1 Introduction

As the largest carbon reservoir in terrestrial ecosystems, soil has a significant impact on the maintenance of ecosystem multifunctionality and mitigation of climate change (Stockmann et al. 2013; Yang et al. 2020a; Walker et al. 2022). Nitrogen (N) enrichment has a substantial influence on nutrient cycling and essential functions in terrestrial ecosystems (Liu et al. 2011; Zhang et al. 2021). In recent years, numerous studies have assessed the impact and potential mechanisms of N addition on soil organic carbon (SOC) storage. Studies have shown that N enrichment promotes soil carbon sequestration by enhancing plant carbon input and weakening microbial decomposition effects (Entwistle et al. 2018; Xu et al. 2021; Lu et al. 2021; Eastman et al. 2021). Studies have also demonstrated that interactions among plants, soil minerals, and microorganisms under N addition can modify the composition and stability of SOC, leading to no substantial changes in SOC pools (Lu et al. 2011; Song et al. 2014; Ye et al. 2018). Furthermore, excessive N addition has been shown to have a significant negative effect on SOC sequestration and biodiversity (Zhang et al. 2018; Tian et al. 2019; Tang et al. 2023). However, most studies have only focused on the overall or physical composition of SOC, and the understanding of molecular level changes in SOC under N addition remains very limited.

SOC is mainly a continuum composed of plant and microbial-derived compounds with different properties, sizes, and stages of decomposition (Grandy and Neff 2008; Lehmann and Kleber 2015). With the development of biomarker technology, sugars, lipids, and lignins derived from plants and microorganisms have been discovered in soil, allowing monitoring of the biochemical cycling process of SOC molecules in intricate soil systems (Feng and Simpson 2007, 2011; Mueller et al. 2012; Sun et al. 2022). Recent research has indicated that microbial residues have a more significant impact on the accumulation of SOC than plant residues (Joergensen and Wichern 2008; Liang et al. 2017). However, most studies have only used lignin phenols to characterise plant-derived organic carbon (OC), a practice that may lack certainty in elucidating the formation and turnover processes of plant residues with respect to SOC (Zhang et al. 2022; Xia et al. 2023). Research has indicated that, within forest ecosystems, plant-derived lipids (free lipids, cutin, and suberin) have a greater impact on SOC accumulation than lignin phenols (Dai et al. 2022). Additionally, research has demonstrated that OC bonded to soil minerals contains a higher proportion of plant-derived lipids compared to microbial sources (Angst et al. 2021). Consequently, notable variations exist in the composition of plant and microbial-derived OC and their contributions to soil organic matter across diverse land-use types and hydrothermal environments. Currently, there is considerable interest in carbon derived from soil microbial residues in SOC research, with an increasing number of studies elucidating its ecological impacts and roles. However, there is still a paucity of research regarding the regulatory role of plant-derived carbon components in SOC accumulation.

The composition of plant-derived carbon molecules in the soil can provide important clues regarding the origins and turnover mechanisms of SOC. As an important component of plant residues, lignin is widely regarded as the main component of plant-derived carbon in the soil because of its stable chemical structure, which make it difficult for microorganisms to degrade rapidly (Grünewald et al. 2006; Clemente et al. 2011). An increasing number of studies have shown that lignin is mainly retained in plant residues and soil particulate organic carbon, but accumulates less in soil minerals (Schmidt and Kögel‐ Knabner 2002; Kiem and Kögel-Knabner 2003). In contrast, plant-derived lipid compounds have a higher preservation efficiency in the soil (Li et al. 2018; Man et al. 2023). Plant-derived lipids in the soil mainly include plant wax and suberin derived from leaf cutin and roots. Plant wax is a layer of hydrophobic lipid compounds on the surface of plants, which is the outer barrier between plants and the environment (Wertz 1996). It is mainly composed of long-chain normal alkanes, fatty acids, and fatty alcohols and has a stable chemical structure and large molecular weight (Wertz 1996; Wiesenberg et al. 2008). Cutin and suberin are important barriers to water loss in plants. Cutin is mainly composed of short chain (C14-18) hydroxy-epoxy acids, whereas suberin is mainly composed of long chain (C20-32) hydroxy-acids and fatty acids (Carrington et al. 2012; Angst et al. 2016). Compared to microbial-derived short-chain lipids, plant-derived lipids are not easily decomposed by microorganisms because of their stable chemical structures (Riederer et al. 1993). There is direct evidence that plant-derived lipids contribute more to SOC accumulation than lignins (Dai et al. 2022). Plant-derived lipids demonstrate a greater propensity for adsorption and binding to mineral surfaces and metal oxides (Quenea et al. 2004; Yang et al. 2020b). Therefore, the importance of plant-derived lipids in regulating SOC accumulation may have been underestimated.

Addition of N can affect the input and transformation processes of plant-derived carbon by modifying the biochemical properties of plants and soil. Research has indicated that the concentration of lignin phenols in the soil of subalpine forests decreases following nutrient addition, primarily because of reduced root C input (Luo et al. 2022). Additional research has demonstrated that fungi, as the main decomposers of lignin, can increase the retention of plant lignin in the soil by reducing fungal biomass (Entwistle et al. 2018). In addition, as an important catalyst in the lignin decomposition process, decrease phenolic oxidase activity with N addition inhibited the lignin decomposition process (Chen et al. 2018b). Studies in agricultural systems have demonstrated that fertilisation leads to decreases in easily decomposable short-chain lipids and enhances the preservation of long-chain lipids associated with minerals (Zou et al. 2023). Cutin and suberin were used to characterise of OC sources in plant leaves and roots. Research has shown that N addition may decrease the concentration of suberin from root sources in soil by diminishing the allocation of photosynthetic carbon in roots (Li et al. 2015; Zou et al. 2023). Consequently, alterations in carbon components derived from plants in the soil due to N addition may depend on the equilibrium between plant carbon inputs and processes driven by soil biochemical properties. In addition, the effects of different N addition levels on plant carbon input and soil properties are different, which may lead to differences in the intensity of the changes in plant-derived carbon. Nevertheless, the effects of plant carbon input and soil property changes under different N addition levels on plant-derived carbon components in meadow grasslands remain uncertain, which hinders the ability to accurately forecast changes in SOC in grassland ecosystems under future N enrichment conditions.

In this study, we analysed the effects of N addition on plant-derived organic carbon components (C>20 long-chain free lipids, bound lipids, and lignin-derived phenols) in the soil of a meadow grassland in eastern Inner Mongolia, China. We also examined the changes in plant biomass, diversity, soil chemical properties, and enzyme activity. This study aimed to reveal the primary driving factors and processes of carbon changes in plant sources under N addition. We hypothesised that (1) owing to the differences in the chemical stability of carbon components from different plant sources, lipid substances would respond differently to N addition than lignin phenols. Specifically, we expected that lignin phenol concentrations would increase, whereas lipid concentrations would decrease. (2) The stability of lignin phenols and lipids in soil depends not only on their chemical structure, but also on the mineral stabilization process related to biochemical processes. According to studies in grasslands in northern China, the turnover time of lipids in soil is longer than that of lignin (Jia et al. 2023). Therefore, we hypothesized that lignin phenols are mainly affected by plant biomass input under N addition, whereas lipids are mainly controlled by soil biochemical properties.

2 Materials and methods

2.1 Site description and experimental design

The experimental site was located in the temperate grassland of Hulunbuir City, Inner Mongolia (48°27′–48°35′ N, 119°35′–119°41′ E). The altitude is approximately 765 m. The annual average temperature is -2.4 to 2.2 ℃, with an average annual precipitation of 350–400 mm. Most of the precipitation occurs from June to September. The soil type is dark chestnut. Before the experiment, the basic physical and chemical properties of the 0–20-cm soil layer were determined to be as follows: soil pH 7.1, organic carbon 27.9 g/kg, total nitrogen 1.85 g/kg, and total phosphorus 0.45 g/kg. Stipa baicalensis was a group species, whereas Leymus chinensis was the dominant plant species.

To simulate the grassland ecosystem response to future increases in N deposition, an experimental plot was established in June 2010 using a randomized block design. Six N application levels were set: 0 (CK), 15 (N15), 30 (N30), 50 (N50), 100 (N100), and 150 (N150) kg N/ha/yr, with four replications per treatment. Each plots measured 8 m × 8 m. The N fertilizer (NH4NO3) was split into two equal quantities and applied twice a year in mid-June and mid-July. To minimize the volatilization of fertilizer N, the fertilizer was dissolved in water before application. The solution was sprayed evenly onto the plots using a watering canon, and the same volume of water was sprayed onto all plots, including the control.

2.2 Plant and soil sample collection

Plant and soil samples were collected during the vigorous plant growth stage in August 2022. A vegetation survey was conducted by randomly selecting two 1 m × 1 m quadrats in each plot, and aboveground plant biomass and floor litter quantity were measured. Roots were measured using the soil auger method, and four 0–20-cm soil cores were taken from each plot using an auger with a diameter of 7.5 cm. After collecting soil cores, the roots were collected by washing and dried in an oven at 65 °C until their weight was consistent, and the underground biomass was weighed and calculated.

In August 2022, 10 soil cores were randomly collected from each plot and mixed to form a composite sample. The diameter of the soil drill was 3.5 cm, and the sampling depth was 0–20 cm. The soil samples were stored on ice and returned to the laboratory for treatment. After removing plant roots and other soil intrusions, the soil was sieved through a 2-mm mesh and divided into two portions. A portion of the sample was air-dried and ground to determine the pH, SOC, TN, lipids, and lignin phenols. The remaining fresh samples were stored at –4 ℃ to determine the soil nitrate N, ammonium N content, and enzymatic activity.

2.3 Soil property analysis

SOC and TN were analysed using an elemental analyser (Vario El Cube, German Elemental Company). Before sample analysis, 0.1 M hydrochloric acid was used to remove inorganic carbon, and inorganic N was determined using a flow injection analyser following extraction with 2 M KCl (Li et al. 2021). Soil pH was measured using a pH meter (water-to-soil ratio of 5:1) (Wang et al. 2023). The soil polyphenol oxidase activity was assessed using phloroglucinol colorimetry. The activity of soil α-glucosidase, soil β-glucosidase, and soil β-1,4-glucosidase was measured using a spectrophotometer with one blank and three replicates for each sample. The samples and blank filtrate were analysed for colorimetric absorbance at 400 nm.

2.4 Biomarker analysis

2.4.1 Lignin phenols analysis

The lignin phenols were extracted using an improved alkaline copper oxide oxidation method (Feng and Simpson 2007; Thevenot et al. 2010; Sun et al. 2015). In this method, the ether bond in the lignin macromolecule is broken through a catalysed reaction of copper oxide under alkaline conditions and high temperatures, releasing lignin phenols. Briefly, 0.5 g of the soil sample was weighed into the PTFE liner of the digestion tank, followed by the addition of 0.5 g of CuO, 0.1 g of ferrous ammonium sulfate [Fe(NH4)2 (SO4)2 6H2O], and 10 mg of glucose. Subsequently, 5 mL of argon-displaced 2 M NaOH solution was added, and finally the digestion tank was placed in an oven at 150 °C for 3 h. After cooling, 20 μL of 5.49 mmol L-1 3-ethoxy-4-hydroxybenzaldehyde (ethyl vanillin) was added, and the digestant was filtered with a 0.45 μm filter membrane, placed in a 50 mL centrifuge tube, washed twice with 5 mL H2O, adjusted to pH 2 with diluted hydrochloric acid, and kept in the dark for 1 h. The supernatant was passed through an activated C18 extraction cartridge at a flow rate of 4–5 mL/min after centrifugation. It was then eluted with 2 mL of ethyl acetate, and the eluted sample was dried using N2. Subsequently, 2 mL of a 10% methanol solution was added, and the mixture was analysed using an Agilent 1260 high-performance liquid chromatography (Agilent, USA).

The sum of the vanillyl, syringyl, and cinnamyl phenols was used to determine the total lignin phenols in the soil (Bahri et al. 2006; Dai et al. 2022). The ratios of the acid-to-aldehyde forms of the vanillyl-based ((Ad/Al)V) and syringyl-based ((Ad/Al)s) phenols were used as an index to characterize the degree of lignin phenol degradation (Bahri et al. 2006; Thevenot et al. 2010).

2.4.2 Bound lipids

Residual components of cutin and suberin were determined according to the method described by Mendez-Millan et al. (2010). The characteristics of the alkanoic acid were subsequently subjected to the BSTFA derivatisation reaction, and the resulting derivative was analysed by gas chromatography. In short, approximately 1.0–2.0 g of soil sample was weighed in a tetrafluoroethylene reactor, 5 mL of 1 mol/L methanol sodium hydroxide was added, and the mixture was boiled in water for 3 h. After the hydrolysis solution was cooled to room temperature, the hydrolysis tube was rinsed with a 10 ml mixture of methanol/dichloromethane (1:1) and sonicated for 15 min. The supernatant was acidified with HCl to pH < 1, 15 mL of deionized water was added, and the organic phase was collected in a 5-mL derivative bottle. Nitrogen gas was carefully passed through the sample at 38 ℃ until it was completely dried. We then added 100 μL of pyridine to the dried derivative bottle, followed by 400 μL of BSTFA, ensuring that the back cover was sealed tightly. The mixture was vortexed for 30 s, mixed well, allowed to react at 70 ℃ for 3 h, and measured on the gas chromatograph after cooling to room temperature. In this study, we did not distinguish between cutin and suberin because we did not analyse specific biomarkers in the aboveground litter or underground roots. Therefore, we only present the total quantity of lipids obtained from the alkaline hydrolysis process.

2.4.3 Free lipids

In short, we weighed approximately 0.5–1.0 g of soil sample into a 10-mL centrifuge tube, added 5 mL of a mixture of acetone and dichloromethane (1:1) for ultrasonic extraction for 20 min, and collected the supernatant by centrifugation. This process was repeated twice to combine the supernatants, and nitrogen gas was used to evaporate the solution. We then added 100 uL pyridine and 400 uL BSTFA to the dried sample and standard derivative bottle and covered it tightly. The mixture was vortexed for 30 s, mixed well, allowed to react at 70 °C for 3 h, and then measured using a gas chromatograph after cooling. As short-chain free lipids (< C20) are considered to be primarily derived from microorganisms, whereas long-chain free lipids (> C20) are primarily derived from plants (Feng and Simpson 2011; Dai et al. 2022), we only analysed long-chain free lipids (> C20) from plant sources.

2.5 Statistical analysis

All analyses were performed in R 4.2.3 (R Core Team 2021). Before analysing the data, we conducted normality and homogeneity tests, which showed that all data conformed to a normal distribution and homogeneity of variance. One-way analysis of variance (ANOVA) was used to analyse the changes of plant biomass and diversity, soil chemical properties, and plant-derived carbon components under N addition. Variance partitioning analysis (VPA) was used to analyse the extent to which soil and plant variables explained the changes in plant-derived carbon components (using the vegan package). Secondly, Pearson’s correlation was used to analyse the relationship between plant-derived carbon components and environmental variables (using the corrplot package). Finally, a structural equation model (SEM) was used to analyse the pathways through which N addition affected the carbon components of the plants (using the lavan package).

3 Results

3.1 Plant and soil properties

Long-term N addition significantly changed soil chemical properties, which manifested as a significant decrease in soil pH and C/N and a significant increase in soil inorganic N (ammonium N and nitrate N) content (P < 0.05; Table 1). The effect on plant communities primarily presented as a significant increase in aboveground biomass (13.91–56.87%) and litter biomass (37.73–264.8%) under the high N treatment and a significant decrease in plant species diversity (P < 0.05; Table 1). There were significant differences in the effects of long-term N addition on soil carbon cycling enzymes with different functions. Compared with CK, β-1,4-glucosidase, β-glucosidase, and α-glucosidase activity under high N treatment was significantly reduced, but polyphenol oxidase activity increased significantly (P < 0.05; Table 1).

Table 1 Vegetation and soil characteristics under different nitrogen application levels
Full size table

3.2 Plant-derived carbon components

The responses of carbon components derived from different plant sources to N addition varied in meadow grassland soils. Compared with CK, N100 significantly increased the OC-normalized concentration of free lipids (15.86%) and bound lipids (0.73–18.67%), whereas N150 treatment significantly reduced the OC-normalized concentration of lignin phenols (20.24%) (P < 0.05; Fig. 1a–c). Under N addition, the increased OC-normalized concentration of free lipids was mainly manifested in the increased C22-30 N-Alkanoic acids, while the increase in OC-normalized concentration of bound lipids was mainly exhibited in increased C16-22 ω-Hydrogen alcoholic acids and C16-22 α,ω-alkanedioic acids (P < 0.05; Table 2).

Fig. 1
figure 1

The OC-normalized concentrations (a–c) and proportions (d) of free lipids, bound lipids, and lignin phenols under different nitrogen application levels. Data are mean ± standard error (n = 4). Different lowercase letters represent significant differences among the different levels of nitrogen treatment

Full size image
Table 2 OC-normalized concentration of soil biomarkers under different nitrogen application levels
Full size table

As the N application level increased, the proportion of lignin phenols (1.43–17.77%) in the plant-derived carbon components in soil decreased significantly, while the proportion of free lipids (0.73–10.80%) increased significantly (P < 0.05; Fig. 1d). We used (Ad/Al)V and (Ad/Al)S to characterize the degree of lignin phenol degradation and found that N addition significantly increased (Ad/Al)V and (Ad/Al)S (P < 0.05; Fig. 2a, b), and (Ad/Al)V and (Ad/Al)S were significantly positively correlated with the OC-normalized concentration of free and bound lipids and significantly negatively correlated with the OC-normalized concentration of lignin phenols (P < 0.05; Fig. 2c and d). The addition of N promoted lignin phenol degradation.

Fig. 2
figure 2

Soil lignin phenols (Ad/Al)V (a) and (Ad/A)S (b) and their relationship between free lipids, bound lipids, and lignin phenols OC-normalized concentrations and lignin phenols (Ad/Al)V (c) and (Ad/Al)S (d) under different nitrogen application levels. Data are mean ± standard error (n = 4). Different lowercase letters represent significant differences among different levels of nitrogen treatments. Linear regression lines with 95% confidence intervals reflect the predictive effects of fixed factors. *P < 0.05, **P < 0.01

Full size image

3.3 Driving factors of carbon components from plant sources

The VPA analysis showed that although there were significant changes in plant biomass and diversity, the changes in plant-derived carbon components under N addition were more strongly regulated by soil properties (P < 0.05; Fig. 3a–c). The OC-normalized concentrations of free and bound lipids were significantly negatively correlated with SOC, TN, C/N, and β-1,4-glucosidase. In addition, the concentration of bound lipids showed a significant positive correlation with the aboveground biomass and litter biomass of plants (P < 0.05; Fig. 3d). The OC-normalized concentration of lignin phenols was significantly positively correlated with soil C/N, pH, β-1,4-glucosidase, and β-glucosidase, while it was significantly negatively correlated with inorganic N and polyphenol oxidase (P < 0.05; Fig. 3d).

Fig. 3
figure 3

Variance partitioning analysis (a-c) and Pearson correlation analysis (d) of vegetation and soil characteristics with soil free lipids, bound lipids, and lignin phenols under different nitrogen application levels. SOC, soil organic carbon; TN, total nitrogen; C/N, carbon to nitrogen ratio; AGB, aboveground biomass; LB, litter biomass; RB, root biomass; PSD, plant species diversity; PPO, phenoloxidase

Full size image

The SEM analysis indicated that the increased polyphenol oxidase activity and decreased C/N ratio under N addition had a significant direct effect on the OC-normalized concentration of lignin phenols. Additionally, the pH reduced the accumulation of lignin phenols by promoting their degradation (P < 0.05; Fig. 4a). The addition of N was a consistent factor in the driving process of free and bound lipids, with N addition mainly increasing the OC-normalized concentrations of free and bound lipids by reducing soil pH (P < 0.05; Fig. 4b, c).

Fig. 4
figure 4

The structural equation model representing the pathways through which nitrogen addition affects soil free lipids (a), bound lipids (b), and lignin phenols (c). The blue and red arrows represent positive and negative relationships, respectively. A line’s thickness represents the strength of the relationship, and the numbers adjacent to the arrows represent the normalized path coefficients. Solid lines indicate a significant relationship at P < 0.05, while dotted lines indicate non-significant relationships

Full size image

4 Discussion

4.1 Source of lipids in soil

Plant-derived free lipids extracted using solvents are primarily composed of long-chain (> C20) n-alkanes, alkanols, and alkanoic acids (Feng and Simpson 2007; Rushdi et al. 2016). Our findings indicated that the OC-normalized concentration of n-alkanoic acids in soil free lipids was the highest, followed by that of n-alkane and alkanol compounds (Table 2). Previous studies indicated that n-alkanoic acids and alkanols are among the primary organic compounds found in forest and grassland soils (Otto and Simpson 2006; Rushdi et al. 2016). Free long-chain (> C20) n-alkanes, alkanols, and alkanoic acids are predominantly derived from plant wax components, and their levels are associated with the biochemical degradation of plant wax (Wertz 1996; Otto and Simpson 2007).

Cutin and suberin derived from alkaline hydrolysis have been extensively utilised to characterise carbon sources from plant leaves and roots (Winkler et al. 2005; Zhu et al. 2019). Nevertheless, specific biomarkers for these compounds in meadow grassland soils were not identified in this study. The bound lipids in the experimental soil mainly include ω-hydrogen alcoholic acids, α,ω-alkanedioic acids, and mid-chain substituted hydroxide alcoholic acids (Table 2). Previous studies have indicated that the composition of ω-hydroxyl alkanoic acids in soil is consistent with the composition of suberin in plant roots and bark (Riederer et al. 1993; Bernards 2002). Therefore, ω-hydroxy alkanoic acids in soil may originate from the hydrolysis products of suberin. C12-30 alkanedioic acids are mostly derived from lignin hydrolysis products (Bernards 2002; Otto and Simpson 2006; Carrington et al. 2012). Mid-chain substituted hydroxy alkanoic acids are mainly abundant in forest litter and are low in herbaceous plants (Otto and Simpson 2006), and this study further validates the finding that the proportion of these acids in grassland soil bound lipids is the lowest among the compounds discussed above (Table 2). Furthermore, mid-chain substituted hydroxy alkanoic acids mainly originate from the hydrolysis products of plant cutin and suberin (Carrington et al. 2012; Angst et al. 2016).

4.2 The effect of N addition on soil lipids

Lipid changes in soils are closely related to the biochemical degradation of plant residues. Our study findings indicated that N addition significantly elevated the OC-normalization concentration of free and bound lipids (Fig. 1a and b). Currently, there is some controversy regarding the factors that regulate lipid concentrations in soil. Several studies have indicated that the soil lipid composition is mainly influenced by the lipid composition of plant leaves and roots, with consistent patterns in their composition in plants and soil are consistent (Santana et al. 2015; Rushdi et al. 2016). In Canadian grasslands, the abundance of alkanoic acids and alkanols in the leaves of aboveground herbaceous plants leads to the enrichment of the corresponding organic compounds in the soil (Otto and Simpson 2006).

Studies have also demonstrated that variations in the chemical stability and functional group activity of lipids with different chain lengths significantly affect the composition of soil lipids, with microbial decomposition and mineral protection being important contributing factors (Wiesenberg et al. 2008; Li et al. 2018; Yang et al. 2020b). For example, high concentrations of plant-derived lipids have been found in mineral components, mainly because of the strong binding ability of long-chain fatty acids with minerals and metal oxides, which selectively preserves these acids in the soil (Mueller et al. 2013). Compared to changes in vegetation characteristics, we found that soil properties (chemical properties and enzyme activity) play a more significant role in explaining variations in free and bound lipids following N addition (Fig. 3a–c). This indicates that changes in plant-derived lipids in the soil under N addition were mainly driven by soil environmental factors.

The SEM results showed that a decrease in pH under N addition was an important factor causing an increase in free and bound lipids (Fig. 4b and c). Plant-derived lipids play a significant role in the accumulation of SOC in forest ecosystems. The preservation of plant-derived lipids is promoted under low-pH conditions because of the decrease in hydrolytic enzyme activity and the increase in iron and aluminium oxides (Dai et al. 2022). Research on soil lipid analysis across various pH gradients has demonstrated that microbial-derived short-chain lipids significantly decrease with decreasing soil pH, whereas plant-derived long-chain lipids significantly increase (Nierop et al. 2005). Currently, a large body of research has confirmed that the hydrophobicity of lipids and the reactivity of multivalent cations give them a higher affinity for minerals and their complexes, leading to lipid enrichment on mineral surfaces (Li et al. 2018; Zhu et al. 2019; Yang et al. 2020b). These results indicate that a low pH enhances the preservation of plant-derived lipids in soil by reducing the biodegradation of plant-derived lipids and promoting the stability of lipids on mineral surfaces.

Although the increase in plant carbon input under N addition benefited the accumulation of plant-derived lipids in the soil, we believe that changes in vegetation community composition and the inhibition of litter decomposition under N addition may offset the positive effects of increased plant biomass. Analysis of lipid abundance in dominant grass species in temperate grasslands showed that the addition of N is beneficial for the growth of grass species with a low lipid composition (such as Leymus chinensis), whereas it inhibits the growth of lipid-rich grass species (Ma et al. 2019). In addition, soil acidification and nutrient imbalance caused by N enrichment can inhibit microbial activity, thereby slowing litter decomposition and reducing the efficiency of litter carbon decomposition and transformation (Zhang et al. 2018; Peng et al. 2022). In summary, intensified competition among grass species with low lipid abundance and a reduced degradation rate of plant residues under N addition may not significantly impact of change plant biomass and the consequent effects on the concentration of plant-derived lipids in the soil.

4.3 The effect of N addition on lignin phenols

Lignin has traditionally been recognized as an important substrate for SOC formation. However, recent studies have shown that lignin undergoes faster turnover in soil than was previously considered (Dignac et al. 2005; Heim and Schmidt 2007). Warming experiments have validated the accelerated degradation of lignin compared to that of lipids (Chen et al. 2020), indicating that its stability in soil is not universal. Although the distribution characteristics of lignin in soil cannot be directly observed, the alkaline copper oxidation method can be used to analyse the characteristics of lignin-derived phenols in soil. Our findings indicate that vanillyls constitute the predominant portion of lignin-derived phenols (Table 2), which can be attributed to its stable chemical properties and the longer turnover times of vanillyl-based monomers (Hedges et al. 1988).

Owing to the complex chemical composition of lignin, only a limited number of microorganisms and specific extracellular enzymes in the soil can break it down (Derenne and Largeau 2001; Bahri et al. 2006), and its degradation is affected by various environmental factors, including soil pH, temperature, and biological activity. Our findings indicated that N addition promoted the degradation of lignin phenols and reduced their accumulation (Figs. 1c and 2a and b). However, the effect concerning the impact of soil N availability on lignin degradation remains controversial. Several studies have indicated that N enrichment inhibits lignin degradation by reducing the activity of lignin-degrading fungi and extracellular enzymes (Entwistle et al. 2018; Chen et al. 2018b). Other studies have indicated that N addition promotes lignin degradation by Basidiomycota and Ascomycota (Song et al. 2022). Lignin-derived phenols are mainly stabilized in the soil for an extended period through physical and chemical adsorption to minerals, rather than relying on their chemical stability (Clemente and Simpson 2013; Yang et al. 2020b). Studies have demonstrated that lignin-derived compounds can stably bind to iron minerals (Liao et al. 2022). A mineral cultivation experiment confirmed the fixation effect of minerals on lignin-derived phenols (Li et al. 2019). Changes in biochemical properties under N addition appear to promote lignin decomposition. The decrease in lignin phenols we found in the soil was primarily associated with a decline in the C/N ratio and an increase in phenolic oxidase activity (Figs. 3d and 4a). The reduction of lignin phenols under N addition may not be related to mineral protection, but rather to changes in plant residues and specific biochemical decomposition processes. Although previous studies have shown that a decrease in pH can inhibit the activity of phenolic oxidases (Chen et al. 2018a), plant residue input and microbial activity also have significant impacts on phenolic oxidases. This is because phenolic oxidases primarily originate from plant and microbial residues released during decomposition processes (Burns et al. 2013). Therefore, we believe that N enrichment promotes the degradation of lignin phenols in soil via the following two pathways. On the one hand, the increased input of plant residues stimulates the synthesis of phenolic oxidase, leading to lignin degradation. On the other hand, a decrease in the C/N ratio between plants and soil under N enrichment conditions can lead to microbial carbon limitation, thereby accelerating the decomposition of plant-derived carbon to acquire additional energy. Research has indicated that N addition stimulates the accumulation of microbial residues (Ma et al. 2021, 2023), indicating an increase in microbial assimilation.

Although lipids, lignin phenols, and their derivatives are protected by minerals, their specific protective mechanisms may differ. We observed varying responses in lignin phenols and lipids to N enrichment. A recent study found that plant C stabilized by iron minerals mainly originates from small-molecule carbon degraded by lignin (Liao et al. 2022). Lipids are stabilized in soil through direct adsorption between their active functional groups (hydroxyl/carboxyl) and mineral surfaces (Li et al. 2018; Yang et al. 2020b). The next step should be to combine molecular microbiology and chemical characterization techniques to further clarify the specific biochemical mechanisms underlying the different responses of lignin phenols and lipids to N addition.

5 Conclusion

The study findings indicate that long-term N addition alters the molecular composition of plant-derived carbon in the soil. N addition promoted the biodegradation of lignin-derived phenols by reducing the soil C/N ratio and increasing phenolic oxidase activity. In addition, soil pH reduction promotes the retention of lipid compounds, which are closely associated with mineral protection pathways. In general, lipids and lignin phenols exhibited different responses to N addition, which reflects the complex processes and pathways the response of the soil carbon pool to N addition. Future research should evaluate the impact of changes in plant-derived components on SOC function and stability following N addition.