Introduction

Mycorrhizal associations of trees with either ectomycorrhizal (ECM) or arbuscular mycorrhizal (AM) fungi are frequently used to predict stand-level nitrogen (N) dynamics (Phillips et al. 2013; Lin et al. 2017; Averill et al. 2019; Zak et al. 2019), but the characterization of mycorrhizal nutrient syndromes has predominantly relied on quantification of net N transformations that can lead to an incomplete understanding of ecosystem N cycling. Net N mineralization rates are often regarded as a proxy for plant available N and the potential for ecosystem N loss via nitrification and downstream N transformations by soil microbes. However, net mineralization rate measurements conflate production and consumption of inorganic N, with greater net N mineralization rates potentially resulting from greater gross N mineralization or lesser microbial N assimilation. Therefore, while net N mineralization rates can be informative, quantification of gross N mineralization rates is necessary to investigate the mechanisms driving inorganic N availability and the potential for ecosystem N loss in ECM- versus AM-dominated forest stands.

ECM-associated trees are thought to create stands with closed, 'organic' nutrient economies while AM-associated trees create stands with open, 'inorganic' nutrient economies (Phillips et al. 2013). Specifically, stands dominated by ECM-associated trees tend to have lower net N nitrification rates, nitrate (NO3−) leaching, and gaseous N losses than stands dominated by AM-associated trees (Phillips et al. 2013; Midgley and Phillips 2016; Lin et al. 2017). These distinct N cycling syndromes have been largely attributed to the contrasting N acquisition strategies of ECM versus AM fungi and production of chemically distinct leaf litter by ECM and AM trees which can influence free-living microbes to alter decomposition and N mineralization rates that, in turn, regulate N losses. However, not all studies have reported evidence of these N cycling syndromes, with some reporting lower (Midgley and Phillips 2016; Lin et al. 2017), higher (Mushinski et al. 2021) and similar (Phillips et al. 2013; Midgley and Sims 2020) net mineralization rates in ECM compared to AM stands. Therefore, while it is clear that ECM and AM stands exhibit distinct N economies, the drivers of these patterns remain uncertain, making it difficult to predict the manifestation of mycorrhizal nutrient economies across diverse ecological contexts.

Differing nutrient economies between ECM and AM stands have been interpreted to reflect direct mediation of N mineralization by the contrasting N acquisition strategies of ECM and AM fungi and their associated trees (Wurzburger and Hendrick 2009; Averill et al. 2019). Low N and high lignin of ECM leaf litter (Cornelissen et al. 2001; Midgley et al. 2015; Keller and Phillips 2019) combined with competition between ECM and saprotrophic fungi for N may suppress organic matter decomposition (Bending 2003; Fernandez and Kennedy 2016), leading to low gross N mineralization rates in ECM stands. At the same time, higher quality litter and AM stimulation of saprotrophs may lead to high gross N mineralization rates in AM stands. However, lower net N mineralization rates in ECM stands can also result from higher ammonium (NH4+) assimilation by microbes rather than lower gross N mineralization. Acidic ECM soils with high C:N ratios favor N-limited ECM fungi over typically C-limited saprotrophic fungi (Lindahl et al. 2007). Fueled by host-supplied C, ECM fungi may rapidly assimilate mineralized N (Langley and Hungate 2003). However, these hypothesized patterns have yet to be empirically tested. Therefore, gross N cycling rate measurements are needed to clarify the role of ECM and AM N acquisition strategies in mediating N mineralization.

Several direct and indirect mechanisms could lead to greater rather than lower gross N mineralization rates in ECM stands compared to AM stands. First, greater C fluxes from ECM roots compared to AM roots may stimulate N mineralization in the rhizosphere of ECM trees (Norton and Firestone 1996; Phillips and Fahey 2006; Meier et al. 2015). Second, low soil pH, which is characteristic of ECM stands (Lin et al. 2022a), could indirectly stimulate gross N mineralization by selecting for a fungal-dominated community with higher biomass C:N ratios and therefore greater C requirements compared to a bacterial-dominated microbial community (Riggs and Hobbie 2016; Li et al. 2021). Fungal uptake of organic monomers to meet increased C requirements leads to N uptake in excess of stoichiometric requirements, resulting in greater N mineralization (Riggs and Hobbie 2016; Li et al. 2021). Third, when pH drops below a threshold value of 5.5, metal toxicity, rather than nutrient stoichiometry can regulate pH effects on C and N use efficiency (Li et al. 2021) with C and N use efficiency decreasing as microbes expend C and N to overcome metal stress (Malik et al. 2017; Jones et al. 2019). These potential mechanisms suggest that slower net N mineralization rates in ECM stands may not necessarily reflect slower gross N mineralization rates.

Soil acidification by ECM trees can lead to closed N cycling independently of ECM effects on N mineralization. Low soil pH can directly and indirectly inhibit chemoautotrophic growth of ammonia oxidizers, suppressing nitrification and downstream N loss pathways (Mushinski et al. 2021). Higher acidity in ECM soils can cause protonation of ammonia to ammonium which decreases substrate availability and may ultimately select for lower abundances of ammonia oxidizers (Xiao et al. 2020; Mushinski et al. 2021; Lin et al. 2022b). Likewise, ammonia oxidizer abundance can be suppressed by aluminum (Al) toxicity, which can result from greater solubility of Al3+ in low pH soils (Prosser and Nicol 2012). Inhibition of NO3− production by nitrification can have cascading effects on ecosystem N loss through leaching and denitrification by limiting NO3− availability for N cycling processes downstream of nitrification.

Here we investigated the roles of gross N mineralization and microbial N assimilation in driving mycorrhizal nutrient syndromes in ECM- versus AM-dominated temperate forest stands. We measured gross N cycling rates in addition to the more commonly measured net and potential N cycling rates used to characterize the mycorrhizal nutrient syndromes (e.g.Lin et al. 2017; Midgley and Phillips 2016; Phillips et al. 2013), and used these data to test three hypotheses. First, we hypothesized that net N mineralization rates mask patterns in gross N cycling in ECM versus AM stands. In support of this hypothesis, we expected to observe greater gross N mineralization rates in ECM stands despite greater net N mineralization rates in AM stands. Second, we hypothesized that nitrification is inhibited in ECM soils by mechanisms other than limited NH4+ production. In support of this hypothesis we predicted lower net nitrification rates in ECM soils relative to AM soils regardless of NH4+ supply from N mineralization or experimental N addition. We expected that inorganic N addition would stimulate nitrification in AM-dominated stands but not ECM-dominated stands because the other factors would continue to suppress nitrification in ECM soils even with increased NH4+ supply. Finally, we hypothesized that inhibition of NO3− production in ECM soils will have cascading effects on gaseous N losses via denitrification. In support of this hypothesis, we predicted lower denitrification-derived net nitrous oxide (N2O) fluxes, corresponding to lower nitrification rates, in ECM stands relative to AM stands.

Materials and methods

Site description

We conducted this study at Moores Creek, which is part of Indiana University’s Research and Teaching Preserve (39°05ʹ N, 86°28ʹ W). Moores Creek is an 85-year-old deciduous hardwood forest in south-central Indiana that has a mean annual precipitation of 1200 mm and a mean annual temperature of 11.6 °C (Midgley and Phillips 2016). The soils are thin, unglaciated Inceptisols, derived from sandstone (Midgley and Phillips 2016). While the forest is comprised of a mix of AM and ECM trees, we selected forest plots where > 85% of the basal area was from a single mycorrhizal type (Midgley and Phillips 2016). Stands dominated by AM-associated trees are largely composed of Acer saccharum Marsh, Liriodendron tulipifera L, Prunus serotina Ehrh., and Sassafras albidum Nutt. whereas ECM-dominated stands are largely composed of Quercus rubra L., Quercus velutina Lam., Quercus alba L., Carya glabra P. Mill., and Fagus grandifolia Ehrh (Midgley and Phillips 2016).

We sampled in plots that were established as part of a long-term N fertilization experiment in May 2011. Paired plots (N addition vs. control; 20 m × 20 m) were established in seven ECM-dominated stands and seven AM-dominated stands. Beginning in 2011, N in equal parts from NH4SO4 and NaNO3 was applied monthly from May to October in the N addition plots at a total rate of 50 kg N ha−1 yr−1 (Midgley and Phillips 2016).

Soil Sampling

Soil was collected from each plot at four dates throughout the 2018 growing season (May, July, September, and October) to account for temporal variability in N cycling process rates. To avoid transient stimulatory effects of fertilizer addition on N cycling processes, sampling was conducted two weeks after fertilizer was applied to the N addition plots. Within each plot we randomly sampled five cores, one from each quadrant of the plot and one from near the center of the plot. We used a 10 cm diameter corer to collect soil from 0–5 cm depth beneath the litter layer. The five cores from each plot were combined such that a single composite soil sample was analyzed from each replicate plot. Soils were stored in gas permeable plastic bags at ambient laboratory temperature overnight. Before soil assays were conducted, we measured soil pH in a 2:1 (water: soil, by mass) slurry. We oven-dried a root-free subsample of homogenized soils at 105 °C for 24 h to determine gravimetric soil moisture. To determine organic matter content, we ashed root-free air dried soils in a muffle furnace for 3 h at 550 °C.

Net N mineralization and nitrification rates

In order to relate our measurements of gross N process rates with more commonly measured indices of net process rates that were previously measured at this site (Phillips et al. 2013; Brzostek et al. 2015; Midgley and Phillips 2016), we measured net N mineralization and nitrification rates. Briefly, we measured 2 M KCl extractable NH4+ and NO3− concentrations before and after a three week aerobic incubation at ambient temperature in the laboratory (~ 21 °C). We calculated net mineralization rates from the change in NH4+ plus NO3− concentrations and net nitrification rates from the change in NO3− concentrations only. NH4+ and NO3− concentrations in the extracts were determined through colorimetric analysis using a SmartChem 200 discrete analyzer (KPM analytics, Westborough, MA). Potassium chloride extracts from ECM soils often had NO3− concentrations below the detection limit of the Smartchem analyzer; these samples were assigned a concentration of the minimum detection limit of the analyzer in order to achieve a low, non-zero value for subsequent net rate calculations and data analyses. Negative net rates, which are biologically possible over the time scale of the incubations, were retained to avoid biasing the dataset toward the positive values.

Gross mineralization

To determine if gross N cycling rates drive patterns in net N cycling rates, we measured gross rates of N mineralization and nitrification using the 15N pool dilution technique (Kirkham and Bartholomew 1954) as described by Hart et al. (1994). Briefly, two 150 g subsamples of root- and rock-free soil samples were placed in separate plastic bags, with one bag receiving 15NH4Cl to quantify gross N mineralization, microbial NH4+ assimilation rates, and nitrification-derived net N2O fluxes and the other bag receiving K15NO3 to quantify gross nitrification rates, NO3− assimilation rates, and denitrification-derived net N2O fluxes. Two mL of 99 atom% 15N (Cambridge Isotopes, Tewksbury, MA) label solution in DI water was pipetted onto each subsample and gently mixed by hand to homogenously distribute the 15N label. Label solution concentrations ranged 0.58–5.85 µgN mL−1 among treatments (unfertilized ECM, fertilized ECM, unfertilized AM, and fertilized AM) in order to target less than 10 atom% 15N enrichment based on previously reported soil NH4+ and NO3− concentrations across treatments (Midgley and Phillips 2016); actual 15N enrichment ranged 0.27–5.90 atom% across all samples from the three sampling dates. We extracted 50 g of soil in 150 ml 2 M KCl at 15 min and 4 h after the addition of 15N label solution to represent the initial and final time points. The KCl extracts were analyzed colorimetrically for NH4+ and NO3− concentrations on a SmartChem 200 discrete analyzer (KPM analytics, Westborough, MA). The 15N isotopic composition of the KCl-extractable NO3− and NH4+ pools were determined using the acid trap diffusion method (Herman et al. 1995) followed by analysis on a Vario Micro Cube elemental analyzer (Hanau, Germany) interfaced with an IsoPrime 100 isotope ratio mass spectrometer (Cheadle Hulme, UK). We calculated gross mineralization rates using NH4+ concentrations and atom percent excess 15N (APE) values after 15 min (ti) and four hours (tf) of incubation using the following equation:

$$Gross\,mineralization (\mu g\,N {g}^{-1}\,{d}^{-1})=\frac{\frac{{{[NH}_{4}]}_{{t}_{i}}- {{[NH}_{4}]}_{{t}_{f}}}{{t}_{f}-{t}_{i}}*\mathrm{log}\left(\frac{{APE}_{{t}_{i}}}{{APE}_{{t}_{f}}}\right)}{\mathrm{log}\left(\frac{{{[NH}_{4}]}_{{t}_{i}}}{{{[NH}_{4}]}_{{t}_{f}}}\right)}$$

Small NO3− pools characteristic of ECM soils at our study site made it difficult to accurately determine changes in 15N enrichment of the NO3− pool to reliably calculate gross nitrification rates. Therefore, we only report gross N mineralization rates despite our attempt to measure gross nitrification rates as described here. We report gross N mineralization rates per gram of dry soil, hereafter referred to as “gross N mineralization”, and gross N mineralization per gram of organic matter, hereafter referred to as “specific gross N mineralization”. The latter metric of N mineralization was included to aid in our interpretation of patterns in gross N mineralization by controlling for the role of soil organic matter (SOM) content in driving gross N mineralization rates.

Microbial biomass and N assimilation rates

To determine the role of microbial N assimilation in driving net mineralization and nitrification rate patterns, we measured microbial NH4+ and NO3− assimilation rates as well as microbial biomass C and N. For these measurements, we used direct chloroform extraction (Setia et al. 2012). At 15 min and 4 h of the soil incubation with the 15N label, two soil subsamples were extracted in 0.5 M K2SO4, with one subsample incubated with chloroform for 1 h before filtration (Brookes et al. 1985). Subsamples of the 0.5 M K2SO4 extracts were analyzed for total organic C and total N on a Shimadzu TOC-V – TM analyzer (Shimadzu corporation, Columbia, Maryland, USA). To estimate microbial biomass C (MBC) and microbial biomass N (MBN), we subtracted total organic C and total N concentrations of unfumigated extracts from total organic C and total N concentrations of chloroform-fumigated extracts and divided the resulting values by 0.45 and 0.54, respectively, to convert chloroform-labile C and N to MBC and MBN (Brookes et al. 1985; Beck et al. 1997). We divided MBC by MBN to obtain microbial biomass C:N ratios. Subsamples of all extracts were also digested with potassium persulfate (Brookes et al. 1985; Cabrera and Beare 1993) before colorimetric analysis of NO3− on a SpectraMax M2 plate reader (Molecular Devices, San Jose, CA). The 15N isotopic composition of the NO3− in digested control and chloroform-treated extracts was determined using the acid trap diffusion method and EA-IRMS analysis as described earlier. Microbial assimilation rates for NH4+ and NO3− were calculated as the change in 15N recovered in microbial biomass divided by the average atom% 15N of the NH4+ and NO3− pools, respectively, over the 4 h incubation (Templer et al. 2008).

Potential nitrification and denitrification rates

To assess nitrification rates in the absence of direct substrate limitation, we assayed the soil samples for potential nitrification rates as an index of the maximum capacity of the extant soil microbial community to nitrify NH4+. We used the Berg and Rosswall, (1985) methodology as described by Kandeler et al. (1999). Briefly, we added 20 mL of 10 mM NH4+ solution as excess substrate and 0.1 mL of 1.5 M NaClO3 as a biotic NO2− reduction inhibitor to a 5 g sample of soil. We aerated the soil slurry during a 5 h incubation at room temperature (23 °C) by vigorously shaking it on a rotary shaker. During this time, another subsample of each soil sample was mixed with the NH4+ solution and placed in a -20 °C freezer to act as an abiotic control, accounting for changes in background soil NO2− concentrations caused by abiotic NO2− reactions. After the incubation, all samples were extracted in 2 M KCl, and extracts were analyzed for NO2− concentrations colorimetrically (Genesys 20 Visible Spectrophometer, Thermo-Scientific, Waltham, MA, USA). To determine the potential nitrification rate for each sample, the NO2− concentration in the frozen soil subsample was subtracted from the unfrozen shaken subsample to calculate the rate of NO2− production through biotic nitrification over the 5 h incubation period.

To assess denitrification rates in the absence of direct substrate limitation, we assayed the soil samples for potential rates of complete and incomplete denitrification as indices of the maximum capacity of the extant soil microbial community to denitrify NO3− to the gaseous end products of N2O and N2. We used the Environmental Protection Agency protocol # RSKSOP-310. Briefly, 25 mL of 0.015 M NO3− solution was added to two 25 g soil subsamples in sealed 150 mL Wheaton vials which had been flushed with helium (He) to create anaerobic conditions. To measure total denitrification, one 25 g subsample was injected with 15 mL ∼10% acetylene (C2H2), inhibiting N2O reduction to N2. To measure total incomplete denitrification, a second 25 g subsample was injected with 20 mL He, resulting in similar headspace pressure as the C2H2 treatment. Beginning directly after gas was added to the headspace, we shook soil slurries vigorously and sampled 10 mL of headspace gas at four time points over 45 min. Gas samples were stored in sealed pre-evacuated Wheaton vials and analyzed using a gas chromatograph equipped with an electron capture detector (ECD) and a thermal conductivity detector (TCD) for N2O and CO2 analysis, respectively (Shimadzu GC-2014, Colombia, MD). We calculated potential total denitrification rates from the linear change in headspace N2O concentrations from the C2H2 treatment, and potential incomplete denitrification rates from the linear change in headspace N2O concentrations from the He only treatment. To calculate potential complete denitrification rates, we subtracted potential incomplete denitrification rates from potential total denitrification rates. We omitted data from 8 out of 82 samples with non-linear changes in headspace N2O concentrations and CO2 concentrations (i.e., R2 < 0.80) which precluded determination of potential denitrification rates.

Gas fluxes

To assess the effects of upstream N mineralization and nitrification processes on downstream gaseous N losses, we measured nitrification-derived net N2O fluxes and denitrification-derived net N2O fluxes from the 15NH4+ and 15NO3− label additions, respectively, as gaseous N loss pathways. We also measured CO2 fluxes as representative of C mineralization rates from the free-living microbial community in the 4 h 15N pool dilution laboratory incubations. At 15 min of the soil incubation with the 15N label, we weighed 100 g subsamples of the 15N-labeled soil into 490 mL mason jars that were sealed with lids fitted with septum ports. Several room air samples were collected and stored in pre-evacuated Wheaton vials as the soil samples were sealed in the jars to represent the initial time point for gas flux calculations. At 4 h of the soil incubation, we sampled 90 mL of headspace gas and stored the gas sample in a 60 mL pre-evacuated Wheaton vial. We analyzed a 5 mL subsample of the stored gas samples for CO2 and N2O concentrations on the GC as described above. The remainder of each gas sample was analyzed for N isotopic composition of N2O on an IsoPrime 100 isotope ratio mass spectrometer interfaced with an IsoPrime trace gas analyzer (Cheadle Hulme, UK) and Gilson GX-271 autosampler (Middleton, WI). Total net N2O fluxes and CO2 fluxes were calculated from the linear change in N2O and CO2 concentrations over time. Net 15N2O fluxes were calculated from the linear change over time in 15N2O abundance, which was determined from N2O concentrations and the 15N atom% enrichment of N2O. Nitrification-derived net N2O fluxes were estimated by dividing the 15N2O flux from the 15NH4+ label treatment by the average 15N atom% enrichment of the NH4+ pool over the 4 h incubation; denitrification-derived net N2O fluxes were similarly estimated from the 15NO3− label treatment.

Statistical methods

All statistics were carried out in R 3.6.2 (R Development Core Team 2019). Statistical significance was determined based on P < 0.05. All of the data presented in this manuscript is available in the Illinois Data Bank (Seyfried et al. 2022). To test for differences in soil chemical properties and N process rates, we fit linear mixed models with stand mycorrhizal type (ECM or AM), N addition (control or N addition), and the interaction between stand mycorrhizal type and N addition as the fixed effects and with plot pair and sample dates as random effects. We performed pairwise comparisons between each level of stand mycorrhizal and N addition using the “emmeans” function in the emmeans package (Russell 2021), with Tukey’s adjustment method for multiple comparisons. The following soil chemical properties served as dependent variables: pH, NH4+ concentration, NO3− concentration, potential N2O production rates, potential denitrification rates (total and incomplete), net N mineralization rates, net nitrification rates, gross N mineralization rates, specific gross N mineralization rates, net N2O fluxes (total, nitrification-derived, and denitrification-derived), microbial biomass N, microbial biomass C, microbial biomass C:N ratios, N assimilation rates (of NH4+ and NO3−), and C mineralization rates.

Results

Consistent with past studies conducted in this site (Midgley et al. 2015; Midgley and Phillips 2016), we found that the effect of stand mycorrhizal type on soil and microbial properties was stronger and more consistent than the effect of N addition on soil and microbial properties (Table 1, S1). Specifically, ECM stands were characterized by significantly lower NO3− concentrations (F1, 12 = 169.61, P < 0.0001; Table1, S1) and pH (F1, 13.77 = 31.18, P < 0.0001; Table 1, S1) but significantly higher microbial biomass C (F1, 186.06 = 73.18, P < 0.0001), microbial biomass N (F1, 186.16 = 29.77, P < 0.0001), microbial biomass C:N ratios (F1, 183.29 = 7.44, P = 0.007) and C mineralization (F1, 12.03 = 48.96, P < 0.0001) rates compared to AM stands (Table 1). In contrast to most soil properties measured, NH4+ concentrations were similar between stand types (Table 1, S1). We found that N addition increased soil NH4+ and NO3− concentrations (F1,93 = 17.45, P < 0.0001; F1,93 = 44.09, P < 0.0001), and decreased soil pH (F1,94.90 = 7.05, P = 0.01) with average pH values of 4.57 and 4.71 for N addition and control plots, respectively (Table 1, S1). Although N addition had no effect on microbial biomass C or microbial biomass N, microbial biomass C:N ratios were significantly greater in control compared to N addition plots in ECM-dominated stands (P = 0.0031), but not AM-dominated stands. The effect of N addition on C mineralization rates was marginally significant with control plots trending towards greater C mineralization rates relative to N addition plots (F1, 204.08 = 3.07, P = 0.08; Table 1, S1).

Table 1 Soil and microbial properties in ectomycorrhizal (ECM) and arbuscular mycorrhizal (AM) dominated forest stands at Moores Creek Research and Teaching Preserve in South Central Indiana (Mean ± SE)
Full size table

We found that gross N mineralization and specific gross N mineralization differed significantly between ECM and AM forest stands, but the effect of stand mycorrhizal type was opposite for the two metrics of mineralization (Fig. 1a,b). Gross N mineralization was greater in ECM stands compared to AM stands (F1,102.03 = 8.95, P = 0.003; Fig. 1a) whereas specific gross N mineralization was greater in AM stands compared to ECM stands (F1,99.06 = 14.44, P = 0.0003; Fig. 1b). Opposite to patterns in gross N mineralization, net N mineralization rates were significantly greater in AM stands compared to ECM stands (F1, 12 = 49.41, P < 0.0001; Fig. 2a). In contrast, gross NH4+ assimilation aligned with patterns of gross N mineralization, with greater rates in ECM relative to AM stands (F1,65.23 = 21.29, P < 0.0001; Fig. 2b). Net N mineralization (F1,93 = 16.60, P < 0.0001) and NH4+ assimilation (F1,65.09 = 8.35, P = 0.005) increased with N addition, but there was no effect of N addition on gross N mineralization or specific gross N mineralization (Figs. 1a, b and 2a, b).

Fig. 1
figure 1

Boxplot showing the effects of stand mycorrhizal type and nitrogen (N) addition on a gross N mineralization and b specific gross N mineralization rates in arbuscular mycorrhizal (AM)- dominated stands (purple) and ectomycorrhizal (ECM)-dominated stands (yellow). Stand mycorrhizal type significantly affected gross N mineralization rates (F1,102.03 = 8.95, P = 0.003) and specific gross N mineralization rates (F1,99.06 = 14.44, P = 0.0003). Asterisks denote significant differences between stand types (*p < 0.05; **p < 0.01)

Full size image
Fig. 2
figure 2

Boxplots showing the effects of stand mycorrhizal type and nitrogen (N) addition on a net N mineralization rates, b ammonium assimilation rates, c net nitrification rates and d nitrate assimilation rates in arbuscular mycorrhizal (AM)- dominated stands (purple) and ectomycorrhizal (ECM)-dominated stands (yellow). The effects of nitrogen addition and stand mycorrhizal type were significant for net N mineralization rates (F1,93 = 16.60, P < 0.0001; F1,12 = 49.41, P < 0.0001, respectively) and net nitrification rates (F1,93 = 21.96, P < 0.0001; F1, 12 = 253.89, P < 0.0001, respectively). Stand mycorrhizal type significantly affected NH4+ assimilation rates (F1,65.23 = 21.29, P < 0.0001) but not NO3− assimilation rates (F1,74.02 = 1.96, P = 0.17). Nitrate and NH4+ assimilation rates were significantly affected by N addition (F1,74.02 = 23.95, P < 0.0001; F1,65.09 = 8.35, P = 0.005 for NO3− and NH4+ assimilation respectively) though this effect was only significant for NH4.+ assimilation in ECM-dominated stands (p = 0.003). Asterisks denote significant differences between stand types (*p < 0.05; ***p < 0.001)

Full size image

The effect of stand mycorrhizal type on nitrification rates was consistent across both nitrification assays and control versus N addition treatments, with lower rates in ECM stands. Potential and net rates were lower in ECM compared to AM-dominated stands, though this effect was only marginally significant for potential nitrification rates (Potential, F1, 12.25 = 3.68, P = 0.08; Fig. 3a; net, F1,12 = 253.89, P < 0.0001; Fig. 2c), with similar patterns across N addition and control treatments. Nitrogen addition resulted in significantly greater net nitrification rates (F1,93 = 21.96, P < 0.0001), but had no effect on potential nitrification rates (Fig. 2c and 3a). In contrast to NO3− production processes, NO3− assimilation rates did not differ between ECM and AM stands and increased with N addition (F1,74.02 = 23.95, P < 0.0001; Fig. 2d).

Fig. 3
figure 3

Boxplots showing the effects of stand mycorrhizal type and nitrogen (N) addition on a potential nitrification rates, b potential total denitrification and c potential incomplete denitrification in arbuscular mycorrhizal (AM)- dominated stands (purple) and ectomycorrhizal (ECM)-dominated stands (yellow). Potential nitrification rates were marginally affected by stand mycorrhizal type (F1, 12.25 = 3.68, P = 0.08). Potential total denitrification rates and potential incomplete denitrification rates were significantly affected by stand mycorrhizal type (F1,12.11 = 27.35, P = 0.0002; F1,12.05 = 41.63, P < 0.0001, respectively). There were no effects of N addition on potential nitrification or potential denitrification rates. Asterisks denote significant differences between stand types (*p < 0.05; ***p < 0.001)

Full size image

We found that total (nitrification- plus denitrification-derived), nitrification-derived, and denitrification-derived net N2O fluxes exhibited similar patterns to nitrification rates with lower fluxes in ECM compared to AM stands (total, F1,12.02 = 32.56, P < 0.0001; nitrification-derived, F1,11.93 = 7.46, P = 0.02; denitrification-derived, F1,79 = 10.34, P = 0.002; Fig. 4a–c). However, for nitrification-derived N2O fluxes, this effect was only significant in N addition plots (Fig. 4a). Potential total denitrification and potential incomplete denitrification were also significantly lower in ECM compared to AM stands (F1,12.10 = 27.35, P < 0.0002; F1,12.05 = 41.63, P < 0.0001; Fig. 3b, c). Total net N2O fluxes were significantly greater in N addition compared to control plots (F1,204.07 = 21.87, P < 0.0001) (Fig. 4c). The effect of N addition on denitrification-derived N2O fluxes was marginally significant with greater fluxes in N addition compared to control plots (F1,79 = 3.08, P = 0.08; Fig. 4b). However, for nitrification-derived net N2O fluxes and for potential total and potential incomplete denitrification, there was no effect of N addition (Figs. 3b, c and 4a).

Fig. 4
figure 4

Boxplots showing the effects of stand mycorrhizal type and nitrogen (N) addition on a nitrification-derived net N2O fluxes, b denitrification-derived net N2O fluxes and c total net nitrous oxide (N2O) fluxes in arbuscular mycorrhizal (AM)- dominated stands (purple) and ectomycorrhizal (ECM)-dominated stands (yellow). Stand mycorrhizal type significantly affected nitrification-derived net N2O fluxes (F1,11.93 = 7.46, P = 0.02), denitrification-derived net N2O fluxes (F1,79 = 10.34, P = 0.002) and total net N2O fluxes (F12.02 = 32.55, P < 0.0001). Although total net N2O fluxes were significantly affected by N addition (total, F1,204.07 = 21.87, P < 0.0001), there was no effect of N addition on source-partitioned N2O fluxes. Asterisks denote significant differences between stand types (*p < 0.05; **p < 0.01; ***p < 0.001)

Full size image

Discussion

Tree association with ECM versus AM fungi clearly mediates distinct nutrient syndromes (e.g. Averill et al. 2014; Corrales et al. 2016; Lin et al. 2017; Phillips et al. 2013; Zhu et al. 2018), yet the mechanisms driving these mycorrhizal type patterns have remained unclear (Lin et al. 2017; Averill et al. 2019; Keller and Phillips 2019). Suppressed N mineralization due to slow decomposition of low quality ECM litter and organic N uptake by ECM fungi may initiate formation of an organic nutrient economy with closed N cycling in ECM stands (Phillips et al. 2013; Brzostek et al. 2015). However, lower net rates of N mineralization quantified in past studies conflate gross production and consumption of inorganic N (e.g. Lin et al. 2017; Midgley and Phillips 2016; Mushinski et al. 2021). In this study, we aimed to address the following hypotheses: (1) net N mineralization rates mask patterns in gross N cycling; (2) ammonium (NH4+) supply does not limit nitrification in ECM soils; (3) N cycling processes downstream of nitrification are limited by nitrate (NO3−) availability. We demonstrated that gross N mineralization rates and microbial inorganic N assimilation rates can be greater while net N mineralization rates are lower in ECM stands compared to AM stands. Strikingly, despite higher gross N mineralization rates in ECM stands and similar soil NH4+ concentrations between stand mycorrhizal types, we observed lower nitrification and denitrification rates in ECM stands. This suggests that ecosystems with closed N cycles do not necessarily cycle inorganic N slowly, as has been presumed for ECM stands. Here we discuss how our findings improve understanding of the potential mechanisms driving mycorrhizal nutrient syndromes.

Net N mineralization rates mask patterns in gross N cycling in ECM versus AM stands

Controls on N mineralization in ECM versus AM soils have previously been considered in the context of litter chemical quality and microbial competition for limited N, but these mechanistic controls may not accurately predict ecosystem scale N cycling dynamics. Slower leaf litter decomposition and greater microbial N limitation in ECM stands, which have previously been documented at our study site in Moores Creek, Indiana (Phillips et al. 2013; Midgley et al. 2015; Midgley and Phillips 2016), are predicted to suppress gross N mineralization rates in ECM compared to AM stands. We found gross N mineralization was indeed greater in AM compared to ECM stands, but only when normalized by organic matter content. This supports the hypothesis that lower quality, slower decomposing ECM leaf litter can suppress N mineralization in ECM surface soils. However, higher organic matter content in ECM soils means that gross N mineralization rates per gram of dry soil were greater in ECM stands. Furthermore, greater microbial biomass C and N in ECM stands suggests that greater OM availability may support a larger microbial community and increased N transformations. Therefore, suppressed decomposition of lower quality ECM leaf litter, driving proliferation of fine roots and mycorrhizal hyphae and accumulation of particulate organic matter (POM) in ECM surface soils, may result in greater surface soil NH4+ availability as compared to faster decomposition of higher quality AM leaf litter. Importantly, this balance between organic matter quality and quantity may vary across systems with lower quality litter produced by conifer species in boreal forests potentially suppressing decomposition and N mineralization more than broadleaf temperate species present at our study site.

Despite the putative capacity of ECM fungi to directly uptake organic molecules and bypass inorganic N cycling (Read and Perez-Moreno 2003; Lindahl and Tunlid 2015), greater gross N mineralization rates in ECM stands may indicate that ECM fungi stimulate mineralization of low quality ECM substrate by free-living decomposers (Phillips and Fahey 2006; Meier et al. 2015; Sulman et al. 2017). Positive priming effects on gross N mineralization have been shown to correlate with N-acquiring hydrolytic enzyme activities in ECM soils (Yin et al. 2021) such that exclusion of ECM hyphae decreases enzyme activities (Brzostek et al. 2015; Yin et al. 2021). Exudation of labile C compounds can stimulate N transformations (Dijkstra et al. 2013; Meier et al. 2017) and increase N availability, such that C exudation rates are often greater in low-N soils (Pausch and Kuzyakov 2018). In our temperate study site, lower N ECM leaf litter decomposes slower than higher N AM leaf litter (Phillips et al. 2013; Midgley et al. 2015), resulting in N-limited free-living microbes in ECM soils where N return from decomposition of ECM leaf litter is lower. Nitrogen limitation may explain significantly greater C exudation rates in ECM compared to AM soils (Phillips & Fahey 2005; Yin et al. 2014). Furthermore, physically accessible particulate organic matter that accumulates in ECM surface soils (Craig et al. 2018; Averill et al. 2019) may be vulnerable to priming effects (Kuzyakov 2010). Specifically, standing fungal biomass, which can be up to 2.5 times greater in ECM stands relative to AM stands (Cheeke et al. 2017), represents a relatively high turnover SOM pool that has been shown to undergo accelerated decay in the presence of primed microbial communities (Meier et al. 2017). In contrast, mineral-associated organic matter that dominates SOM pools in AM soils (Craig et al. 2018; Cotrufo et al. 2019) is spatially diffuse and physically protected such that priming SOM decomposition may not be worth the C cost (Brzostek et al. 2015; Sulman et al. 2017). Therefore, rather than suppressing N mineralization, we speculate that slow leaf litter decomposition rates in ECM stands may indirectly stimulate mineralization by driving ECM trees to allocate more C belowground to prime SOM decomposition by rhizosphere microbial communities.

Potential drivers of nitrification and denitrification in ECM versus AM stands

We found that inorganic N supply from N mineralization does not regulate nitrification rates in ECM versus AM stands, but rather, other mechanisms could drive suppression of nitrification in ECM stands despite high rates of NH4+ production. First, heterotrophs may outcompete nitrifiers for NH4+ when substrate C:N ratios are high and heterotrophs are relatively more N-limited than C-limited. This mechanism may be particularly relevant in ECM surface soils that are characterized by accumulation of high C:N particulate organic matter (Craig et al. 2018; Averill et al. 2019). In support of this, we found that addition of inorganic N increased microbial assimilation of NH4+ and NO3− in ECM, but not AM soils, demonstrating the stronger demand by free-living microbes for inorganic N in ECM soils. This competitive advantage of heterotrophs over nitrifiers could lead to the lower abundance of ammonia oxidizers often observed in ECM stands relative to AM stands (Mushinski et al. 2019; Lin et al. 2022b). Alternatively, protonation of ammonia to ammonium in acidic ECM soils and heterotrophic N demand may decrease the availability of NH4+ for chemoautotrophic growth of ammonia oxidizers. Therefore, low pH and high C:N ratios in ECM soils may result in decreased nitrifier abundance and suppressed nitrification (Scharko et al. 2015; Mushinski et al. 2019; Xiao et al. 2020; Lin et al. 2022b). In support of this, we found greater potential nitrification rates in AM- relative to ECM-dominated stands, suggesting that extant soil microbial communities in higher pH, lower C:N ratio AM soils have a greater capacity to nitrify NH4+ than extant microbial communities in lower pH, higher C:N ratio ECM soils. Leaf litter decomposition dynamics may play a role in forming low pH, high C:N ECM soils that can indirectly suppress nitrification. However, we provide evidence that slower decomposition of lower quality leaf litter inputs in ECM compared to AM stands does not drive nitrification via suppressed mineralization. Instead, separate mechanistic pathways may govern mineralization and nitrification in ECM soils.

Gaseous nitrogen losses via denitrification generally correlated with patterns in NO3− availability across ECM and AM forest stands, suggesting that suppressed nitrification and small NO3− pools in ECM stands limit denitrification. However, opposite to what we would have expected based on substrate-limited denitrification in ECM stands, long-term N addition did not stimulate denitrification derived N2O fluxes. Denitrifiers, which are facultative anaerobes, will only respire NO3− in anoxic microsites within unsaturated forest soils. Therefore, although addition of NO3− to denitrifiers in anaerobically incubated soils can stimulate N2O production (Mushinski et al. 2021), addition of NO3− to denitrifiers in oxic soils may have no effect on denitrification rates. Our results suggest that nitrification-derived NO3− availability, which is greater in AM compared to ECM stands, has distinct effects on N2O fluxes compared to fertilizer-derived NO3− availability. Nitrification-derived NO3− may be produced in closer proximity to anoxic microsites, stimulating greater N2O fluxes in AM stands. However, long-term N addition to the soil surface may not reach anoxic microsites to alleviate substrate-limited denitrification in ECM stands. Instead of fueling denitrification, microbial N assimilation or leaching could be more important fates for added NO3−. Consistent with this hypothesized mechanism, we found increased microbial assimilation of NO3− and NH4+ in response to N addition in ECM-dominated stands. Furthermore, N addition has been demonstrated to stimulate NO3− leaching particularly in ECM stands (Midgley and Phillips 2014), such that added inorganic N could be largely lost from the system. Our data suggest that patterns in denitrification-derived net N2O fluxes will mirror patterns in nitrification due to nitrification-derived NO3− limitation of denitrification in unsaturated forest soils.

Conclusion

Our study has advanced mechanistic insight into formation of an organic nutrient economy in ECM stands and an inorganic nutrient economy in AM stands. First, we show that gross N mineralization and consumption rates can be greater in ECM compared to AM soils, resulting in similar NH4+ concentrations across forest types. Second, we found that nitrification may be limited by factors other than NH4+ concentrations such that nitrification patterns do not necessarily correspond to mineralization patterns. Instead, nutrient conservative traits of ECM trees and associated mycorrhizal fungi may indirectly suppress nitrification rates by decreasing soil pH and increasing soil C:N ratios. Third, we observed that downstream gaseous N losses may be limited by NO3− availability such that mechanistic drivers of nitrification may also control ecosystem N2O fluxes. This suggests that strong inorganic N demand by free-living microbes and soil acidity effects on nitrification may lead to the closed ecosystem N cycle characteristic of ECM forest stands compared to the open ecosystem N cycle of AM-dominated forest stands. Overall, we conclude that N mineralization does not play a central role in forming mycorrhizal nutrient syndromes as previously thought, and that soil pH may ultimately control nitrification and the potential for ecosystem N loss.