Kinetic Parameters of Denitrification in a River Continuum

Abstract

Kinetic parameters for nitrate reduction in intact sediment cores were investigated by using the acetylene blockage method at five sites along the Swale-Ouse river system in northeastern England, including a highly polluted tributary, R. Wiske. The denitrification rate in sediment containing added nitrate exhibited a Michaelis-Menten-type curve. The concentration of nitrate for half-maximal activity (K_map) by denitrifying bacteria increased on passing downstream from 13.1 to 90.4 μM in the main river, but it was highest (640 μM) in the Wiske. The apparent maximal rate (V_maxap) ranged between 35.8 and 324 μmol of N m⁻² h⁻¹ in the Swale-Ouse (increasing upstream to downstream), but it was highest in the Wiske (1,194 μmol N m⁻² h⁻¹). A study of nitrous oxide (N₂O) production at the same time showed that rates ranged from below the detection limit (0.05 μmol of N₂O-N m⁻² h⁻¹) at the headwater site to 27 μmol of N₂O-N m⁻² h⁻¹ at the downstream site. In the Wiske the rate was up to 570 μmol of N₂O-N m⁻² h⁻¹, accounting for up to 80% of total N gas production.

The massive use of nitrogen-based fertilizer has inevitably increased the concentrations of nitrate in many rivers (14, 15). This has led to the acceleration of the nitrogen cycle, especially those processes that use nitrate as the substrate, such as denitrification, the biological reduction of nitrogen oxides (nitrate and nitrite) to N-containing gases (N₂ and N₂O) in environments with relatively low levels of oxygen (9). Interest in denitrifying bacteria has also been stimulated by concern over the recognition of the role of N₂O in the destruction of stratospheric ozone (5) and in the radiative heat budget of the atmosphere (19).

Although rivers are usually the first aquatic system to receive the excess of nitrate, fewer studies have been undertaken on them than on soils or estuaries. Only one (2, 27) or a few sites down the river (4) have been sampled, and it is therefore difficult to generalize from the results. According to the river continuum concept of Vannote et al. (28), the physical and chemical variables of water and sediments should present a gradient from the headwaters to the estuary, which should elicit a series of responses within the denitrifying bacterial population. For instance, in the Swale-Ouse river system in northeastern England we have observed (18) a consistent increase (up to 30 times) in the denitrification rate of sediment cores on passing downstream.

There have apparently been no studies to determine the apparent affinity (K_map) and maximal capacity (V_maxap) for nitrate utilization of denitrifying bacteria down a whole river; however, this is important for clarifying the relationship between nitrate availability in the environment and nitrate use by denitrifying bacteria. Previous attempts to measure the kinetics of nitrate utilization in soils and sediments have been hampered by a number of problems. The earliest studies employed slurry techniques (7, 12, 17), where the structure of the sediment is destroyed and hence the in situ oxygen and nitrate gradients. The values of K_map and V_maxap obtained are only meaningful under conditions of no limitation and cannot be incorporated into predictive models. Other researchers have used intact sediment cores to study the kinetics of nitrate reduction through measuring the decrease in nitrate concentration in the overlying water (1). The long incubation period, possible errors in measuring nitrate, and the unknown fate of the depleted nitrate make uncertain the significance of the reported kinetic parameters.

The aim of the present study was to investigate the effect of nitrate concentration on the denitrification rate and N₂O production in intact sediment cores taken from a river continuum.

MATERIALS AND METHODS

Study area.

The study was performed on the Swale-Ouse river system. A detailed description of the geography of the area has been provided by Jarvie et al. (8), the vegetation has been surveyed by Holmes and Whitton (6), and biological studies have been reviewed by Whitton and Lucas (29). The Swale rises on the Northern Pennines, running 117 km to its confluence with the Ure, where it becomes the Ouse (Fig. 1). The river becomes tidal (but still freshwater) at km 145.0. (Main river distances are measured downstream from the junction of two upland tributaries; the location of a site on a tributary is given by a negative value indicating the distance upstream, together with the point at which it enters the main river.)

FIG. 1.

Location of the sampling sites in the Swale-Ouse and R. Wiske.

The Swale-Ouse shows a consistent trend of changes in many physical and chemical features on passing downstream, while the whole river system is subject to marked variations in discharge. An important tributary is the Wiske, a lowland and highly organically polluted stream, that receives treated sewage effluent from a town of 10,700 inhabitants.

The oxygen concentration in water at sites on the main river was always over 93% saturation (based on monthly data taken between 0900 to 1200 h), but this value fell to 64% in the Wiske.

Sediment and water collection.

Samples were taken from upstream (km −2.5 [0.0] and km 10.9), midstream (km 49.9), and downstream (km 107.9 and km 145.0, i.e., freshwater tidal) sites (Fig. 1). In addition, sediment was collected from the Wiske (km −1.7 [86.1]). Samples were collected between 18 and 24 April 1997, with repeat sampling at two sites, upstream (km 10.9) and the Wiske, on 18 June and 4 August 1997.

Intact sediment cores were taken in Plexiglas cylinders (3.5-cm inner diameter, 25-cm height, with a vertical series of silicone rubber inserts to allow injection at different positions) from a defined 2-m² area at positions where sediment was accumulating. Each tube was sealed with a rubber bung, and the core was removed carefully. The bottom of the core was sealed with an additional rubber bung. Special care was taken to preserve the sediment structure. The pH of the water was determined with a glass combination electrode (Russell CW76) and a Wissenschaftliche Technische Werkstätten meter. Water samples were taken at the same time and filtered through 0.45-μm (pore size) membrane filters. Nitrite and nitrate were determined according to the method of Stainton et al. (26).

Depletion of in situ nitrate.

In order to establish the required nitrate concentration during experimental studies, nitrate and nitrite were removed from the sediment by using natural denitrification activity, as adapted from the method of Murray et al. (12), with river water being substituted by a medium lacking N or P. This medium was modified from that used by Chu (3) and contained (per liter) 4.28 mg of KCl, 54.75 mg of CaCl₂ · 6H₂O, 25 mg of MgSO₄ · 7H₂O, 15.85 mg of NaHCO₃, 2.44 mg of FeCl₂ · 6H₂O, 3.33 mg of Na₂EDTA, and 0.25 ml of AC microelement stock (11). The medium was buffered with HEPES (0.6 g liter⁻¹) to the natural pH. All samples were incubated in the dark for 24 h at 15°C.

Assay for denitrification activity and N₂O production.

The acetylene (C₂H₂) block method (24) was used to determine the denitrification rate. This method is based on assessing the rate of N₂O accumulation in cores to which C₂H₂ has been added to block the bacterial reduction of N₂O to N₂. Net N₂O production was also determined in intact sediment cores incubated without C₂H₂. After incubation to remove environmental nitrate and nitrite, the supernatant water was drained off and replaced with the freshwater medium with the required nitrate concentration supplied as KNO₃. Control experiments with KCl showed that K⁺ had no influence on the results. Nitrate concentrations used in the assays were modified to reflect the expected denitrification rate and ranged from 0 to 140 μM for the upstream sites, 357 μM for downstream sites, and 11,420 μM for the Wiske. Two triplicate sets of cores were used for each concentration assayed at each site. The first set was incubated without C₂H₂ to measure net N₂O production, while the second was incubated with C₂H₂ to determine the denitrification rate (N₂O plus N₂ production). For the latter, C₂H₂ saturated medium with the required nitrate concentration was added in the overlying water to obtain 10% C₂H₂ (final concentration). C₂H₂ saturated medium was prepared by flushing for 20 to 30 min with pure C₂H₂ (previously passed for 30 min through 0.1 N phosphoric acid to remove ammonium contamination). Then 1.0 ml of this solution was injected (where required) into the sediment through the holes in the cylinder at each centimeter depth for the top 3 cm. Cores were filled with the test medium, sealed without gas space, and incubated at 15°C in the dark for between 3 and 5 h. Immediately before the assay, another three cores from each site were used to determine the initial N₂O concentration. In addition, another two sets of three cores were used to determine denitrification and N₂O production by using river water from each site. For this analysis, cores were incubated at the same time and under similar conditions to those described above but with C₂H₂ saturated river water (denitrification rate) or natural river water (N₂O production).

At the end of incubation, the cores were vigorously shaken, and 5 ml of the resulting slurry was transferred to a 12.5-ml gas container, which included 0.1 ml of 40% formaldehyde, and immediately frozen until N₂O analysis. After the gas container was shaken for 2 min to equilibrate gas in the headspace and sediments, 2 ml of the headspace was removed for N₂O measurement. This was done by using a gas chromatograph equipped with ⁶³Ni electron capture detector (Perkin-Elmer 2000; detector temperature 350°C; flow of 20 ml min⁻¹). The N₂O concentration was determined by using a standard curve generated with purified N₂O (BOC, Ltd.). The N₂O concentration in the headspace was used to back-calculate the amount of N₂O in the sediment by using Henry’s law. N₂O production (expressed as micromoles of N₂O-N per square meter per hour) and denitrification rate (N₂O plus N₂, expressed as micromoles of N per square meter per hour) were calculated after subtracting the initial N₂O in the core and taking into account the incubation period and the area of the core.

The kinetic parameters for denitrification and N₂O production were computed from the Lineweaver-Burk transformation (1/V versus 1/S) of the Michaelis-Menten equation, where V is the denitrification rate or N₂O production, and S is the concentration of nitrate. Because the denitrifying enzyme activity was assayed in a complex system, the kinetic parameters were termed as apparent concentration of NO₃⁻ for half-maximal activity (K_map) and apparent maximal activity (V_maxap).

RESULTS

Environmental data.

The data for sediments obtained at the six sites are shown in Table 1. With respect to sediment particle size, there was an increase in the silt-plus-clay fractions and a decrease in the sand fraction on passing downstream. In general, total carbon and nitrogen in the sediment increased on passing downstream.

TABLE 1.

Sediment characteristics of the top centimeter for sampling sites on the Swale-Ouse and R. Wiske

River and site	Distance (km)	Carbon and nitrogen contenta			Composition (%)b
		% C	% N	C/N	Silt + clay	Fine sand	Sand
Swale
Ravenseat	−2.5, 0.0c	0.17	0.009	19.5	8.3	36.1	43.5
Ivelet Bridge	10.9	0.43	0.037	14.8	11.2	50.6	20.0
Catterick Bridge	49.9	0.56	0.034	17.4	23.4	41.3	13.6
Thornton Manor	107.9	0.26	0.017	17.6	31.7	35.0	1.6
Naburn Weir	145	0.51	0.034	16.8	38.1	11.2	0.6
Wiske (Wiske)	−1.7, 86.1c	0.40	0.031	12.5	10.0	58.7	14.1

^aMonthly average values (excluding February) for 1997. 

^bAverage obtained from January, April, July, and October 1997. 

^cThe first value is the distance upstream from the main river. The second value is the distance down the main river where the tributary enters. 

Nitrate depletion.

The denitrification rate and N₂O production in the cores after incubation for 24 h to remove sediment nitrate and nitrite are shown in Table 2. Denitrification was still detectable, though rates were low: the rates were 5 μmol of N m⁻² h⁻¹ in the headwater site and 31.4 μmol of N m⁻² h⁻¹ in the Wiske. These values represent, respectively, 16 and 3% of the maximum rates found during assays with the saturation level of nitrate (see below). N₂O production was very low in the Wiske (2.7 μmol of N₂O m⁻² h⁻¹) and lower or negative (net consumption of N₂O) at the other sites (Table 2).

TABLE 2.

Net N₂O production and denitrification rate with no nitrate after 24 h at 15°C with nitrogen-free freshwater medium

River and site	Distance (km)	N2O production (μmol of N2O-N m−2 h−1 ± SD)	Denitrification (μmol of N m−2 h−1 ± SD)
Swale
Ravenseat	−2.5, 0.0	0.74 ± 0.62	5.0 ± 4.0
Ivelet Bridge (April)	10.9	−1.42 ± 0.93	5.5 ± 0.6
Ivelet Bridge (June)		−2.34 ± 0.71	3.5 ± 0.7
Ivelet Bridge (August)
Catterick Bridge	49.9	0.25 ± 0.61	13.4 ± 1.6
Thornton Manor	107.9	1.23 ± 0.56	12.2 ± 3.4
Naburn Weir	145	0.83 ± 0.86	9.2 ± 0.36
Wiske
Wiske (April)	−1.7, 86.1	2.72 ± 2.52	23.1 ± 9.4
Wiske (June)		−1.57 ± 0.13	31.4 ± 3.8
Wiske (August)		−9.70 ± 2.35	6.6 ± 5.8

Kinetics for denitrification and N₂O production.

Nitrate addition in the overlying water resulted in an increase in the denitrification rate (Fig. 2a). The response of denitrification to nitrate at all sites could be fitted successfully (minimum regression coefficient of 0.9) to a Michaelis-Menten-type curve (Fig. 2a and Table 3). During the April survey (Table 3), V_maxap increased on passing down the river (from 35.8 to 324 μmol of N m⁻² h⁻¹), but it was highest on the Wiske (758 μmol of N m⁻² h⁻¹). The value for V_maxap in the Wiske in August was still higher (1,194 μmol of N m⁻² h⁻¹). K_map constants increased in April on passing down the main river from 13.1 to 90.4 μM nitrate (Table 3). The overall highest value for K_map (640 μM) was for the Wiske in August. Where repeat measurements were made (km 10.9 and Wiske), values for V_maxap and K_map were similar in April and June, but they were about one-third higher in August (Table 3).

FIG. 2.

Responses of denitrification rate (a) and N₂O production (b) to nitrate additions at the sampling sites in April 1997 (•). For the Ivelet Bridge (km 10.9) and Wiske sites the assays were repeated in June (○) and August (▴) 1997. The bars denote the standard deviations of three replicates. Note the different scales for denitrification rate, N₂O production, and nitrate concentration.

TABLE 3.

Kinetic parameters for denitrification rate and N₂O production estimated according to the Lineweaver-Burk transformation of the Michaelis-Menten equation

River and site	Distance (km)	Denitrification			N2O production
		Vmaxap (μmol of N m−2 h−1)	Kmap (μM)	r2a	Vmaxap (μmol of N2O m−2 h−1)	Km (μM)	r2
Swale	−2.5, 0.0
Ravenseat	10.9	35.8	13	0.95
Ivelet Bridge (April)		52	21	0.98
Ivelet Bridge (June)		51	23	0.98
Ivelet Bridge (August)		74	23	0.98
Catterick Bridge	49.9	144	18	0.90	70.2	517	0.98
Thornton Manor	107.9	197	29	0.98
Naburn Weir	145	324	90	0.99	15.5	138	0.93
Wiske
Wiske (April)	−1.7, 86.1	758	351	0.99	570	543	0.97
Wiske (June)		860	460	0.99	626	532	0.98
Wiske (August)		1,194	640	0.98	496	1,952	0.98

^aCoefficient of determination. See Results. 

The production of N₂O in the cores without C₂H₂ could only be fitted to a Michaelis-Menten curve for three sites (km 49.9, km 145.0, and Wiske [Fig. 2b]). At the two upstream sites (km −2.5 [0.0] and km 10.9), N₂O production was less than 15 μmol of N₂O-N m⁻² h⁻¹ and without any clear trend with increasing nitrate concentration. The maximum values for N₂O production were 70.2, 15.5, and 570 μmol of N₂O-N m⁻² h⁻¹ in April for km 49.9, km 145.0, and the Wiske, respectively, and 626 and 496 μmol of N₂O-N m⁻² h⁻¹ in June and August for the Wiske (Table 3). The K_map values were 517, 138, and 542 μM in April for the same sites and 532 and 1,952 μM in June and August for the Wiske (Table 3). The K_map value for N₂O production for these sites was higher than the K_map value for the denitrification rate, while the V_maxap values were lower.

The proportion of N₂O to total N gases (N₂O + N₂) ranged between 2.4 and 38% and between 2 and 58% at the headwater and km 10.9 sites, respectively (Table 4), but there was no clear trend with increasing nitrate concentration. The values increased on addition of nitrate to the km 49.9 and km 107.9 samples but not to the km 145.0 sample. In the Wiske the values exceeded 50% at all nitrate concentrations in April and June (Table 4).

TABLE 4.

Percentage of N₂O related to total N gases evolved (N₂O + N₂) for the nitrate concentration assayed

Nitrate concn (μM)	% N2O in samples from:
	Ravenseat	Ivelet Bridge			Catterick Bridge	Thornton Manor	Naburn Weir	R. Wiske
		April	June	August				April	June	August
17.8	21	6	7.5	58		4.3
37.7	38	2	3.4	21	1.0	3.0	<0.0
53.6	19	3	6.4	25
71.4					6.0	5.1	6.0
107.1	2.3					5.3
142.8		4	12	15	17	10	6.4
214.3					15	14	3.2
357.1					17		11	50	68	18
714.3								65	80	20
1,428								71	62	22
2,857								76	73	29
5,714								68	73	33
11,429								72	75	43

When river water (with its natural nitrate concentration) was used instead of medium, the denitrification rate was linearly related (r² = 0.89; slope = 0.92) to the rate predicted from the Michaelis-Menten curve for the nitrate concentration in the river.

DISCUSSION

Nitrate depletion and freshwater medium.

The successful application of kinetic studies to the sediments required removal of nitrate and nitrite, and this was largely achieved. This suggests that any nitrate produced via nitrification of the ammonium in the sediment was also consumed by natural denitrification during this period. The highest denitrification persisting after 24 h of incubation in the absence of nitrate was at the headwater site (Ravenseat).

Influence of nitrate addition on denitrification and N₂O production.

The denitrification rate in intact sediment cores responded strongly to nitrate addition and followed a Michaelis-Menten-type curve, as found by other researchers for freshwater (1) and estuarine (10, 16, 17) sediments. Repeat measurements on different dates at particular sites showed only moderate differences.

The V_maxap values for denitrification rate were mostly similar to the maximum rates reported by Pattinson et al. (18) for field conditions in a 17-month survey of intact sediment cores from the same sites. However, comparisons with other studies are complicated by the fact that these have been made at a range of temperatures. Ideally, results would be presented for both the field temperature and a standard laboratory temperature. In the case of the Swale-Ouse, the values are comparable because the temperature at the time of the maximum value was always within 5°C of the standard temperature of 15°C.

Other published kinetic data apparently refer only to experiments with slurries, so caution is needed in comparing our data with previous data. However, the values for V_maxap at the headwater site and at the main river (35.8 to 324 μmol of N m⁻² h⁻¹) are well within the range of denitrification rates measured at near-ambient conditions for rivers (0 to 345 μmol of N m⁻² h⁻¹) and for coastal marine sediments (0 to 888 μmol of N m⁻² h⁻¹) (23), although denitrification was always detectable in our studies. The uppermost value on the Wiske (1,194 μmol of N m⁻² h⁻¹ in August) exceeded this range. The K_map values may be compared with the lower (8 μM) and upper (344 μM) values found in previous studies (10, 12, 16, 17) with the slurry technique.

Another problem in assessing results is the fact that denitrification in the sediment has been related to the nitrate concentration in the overlying water. Although transport of nitrate into the sediment is likely to be rapid, the lower and variable nitrate concentration in the sediment means that the K_map value has inevitably been overestimated. However, we advocate the use of cores to obtain realistic measurements of kinetic parameters to be incorporated into integrating models.

Net N₂O production occurred at all sites, but only at some sites was there an increase with increasing nitrate concentration. Compared to other studies, the proportions of N₂O related to total N gases (N₂ + N₂O) for the highest nitrate concentrations (Table 4) were high, reaching 80% in the Wiske in June. The published values for N₂O in aquatic sediments have been obtained from pore water profiles (25) and direct N₂O flux measurements from cores (13, 21); net N₂O flux is generally less than 2 μmol of N₂O m⁻² h⁻¹ and less than 5% of N₂ production (23). As N₂O production is an intermediate product in at least three processes in the nitrogen cycle, each influenced by a range of environmental factors, it is difficult to explain these high values. However, eutrophication is one factor reported to lead to increased N₂O production, as shown in Narragansett Bay (20), where N₂O fluxes increased around 10-fold from the relatively unpolluted lower- and mid-bay sediments to the eutrophic upper-bay sediments. A study (22) on a marine mesocosm showed that N₂O flux from sediment and the N₂O/N₂ ratio increased markedly in relation to nitrate input.

Ecological significance.

The V_maxap value showed a marked increase coincident with increasing nitrate concentration on passing downriver, and it seems almost certain that this is a key factor. However, other factors must be considered when assessing the results shown in Fig. 2. These factors include the number of denitrifying bacteria, their genetic and physiological properties, and the optimal conditions of the process (e.g., the concentrations of suitable carbon substrates, phosphate, and oxygen). Downstream, the sediment particles were finer, presumably decreasing oxygen penetration, and the carbon content of the upper 2 cm of sediment was higher (Table 1), all features likely to favor denitrification. Although the changes in oxygen concentration were not investigated, the high O₂ saturation usually present in the river water and the relatively low incubation period minimized the disturbance of sediment in our experimental setup.

The K_map values increased on passing downstream and were maximum in the Wiske. This trend agrees with the observed increase in nitrate concentration on passing downstream. In environments with high affinity (low K_map) and high nitrate concentration or in those with low affinity (high K_map) and low nitrate concentration, nitrate reduction by denitrifying bacteria would be inefficient. The K_map values calculated in this study for each site were compared with the nitrate data available for 1995 to 1996 from other sources (Environment Agency and the LOIS data base). For the upland sites, the K_map values were higher than the nitrate concentration in the river, and therefore the response of denitrification to the nitrate concentration is likely to be linear for most of the year. However, the K_map values for all sites on the main river and the Wiske were lower than the nitrate concentration for most of the year, and therefore nitrate reduction by denitrifying bacteria is probably nitrate saturated, especially during winter and spring. This suggests that other variables, such as temperature, organic carbon, and oxygen, must be considered when studying the seasonal changes in denitrification (18). As denitrifying bacteria are taxonomically heterogeneous (10), it is likely that these factors also exert a selective pressure on groups that are dominant at particular times of the year.