Atmospheric nitrogen (N) deposition in forests can affect soil microbial growth and turnover directly through increasing N availability and indirectly through altering plant-derived carbon (C) availability for microbes. This impacts microbial residues (i.e., amino sugars), a major component of soil organic C (SOC). Previous studies in forests have so far focused on the impact of understory N addition on microbes and microbial residues, but the effect of N deposition through the plant canopy, the major pathway of N deposition in nature, has not been explicitly explored. In this study, we investigated whether and how the quantities (25 and 50 kg N ha^-1 yr^-1) and modes (canopy and understory) of N addition affect soil microbial residues in a temperate broadleaf forest under 10-yr N additions. Our results showed that N addition enhanced the concentrations of soil amino sugars and microbial residual C (MRC) but not their relative contributions to SOC, and this effect on amino sugars and MRC was closely related to the quantities and modes of N addition. In the topsoil, high-N addition significantly increased the concentrations of amino sugars and MRC, regardless of the N addition mode. In the subsoil, only canopy N addition had a positive effect on amino sugars and MRC, implying that the indirect pathway via plants plays a more important role. Neither canopy nor understory N addition significantly affected soil microbial biomass (as represented by phospholipid fatty acids), community composition, and activity, suggesting that enhanced microbial residues under N deposition likely stem from increased microbial turnover. These findings indicate that understory N addition may underestimate the impact of N deposition on microbial residues and SOC, highlighting that the processes of canopy N uptake and plant-derived C availability to microbes should be taken into consideration when predicting the impact of N deposition on the C sequestration in temperate forests.

Soil physicochemical properties Soil pH was measured in a slurry (soil: water = 1: 2.5, w/v) with a pH meter (FiveEasy Plus^TM FE28, Mettler Toledo). SOC, soil total nitrogen (TN), and total phosphorus (TP) concentrations were determined by the concentrated sulfuric acid-potassium dichromate external heating method, the concentrated sulfuric acid digestion-phenol blue colorimetric method, and the concentrated sulfuric acid digestion-molybdenum antimony anti-colorimetric method, respectively (Lu, 1999).

Soil microbial PLFAs The Soil microbial community was characterized by the phospholipid fatty acids (PLFAs) method (Bossio & Scow 1998). The concentration of each PLFA was calculated based on the 19:0 internal standard concentration. The PLFAs i14:0, i15:0, a15:0, i16:0, a16:0, i17:0, a17:0, a18:0, i18:0, a19:0, 16:1ω7c, 16:1ω9c, 17:1ω8c, 18:1ω7, cy17:0, and cy19:0 were used to indicate bacterial biomarkers. The PLFAs 18:1ω9c, 18:2ω6,9c, and 18:3ω6,9,12c were applied to denote fungal biomarkers. The PLFA 16:1ω5c was considered an arbuscular mycorrhizal fungal (AMF) biomarker. The PLFAs 10Me 16:0, 10Me 17:0, and 10Me 18:0 were used as actinomycetes biomarkers. Total microbial biomass was represented by the sum of bacterial, fungal, AMF, and actinomycetes biomarkers. Soil microbial community structure was represented by the ratio of fungal to bacterial PLFAs (F:B ratio) (Frostegård & Bååth 1996).

Soil microbial activity Soil microbial activity was assessed through determining microbial respiration under incubation (Hu & van Bruggen, 1997). A 50 ml beaker containing 30.00 g (fresh weight) soil sample was placed in a 1000 ml jar. Inside this jar, a 25 ml beaker containing 5.00 mL 1 M NaOH solution was placed to trap CO₂ produced by incubated soil sample. Four replicates of each treatments and four control jars with no soil samples were prepared. Each jar was sealed and then kept at 25℃ for 10 days in one incubation period, 44 jars in total. Subsequently, the beaker containing NaOH solution was taken out and the NaOH solution was titrated by 0.20 M HCl solution. After each measurement, each sample was returned to ambient CO₂ level by leaving the incubation jars open for 1 h and NaOH was refreshed. Meanwhile, the soil moisture content was maintained by weighing the 50 ml beaker and adding the distilled water. These soil samples were subjected to six incubation periods (60 days of incubation period in total). The released CO₂ in each incubation period was calculated as the follows: R_CO2 = (V₀ - V_k) × C_HCl × 44 / (2 × w) / 10, where R_CO2 is the rate of CO₂ release (mg g^-1 d^-1). V₀ and V_k are the volumes of consumed HCl in the blank controller (without soil) and the incubated soil treatments (ml), respectively. C_HCl is the calibrated concentration of HCl (mol L^-1). 44 is the molar mass of CO₂; w is the dry weight of the fresh soil (g); 10 is the days of each incubation period (days).

Soil amino sugars The concentrations of soil amino sugars, including muramic acid (MurN), galactosamine (GalN), and glucosamine (GluN), were determined as described by Indorf et al. (2011). In brief, amino sugars were hydrolyzed, extracted, and derivatized with ortho-phthaldialdehyde, determined by high-performance liquid chromatography (Dionex Ultimate 3000, Thermo Fisher Scientific). The detailed relevant information was described by Yuan et al. (2021).

Data calculation and statistical analysis

Microbial residual C (MRC) was calculated by the following formulas (1)-(4): (1) F-GluN (μg g⁻¹) = total GluN (μg g⁻¹) – 2 × MurN (μg g⁻¹) × (179.2/251.2), where F-GluN is fungi-derived GluN. It was assumed that MurN and GluN occurred at a molar ratio of 1 to 2 in bacterial cell walls (Engelking et al., 2007). Where 179.2 and 251.2 are the molecular weights of GluN and MurN, respectively (Shao et al., 2017). (2) Fungal MRC (μg g⁻¹) = F-GluN × 9, (3) Bacterial MRC (μg g⁻¹) = MurN × 45, (4) Total MRC (μg g⁻¹) = Fungal MRC + Bacterial MRC, where 9 and 45 are conversion factors (Appuhn & Joergensen, 2006). Fungal MRC, bacterial MRC, and total MRC are fungi-derived, bacteria-derived, and total microbial residual C, respectively.

A univariate general linear model (GLM) with a random factor (blocks) was used to examine the effects of N deposition treatments on concentrations of soil total amino sugars, MurN, GalN, GluN, fungal MRC, bacterial MRC, total MRC, and the ratio of fungal to bacterial MRC, while the effect of blocks was not significant. The effects of N deposition on the mean of SOC mineralization rate, relative contributions of total amino sugars, fungal MRC, bacterial MRC, and total MRC to SOC were tested by a general linear model as well. The impacts of N deposition on soil physicochemical properties (pH, SOC, TN, and TP) and soil microbial parameters (the PLFAs of bacteria, fungi, AMF, actinomycetes, and total microbes, and the F:B ratio) were examined by one-way ANOVA. Multiple comparison analyses (LSD) were used after general linear model and one-way ANOVA. Pearson correlation analysis was performed to assess the relationships of measured soil microbial residues (amino sugars and residual C) with soil physicochemical properties and soil microbial parameters (microbial PLFAs and community structure). All statistical analyses were carried out with SPSS 18.0 (SPSS, Chicago, Illinois, USA), and results were considered statistically significant at p < 0.05.