO₃⁻:

(1)

Further work, however, showed that the oxidation of ammonium was actually coupled to the reduction of nitrite rather than nitrate (34, 35):

(2)

The application of this process to the treatment of nitrogenous waste has received a great deal of attention (9, 10), and more recently, the organism responsible has been classified as a new autotrophic planctomycete (28).

Anaerobic ammonium oxidation was recently reported to account for as much as 24 and 60% of N₂ formation in continental shelf sediments in relatively deep water (380 and 695 m, respectively) but less than 2% of N₂ formation in eutrophic shallow coastal bay sediments (30). The drop in the significance of anaerobic ammonium oxidation for N₂ formation relative to denitrification was attributed to changes in organic matter availability and hence sediment reactivity. If this trend continues inshore, where organic content and reactivity further increase, anaerobic ammonium oxidation might be assumed to be insignificant in organically enriched estuarine sediments (33).

The purpose of this research was to look for direct evidence of environmental anaerobic ammonium oxidation in estuarine sediments. Initial trials were designed to simply assay sediment at one site in the Thames estuary. Having established that anaerobic ammonium oxidation was detectable in estuarine sediments, the trials were extended downstream to reflect the changing sediment characteristics of both organic content and reactivity.

MATERIALS AND METHODS

Sampling sites.

The Thames estuary is a major macrotidal estuary on the southeastern coast of England which flows into the North Sea at Southend-on-Sea (Fig. 2). Oxygen saturation in the Thames estuary generally drops to 20% in the summer at site 1 and does not return to full saturation until some 50 km seaward at Southend-on-Sea (33). In contrast, the waters are highly enriched with NO₃⁻ and NO₂⁻, and their concentrations are on average for the year ∼550 and 9 μM, respectively, at 0 salinity. At the outset of this study, no other studies of anaerobic ammonium oxidation in the environment had been carried out, and until techniques were proven to be reliable, experiments were restricted to the most-NO₃⁻-enriched and -oxygen-depleted sediments near the major sewage discharge point at Crossness (site 1). Sediments from along the estuary were collected from sites close to those described in detail earlier (33). At sites 1 and 2, the sediments have a high organic and silt-clay content, both of which then decrease seaward along the estuary from sites 3 to 6 (Table 1 and see below).

FIG. 2.

Map of the Thames estuary on the southeastern coast of England showing locations of the six sample sites (filled circles) and the two major STWs (open circles).

TABLE 1.

Sediment characteristics and representative salinity of the overlying water at the six sites in the Thames estuary

Site	Particle sizea (0-5 cm [≥50%, wt/wt])	Organic carbon (% of dry wt)	Porosity	Specific gravity (g cm−3)	Salinity (%)
1 Crossness	Silts and clays	2.91	0.76	1.21	2
2 Grays	Silts and clays	3.59	0.76	1.24	10
3 Gravesend	Very fine sands	2.70	0.67	1.45	10
4 Allhallows	Very fine sands	1.18	0.65	1.46	26
5 Southend	Fine sands	0.45	0.50	1.68	30
6 Grain	Fine sands	0.26	0.47	1.74	30

^aFor detailed descriptions, see reference 33.

Sediment preparation.

Sediment samples (oxic and suboxic layers, 0 to 2 cm) were collected from intertidal flats along the Thames estuary at low tide (Fig. 2), stored in plastic bags, and returned to the laboratory within 1 h. Sediment slurries (30 ml; 50%, vol/vol) were prepared in serum bottles (37 ml) with low-nutrient seawater, adjusted to in situ salinity with distilled water, sealed, degassed (oxygen-free nitrogen, 20 min), and preincubated on rollers in the dark at 15°C for 6 h in a constant-temperature room. All subsequent experiments were carried out in a constant-temperature room at 15°C. Anoxia was checked by sampling the headspace in each serum bottle and measuring the O₂ with a gas chromatograph fitted with an electron capture detector. Preliminary trials showed that ambient NO₃⁻ and NO₂⁻ were reduced to less than 1 μM during the preincubations.

Measuring anaerobic ammonium oxidation.

Slurries from site 1 were enriched to approximately 10% above the ambient level with concentrated stocks of labeled ¹⁵NH₄⁺ (18 mM ¹⁵NH₄Cl [99.3 ¹⁵N atom%]; Sigma-Aldrich, Poole, United Kingdom) and either ¹⁴NO₂⁻ or ¹⁴NO₃⁻ (or both) at either 800 μM or a range up to 3,200 μM and incubated as described above. Some additional trials were carried out with autoclaved sediment in serum bottles to confirm that the N transformations were biologically mediated. Initial trials at 200 μM NO₂⁻ or NO₃⁻ had demonstrated a very rapid turnover (1 to 2 h) in these sediments, and concentrations were increased to enable production of ²⁹N₂ to be measured with time. Under anaerobic conditions and the above-described isotopic labeling, any anaerobic ammonium oxidation coupled to the reduction of NO₂⁻ or NO₃⁻ would yield ¹⁵N-labeled gas according to the following formulas:

(3)

(4)

The sole production of ²⁹N₂ would confirm anaerobic ammonium oxidation coupled to the reduction of NO₂⁻ rather than NO₃⁻ in these sediments. Assuming that the pool of ¹⁴NH₄⁺ turned over at the same rate as ¹⁵NH₄⁺, the total rate of anaerobic ammonium oxidation could then be calculated with the equation (29)

(5)

(where dN₂ is total N₂ production by anaerobic ammonium oxidation per unit of time [dt], d²⁹N₂ is the measurement of ²⁹N₂ per unit of time, NH₄⁺_total is the total soluble NH₄⁺ pool, and [¹⁵NH₄⁺] is the concentration of ¹⁵NH₄⁺ determined by difference from the nonenriched reference samples) and expressed as a proportion of the total NO₂⁻ or NO₃⁻ reduction in the slurries. Separate slurries were enriched with their respective analogues, e.g., ¹⁴NH₄⁺, ¹⁵NO₂⁻, and ¹⁵NO₃⁻ (>99% of the ambient level, 93 mM stock Na¹⁵NO₃⁻/NO₂⁻ [99.2 ¹⁵N atom%]; Sigma-Aldrich), which gives a measure of total ¹⁵N gas production (²⁹N₂ and ³⁰N₂) from either ¹⁵NO₂⁻ or ¹⁵NO₃⁻. With ¹⁵NO₂⁻ or ¹⁵NO₃⁻, the production of ²⁹N₂ can be due to both anaerobic ammonium oxidation (A₂₉) and denitrification (D₂₉) (30). Essentially, the proportion due to each respective process can be calculated from the equation A₂₉ = P₂₉ − D₂₉, where A₂₉ equals the production of ²⁹N₂ gas due to anaerobic ammonium oxidation, D₂₉ equals the production of ²⁹N₂ gas due to denitrification, and P₂₉ equals the total production of ²⁹N₂ measured by isotope ratio mass spectrometry (see below). In contrast, the production of ³⁰N₂ (P₃₀) is assumed to be due solely to denitrification of ¹⁵NO₂⁻/NO₃⁻; hence, P₃₀ equals D₃₀ and, from this, D₂₉ can be calculated, assuming random pairing of ¹⁴N and ¹⁵N from the labeled NO₂⁻ or NO₃⁻ pools, which, with only trace amounts of ¹⁴NO₂⁻/NO₃⁻ and enrichment with 99.2% atom ¹⁵NO₂⁻/NO₃⁻, would usually be less than 1.5% (6). Overall, the calculations (30) give an estimate of total anaerobic ammonium oxidation and denitrification in the sediment slurries. These estimates can then be compared to ²⁹N₂ production in the presence of ¹⁵NH₄⁺ after taking into account the proportionate labeling of the ¹⁴N- and ¹⁵N-NH₄⁺ pools. Trials were extended along the estuary by using labeling with ¹⁵NO₃⁻, as this gives a measure of total denitrification and anaerobic ammonium oxidation in the same serum bottle.

During the incubations, independent slurries were sacrificed with time, and a gas sample (4 ml) was collected from the slightly pressurized headspace with a gastight syringe (SGE gastight Luer lock syringe; Alltech Associates Ltd., Carnforth, Lancashire, United Kingdom), with allowance for equilibration to atmospheric pressure, and transferred through a septum to an inverted water-filled vial (12-ml Exetainer vial; Labco Ltd., High Wycombe, United Kingdom), with venting of water to atmosphere through a small-bore needle. Microbial activity was then inhibited by injecting ZnCl₂ (250 μl; 50%, wt/vol) through the septum. Pore waters were recovered by centrifugation of the slurry, filtered (0.2-μm-pore-size Minisart Plus filter; Sartorius UK Ltd.), and frozen (−20°C) until analysis. Exchangeable ammonium was recovered by using a KCl extraction technique, which recovered >95% of adsorbed NH₄⁺ by a double extraction with 2 M KCl (16).

Analytical procedures.

All nutrient analyses (NO₃⁻, NO₂⁻, and NH₄⁺) were carried out with a continuous-flow autoanalyzer (SAN⁺⁺; Skalar, De-Breda, The Netherlands) and standard colorimetric techniques (13). Salinity was measured with a handheld refractometer. Water content, specific gravity, and porosity were determined from the dry weights and wet weights of known volumes of sediment. Acidified dried sediment samples (with allowance for hydroscopic adsorption [see reference 7]), were analyzed for organic carbon with an elemental analyzer coupled to a continuous-flow isotope ratio mass spectrometer calibrated with known quantities of urea (Delta Matt Plus; Thermo Finnigan, Bremen, Germany).

Mass/charge ratios for m/z 28, m/z 29, and m/z 30 nitrogen (²⁸N₂, ²⁹N₂, and ³⁰N₂) were measured by the Natural Environment Research Council (NERC) ¹⁵N stable isotope facility (Centre for Ecology and Hydrology-Merlewood, Cumbria, United Kingdom). The headspace of each respective vial was sampled (20 μl) with a gastight precision syringe and directly injected into a N₂ preparative interface coupled to an Isoprime isotope ratio mass spectrometer (Micromass U.K., Wythenshawe, United Kingdom) via an open split. Instrument stability checks were performed prior to each analysis by running a series of 10 reference pulses of N₂ until a standard deviation of better than 10⁻⁶ was achieved.

RESULTS

Anaerobic ammonium oxidation at site 1.

Production of ²⁹N₂ from the oxidation of ¹⁵NH₄⁺ occurred at equal rates in the presence of either ¹⁴NO₃⁻, ¹⁴NO₂⁻, or both (Fig. 3) but could not be measured in control slurries enriched with only ¹⁵NH₄⁺ (data not shown). Production of ²⁹N₂ was linear with time until the pools of either NO₂⁻ or NO₃⁻ were depleted and, hence, was independent of the decreasing concentration of either acceptor. Measurable production of ²⁹N₂ ceased after 8 h, when either electron acceptor had been depleted, reconfirming the production measured in the controls. When the slurries were enriched with NO₃⁻, little, if any, transient accumulation of NO₂⁻ could be detected as the NO₃⁻ diminished. When spiked together with ¹⁴NO₂⁻ and ¹⁴NO₃⁻, production of ²⁹N₂ occurred at the same rate as in the independent trials, but NO₃⁻ was consumed preferentially to NO₂⁻ by the NO₃⁻-reducing community (Fig. 3c). In addition, the production of ²⁹N₂ continued to the end of the trial (12 h), being sustained by the NO₂⁻ pool, which remained largely untouched until after 8 h, when the concentration of NO₃⁻ had been reduced to approximately 200 μM. The rates of turnover were consistently equal for NO₃⁻ and NO₂⁻, being equivalent to 0.11 and 0.12 μmol ml of wet sediment⁻¹ h⁻¹, respectively, at a starting concentration of 800 μM. Anaerobic ammonium oxidation produced 0.16 and 0.15 nmol of ²⁹N₂ ml of wet sediment⁻¹ h⁻¹ for NO₃⁻ and NO₂⁻, respectively. With 10.3% labeling of the NH₄⁺ pool, this gives an average total anaerobic ammonium oxidation for both acceptors of 1.5 nmol of N₂ ml of wet sediment⁻¹ h⁻¹, equivalent to 1.3% of the total NO₃⁻ or NO₂⁻ reduction.

FIG. 3.

Production of ²⁹N₂ from labeled ¹⁵NH₄⁺ oxidation in the direct presence of ¹⁴NO₃⁻ (a), ¹⁴NO₂⁻ (b), or both (c). Values are means ± 1 standard error of the mean (SEM) (n = 4). The same symbols are used for each panel. Data were collected July 2002.

The sediments at site 1 reached an average maximum NO₃⁻ or NO₂⁻ reduction rate of 383 μM h⁻¹ at 3,200 μM (Fig. 4a). Because the pool of either substrate from 200 to 800 μM was completely turned over during the incubation, the data presented in Fig. 4 are for the whole 4-h incubation period, and rates are estimated from data points above 800 μM. The yield of ²⁹N₂ from anaerobic ammonium oxidation increased with the turnover of substrate, up to a maximum of 0.8 or 1 nmol of ²⁹N₂ ml of wet sediment⁻¹ for approximately 1.5 μmol of NO₃⁻ or NO₂⁻ ml of wet sediment⁻¹ (Fig. 4b). Though NO₂⁻ accumulated in the NO₃⁻-enriched slurries at a constant 20 μM between 800 and 3,200 μM NO₃⁻, this had no effect on the yield of ²⁹N₂. Although total NO₃⁻ or NO₂⁻ reduction increased above 1.5 μmol ml of wet sediment⁻¹, production of ²⁹N₂ did not change significantly, and in turn, the yield of ²⁹N₂ as a proportion of total NO₃⁻ or NO₂⁻ reduction decreased (Fig. 4b and c).

FIG. 4.

(a) Total rate of turnover for either NO₃⁻ or NO₂⁻ as a function of the initial enrichment concentration; (b) production of ²⁹N₂ from labeled ¹⁵NH₄⁺ oxidation; (c) ratio of ²⁹N₂ per mole of NO₃⁻ or NO₂⁻ reduced as a function of total NO₃⁻ or NO₂⁻ reduction. Values are means ± 1 SEM (four replicates were performed). The same symbols were used for each panel. Data were collected August 2002.

The addition of ∼9 μmol of ¹⁵NH₄⁺ to the slurries increased the soluble NH₄⁺ concentration at site 1 by 150 μM, i.e., 10% (or half that predicted), which indicated that half of the label was exchanged and bound to the sediments. Throughout all the slurry experiments, the total NH₄⁺ pool remained essentially constant, suggesting that NH₄⁺ oxidation (and/or assimilation) was balanced by ammonification (data not shown). The ¹⁵NH₄⁺ pool was diluted during these incubations, but as ²⁹N₂ production remained linear, this dilution was not significant and did not affect our measure of anaerobic ammonium oxidation. There was no significant production of ³⁰N₂ in the ¹⁵NH₄⁺ incubations and, in addition, no ¹⁵N gases were produced in slurries that had previously been autoclaved, confirming the theory that anaerobic ammonium oxidation is a biological process.

Anaerobic ammonium oxidation along the estuary.

Slurries spiked with ¹⁴NH₄⁺ and ¹⁵NO₃⁻ produced >90% ³⁰N₂ (Table 2). With >98 atom% ¹⁵N labeling of the NO₃⁻ pool, the predicted proportion of ²⁹N₂ was <2%; but under these conditions, ²⁹N₂ contributed up to 9.42% of the total ¹⁵N gas production at site 2, and anaerobic ammonium oxidation could account for 7.82% of total ¹⁵N gas formation. Overall, there was good agreement between the rates of anaerobic ammonium oxidation measured with ¹⁵N labeling of either the NH₄⁺ or the NO₃⁻ pool (Table 2) (P = 0.497).

TABLE 2.

Anaerobic ammonium oxidation measured by production of ²⁹N₂ with ¹⁵N labelling of either the NO₃⁻ or NH₄⁺ pools

Site	Actual 30N2 (% 15N gas)	Predicted 29N2 (% 15N gas)a	Actual 29N2 (% 15N gas)b	A29 (% 15N gas)c	Probability of 15N from 15NH4+d	AAOe from:
						15NO3− (μmol)	15NH4+ (μmol)
2 Grays	90.58	1.86	9.42	7.82	0.49	0.261 (0.019)	0.296 (0.0289)
3 Gravesend	93.77	1.99	6.23	4.37	0.56	0.120 (0.0228)	0.160 (0.0164)
4 Allhallows	95.44	1.93	4.56	2.71	0.27	0.040 (0.0045)	0.026 (0.0033)
6 Grain	97.60	1.85	2.40	0.57	0.54	0.004 (0.0006)	0.003 (0.0005)

^aFrom >98 atom % ¹⁵N-NO₃⁻ and random pairing of ¹⁴N and ¹⁵N.

^bAs a proportion of total measured ¹⁵N gases.

^cA₂₉ = P₂₉ − D₂₉ (see the text and reference 30).

^d[¹⁵NH₄⁺]/[total NH₄⁺].

^eAnaerobic ammonium oxidation (AAO) with ¹⁵N-labeled NO₃⁻ (n = 4) or NH₄⁺ (n = 3). Figures in parentheses are SEM.

Maximum rates of anaerobic ammonium oxidation were measured at site 2 in October and November 2002; the rates then decreased significantly between sites (by analysis of variance, P was <0.05) to a minimum at site 6 at the mouth of the estuary (Fig. 5a). Denitrification rates followed a similar pattern, though the peak was found further downstream at site 3, and were significantly positively correlated with rates of anaerobic ammonium oxidation (P < 0.05). In a similar pattern, maximum N₂ formation via anaerobic ammonium oxidation was measured at site 2 (8%); the proportion then decreased steadily to <1% at site 6 at the mouth of the estuary (Fig. 5b). Overall, there was good positive correlation (P < 0.05) between sediment organic carbon content and the significance of anaerobic ammonium oxidation to N₂ formation along the estuary, although site 1 was a distinct outlier in this relationship (Fig. 5c; Table 1). In November 2002, when all six sites were visited, there was a distinct decrease in the rate of NO₃⁻ reduction along the estuarine gradient and a distinct shift in the fate of reduced NO₃⁻ (Fig. 5d). At sites 1 and 2, 60 to 70% recovery of reduced NO₃⁻ as ¹⁵N gas (as in the initial trials at site 1 [data not shown]) suggested significant DNRA, while 100% recovery at sites 3 to 6 suggested the dominance of denitrification. Nitrite accumulated in the sediments to 3 μM during the incubations with ¹⁵NO₃⁻ (800 μM) at all sites (except site 1, as in the initial trials). A large accumulation of NO₂⁻ (150 μM) at site 5 explains the difference between NO₃⁻ reduction and ¹⁵N gas formation but does not necessarily suggest significant DNRA.

FIG. 5.

(a) Anaerobic ammonium oxidation (AAO) and denitrification (Den) at six sites along the Thames estuary in November 2002; (b and c) anaerobic ammonium oxidation as a proportion of total ¹⁵N gas formation in October and November 2002 (b) and as a scatter against sediment organic carbon content (c); (d) total NO₃⁻ reduction and ¹⁵N gas formation in November 2002. Values are means ± 1 SEM (n = 4). sed, sediment. Note that a large accumulation of NO₂⁻ at site 5 explains differences between reduction and gas formation (see the text).

DISCUSSION

Biogeography of anaerobic ammonium oxidation.

The results presented here confirm the presence of anaerobic ammonium oxidation in another aquatic system and extend the findings recently reported for marine sediments and anoxic marine basins (2, 15, 30). However, the pattern reported here is in stark contrast to that reported for marine sediments. For example, the significance of anaerobic ammonium oxidation to N₂ production decreased along the estuarine gradient with decreasing organic content to <1% at the most seaward site (Fig. 5c; Table 1). Though a similar low proportion of anaerobic ammonium oxidation was reported at an inshore coastal site (30), an increase in the significance of anaerobic ammonium oxidation was measured with decreasing organic content and decreasing sediment reactivity in an offshore direction, i.e., in the opposite direction to that of our estuarine gradient. In addition, the maximum proportion of anaerobic ammonium oxidation to N₂ production in the Thames sediments was 8%; therefore, denitrification still dominated N₂ production overall, while up to 60% of N₂ production via anaerobic ammonium oxidation was reported to occur in offshore marine sediments (30).

Anaerobic ammonium oxidation via nitrite reduction.

The sole production of ²⁹N₂ from sediments incubated with ¹⁵NH₄⁺ and either ¹⁴NO₃⁻ or ¹⁴NO₂⁻ agreed with the 1:1 stoichiometry of equation 3. With the stoichiometry of equation 4, if any oxidation of ammonium was directly coupled to the reduction of NO₃⁻, then at least 25% of the ¹⁵N-labeled gases should have been ³⁰N₂; as this was not the case, it was concluded that the anaerobic ammonium oxidation in these estuarine sediments was coupled to the reduction of NO₂⁻, which agrees with the original hypotheses and that recently reported for marine sediments (3, 30, 34).

In the presence of both electron acceptors, the concentration of NO₂⁻ remained essentially the same (Fig. 3c), which can be explained only if the rate of NO₂⁻ reduction was balanced by the supply of NO₂⁻ from NO₃⁻ reduction. Given that anaerobic ammonium oxidation could account for only 1% of NO₂⁻ reduction at site 1, this change remained unnoticeable. Even if obligate NO₂⁻ respirers, as well as anaerobic ammonium oxidizers, were present (5), their impact on the NO₂⁻ pool while in the presence of NO₃⁻ was minimal. After 8 h, NO₃⁻ became limiting, and the community that reduced NO₃⁻changed to a community that directly reduced NO₂⁻. Despite the increase in the rate of turnover, and therefore competition, for NO₂⁻, the rate of anaerobic ammonium oxidation remained unaffected. Saturation of anaerobic ammonium oxidation at low concentrations of NO₂⁻ would explain why there was no measurable difference between the rates measured with either trace or significant (∼20 μM) accumulation of NO₂⁻ from NO₃⁻ (or in the direct presence of NO₂⁻) but also suggests that the availability of NO₂⁻ would seldom be limiting for anaerobic ammonium oxidation in estuarine sediments. Anaerobic ammonium-oxidizing bacteria are reliant on the in situ formation of NO₂⁻ via NO₃⁻ reduction, and as concentrations of NO₂⁻ are usually less than 5 μM in upper sediment layers (M. Trimmer, unpublished results), such an affinity is suited to scavenging NO₂⁻ at low concentrations. Whereas the NO₃⁻- and NO₂⁻-reducing community has a high capacity to reduce NO₂⁻ (up to 383 μM h⁻¹), anaerobic ammonium oxidation appears to be saturated at a comparatively low NO₂⁻ concentration, which agrees with K_m estimates (3 μM) for NO₂⁻ consumption via anaerobic ammonium oxidation in marine sediments (3).

In the only other published environmental study of anaerobic ammonium oxidation in aquatic sediments, the vast majority (>91%) of NO₃⁻ or NO₂⁻ reduced was recovered as N₂ gas (30). This finding suggests that DNRA was insignificant and that anaerobic ammonium oxidation was reliant on NO₂⁻ formed intracellularly by the denitrifying community “leaking” into the sediment. In contrast, both the denitrification and DNRA pathways may support anaerobic ammonium oxidation at sites 1 and 2. In previous studies on the end products of NO₃⁻ reduction in organically enriched sediments, it was speculated that NO₂⁻ formed by DNRA may later be respired by denitrifying bacteria (11). Yet this may be an additional pathway for NO₂⁻ formation and, in turn, a pathway for NO₂⁻ coupling to anaerobic ammonium oxidation in estuarine sediments. Hence, N transformations in estuarine sediments may actually be more complex than originally thought, with potential interplay between DNRA, denitrification, and anaerobic ammonium oxidation. At the remaining sites, however, anaerobic ammonium oxidation must have been coupled to the in situ formation of NO₂⁻ via the denitrifying community.

The quantification of anaerobic ammonium oxidation using ¹⁵NO₃⁻ is based on the assumption that the process produces only ²⁸N₂ and ²⁹N₂ (30). This assumption may be violated in the presence of DNRA, which, via the reduction of ¹⁵NO₃⁻ to ¹⁵NH₄⁺, may in turn generate ³⁰N₂ by anaerobic ammonium oxidation. At site 2, where the contribution of anaerobic ammonium oxidation to N₂ formation was greatest, DNRA may account for approximately 30% of the reduced ¹⁵NO₃⁻. The starting concentration for ¹⁵NO₃⁻ was 800 μM, and with a 50% turnover and 30% DNRA, these conditions would produce 120 μM ¹⁵NH₄⁺. As only approximately 50% of ¹⁵NH₄⁺ remains in the soluble pool, 60 μM ¹⁵NH₄⁺ was left, which represents about 5% of the total soluble NH₄⁺ pool of ∼1,200 μM. At the start, the probability of ²⁹N₂ production with ¹⁵NO₃⁻ and ¹⁴NH₄⁺ was 100%, but with a 50% turnover, this estimate dropped to 95%, and in turn, our estimate of anaerobic ammonium oxidation from our measured production of ²⁹N₂ is underestimated by ∼5%. The magnitude of this violation via DNRA depends upon its contribution to NO₃⁻ reduction and the relative size of the nonlabeled NH₄⁺ pool.

Anaerobic ammonium oxidation along the estuary.

The distribution of anaerobic ammonium oxidation along the Thames estuary may simply reflect changes in the abundance of bacteria per unit of wet sediment. At site 1, despite having the highest concentrations of NO₂⁻ in the water column throughout the estuary, the sediments are highly reduced (visible observations) and concentrations of NO₂⁻ are permanently low (<1 μM), which may in part be due to denitrification, substantial DNRA, and the reoxidation of sulfides. Such a low NO₂⁻ supply may be capable of sustaining only a relatively small population of anaerobic ammonium-oxidizing bacteria; hence, anaerobic ammonium oxidation accounted for only ∼1% of N₂ formation. Though anaerobic ammonium oxidation was saturated at low concentrations of NO₂⁻, it accumulated at all sites except for site 1, which was a clear outlier in the overall relationship between organic content and anaerobic ammonium oxidation (Fig. 5c). At site 2, conditions must have been favorable for both anaerobic ammonium-oxidizing bacteria and denitrifying bacteria, e.g., sustained though low supplies of NO₂⁻ and organic substrates and NO₃⁻, respectively. Despite the saturation of anaerobic ammonium oxidation at a low NO₂⁻ concentration, a sustained population would still be reliant on an in situ supply of NO₂⁻, hence the correlation between sediment organic C and the significance of anaerobic ammonium oxidation. Higher organic loadings, i.e., the loadings at site 2, stimulated NO₃⁻ reduction (26) and, in turn, the in situ availability of NO₂⁻, which sustains a greater population of anaerobic ammonium-oxidizing bacteria. The overall positive correlation between rates of denitrification and anaerobic ammonium oxidation supports this hypothesis, though DNRA may help support the maximum rates of anaerobic ammonium oxidation at site 2. The marked decrease in anaerobic ammonium oxidation and denitrification seaward of site 2 suggested a decline in both populations as organic C, NO₃⁻, and NO₂⁻ decreased along the estuary.

The organism responsible for anaerobic ammonium oxidation in the original bioreactor studies has been identified as a novel planctomycete (17, 28), but the organism(s) that is responsible for anaerobic ammonium oxidation in the Thames estuary, as well as what conditions regulate its significance relative to denitrification, remains unknown. Given that the anaerobic ammonium oxidation reaction was first discovered in a sewage bioreactor (17) and that two very large sewage treatment works (STWs) (among many others) discharge into the Thames at site 1 (combined output, 1.3 × 10⁶ m³ day⁻¹), the sediments along the estuary may have been seeded by these STWs, yet the activity was most significant some 13 km downstream from the main discharge area. In addition, the well-studied classic nitrifier Nitrosomonas eutropha can perform anaerobic ammonium oxidation under anaerobic or O₂-limiting conditions and, as such, is distinct from the anaerobic ammonium-oxidizing bacteria (1). Indeed, anaerobic ammonium oxidation may in fact be due to a combination of both of these metabolisms (24).

Although it is possible to measure the proportion of N₂ production from either anaerobic ammonium oxidation or denitrification in anaerobic slurries, regulation of either process maybe more complex under in situ conditions, and such measurements should be taken as estimates of potential rates. For example, anaerobic ammonium oxidation is an obligate anaerobic process, while denitrification is facultative, and because of this, denitrification is likely to occur along the gradient of decreasing oxygen concentrations and below the oxic-anoxic interface within sediments (14). Oxygen penetration (typically 0.2 to 2 mm) into intact estuarine sediment cores will directly affect the demand for NO₃⁻ from the overlying water and, in turn, the in situ rates of NO₂⁻ formation via nitrification and reduction (21, 22, 23). Oxygen penetration is likely to be more significant where the dominant source of NO₂⁻ and NO₃⁻ within sediments is that of nitrification, as in unpolluted estuaries or coastal seas (25, 32). While total anaerobic ammonium oxidation is unlikely to be limited by the availability of NO₂⁻, oxygen penetration will affect the source of NO₂⁻ and overall rates of denitrification.

The organically enriched sediments of the Thames estuary, although representing only 18% of the total sediment area, are responsible for 34% of the total mineralization of organic material (33). While the maximum contribution from anaerobic ammonium oxidation to N removal was only 8%, this maximum activity occurred within these organically enriched sediments, which again highlights the significance of the depositional zone (i.e., around the turbidity maxima) to N processing—including novel pathways—in estuarine sediments.