Biosensor Determination of the Microscale Distribution of Nitrate, Nitrate Assimilation, Nitrification, and Denitrification in a Diatom-Inhabited Freshwater Sediment

Abstract

High-resolution NO₃⁻ profiles in freshwater sediment covered with benthic diatoms were obtained with a new microscale NO₃⁻ biosensor characterized by absence of interference from chemical species other than NO₂⁻ and N₂O. Analysis of the microprofiles obtained indicated no nitrification during darkness, high rates of nitrification and a tight coupling between nitrification and denitrification during illumination, and substantial rates of NO₃⁻ assimilation during illumination. Nitrification during darkness could be induced by purging the bulk water with O₂ gas, indicating that the stimulatory effect on nitrification by illumination was caused by algal production of O₂. NH₄⁺ addition did not stimulate nitrification during darkness when O₂ was restricted to the upper 1-mm layer, and there was thus a low nitrification potential in the permanently oxic top 1 mm of the sediment.

The nitrogen cycle and the bacteria mediating its various transformations have been studied extensively since Winogradsky isolated the first nitrifying bacteria in the late-19th century. Although a tremendous effort has been spent on quantification of the processes, new findings and developments continue to give us new and important insights. One such important finding was the recent discovery of extensive nitrogen fixation in the oligotrophic oceans by the filamentous but heterocyst-free cyanobacterium Trichodesmium sp. (4). An important methodological development is represented by the isotope pairing method by Nielsen (17), which compared to previous methods gives us much better estimates of denitrification in most aquatic environments. However, there are still problems in the quantification of nitrogen transformations in a variety of environments. One such environment is sediments inhabited by benthic microphytes, where assimilation processes occur simultaneously with nitrification and denitrification. The influence of microphytobenthos on sediment-water fluxes of combined, inorganic nitrogen has been investigated by a number of authors. The presence of benthic microalgae generally reduce the efflux of NH₄⁺ and NO₃⁻ from the sediment (33), with a minimum efflux during illumination due to algal assimilation (32). Also nitrification and denitrification have been shown to be affected by light-dark cycles. Denitrification of NO₃⁻ supplied from the overlying water was reduced during light periods (18, 27), while denitrification coupled to nitrification in the sediment was shown to be stimulated during daytime due to a stimulation of nitrification (27). However, other data indicate lower rates of potential nitrification in the top 1 cm of sediment inhabited with microalgae compared to that in sediment without algae (12), which is to be expected when there is competition for NH₄⁺. The diurnally integrated rates of nitrification may thus be lowered by the presence of microalgae even when oxygen production by the microalgae stimulates nitrification during illumination compared to dark incubation.

Microsensors for NH₄⁺ and NO₃⁻ based on liquid ion exchangers have been used to measure microprofiles of NH₄⁺ (7) and NO₃⁻ (8, 13) in sediments, and such sensors could in principle also be used in photosynthetically active sediments. By modelling based on the measured microprofiles it would then be possible to calculate vertical profiles of production and consumption processes for the two chemical species. The NH₄⁺ consumption rates, however, would not be very informative for layers with an abundance of microphytes where assimilation cannot be distinguished from nitrification. The NO₃⁻ microsensors used until now suffer from bicarbonate interference so that the extremely steep concentration gradients of bicarbonate in communities of benthic microphytes (26) result in a substantially inaccurate determinations of low NO₃⁻ concentrations.

In the present study we used a newly developed microscale NO₃⁻ biosensor (15), characterized by the absence of interference from species other than NO₂⁻ and N₂O, to resolve the vertical and temporal distribution of NO₃⁻ in diatom-covered sediment during light-dark cycles. By using diffusion reaction models the vertical distribution of NO₃⁻ assimilation, nitrification, and denitrification can be calculated from the measured steady-state NO₃⁻ profiles.

MATERIALS AND METHODS

Sediment sampling.

Sediment cores covered with a thin layer of pennate diatoms were collected by using 44-mm inside diameter Plexiglas tubes in the creek inlet of Vilhelmsborg Sø, a small pond situated 15 km south of Aarhus, Denmark. The pond was created by damming a creek draining agricultural areas, and the water is therefore rich in NO₃⁻ (up to 1 mM) during periods of high water discharge, but the NO₃⁻ concentration may also decline to very low levels during dry summer periods.

After the return to the laboratory, the sediment cores were manipulated so that the sediment surface became flush with the upper edge of the Plexiglas cylinder. The cores were incubated in a constant-temperature (20°C) water bath containing approximately 43 liters of constantly aerated tap water (about 5 mM HCO₃⁻, comparable to the lake water) with NO₃⁻ and NH₄⁺ concentrations adjusted to meet the demands of the specific experiment. No attempt was made to add extra CO₂ during the aeration, and the overlying water therefore reached a pH of 8.6 compared to a pH of 8.2 in the lake water. Water movement across the sediment surface was facilitated by directing a small air jet at the water surface adjacent to the sediment core. The samples were subjected to a 12-h/12-h light-dark cycle by using a slide projector giving a light intensity of 110 μmol of photons m⁻² s⁻¹ in the 400 to 700 nm range to simulate the low-light conditions of the densely forested sampling site. The sediment cores were subjected to one or two light-dark cycles before analysis with microsensors.

Microsensors.

Microprofiles of O₂ were obtained with a Clark-type O₂ microsensor (Unisense, Aarhus, Denmark) mounted on a micromanipulator with computerized depth control and data acquisition (25). A 2-point linear calibration for the O₂ sensors (23) was obtained by reading the signal in well-aerated overlying water and anoxic sediment. The O₂ content of air-saturated water was calculated by using the formula of Garcia and Gordon (10).

Depth profiles of NO₃⁻ were obtained by using microscale NO₃⁻ biosensors with a sensing tip of 30 μm in diameter (15). The NO₃⁻ sensors were constructed by placing immobilized, nitrous oxide reductase-deficient denitrifying bacteria in front of a N₂O microsensor. The bacteria reduce NO₃⁻ and NO₂⁻ to N₂O, which in turn is detected by the built-in N₂O sensor. The signal from the sensor thus represents the sum of NO₃⁻ and NO₂⁻, and if any N₂O is present it would result in a signal corresponding to 2.4 times the signal of an equivalent concentration of NO₃⁻ or NO₂⁻ (N₂O contains the N atoms from two NO₃⁻ or NO₂⁻ ions but as the diffusivity in water is about 1.2 times that for the ions the contribution to the signal is increased by a factor of 2.4). Readings performed with a N₂O microsensor (24) indicated, however, that N₂O concentrations were below 1 μM and therefore no corrections for N₂O were performed. In the following we only write NO₃⁻, although the actual parameter being measured is NO₃⁻ plus NO₂⁻. The bacterial growth medium within the sensor contained LiCl as the dominating salt to ensure a positive tip potential of the sensor so that entry of negatively charged ions including NO₃⁻ and NO₂⁻ was facilitated.

The NO₃⁻ sensors were mounted on a computer-controlled system which was similar to that used for the O₂ microsensors. Before the experiments the linearity of the NO₃⁻ sensor signals in the range of 0 to 250 μM was checked by a 5-point calibration in a test chamber. Calibration of the NO₃⁻ sensors during incubations was done by linear 2-point calibrations by using the signal in the bulk water and the signal at a depth of 1 cm in the sediment, where the concentration was assumed to be zero. Reference NO₃⁻ and NO₂⁻ concentrations in bulk water samples were determined by high-pressure liquid chromatography (anion separation column LCA A14; SYKAM, Gilching, Germany) with 20 mM NaCl as eluent. Peaks were detected at 220 nm with a SpectroMonitor 3200 spectrometer. In order to correct for minor drift in the signal of the NO₃⁻ sensors we used the average of two sets of calibration points for the calibration of each profile, one set obtained just before and the other just after profiling. One of the two applied sensors had a downward drift in sensitivity of about 4% per hour, while the other sensor exhibited a drift of less than 1% per hour. When signal drift has resulted in too low a signal, a high sensitivity can be reestablished by reversing the polarization of the nitrous oxide sensor for about 1 min (15).

Microprofiles were obtained at randomly chosen locations in a single sediment core. Sediment porosity was determined for the top 5.5 mm by weighing and drying four subcores of known volume taken at the end of the experiment and was found to be 0.8 ± 0.04 (mean ± standard deviation). As an estimate of the effective sediment diffusion coefficients (D_s) of O₂ and NO₃⁻, we used the product of the diffusion coefficient in free water and the squared porosity (35). Diffusion coefficients for O₂ (2.1 × 10⁻⁵ cm² s⁻¹) and NO₃⁻ (1.7 × 10⁻⁵ cm² s⁻¹) in water were found in the literature (3, 16). The D_s estimated for O₂ and NO₃⁻ were thus 1.3 × 10⁻⁵ cm² s⁻¹ and 1.1 × 10⁻⁵ cm² s⁻¹, respectively.

Estimation of consumption and production profiles.

As the basis for the calculation of net consumption and production rates of O₂ and NO₃⁻ from the measured microprofiles we used Fick’s second law of diffusion including a production and a consumption term (26):

1

where C(z,t) is the concentration at time t and depth z, R is the respiration rate, and P is the production rate. Assuming steady state we have

2

so Eq. 1 can be reduced to

3

Defining A(z) = [R(z) − P(z)]/D_s and using Euler’s formula for numeric integration we find

4

where h determines the step size used for numerical integration. After further integration we have

5

Substituting ∂C/∂z_n with equation 4 we find

6

Using equation 6 we can calculate concentration profiles on the basis of net activities (A_n × D_s) and by altering these net activities we can minimize the sum of squared deviations of the calculated profile from the measured profile. We chose to use Microsoft EXCEL Solver to achieve this goal and as a boundary condition we introduced a point below the deepest measuring point with concentration and activity equal to zero. Starting from this point we stepwise integrated upwards toward the sediment surface with h = 0.1 (the actually used step size during the measurement of the profile). Activities were not allowed to vary within a measuring step. This procedure on its own often yields oscillating activities with no biological relevance because of noise in the measured pore water concentration profiles. To compensate for this effect we also minimized the sum of squared first derivatives of the guessed activities, Σ(∂A/∂z)². These two minima were weighted equally in order to smooth the oscillations in the activities while still giving a good fit between the measured and the calculated concentrations. The activities obtained this way are per unit volume of pore water, and integrated activities expressing the activity per unit of surface area of sediment must thus be multiplied with the porosity.

RESULTS

Profiles and activities in the dark.

The oxygen penetration into the sediment was only 0.9 mm when the sediment was incubated in the dark with 27 μM NO₃⁻ (actually 25 μM NO₃⁻ and 2 μM NO₂⁻) in the overlying water (Fig. 1A). The penetration of NO₃⁻ extended to a depth of about 1.5 mm, with some inaccuracy in the determination of concentrations below 1 μM as evidenced by the scatter of the data points at these low concentrations. The decrease in NO₃⁻ concentration in the upper part of the oxic zone was linear and indicated neither NO₃⁻ production (nitrification) nor consumption (assimilation or denitrification).

FIG. 1.

Measured O₂ (○, mean values with bars indicating standard deviations, n = 6) and NO₃⁻ (□, mean values with bars indicating standard deviations, n = 6) concentrations, calculated profiles (lines) and NO₃⁻ assimilation (gray [top]), nitrification (light grey), and denitrification (gray [bottom]) rates during darkness (A) and illumination (B). Both the O₂ and the NO₃⁻ profiles were measured at different sites in the core.

Modelling of the NO₃⁻ profile in the dark indicated moderate denitrification activities up to 0.03 nmol cm⁻³ s⁻¹. The total rate of oxygen consumption in the dark for the core analyzed in Fig. 1 amounted to 2.4 mmol m⁻² h⁻¹ (Table 1) while the denitrification amounted to 0.08 mmol NO₃⁻ m⁻² h⁻¹ (Table 2).

TABLE 1.

Depth-integrated net rates of O₂ consumption and production

Treatment conditions	Depth interval (mm)	Rate of production or respiration (mmol of O2 m−2 h−1)	Activity
104 μM NO3−, light	0.0–1.2	3.9	Net production
104 μM NO3−, light	1.3–4.3	1.5	Respiration
104 μM NO3−, dark	0.0–1.2	2.3	Respiration
27 μM NO3−, light	0.0–0.3	7.4	Net production
27 μM NO3−, light	0.4–2.2	2.1	Respiration
27 μM NO3−, dark	0.0–1.0	2.4	Respiration

TABLE 2.

Depth-integrated net rates of NO₃⁻ consumption and production

Treatment conditions	Depth interval (mm)	Rate of nitrification, denitrification, or assimilation (mmol of NO3− m−2 h−1)	Activity
104 μM NO3−, light	0.0–1.8	0.35	Assimilation
104 μM NO3−, light	2.0–4.2	0.36	Nitrification
104 μM NO3−, light	4.4–6.4	0.26	Denitrification
104 μM NO3−, dark	0.7–2.4	0.20	Denitrification
27 μM NO3−, light	0.0–0.7	0.24	Assimilation
27 μM NO3−, light	0.7–2.1	0.23	Nitrification
27 μM NO3−, light	2.1–2.7	0.18	Denitrification
27 μM NO3−, dark	0.4–1.8	0.08	Denitrification

Profiles and activities during illumination.

When the sediment was illuminated both O₂ and NO₃⁻ profiles changed dramatically compared to the dark profiles (Fig. 1B). An O₂ peak was created in the upper 1 mm of the sediment due to a net production of 7.4 mmol of O₂ m⁻² h⁻¹ in the upper 0.3 mm while a net O₂ consumption of 2.1 mmol m⁻² h⁻¹ below 0.3 mm in depth decreased the oxygen concentration to zero at a depth of 2.2 mm. The nitrate concentration decreased in the 0.3-mm surface layer characterized by photosynthetic oxygen production and reached a minimum of 15 μM before it increased again in the oxic layer below the photosynthetic zone, where a peak concentration of 28 μM was found at a depth of 1.8 mm. A concentration of zero was reached at 2.7 mm. Calculations based on the profile showed an upper net NO₃⁻ consumption zone from 0 to 0.7 mm, a net production zone from 0.7 to 2.1 mm, and a lower net consumption zone from 2.1 to 2.7 mm. The integrated rates are listed in Table 2.

Higher NO₃⁻ concentration.

Analysis of another sediment core incubated with 104 μM NO₃⁻ showed comparable data for oxygen penetration and oxygen consumption in the dark (Table 1), but denitrification was increased to 0.20 mmol of NO₃⁻ m⁻² h⁻¹ (Table 2). The sediment core was, however, quite different from the core analyzed in Fig. 1, as the oxygen penetration in the light was about double, although the rate of net oxygen production in the photic layer was only 3.9 mmol m⁻² h⁻¹. The values for nitrate transformations at 104 μM NO₃⁻ presented in Table 2 therefore cannot be directly compared with the values for 27 μM NO₃⁻. The general tendencies are, however, identical between the two experiments, with a pronounced assimilation zone in the top, a nitrification zone in the oxic layer below the diatom layer, and a largely anaerobic denitrification zone. The maximum nitrification activities in both experiments were found very close to the oxic-anoxic interface where the nitrifiers can be supplied by ammonium from below.

High O₂ concentration in water.

It seemed obvious that the induction of oxic conditions in deeper layers by photosynthetic oxygen production caused the relatively high rates of nitrification during illumination. To verify this assumption an experiment was conducted where the overlying water was purged with pure O₂. This caused the oxygen penetration to increase from 1.2 to 3.2 mm after purging for 200 min (Fig. 2A). The increased oxygen penetration caused a gradual change in NO₃⁻ profile so that a situation with a subsurface NO₃⁻ peak was reached after 150 min. Modelling of this nitrate profile resulted in a distribution of nitrification and denitrification similar to that observed for the illuminated sediment (Fig. 3), i.e., significant nitrification activities in the oxic layers next to the oxic-anoxic interface but no activity in the layer with diatoms.

FIG. 2.

(A) ○, measured O₂ profile with atmospheric O₂ concentration in the bulk water at time −40 to −5 min (means ± standard deviations [error bars], n = 6); □, measured O₂ profile with 930 μM O₂ in the bulk water at time 200 to 250 min (means ± standard deviations [error bars], n = 6). (B) NO₃⁻ profiles obtained in the period between the O₂ profiles shown in panel A after oxygenation for 5 min (□), 30 min (○), 60 min (▵), 80 min (▿), 100 min (◊), and 150 min (×). Oxygen profiles were measured at the same site whereas the nitrate profiles were measured at different sites in the core.

FIG. 3.

Measured NO₃⁻ concentrations (□) at time 150 min (cf. Fig. 2), modelled NO₃⁻ profile (line), and nitrification (grey bars) and denitrification (dark bars) rates with 930 μM O₂ in the bulk water.

NH₄⁺ addition.

The NO₃⁻ profile in a dark-incubated core did not change upon addition of 1 mM NH₄Cl to the overlying water (data not shown), indicating that no active nitrifiers were present in the uppermost 1-mm layer that was oxic during dark incubation.

DISCUSSION

Nitrification.

The present study shows that nitrification in sediment covered by benthic microalgae is dependent on the O₂ produced by microphytobenthic photosynthesis. During darkness nitrification could not be detected in the 1-mm thick oxic surface layer (Fig. 1A), whereas increased oxygen penetration down to a depth of 2.2 mm, due to photosynthesis during illumination, resulted in considerable nitrifying activity in the deeper oxic layers devoid of microalgae. A high O₂ concentration in the overlying water also resulted in nitrification near a deeply positioned oxic-anoxic interface but with no activity in the surface zone containing microalgae. Even addition of 1 mM NH₄⁺ to the overlying water did not induce nitrifying activity in the surface layer, and there is thus evidence that nitrifiers were virtually absent from these layers.

Several reasons might explain the absence or low levels of nitrification and nitrifying bacteria in the surface layers of microalgae-inhabited sediments including low NH₄⁺ concentration, high O₂ concentration, high pH, low CO₂ concentration, high light intensity, and possibly an excretion of chemical agents with activity against nitrifiers (12). To these factors, which are discussed below, could be added a high grazing pressure in the surface layers where the input of high-quality organic matter by photosynthesis is high.

It is evident that the phototrophs with their abundant supply of energy must be very potent competitors for ammonium. In planktonic environments microalgae may decrease the NH₄⁺ concentrations to nanomolar levels, whereas the half-saturation concentration is reported to be in the range of 5 to 200 μM NH₄⁺ for nitrifiers (12, 31). Nitrifiers in layers with an abundance of microalgae should thus be active only if the supply of ammonium to the layer is in excess of the demand for the phototrophs. An excess of nitrate will not improve the competitiveness of the nitrifiers as microalgae generally prefer NH₄⁺ over NO₃⁻ (36). The high O₂ concentrations present in highly active benthic communities of microphytes, which may represent O₂ partial pressures above 1 atm (25), may inhibit nitrifiers (29), as may the pH values around 10 often observed in such communities (25). High light intensities are assumed to inhibit nitrification (11, 34) in the surface waters of oceanic environments, and the sediment surface layers are exposed to similar high light intensities. Judged from our data, however, light is not likely to be the main factor responsible for low nitrification in the surface layers, as nitrification was also absent in the bottom of the diatom mat (Fig. 1) where the light intensity must have been very low.

Whatever the reasons are for the inhibition of nitrification in the top sediment layer such inhibition has some potential advantages for the benthic microalgae. The algae do not have to compete for NH₄⁺ with the nitrifying bacteria and nitrogen loss from the system is reduced by the lower amount of NO₃⁻ available for denitrification compared to a situation with nitrification going on in the whole oxic layer. It is thus also possible that the microalgae produce chemical agents which inhibit nitrification. Nitrification is easy to inhibit as the ammonium monooxygenase is sensitive to a wide range of chemical species (20).

Denitrification.

The NO₃⁻ profiles obtained during darkness indicate that denitrification occurs at O₂ concentrations of up to 20 μM (Fig. 1A). During illumination, on the other hand, net denitrification was not detected until O₂ concentrations had dropped below 1 μM, above which nitrification dominated. It is, however, likely that the zones of nitrification and denitrification overlapped during the day, so that the low denitrification rates at O₂ concentrations below 20 μM were masked by nitrification. Aerobic denitrification has been observed in pure cultures of denitrifying bacteria (28), and it is possible that the bacteria mediating the nitrification also participate in an aerobic denitrification as denitrification has been observed in both Nitrosomonas spp. (2) and Nitrobacter spp. (9). Repeated analysis of denitrification in a biofilm by use of N₂O microsensors also indicated denitrification when the O₂ concentration was <20 μM (6). Denitrification in sediments in the presence of up to 10 μM O₂ has been suggested by Blackburn et al. (1), who used a computer model to simulate NO₃⁻ profiles obtained with liquid ion exchanger-type NO₃⁻ microelectrodes (14).

Comparison of data on nitrogen cycling obtained by the NO₃⁻ biosensor and other techniques.

The depth-integrated rate of denitrification was highest in the illuminated sediments (Table 2). This is contradictory to early reports on denitrification in illuminated sediments (5), where it was found that photosynthesis increased the O₂ penetration and thereby lowered denitrification by causing a reduced flux of nitrate from the overlying water to the denitrifying layers. This earlier work was based on acetylene inhibition of N₂O reductase, but the addition of acetylene also blocked nitrification and the method thus only gave reliable results in environments where tightly coupled nitrification-denitrification did not occur. Later work based on the isotope pairing method (17) showed, however, that coupled nitrification-denitrification could be stimulated by light-induced oxygen production (27) and that this coupled process was almost independent of the NO₃⁻ concentration in the overlying water. The data presented here thus support the findings obtained by the isotope pairing method. However, in contrast to the isotope pairing method the NO₃⁻ microprofiles also yield information about the exact location of the NO₃⁻-transforming processes. Analysis of NO₃⁻ microprofiles may also result in good estimates of nitrifying activity, even when the isotope pairing method cannot yield such information due to high rates of NO₃⁻ assimilation, but the zones of NO₃⁻ assimilation and nitrification should then be well separated as they were in this study. An overlap between the nitrifying and denitrifying layers may also result in minor underestimations of both nitrification and denitrification.

By use of the microscale NO₃⁻ biosensors it is possible to get very detailed information about the processes producing or consuming NO₃⁻. Due to the absence of interference it is now possible to analyze such processes in all environments with sufficient water content. Coupled nitrification-denitrification in some environments such as the plant rhizosphere (22) and manure-soil interface (19) is difficult to study by conventional techniques, and the microsensors may prove useful for the study of these ecologically very important microenvironments. With the rise of molecular microbial ecology any new microsensor for an ecologically important chemical species also improves our possibility of linking species distribution and in situ gene expression, etc., with microscale chemical transformations, such as has already been done in a few studies (21, 30).