Transformation pathways for NO₃^- under the four incubation conditions are presented schematically in Fig 4. Our results indicate both redox and abiotic conditions govern the dynamics of NO₃^- transformations in organic matter systems. Specifically, oxic incubations indicate that ¹⁵NO₃^- was not recovered in the NH₄⁺ pool; the dominant pathway for NO₃^- transformation is conversion to DON, accounting for 20 and 17% of the spiked ¹⁵NO₃^- in biotic-oxic and γ-irradiated abiotic-oxic systems, respectively (Table A in S1 File, Fig 4). Hence, under oxic conditions NO₃^- is more prone to being incorporated abiotically into DOM than into SOM when incubations were conducted in suspension. A similar trend was also found in several studies [15,20,34]. Davidson et al. (2003) [25] proposed that NO₃^- is first reduced to NO₂^- prior to its incorporation into DOM. Their hypothesis is based on the results of two discrete experiments. The first experiment showed nitrate reduction by Fe(II) in the presence of a Cu catalyst under anoxic conditions; the second experiment showed nitrite incorporation into dissolved organic matter (i.e., solutions of several organic compounds). In the first experiment, a decrease in nitrate concentration with time, not an increase in nitrite concentration, was measured; therefore, it was not demonstrated that NO₂^- is the intermediary between NO₃^- and DOM. In addition, Schmidt and Matzner (2009) [45] showed that NO₂^- transformation to DOM under oxic conditions did not occur after NO₂^- was added to sterilized DOM. Matus et al. (2019) [23] however demonstrated that under abiotic-anoxic conditions, and in the presence of Fe²⁺, DOM reacts with NO₃^- to form DON. Based on our experimental results we cannot confirm nor completely refute whether NO₃^- was reduced to NO₂^- under the oxic conditions in which NO₃^- was converted to DOM (NO₂^- was not detected in oxic incubations at values above 1 μM). However, we can use the Nernst equation to calculate the theoretical concentration of nitrite (NO₂^-) that would result from the NO₃^-/NO₂^- redox couple: ½ NO₃^- + H⁺ + e^- → ½ NO₂^- + ½ H₂O (E_h^o = 0.834 V). With parameter values of pH = 6.5, E_h = 500 mV and [NO₃^-] = 200 μM (i.e., NO₃^- initially added), we calculated [NO₂^-] = 4.1 μM, which is above our detection limit of 1 μM. Higher NO₂^- concentrations are expected at lower redox potentials whereas lower NO₂^- concentrations are expected at higher redox potentials. Our oxic incubations, with a redox potential of ~300 mV, should have in theory resulted in a nitrite concentration of 199 μM, a concentration easily detectable. Therefore, the conversion of NO₃^- to DOM that we observed in experiments conducted under oxic conditions does not seem to follow the path hypothesized by Davidson et al. (2003) [25]. Although NO₂^- was not detected, it is possible the kinetics of NO₂^- consumption were as fast (or faster) than those of NO₂^- production, which would have prevented NO₂^- accumulation in the system. A potential pathway is that NO₃^- undergoes an electrophilic aromatic substitution (aromatic nitration) in which a nitro group (R-NO₂) is introduced into an organic chemical compound. Yet, we do not have conclusive evidence in support of this reaction as a legitimate possibility. Recent studies have demonstrated that UV irradiation effects the incorporation of nitrate and nitrite into natural organic matter via nitration and nitrosation yielding a variety of organic-N functionalities [50]. These reactions present additional evidence in support of abiotic pathways leading to the incorporation of inorganic N species into organic matter [51].

Under anoxic conditions, all of the spiked ¹⁵N-NO₃^- was transformed within 24 h (Fig 1B). About 8.6% of the spiked ¹⁵N-NO₃^- was recovered as ¹⁵N-NH₄⁺ (Table A in S1 File, Fig 4) by the end of the biotic-anoxic and γ-irradiated abiotic-anoxic system incubations, with no ¹⁵N recovery in the DOM pool. Given that ~0.8% of ¹⁵N-NO₃^- was immobilized into SOM and that relatively large concentrations of NO₂^- were detected in anoxic incubations, we infer ~90% of the spiked ¹⁵N-NO₃^- was reduced to gaseous N species. Our inference is supported by the work of Zhang et al. (2010) [34] in which the authors found that up to 51% of the spiked NO₃^- was converted to N₂ gas when NO₃^- was incubated with forest soils anaerobically. Although gaseous N species were not measured in this study, the low E_h (-330 mV) in all anoxic incubations suggests a very low oxygen level conducive to N₂ gas production.

Although Fe(II) is expected to play a role in NO₃^- reduction, the results from biotic-oxic incubations show the addition of NO₃^- had no influence on Fe(II) concentrations (Fig 1C), therefore indicating Fe(II) did not act as an electron source in biotic-oxic incubations (but perhaps as a catalyst). However, the addition of nitrate under anoxic conditions affected Fe(II) concentrations. Fe(II) concentrations were consistently lower in experimental compared to blank anoxic incubations (except at 120 h) suggesting Fe(II) was directly involved in NO₃^- reduction (Fig 1D). These findings lend support for Davidson’s hypothesis that postulates NO₃^- reduction by Fe²⁺ under anoxic conditions [25]. Potential electron sources that would support NO₃^- reduction under anoxic conditions are reducing organic functional groups and Fe(II) present in the leaf compost (Fe(II) = 1.2 g Fe(II)/kg = 21.5 μmole Fe(II)/g = 43 μmole Fe(II) in reaction vessel; total Fe = 4.8 g Fe/kg = 85.9 μmole Fe/g = 171.8 μmole Fe in reaction vessel). Since 100 μmole NO₃^- were added to each experimental incubation, and given the fact that all of the added NO₃^- was transformed (reduced) to other species under anoxic conditions, reduction of NO₃^- to NO₂^- alone would require 100 μmole of electrons. Using the concentration of NO₃^- (~100 μM in solution), NO₂^- (~40 μM in solution) and Fe(II) (~4 μmole Fe(II)/g = difference between blank and experimental values) at 8 h in biotic-anoxic incubations, and our experimental parameters (2 g LC in 500 ml suspension), we calculate that ~50 μmole of NO₃^- were consumed, ~20 μmole of NO₂^- were produced and that Fe(II) was reduced by ~8 μmole in experimental compared to blank incubations. These calculations indicate Fe(II) provided electrons for NO₃^- reduction but was not the sole electron source for NO₃^- reduction in anoxic incubations, otherwise the Fe(II) concentration would have decreased considerably more. If Fe(II) had served as a catalyst (i.e., a substance that increases the rate of a chemical reaction without itself undergoing any permanent chemical change), the reduction of NO₃^- to NO₂^- would take place but the concentration of Fe(II) would remain constant. We suggest organic matter (i.e., LC in this work) provided electrons for NO₃^- reduction. Some studies have found that the rate of denitrification is highly correlated to DOC concentration in groundwater [52,53], suggesting microorganisms may prefer DOC over solid organic C (SOM) as the electron source to fuel denitrification. In this study, experimental and blank DOC concentrations in biotic-anoxic systems increased with time, independent of nitrate addition (Figure B in S1 File), thus suggesting the LC may have served as an additional electron source for NO₃^- reduction in anoxic incubations.

¹⁵N-NO₃^- was not recovered in the NH₄⁺ pool of the experimental oxic systems which implies ¹⁵N-NO₃^- was not directly or indirectly converted to NH₄⁺. The contrast in NH₄⁺ production between the biotic-oxic and γ-irradiated abiotic-oxic systems (Fig 2A and 2B) indicate microbial activity was required for NH₄⁺ production. Production of NH₄⁺ can be explained by classic theory, namely, microbial generation of low molecular weight organics from solid-phase OM using extracellular enzymes, and organic N assimilation followed by NH₄⁺ excretion [54,55]. The fact that NH₄⁺ was not labeled but DON in the experimental biotic-oxic system was highly enriched in ¹⁵N due to ¹⁵N-NO₃^- conversion to DON is striking (Table A in S1 File). Although microbial utilization of non-labeled DON might be coincidental, we speculate microbes could release enzymes that convert solid phase OM into NH₄⁺ (e.g., enzymes that cleave amino groups from proteinaceous material). The addition of NO₃^- did not seem to influence the outcome of microbial NH₄⁺ production, as evidenced from the fact that the same amount of NH₄⁺ was produced in experimental and blank oxic incubations. Schmidt et al. (2011) [56] also found that NO₃^- addition exerted no impact on the rate of DOM microbial mineralization.

As mentioned above, NH₄⁺ produced in the experimental biotic-anoxic incubation was twice that in the blank (Fig 2C). This additional NH₄⁺ production can be attributed to two processes, one of which is ¹⁵N-NO₃^- reduction to ¹⁵N-NH₄⁺ as confirmed by increased ¹⁵N enrichment in the NH₄⁺ pool (Table A in S1 File). We cannot confirm, however, whether the reduction was a biotic or abiotic process. Our calculation indicates that ~8.7% of the spiked ¹⁵N-NO₃^- was reduced to ¹⁵N-NH₄⁺ (Table A in S1 File, Fig 4), accounting for ~14% of the total NH₄⁺ production. Besides differences in NH₄⁺ production, the experimental and blank biotic-anoxic systems also differ in DON production: DON concentration remained unchanged in the experimental biotic-anoxic incubation but increased in the blank. This suggests DON was mineralized (reduced) to NH₄⁺ in the experimental biotic-anoxic system but not in the blank. A potential explanation for such difference is that DON mineralization under anoxic conditions was a microbially-driven process enhanced in the experimental incubation due to NO₃^- addition. ¹⁵N-NO₃^- was also reduced to ¹⁵N-NH₄⁺ (~8.6%) in the experimental γ-irradiated abiotic-anoxic system (Table A in S1 File, Fig 4), accounting for ~10% of the total NH₄⁺ production (Fig 2D). The remaining NH₄⁺ production in experimental as well as blank γ-irradiated abiotic-anoxic incubations should be attributed to organic nitrogen chemical reduction. In these γ-irradiated anoxic systems, since a decrease in DON was associated with an increase in NH₄⁺ production, it is most reasonable to conclude that DON was chemically reduced to NH₄⁺ under abiotic-anoxic conditions.

The magnitude of N immobilization within SOM ranged from 4.1 to 6.6 μg ¹⁵N per g LC, accounting for 0.6–0.9% of the total added ¹⁵N. Nitrogen immobilization, expressed as percentage, appears to be lower than figures reported in similar studies. Dail et al. (2001) [15] and Fitzhugh et al. (2003) [20] reported that 5–10% of the total added ¹⁵N-NO₃^- (4–5 μg ¹⁵N per g soil) was immobilized by live or sterilized O-horizon forest soils; such immobilization translates to a magnitude less than 1 μg ¹⁵N per g soil. Thus, our seemingly low % N immobilization by leaf compost can simply be explained by the larger total ¹⁵N-NO₃^- input. Moreover, using a N immobilization value of 6.5 μg N g^-1 leaf compost and a density of 1 g cm^-3, we calculate 5.2 kg N ha^-1 would be immobilized (stored) within an organic matter layer 8 cm in thickness (e.g., O-horizon of a forest soil). The amount of N stored within SOM would be 3.2 kg N ha^-1 using an N immobilization value of 4 μg N g^-1 leaf compost. Our estimates of N immobilization help to explain, at least in part, the observed decline in N export in some forest ecosystems [57].

As NO₃^- input from fossil fuel combustion and fertilizer application continues to bypass the natural N cycle, changes in climatic conditions (e.g., frequency and intensity of precipitation that affects redox gradients and microbial growth rates and metabolism) might enhance the prevalence of abiotic transformations (i.e., chemical processes) that shift pathways and N species concentrations from those controlled by biota.