Correspondence: Jeanette Norton, Department of Plants, Soils and Climate, Utah State University, Logan, UT 84322-4820, USA. Tel.: +001 435 797 2166; fax: +001 435 797 3376; e-mail: firstname.lastname@example.org
An agricultural soil was treated with dairy-waste compost, ammonium-sulfate fertilizer or no added nitrogen (control) and planted to silage corn for 6 years. The kinetics of nitrification were detemined in laboratory-shaken slurry assays with a range of substrate concentrations (0–20 mM NH4+) over a 24-h period for soils from the three treatments. Determined concentrations of substrate and product were fit to Michaelis–Menten and Haldane models. For all the treatments, the Haldane model was a better fit, suggesting that significant nitrification inhibition may occur in soils under high ammonium conditions similar to those found immediately after fertilization or waste applications. The maximum rate of nitrification (Vmax) was significantly higher for the fertilized and compost-treated soils (1.74 and 1.50 mmol N kg−1 soil day−1) vs. control soil (0.98 mmol kg−1 soil day−1). The Km and Ki values were not significantly different, with average values of 0.02 and 27 mM NH4+, respectively. Our results suggest that both N sources increased nitrifier community size, but did not shift the nitrifier community structure in ways that influenced enzyme affinity or sensitivity to ammonium. The Km values are comparable to those determined directly in other soils, but are substantially lower than those from most pure cultures of ammonia-oxidizing bacteria.
Nitrogen (N) is required in large amounts by crops and remains a major limiting nutrient for plant growth in both wildland and agricultural systems in temperate regions (Vitousek & Howard, 1991; Chapin et al., 2002; Meisinger et al., 2008). Fertilizers and organic materials such as animal manures are used to supply N for increased crop yields. In the majority of agricultural soils, NH4+ is rapidly converted to NO3− by the process of nitrification. Nitrification strongly influences the availability and loss of N from soils and also produces trace gases (N2O and NO) important in greenhouse gas and ozone atmospheric chemistry (Hutchinson & Davidson, 1993; Godde & Conrad, 2000). Improved management of nitrification in agricultural systems may aid the effort to increase N fertilizer use efficiency while decreasing the impacts of excess reactive N in the environment.
Short-term nitrification assays are used to describe the relationship among the substrate concentration, environmental conditions and the rate of NH4+ conversion to NO2− and NO3−. This may be considered similar to the analysis of biomass-specific substrate utilization rates assuming nongrowth conditions. In this study, we examine the kinetics of nitrification in an agricultural soil treated repeatedly with contrasting sources of nitrogen supplied from ammonium fertilizer or composted dairy waste (Habteselassie et al., 2006a). Our focus was on the relationship between NH4+ concentration and the rate-limiting step of ammonia oxidation during a short-term assay (<24 h) before ammonia oxidizer growth occurs. This type of kinetic analysis may be used to model the response in the rate of nitrification in environmental samples and in enrichment cultures of ammonia-oxidizing bacteria (AOB) or ammonia-oxidizing archaea (AOA). The parameters determined in this current study are implicitly understood to be apparent half-saturation constants, Km (app), and apparent maximum velocities, Vmax (app), because these were not determined in cell-free enzymatic systems, although the original terms are used for simplicity.
Nitrification kinetics are often modeled using the Michaelis–Menten equation:
where V is the NH4+ oxidation rate, Km is the half-saturation constant and S is the substrate (NH4+) concentration. The Vmax parameter is the maximum reaction rate that can be achieved with nonlimiting substrates in the absence of substrate inhibition; for the process of nitrification, this is also known as the nitrification potential. Because Vmax is indicative of the maximum activity under a nonlimiting substrate, this is often interpreted as an estimate of ammonia-oxidizer population size (Belser & Mays, 1982; Hart et al., 1994). The Km is the substrate concentration at 1/2 Vmax and can be derived from the dissociation constant for the enzyme–substrate complex. Haldane (1965) further developed enzyme kinetic equations to include the situation where the reaction rate is inhibited at very high concentrations of substrate. This is represented by the following equation:
In the Haldane equation, the Km is equal to the minimum substrate concentration that produces a rate of 1/2 Vmax, whereas the Ki, the inhibition parameter, is equal to the maximum substrate concentration that produces a rate of 1/2 Vmax (Haldane, 1965; Stark & Firestone, 1996). For the case of nitrification, as NH4+ concentrations increase, generally, the nitrification rates increase, but at some point, the NH4+ concentrations may become inhibitory to the ammonia oxidation activity. There is evidence that AOB have different sensitivities to high levels of their substrate, ammonium or ammonia (NH4+ or NH3) (Koops et al., 1991). Some sensitive AOB also have a lower Km and thus may have an advantage over NH4+-tolerant AOB under low NH4+ concentrations, while under high NH4+ or fluctuating concentrations, the tolerant AOB may have the advantage (Suwa et al., 1994; Bollmann et al., 2002). Thus, the sensitivity of different AOB to ammonia reflects niche differentiation and their competitiveness under environmental conditions (Bollmann et al., 2002; Norton, 2008). Our objective was to determine whether the ammonia-oxidizer (AO) community in an agricultural soil is sensitive to high concentrations of NH4+ and whether annual applications of contrasting sources of N (fertilizer vs. composted dairy manure) maintained over 6 years result in AO communities with altered kinetic parameters. Understanding nitrification kinetics may help us predict the response of N transformations to agricultural practices under a changing environment (Stark & Firestone, 1996; Norton, 2008) and better parameterize N cycle models (Cabrera et al., 2008).
Materials and methods
The nitrification kinetics were determined for soil sampled from irrigated silage corn field plots on the Utah Agricultural Experimental Station Greenville Farm that have been described previously (Shi et al., 2004; Habteselassie et al., 2006a). The soil is an irrigated very strongly calcareous Millville Silt Loam with a pH1 : 1 of 8.2. There are seven nitrogen treatments replicated four times in a completely randomized block design. We selected three of the field N-treatments for use in this study: the control, with no added nitrogen source; a 200 kg N ha−1 (NH4)2SO4 application rate (AS200); and a high-rate dairy-waste compost application (sufficient to supply approximately 200 kg N ha−1 of available N for the growing season, DC200). Two soil cores (5 cm internal diameter × 15 cm length) were collected in October 2002 from each treatment in each of the four blocks. The soil from each core was mixed and sieved (<2 mm) and 75 g of soil from each core was further homogenized using a soil splitter. The soils were then stored at −20 °C until use to minimize changes in nutrients and AO populations. Because these surface soils are normally subject to freezing conditions annually and resume activity quickly in the spring, we selected this condition as an appropriate option for storage. Soil nitrification assays were performed within 1 year of soil collection. Two days before use, the soils were taken out of the freezer and allowed to thaw slowly at 4 °C and the gravimetric soil moisture content was determined on a subsample. Then 15 g soil (oven dried equivalent) was suspended in 250 mL of phosphate buffer (0.2 M at pH 8.0), shaken at 200 r.p.m. for 5 min, centrifuged (17 300 RCF, 10 min, at 4 °C) and the supernatant was removed. This removed excess inorganic N present in the different soils so that lower starting concentrations and lower variability of substrates and products could be achieved. The soil was transferred to wide-mouthed 250-mL glass flasks with 100 mL of phosphate buffer (0.2 M at pH 8.0). Then 14 different NH4+ concentrations, ranging from 0.0 to 20.0 mM NH4+-N, were generated by adding different amounts of (NH4)2SO4 solutions to the flasks.
All of the soil slurries from one block were randomly placed in a covered orbital shaker maintained at 20 °C and shaken continuously (200 r.p.m.) during a 24-h incubation (Stark, 1996). During the first 2 h, pH was measured and adjusted to pH 8.0 several times with sulfuric acid or potassium hydroxide. The pH was also adjusted to pH 8.0, when needed, on four occasions immediately after slurry samples were collected. Soil slurry samples were collected from the shaking flasks through a two-hole stopper fitted with Tygon tubing (1 cm internal diameter × 20 cm length) that could be attached to a 10 mL pipette. The tubing allowed us to collect slurry samples without interrupting the shaking, while the second hole allowed air exchange. Samples were taken at 2, 8, 20 and 24 h. Each sample was immediately chilled, centrifuged (8500 RCF, 8 min, at 4 °C) and the supernatant was frozen until analyzed.
The NH4+ and NO2−+NO3− concentrations in the samples were determined colorimetrically using a flow injection automatic analyzer (QuikChem Systems, Lachat Chemicals, Mequon, WI). The net nitrification rates were determined by linear regression of NO2−+NO3− accumulation vs. time, while the substrate concentration was calculated as the time-weighted average of the measured NH4+ concentration. Based on the lack of previous detection of NO3− assimilation in soils from this location (Shi et al., 2004; Habteselassie et al., 2006b), we did not expect that NO3− assimilation would be a significant factor and the net rates were expected to be good estimates of the actual rates of nitrification in these shaken slurry assays.
The Michaelis–Menten and Haldane models were fit to the nitrification rate and measured NH4+concentration data for each of the 12 combinations of block and treatment using the nonlinear regression procedure (proc nlin) in sas 9.1 for Windows. Analysis of residuals was used to assess assumptions of normality and homogeneity of variance. The parameter estimates for Vmax, Km and Ki obtained for individual field plots were used as data in an anova to assess the effect of treatment. The anova model was a one-way factorial (treatment with three levels) in a randomized block design. The control was compared with the AS200 and DC200 treatments using contrasts. For anova, significance levels of pairwise comparisons among treatments means were adjusted using the Tukey method to control experiment-wise Type I errors. Computations were performed using proc glm in sas 9.1 for Windows. An F-test was used to evaluate whether the Haldane fit the data better than the Michaelis–Menten model (Neter & Wasserman, 1974; Robinson, 1985). The full model (Haldane) was considered to be significantly better than the reduced model (Michaelis–Menten) when F* had a P<0.05.
The control treatment had a significantly lower Vmax than both AS200 and DC200 treatments, while the Vmax of the AS200 and DC200 treatments were not significantly different from one another (Tables 1 and 2). This was the case when either the Haldane or the Michaelis–Menten models were fit to the data. The values of Km and Ki were not significantly different among the treatments (P=0.30 and 0.35, respectively), and there was no significant block effect. The use of the Haldane model resulted in slightly higher estimates of Vmax and higher estimates of Km than those from the Michaelis–Menten model.
Table 1. Apparent nitrification kinetic parameters and model fits for Haldane and Michaelis–Menten models to data from the different soil treatments
Soil samples assayed were collected from field plots that had received no N amendments (Control), or annual applications of 200 kg N ha−1 as ammonium sulfate (AS200) or 200 kg available N ha−1 as dairy compost waste (DC200) over the past 6 years.
Within a row, means followed by different letters are significantly different at P≤0.05 based on a Tukey's least-squares means comparison.
Model fit compared using an F-ratio calculated within each treatment (d.f. 50, 1). A value of P≤0.05 indicates that the Haldane model provides a significantly better fit than the simpler Michaelis–Menten model.
Table 2. Statistical results (P-values) for the mixed model anova analyzing the significance of treatment (Trt) or block effects on model parameters
Values ≤0.05 indicate significant effects.
Overall, the R2 values and the F-test showed that the Haldane model provided a significantly better fit to the data than the Michaelis–Menten model (Table 1, Fig. 1). F-tests further indicated that the Haldane model fit the data significantly better than the Michaelis–Menten model when all the treatments were compared individually.
Because of the repeated application of soil amendments with high N concentrations, and the possible selection for NH4+-tolerant AO in these soils, we hypothesized that the Haldane model would be a better fit for kinetic data from the control treatment, while the Michaelis–Menten model would provide a better fit for the AS200 and DC200 treatments. However, we found that the Haldane model showed a better fit for all three treatments compared with the Michaelis–Menten model. These results demonstrate that soil AO communities at our site are sensitive to high concentrations of NH4+/NH3 in spite of previous long-term exposure to large quantities of N amendments. While for many agricultural soils, NH4+ concentrations may be close to or slightly above Km, and thus nitrification rates will be in the first-order range, concentrations may be quite high immediately following N-fertilization, or in fertilizer bands, anhydrous ammonia or liquid waste injection zones or urea nest applications, where concentrations often reach over 1000 mg ammonical N kg−1 soil (Kissel et al., 2008). For this reason, the Haldane model may be more suitable for predicting nitrification rates for the range of conditions that exist over the entire growing season.
In studies performed from 1998 to 2002 in the same field plots where our samples were collected, the nitrification potentials were observed to be higher in AS200 and DC200 treatments (11.4 and 14.1 mg N kg−1 day−1, respectively) than in the control treatment (3.6 mg N kg−1 day−1) (Habteselassie et al., 2006a). While we did find that the control treatment had the lowest Vmax, we found no significant difference in Vmax between AS200 and the DC200 treatments, and no difference in Km among any of the three treatments (Table 1). The differences in the nitrification potential suggest a larger AO community size in plots receiving large quantities of ammonium sulfate or dairy compost. This was confirmed recently by real-time PCR quantifying the amoA gene copies of both AOB and AOA, which were 4.7 × 107, 5.9 × 107 and 7.4 × 107 copies g−1 soil for control, AS200 and DC200 soils, respectively (Habteselassie, 2005; M. Y. Habteselassie, L. Xu & J. M. Norton, unpublished data). Molecular methods have also been used to show that nitrogen fertilization increases the numbers of ammonia oxidizers in other studies (Phillips et al., 2000; Wakelin et al., 2007). Given the difference in the size of the AO community, one might question whether the composition of this community would differ as well; however, we found no evidence for this in terms of differences in the kinetic parameters Km and Ki.
The kinetic parameters Km and Ki have been shown to differ among AOB populations, among AOB communities from different soil types and between AOB and AOA (Table 3, Koops & Pommerening-Roser, 2001). The kinetic parameters determined directly in soil samples (13–40 μM NH4+ for Km; Table 3A) are typically lower than the parameters determined from pure cultures or enrichment cultures (27–1780 μM). The most likely explanation for the difference is that the enrichment and isolation steps used to obtain pure or enrichment cultures usually select for the fastest-growing AOB within the community (Stark & Firestone, 1996; Webster et al., 2005) and are not appropriate for isolating ammonia-sensitive AOB or AOA. Molecular methods have been used with soils from the same plots of this study, and the AOB community was found to be dominated by AOB strains similar to cluster 3 nitrosospiras (Habteselassie (2005). The Km for Nitrosospira briensis is relatively high (Table 3BBollmann et al., 2005), but the kinetics of the nitrosospiras found in soils may be distinct from those of this type strain of Nitrosospira (Webster et al., 2005). The marine AOA Nitrosopumilus maritimus has been shown to have an extremely low Km for ammonia and has been shown to exhibit ammonia substrate inhibition (Table 3BMartens-Habbena et al., 2009), and the AOA ‘Candidatatus Nitrososphaera gargensis’ also showed substrate inhibition above 3.0 mM NH3+NH4+ at pH 7.8. (Hatzenpichler et al., 2008), suggesting that low-nutrient environments are the preferred niche of AOA (Erguder et al., 2009). Little is known about the kinetics of ammonia oxidation by the different lineages of AOA present in soil environments because no representative pure cultures are currently available.
Table 3. Kinetic parameters for ammonia oxidation activity in either (A) soil slurries or (B) liquid pure or enrichment cultures of ammonium-oxidizing bacteria or archaea
Niche differentiation and distinct nitrification kinetics of ammonia oxidizers in various environments have been observed (Bollmann et al., 2002; Webster et al., 2005), but our investigation attempted to observe whether agricultural management over multiple seasons might shift the functional characteristics of ammonia oxidizers in a specific soil. Our results suggest that while the size of the AO community is affected within a seasonal timeframe by agricultural management, community characteristics such as substrate affinity (as reflected in Km) and sensitivity to substrate inhibition (as reflected in Ki) are relatively resistant to change over multiple seasons of fertilization. Further investigation into the timing and the mechanisms that alter ammonia oxidizer community composition and function in response to agricultural practices may improve our ability to manage nitrification for improved agricultural sustainability.
This work was supported by grants from the USDA NRICGP (#9600839 and #9935107-7808), the Vice President for Research at Utah State University and Utah Agricultural Experiment Station and approved as journal paper 8032. Special thanks are due to Laurie Fairbanks, Seth Thacker, Rita Nelson, Abe Tanner and Jeff Williams for their assistance in the field and laboratory; Myq Larson, Toby Hooker and Sarah Benanti for their help with Lachat analysis; and Susan Durham and Bill Kemp for their statistical assistance and consultation.