Introduction

The accelerating pace of urbanization and industrialization has led to a substantial increase in the use of nitrogen fertilizers in agriculture and animal husbandry, culminating in the pressing issue of atmospheric nitrogen deposition1,2. Since the 1980s, China’s rapid expansion in agriculture, industry, and urbanization has spawned a range of environmental challenges3, positioning the nation among those facing the most severe nitrogen deposition issues globally4,5. Between 2010 and 2020, the nitrogen deposition rate in China decreased by 14%6. The model predicts that atmospheric nitrogen deposition will drop by nearly 50% by 20507. Nonetheless, the challenge of nitrogen deposition remains a persistent global and national issue in China.

Soil nutrients are essential for maintaining vegetation growth. A rigorous scientific assessment of soil nutrients, which are crucial for sustaining vegetation growth, is imperative8. Soil extracellular enzymes are proteins released into the soil environment by microorganisms (bacteria and fungi), root systems and soil animals. They break down complex, large organic compounds (polymers) into simpler, smaller molecules (monomers) that can be absorbed by living organisms. We generally classify the most relevant nutritional cycles of enzymes into carbon (C), nitrogen (N), and phosphorous (P) cycle enzymes. Soil ecoenzymatic stoichiometry (EES) is the study of the ratios of the activities of key nutrient-acquiring extracellular enzymes in the soil. Soil ecological stoichiometry which involves the study of elemental ratios (C:N:P) that govern microbial growth and nutrient limitations is a valuable method for evaluating soil nutrient status, serving as a reliable indicator9,10and facilitating the quantification of soil nutrient limitations11. Amidst nitrogen deposition, various factors influence forest soil nutrients, which are categorized as abiotic and biotic. Abiotic effects primarily involve environmental factors, such as temperature, moisture, pH, and soil texture12. Enhanced atmospheric nitrogen deposition typically induces a progressive decline in soil pH, which may negatively impact soil microbial communities. Conversely, biotic effects encompass responses from underground vegetation roots, soil animals, and soil microorganisms13, with soil microorganisms prominently shaping this intricate interplay9. Nitrogen enrichment alters nutrient availability, accelerates soil acidification3,4,14, and impacts soil microbial community structure, abundance, and activity7,12,15, thereby influencing the dynamic equilibrium of soil nutrients. The role of soil microbial communities, including fungal and bacterial groups, in mediating nutrient limitation and maintaining the dynamic equilibrium of soil nutrients warrants further investigation.

Advancements in molecular biology have afforded scientists a novel perspective for comprehending the diverse microbial resources present in soil and their indispensable functions in ecosystems16. Soil microorganisms serve as crucial contributors to underground matter cycling and energy flow processes, dynamically regulating their biomass and directly or indirectly affecting soil nutrients and components. Additionally, they contribute to soil nutrient turnover through the secretion of extracellular enzymes17,18. However, our understanding of forest soil microbial communities remains limited. State-of-the-art high-throughput sequencing methods have unveiled specific microbial populations that play pivotal roles in essential carbon and nitrogen cycling processes, including lignin degradation, nitrogen fixation, and nitrification. This finding significantly enhances our understanding of the functions and impacts of microbial communities at various scales. Despite the influence of atmospheric nitrogen deposition, the ecological niches of actinomycetes, bacteria, archaea, fungi, and other microbial groups, along with their classified species, undergo changes. Moreover, their symbiotic, competitive, and other relationships vary. The mechanisms by which the nutrient utilization strategies of microorganisms adapt to atmospheric nitrogen deposition and changes in soil nutrient status remain unclear.

Pinus taiwanensis, an endemic species in southeastern China, is characterized by its preference for light and deep-rooted habits. This species is primarily distributed across the provinces of Anhui, Zhejiang, Jiangxi, and Fujian provinces, as well as Taiwan, China. Flourishing in subtropical humid and infertile areas, this species can form pure forests, thereby fostering soil stability and contributing to ecological restoration19. Consequently, it is a key species for ecological restoration in the medium- and high-altitude mountainous regions of China’s subtropics. Given the prevailing " nitrogen-rich and phosphorous -poor" conditions in subtropical forest soils and the escalating nitrogen deposition resulting from economic growth and urbanization, leading to pronounced soil nutrient limitations, the following issues are discussed in depth: (1) How soil microbial community structure responds to nitrogen deposition and nutrient restriction. (2) Which microbiota emerge as key responders to nitrogen enrichment, and how do they show reactivity? We hypothesized that nitrogen enrichment in subtropical forest soils would alter the microbial community structure and increase phosphorous limitation, as indicated by enzyme stoichiometry and microbial biomass ratios. Nitrogen is frequently the primary limiting nutrient in terrestrial ecosystems. The direct addition of nitrogen mitigates this limitation, resulting in a marked increase in plant productivity, as evidenced by enhanced biomass accumulation and increased litterfall. This accelerated growth intensifies the demand for other essential nutrients, particularly phosphorous , which serves as a critical constituent of nucleic acids (DNA and RNA), phospholipids, and energy transfer molecules such as Adenosine Triphosphate (ATP). The nitrogen-to-phosphorous (N:P) ratio in plant tissues is widely recognized as a reliable indicator of nutrient limitation. The central hypothesis posits that nitrogen enrichment will lead to an increase in the soil and foliar N:P ratio. We further predict that specific microbial taxa will emerge as indicators of nitrogen-induced nutrient imbalances, with characteristic responses to N addition. To address these questions, we established a 3-year field experiment in a subtropical Pinus taiwanensis forest with gradient nitrogen enrichment and evaluated changes in soil nutrients, microbial biomass, enzyme activity, and microbial community composition.

Materials and methods

Site description and sampling

The experimental site is situated within the Jiuxian Mountain Scenic Area, encompassed by the Daiyun Mountain National Nature Reserve (25°38′07’' to 25°43′40’' N, 118°05′22’' to 118°20′15’' E) in Fujian Province, China. The climatic classification is subtropical oceanic monsoon, characterized by an average annual temperature of approximately 15.6–19.5 °C and an average annual precipitation of 1700–2000 mm. The region is distinguished by pure stands of Taiwan pine (Pinus taiwanensis), accompanied by flourishing understory vegetation comprising bamboo, evergreen shrubs (Eurya japonica), camellias, and Paeonia delavayi. The soil type is categorized as Ultisol according to the United States Soil Taxonomy. The experimental site was established in 2017, and nitrogen fertilization commenced in March 2018. Based on the N deposition in the Fujian coastal area (30–40 kg N ha− 1 yr− 1 )15, three N enrichment levels were established in the experiment: control (CT,0 kg N ha-1 year-1), low nitrogen (LN, 40 kg N ha-1 year-1), and high nitrogen (HN, 80 kg N ha-1 year-1), each with four replicates. Each plot is 15 m × 15 m. Urea (CO(NH2)2) was used as the N source. Annual sampling was conducted in May, except for 2020, when sampling was conducted in July because of the pandemic. Subsequent sampling years was conducted in the same plot in May 2019 and May 2021.

Soil sample collection and processing: Within the 12 experimental plots, five sampling points were systematically selected for collection. Using an S-shaped sampling method, a 5-cm diameter soil drill was utilized to extract soil samples at a depth of 20 cm. First, the surface litter was removed, followed by the collection of topsoil (horizon A, approximately 0–10 cm) and subsoil (horizon B, approximately 10–20 cm), both of which were carefully placed in labeled zipper bags. The sample numbers were classified as CT_A, LN_A, and HN_A (A horizon); and CT_B, LN_B, and HN_B (B horizon). The soil samples were processed using the three-point method, where in each treatment sample was subdivided into three portions, each of which was stored in a labeled sample bag. One portion was allocated for nucleic acid extraction for high-throughput sequencing analysis (16 s for bacteria and internal transcribed spacer [ITS] region for fungal analysis). It is sufficient to state that the roots and debris were removed, followed by storing the samples at -20 °C. Concurrently, another portion was designated for the assessment of soil extracellular enzyme activity and soil availability index measurements, which required storage at 4 °C. The third portion of the soil sample was collected, and debris was removed, air-dryed, and passed through a 2-mm soil sieve for subsequent soil property determination.

Soil parameters and microbial biomass

Following the methodology outlined in Carter & Gregorich,20, soil organic carbon (SOC) and total nitrogen (TN) were determined after 100 mg of soil using a carbon–nitrogen element analyzer (Elementar Vario MAX, Germany). Dissolved organic carbon (DOC) was measured using a total organic carbon analyzer (Shimadzu, Japan), and dissolved organic nitrogen (DON) was quantified using a continuous flow analyzer (SKALAR San +  + , Netherlands). Fresh soil extraction for analysis was performed using deionized water. Soil pH measurements were performed using a PHS-3B pH meter, following the national forestry industry standard. To determine the soil moisture content, 10 g of fresh soil was weighed after drying. Mineral nitrogen was extracted using KCl following Liu et al.21, with subsequent analysis of NO3- and NH4+ conducted using a continuous flow analyzer (SKALAR San +  + , Netherlands). Soil phosphorous content was assessed utilizing the M3 extraction method20. The total phosphorous content (TP) was estimated using the H2SO4-HClO4 double acid digestion method22.

Soil microbial biomass carbon (MBC), microbial biomass nitrogen (MBN) and microbial biomass phosphorous (MBP)were measured using the chloroform fumigation extraction method. The correction factor of MBC was 0.3823, while that of MBN was 0.45 and that of MBP was 0.424.

High-throughput sequencing of soil microbes and enzyme activity assays

Following the ecological enzyme stoichiometry method outlined by Sinsabaugh25, we measured the activities of five hydrolytic enzymes: β-glucosidase (BG), cellulose hydrolase (CBH), β-N-acetylglucosaminidase (NAG), leucine aminopeptidase (LAP), and acidic phosphatase (AcP).

A 1 g sample of fresh soil was accurately weighed and transferred to a brown glass bottle . Acetate buffer solution was added and stirred for 20 min using a magnetic stirrer to achieve homogenization. After allowing the mixture to stand undisturbed, the supernatant was carefully removed. Using an 8-channel pipette, transfer the clear liquid was transferred into a 96-well microplate and the corresponding substrate was added (Table S1). The 96-well microplates containing the hydrolases BG, CBH, NAG, LAP, and AcP were incubated in a controlled-temperature incubator at 20 °C for 4 h. Prior to measurement, 10 µL of 1 M NaOH solution was added to each well to terminate the enzymatic reaction. Subsequently, hydrolase activity was measured using a microplate reader (Synergy H4; USA).

Enzyme stoichiometric ratio: ln(BG + CBH): ln(NAG + LAP): ln(AP)25 Vector analysis based on ecological enzyme stoichiometry was conducted by calculating vector length (L) and angle (A), which reflect resource limitations11.

$$VectorL = \sqrt {\left( {\ln \left[ {BG + CBH} \right]/\ln \left[ {NAG + LAP} \right]} \right)^{2} + \left( {\ln \left[ {BG + CBH} \right]/\ln AP} \right)^{2} }$$
(1)
$$VectorA = Degrees(ATAN2((\ln \left[ {BG + CBH} \right]/\ln AP,)(\ln \left[ {BG + CBH} \right]/\ln \left[ {NAG + LAP} \right])))$$
(2)

A relatively longer vector length (L) indicates a greater carbon (C) limitation. Vector angles (A) less than 45° and greater than 45° indicate nitrogen (N) and phosphorous (P) limitation, respectively.

Threshold Elemental Ratios (TER) were calculated as C:N and C:P ratios.

$$TER_{C:N} = (A_{N} /GE)B_{C:N}$$
(3)
$$TER_{C:P} = (AP/GE)B_{C:P}$$
(4)

In this equation, AP and AN represent the assimilation efficiencies of phosphorous and nitrogen, respectively. Both were assumed to be 0.98. GE (microbial growth efficiency) is set at 0.2925, whereas BC:N and BC:P denote the microbial biomass stoichiometric ratios of C:N and C:P, respectively.

Soil DNA was extracted from 250 mg soil samples using the PowerSoil DNA Isolation Kit (MoBio Laboratories, Carlsbad, CA, USA) following the manufacturer’s instructions. Genomic DNA purity and quality of the genomic DNA were evaluated using 0.8% agarose gel electrophoresis. Quantitative PCR (qPCR) was employed to determine the relative abundance of bacterial 16S rDNA genes and fungal ITS. For each soil sample, 10-base barcode sequences were appended to the 5’ ends of the forward and reverse primers (provided by Beijing Allwe Gene Company) and analyzed by the company. The primers 338F (5′-GTACTCCTACGGGAGGCAGCA-3′) and 806R (5′-GTGGACTACHVGGGT WTCTAAT-3′) were designed to amplify the V3–V4 hypervariable regions of the bacterial 16S rRNA gene. Similarly, the primers ITS1F (5′-CTTGGTCATTTAGAGGAAGTAA-3′) and ITS2R (5′-TGCGTTCTTCATCGATGC-3′) were designed for amplifying the ITS1 region of the fungal ITS rRNA gene. For each soil sample, a 10-digit barcode sequence was incorporated at the 5ʹ end of the forward and reverse primers (Allwegene Company, Beijing, China). QPCR was conducted on a Mastercycler Gradient (Eppendorf, Hamburg, Germany) using 25 μl reaction volumes, comprising 12.5 μl 2 × Taq PCR MasterMix, 3 μl bovine serum albumin (2 ng μl-1), 2 μl primers (5 μM), 2 μl template DNA, and 5.5 μl ddH2O. The cycling parameters included an initial denaturation at 94 °C for 5 min, followed by 30 cycles at 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s, concluding with a final extension at 72 °C for 10 min. Three PCR products per sample were pooled to mitigate PCR biases. Subsequently, the PCR products were purified using a QIAquick Gel Extraction Kit (QIAGEN, Germany) and quantified via real-time PCR. The fungal-to-bacterial (F/B) ratio was calculated using 16S rDNA and ITS gene copy numbers. Microbial 16S DNA and ITS analyses were conducted on the MiSeq platform at Beijing Allwegene Company.

Data analysis

Before conducting the variance analysis, the normality of the variances and residual plots were examined, and the assumption of sphericity was assessed. One-way analysis of variance (ANOVA) was used to investigate the impact of nitrogen application on soil nutrient factors, microbial enzyme activity, biomass, and microbial community composition. A repeated-measures ANOVA was used to assess variations in soil physicochemical properties, soil nutrient stoichiometry, microbial enzyme activity, and soil ecological stoichiometry characteristics following nitrogen application.

To analyze the nutrient limitation status, the data were log-transformed and subjected to regression analysis, adhering to the principles of stoichiometric analysis and variance standardization26. Redundancy analysis (RDA) was conducted to investigate the primary environmental factors influencing microbial community structure, soil stoichiometry, and enzyme stoichiometry in response to nitrogen application.

Based on the normalized relative abundance matrix, the LEfSe method was employed to discern potential differential species. LEfSe was applied to identify taxa significantly enriched in each treatment (CT, LN, and HN) across all years. The Kruskal–Wallis test was conducted to identify significant differences across various sampling sites, and the effect size of each feature was assessed using LDA. Additionally, stepwise regression and correlation analyses (paired-tailed) were performed to elucidate the correlation between microbial communities and vector length, angle, and the threshold element ratio (TER). Structural equation modeling (performed using SPSS AMOS 21.0 software) was employed to examine the pathway relationships between nitrogen application, soil ecological stoichiometry, and microbial communities. The data were tidied using Excel 2019, and preliminary statistical analysis was performed using SPSS 21.0. Graphical representations were generated using Origin 2018 and R language 3.6.5 version.

Result

Effect of nitrogen enrichment on soil properties and microbial biomass

Nitrogen enrichment significantly impacted various soil parameters, including phosphorous, nutrient stoichiometry carbon-phosphorous ratio (C:P) and nitrogen-phosphorous ratio (N:P), available nitrogen (AN), available carbon–nitrogen ratio (AC:N), and available nitrogen-phosphorous ratio (AN:P). The effects were observed across different soil layers following nitrogen application, and notable changes in the time were observed effect except for TN (Table 1). The interaction effect of nitrogen application and soil layer significantly influenced TN, C:P, N:P, AC:P, and AN:P ratios (Table 1). Additionally, the interaction effect of nitrogen application and time significantly affected TC, TN, C:P, N:P, DOC, AN, AC:N, and AN:P. Furthermore, the interaction between nitrogen application, time, and soil layer affected TC, TN, C:P, N:P, AN, AC:P, and AN:P. After three years, nitrogen application significantly affected AN (P < 0.05, Table S2). The effects of soil layers A and B were highly significant (P < 0.001; Table S2). The interaction between nitrogen application and soil layer significantly increased the total nitrogen (TN) content (P < 0.05). Additionally, the C:P and N:P ratios increased significantly (P < 0.001), whereas the AC:P and AN:P ratios exhibited a significant increase (P < 0.01).

Table 1 Effectofnitrogenenrichmentandhorizononsoilchemicalpropertiesbyrepeatedmeasures AVONA.
Full size table

Nitrogen application significantly impacted the quantities of MBC, MBN, and microbial biomass phosphorous (MBP), as well as the MBN:MBP ratio (Table S3). The independent effects of time and soil layer significantly impacted soil microbial biomass. The interaction between nitrogen application and soil layer played a significant role in shaping the MBC, MBC:MBP, and MBN:MBP ratios. Similarly, the interaction effect between nitrogen application and time significantly changed the microbial biomass and its ratios, except for MBN. Repeated measurements involving nitrogen application, time, and soil layer collectively altered the MBC, MBP, MBC: MBP, and MBN: MBP ratios.

Effect of nitrogen enrichment on soil enzyme activities and stoichiometry

Repeated-measures ANOVA showed that the interaction effect of time, nitrogen application, and soil horizon caused significant changes in carbon cycling enzymes (P < 0.01), nitrogen cycling enzymes (P < 0.01) and AcP enzymes (P < 0.05 ) (Table S2).

The analysis of the impact of nitrogen application on soil enzyme stoichiometry and nutrient limitation (Table 2) revealed significant alterations in EC/P, EN/P, LnP-enzymes, LnC:P, LnN:P, Vector A, TERC:N, and TERC:P. The interaction effect between nitrogen application and soil layer yielded notable changes in EN/P, LnN-enzymes, LnP-enzymes, LnN:P, Vector A, and TERC:P. Moreover, the cumulative effects of nitrogen application and time impacted EC/P, EN/P, LnC-enzymes, LnP-enzymes, LnC:N, LnC:P, Vector L, TERC:N, and TERC:P. Furthermore, the interaction effect involving time, nitrogen application, and soil horizon significantly influenced all factors except EC/P.

Table 2 EffectsofNadditiononsoilmicrobialenzymesstoichiometrybyAVONA
Full size table

Effect of nitrogen enrichment on soil microbial community composition

The soil bacterial and fungal communities of 72 samples collected between 2019 and 2021 were analyzed. After filtering and eliminating chimeras, 16S rDNA gene amplicon MiSeq sequencing and MiSeq ITS sequencing retained 3,717,219 and 4,838,947 high-quality sequences, respectively (with an average of 51,787 bacterial and 67,207 fungal sequences per sample). These sequences yielded 7,594 bacterial Operational Taxonomic Units (OTUs) and 7,372 fungal OTUs. Following rarefaction, the final OTU counts were 7,329 for bacteria and 6,806 for fungi. To validate the sequencing data volume, a dilution curve was constructed for the samples Good’s coverage, ranged from 95.13% to 97.25% for bacteria and from 96.78% to 98.99% for fungi (FigS1).

Figure 1A illustrates the relative abundance of dominant bacterial and fungal communities in different soil layers following nitrogen application. Notably, Proteobacteria, Acidobacteria, Actinobacteria, Verrucomicrobia, Planctomycetota, and Chloroflexi stand out as the predominant bacterial groups, collectively constituting more than 80% of the total bacterial richness in subtropical forest soils. This finding aligns with previous experimental results27,28. Nitrogen application decreases the relative abundance of oligotrophic groups, including Acidobacteria, Verrucomicrobia, Nitrospira, and Chloroflex, while concurrently increasing the prevalence of eutrophic groups such as Proteobacteria, Actinobacteria, and Firmicutes29,30. ITS sequencing analysis indicates notable modifications in the composition of the soil fungal community in response to nitrogen application. Within the subtropical forest soil of Daiyun Mountain, Basidiomycota and Ascomycota emerge as dominant fungal groups (Fig. 1B).

Fig. 1
figure 1

Relative abundance of dominant phylum of soil bacterial (A) and fungal (B) community in different horizon.

Full size image

Relationship among soil microbial community composition and environmental factors after nitrogen enrichment

Based on the OTU level, a two-dimensional RDA was conducted to analyze the results (Fig. 2). The analysis revealed that the first axis of the RDA accounted for 39.62% of the variation, with the second axis contributing 2.29% (Fig. 2A). Among the various parameters, TC and AP emerged as the most influential factors affecting the composition of soil bacterial communities. Their relationship closely aligns with the first RDA axis (Fig. 2A; Table S4). Additionally, the contents of soil MBC and MBP significantly impact the composition of soil bacterial communities (P < 0.01). RDA results showed that the first and second axes explained 8.85% and 6.90% of the changes, respectively (Fig. 2B). The analysis underscored the pivotal role of soil AN and Total TN concentrations, MBC:MBN, and MBC:MBP as the primary factors contributing to the variation in fungal community composition (P < 0.01).

Fig. 2
figure 2

Redundancy analysis of nitrogen application on soil bacterial and fungal community composition.

Full size image

A significant positive correlation exists between enzyme activity and its ratio (BG, NAG, LAP, AcP, LnC:P, and LnN:P) in the dominant soil bacteria (P < 0.01) (Fig.S2A). Conversely, the relative abundance of Chloroflexi exhibit a negative correlation with BG, NAG, AcP, LnC:N, LnC:P, and LnN:P (P < 0.05). Acidobacteria have a significant negative impact on NAG enzyme activity, while Actinobacteria demonstrate a significant positive correlation with EC/P (P < 0.05). Additionally, Verrucomicrobiota is positively correlated with AcP (P < 0.05), whereas Planctomycetota has a significant negative effect on EC/P (P < 0.05).

Within the dominant soil fungi, Basidiomycota and Ascomycota exhibit significant associations with all enzyme activities and their respective ratios, excluding EC:N. This relationship is negatively correlated with Basidiomycota and positively correlated with Ascomycota. Mortierellomycota has a significant negative effect on BG, NAG, and AcP. Conversely, Rozellomycota is positively correlated with CBH, EC:N, LnC:N, and LnC:P, while displaying a negative correlation with EN:P. Moroever, unidentified is positively correlated with EN:P (P < 0.05).

Effect of nitrogen enrichment on soil microbial nutrients limitations

Threshold element analysis was employed to elucidate the impact of nitrogen application on the relative nutrient limitation status of microbes (Fig. 3). Nitrogen application significantly affected TERC/N and TERC/P across various soil horizons and years (P < 0.05). The average effect following nitrogen application displayed a significant downward trend (P < 0.05). Specifically, nitrogen application induced a significant alteration in vector length L only during the second year of HN treatment in soil horizon A. Additionally, in the soil horizon B layer, vector length L significantly increased in the first year under both LN and HN treatments. The orientation of vector angle A, exceeding 45°, indicated a strong phosphorous limitation for soil microorganisms (Fig. 4).

Fig. 3
figure 3

Effects of nitrogen application on soil microbial element threshold ratio TERC:X in different horizons.

Full size image
Fig. 4
figure 4

Effect of nitrogen enrichment on vector length and vector angle of soil microbial community.

Full size image

This study employed three distinct methods based on the threshold element ratio-based indicator concept to assess limitations: (1) TERC:N and TERC:P represent threshold ratios correlating the metabolic status of extracellular enzyme counts with resource chemical counts. RC:N and RC:P denote the molar ratios of the soil C:N and C:P, respectively. A negative value in RC:N-TERC:N or RC:P-TERC:P indicates no constraint by nitrogen or phosphorous, while a positive value suggests nitrogen or phosphorous limitations26. (2) Soil enzyme stoichiometry was utilized to assess potential nitrogen and phosphorous limitations in the soil25. Four distinct microbial resource limitations (nitrogen limitation, phosphorous limitation, carbon and phosphorous co-limitation, and nitrogen and phosphorous co-limitation) were classified using 1 as the horizontal and vertical baselines along the enzyme activity ratio axis ([NAG + LAP]/AcP as the x-axis, [BG + CBH]/[NAG + LAP] as the y-axis). (3) The vector length (L) and vector angle (A) changes were computed based on the ratios of (BG + CBH), (NAG + LAP), and AP after log transformation31. A larger vector length (L) signifies greater carbon limitation, vector angles less than 45° indicate nitrogen limitation, and angles greater than 45° suggest phosphorous limitation. Throughout the three-year nitrogen application period, the consistent trend observed was that soil microorganisms were limited by phosphorous (Fig. 5A). The indicators derived from the soil enzyme activity stoichiometry analysis further supported the notion that microorganisms experienced phosphorous limitation throughout the decomposition process (Fig. 5B).

Fig. 5
figure 5

Ecological indicators demonstrate stoichiometric constraints of nutrition by microorganisms. (A) Index Analysis Based on Threshold Element Ratio (TER);(B) Stoichiometric Analysis of Soil Enzyme Activity; (C) Analysis of Changes in Vector Length and VectorAngle.

Full size image

Throughout the entire experiment period, phosphorous limitation of soil microorganisms appeared more pronounced than nitrogen limitation, as evidenced by vector angles consistently greater than 45°. Utilizing the ratios of (BG + CBH)/(NAG + LAP) and (BG + CBH)/AcP, we determined that soil microorganisms were consistently limited by phosphorous throughout the experimental process (Fig. 5C), aligning with findings from all three assessment methods. Compared to the initial stage, the difference between the ratios of (BG + CBH)/(NAG + LAP) and (BG + CBH)/AcP decreased in the third year of nitrogen application, suggesting a mitigation of phosphorous limitation with prolonged nitrogen application time. Across all treatments, microbial activity indices (vector angles) were consistently above 45°, confirming pervasive P-limitation in this forest soil. Notably, N addition tended to increase the degree of phosphorous limitation (vector angle increased and TERC:P decreased relative to control), especially in the first year of treatment (Fig. 4 & 5).

Potential biomarkers among nitrogen enrichment and soil microbial abundance with TER

The LDA and LEfSe methods were utilized to identify bacterial and fungal groups exhibiting significant differences in abundance across various sampling points (Fig. 6). In total, 44 bacterial and 20 fungal taxa displayed significant differences. The LDA threshold was set at 4 for bacteria and 5 for fungi in the analysis diagram (Fig. 6 A and B). When the LDA threshold was set at 4.5 for bacteria, eight significantly distinct bacterial taxa were identified, including five phylum-level (Proteobacteria, Actinobacteriota, Chloroflexi, Actinobacteria, Alphaproteobacteria), and three orders (Acidobacteriales, Subgroup_2, and Rhizobiales). For fungi, 11 significantly different taxa were found, including phylum-level (Ascomycota, Basidiomycota), class (Tremellomycetes), order (Cantharellales, Tremellales, Trechisporales), family (Hydnaceae, Trimorphomycetaceae), and genus (Trechispora). These bacterial and fungal taxa are potential biomarkers.

Fig. 6
figure 6

The cladogram of bacterial communities (A) and fungal communities among different treatment sites.

Full size image

Correlation and regression analyses were performed to examine the relationship between key microbial taxa and threshold element ratio (TERC:N or TERC:P), vector length, and angle (Tables 3 and 4). Notably, four fungal biomarkers (Tremellomycetes, Hydnum, Trimorphomycetaceae, and Tremellales) exhibited a significantly positive correlation with TERC:N. Similarly, one bacterial biomarker (Chloroflexi) and three fungal biomarkers (Tremellomycetes, Trimorphomycetaceae, and Tremellales) also showed a significant positive correlation with TERC:P. Chloroflexi displayed a significant negative correlation with vector length L (r = -0.622, P < 0.01) and a positive correlation with vector angle A (r = 0.482, P < 0.01). Chloroflexi abundance was negatively associated with carbon limitation and positively associated with P limitation metrics, suggesting this bacterial group thrived under exacerbated P limitation. Additionally, another fungal marker, Hydnum, was significantly negatively correlated with the vector angle A (Table 3). Trimorphomycetaceae was more abundant in LN plots. These identified biomarker taxa serve as representatives of key microbial groups contributing to the metabolic thresholds within microbial communities.

Table 3 Pearson correlation matrix between the relative abundance of key microbial taxa (biomarkers) and TERC:N or TERC:P.
Full size table
Table 4 Pearson correlation matrix between the relative abundance of key microbial taxa (biomarkers) and TERC:N or TERC:P.
Full size table

To further explore the relationship between key microbial groups and nutrient limitation indicators (threshold element ratio, vector length, and angle), a standardized regression equation was fitted to construct a model to quantify this relationship. The stepwise regression analysis revealed that the model explaining TERC:N was predominantly composed of Tremellomycetes and Hydnum, exhibiting a positive correlation. The model explaining TERC:P included Trimorphomycetaceae, Chloroflexi, and Tremellales, with TERC:P positively correlated with Trimorphomycetaceae and Chloroflexi, and negatively correlated with Tremellales and Vector.

Finally, the structural equation modeling results (CMIN/DF = 1.407, P = 0.162, RMSEA = 0.067, AIC = 63.478) indicate an acceptable model fit. As illustrated in Fig.S4, three years of nitrogen application directly and indirectly influenced soil ecological stoichiometry, thereby explaining the variations in the composition of soil bacterial and fungal communities. Moreover, this study revealed that the availability of soil nutrients significantly affected changes in bacterial community composition (P < 0.001), whereas soil microbial biomass had a significant effect on fungal community composition (P < 0.01). The magnitude of change in bacterial community composition was notably greater than that observed for fungal communities. Enzyme activity was influenced by both total soil content and its associated measurements, which subsequently impacted assessments of soil nutrient availability. Additionally, microbial biomass measurements were affected by available soil nutrient levels, which in turn influenced total soil nutrient dynamics. A significant correlation was also observed between microbial biomass and enzyme activity measurements.

Discussion

Response of soil microbial nutrient limitation to nitrogen application

Intensified nitrogen deposition is the primary driver of soil acidification, causing a sustained pH decline whose rate varies across forest types based on background deposition levels. This acidification inhibits microbial activity and biomass, thereby increasing soil carbon storage21,2⁷. In this study, elevated available nitrogen did not induce leaching of ammonium or nitrate nitrogen, indicating that vegetation and soil microorganisms can effectively utilize increased nitrogen inputs over extended periods. The observed decrease in total soil nutrient ratios (C:P and N:P) is associated with an increase in total phosphorous (TP), whereas changes in available nutrient ratios—specifically a decline in AC:N and an increase in AN:P—are linked to stoichiometric shifts driven by elevated available nitrogen (AN). The C:N ratio is widely recognized as a key predictor of litter decomposition rates, as it reflects the relative proportions of carbohydrates and proteins in leaf litter and represents a fundamental characteristic of litter quality28. Typically, nitrogen enrichment inhibits microbial growth in soils by promoting acidification, rather than facilitating microbial biomass accumulation.

The interaction effect of time and nitrogen application significantly affected vector length, suggesting the potential for increased carbon limitation in future (Table 2). Short-term nitrogen application has no significant effect on vector length. Nitrogen application significantly affects the vector angle, exacerbating soil microbial phosphorous limitation (Figs. 4 and 5). The consistent observation of vector angles exceeding 45° in all data implies the ubiquity of microbial phosphorous limitation in southern forest ecosystems, aligning with results from other global systems31. This finding challenging the long-standing belief that nitrogen is the primary limiting factor, our research reveals that phosphorous is, in fact, the key limiting element in subtropical forests8,32. Recent studies suggest that phosphorous limitation is as crucial as nitrogen limitation in terrestrial ecosystems33, which partially supports our results. The observed change in phosphorous limitation post-nitrogen application may be attributed to the heightened soil phosphatase activity, resulting in an accelerated rate of phosphorous cycling34. In this study, TERC:N was significantly higher than RC:N compared with TERC:P. As nitrogen application intensified and years elapsed, TERC:N showed an upward trend, signifying that microbial communities demonstrated high carbon utilization efficiency, which is consistent with the lack of carbon limitation in this study9,35. Comparing TERC:P with RC:P, high nitrogen caused RC:P to decrease, and the microbial community transitioned from carbon utilization efficiency to nutrient utilization efficiency, aiming to maintain chemical balance and alleviate phosphorous limitation. A severe stoichiometric imbalance indicates that microorganisms require greater nitrogen and phosphorous inputs to maintain nutrient homeostasis. The more pronounced imbalance observed in soil layer B suggests that vertical stratification affects microbial growth efficiency and nutrient availability, thereby indirectly regulating soil enzyme activity and their stoichiometric ratios.

Soil microbial extracellular enzymes play a crucial role in microbial metabolism36. Ecological enzyme stoichiometry reflects microorganisms’ ability to utilize nutrients8,36,37. This microbial shift towards P limitation could eventually feedback to slower organic matter turnover or necessitate greater P mobilization (via phosphatase activity) in these forests. Increased soil nitrogen availability alters microbial resource acquisition strategies via enzyme secretion, with microbes modulating enzyme production according to substrate accessibility38. It suggests that chronic N deposition might require increased P availability to maintain soil fertility and carbon cycling rates. In the early stages of soil development, nitrogen supply is typically limited. However, as soil matures, primary minerals are depleted, and the release of phosphorous from weathering diminishes6. phosphorous is the primary limiting factor in primitive tropical and subtropical forest ecosystems, with soil microorganisms experiencing more severe phosphorous limitation6,19,33. This study reveals that microbial metabolism in the pine forest soil tends to invest more in nitrogen to acquire enzymes rather than phosphorous . As nitrogen application increases, the microbial community in the experiment shifts towards investing in phosphorous-acquiring enzymes rather than carbon and nitrogen acquiring enzymes. This adaptive response supports the resource allocation theory, predicting that microorganisms increase enzyme secretion for mining scarce elements, thus reducing the chemical imbalance of their resources12,38. Nitrogen and phosphorous availability can regulate litter decomposition rates and ecosystem carbon balance8,13. Therefore, soil enzyme stoichiometry serves as a suitable indicator for assessing soil nutrient utilization and microbial metabolism36. In this experiment, the soil C:P ratio exceeding 186 further emphasizes that phosphorous is the limiting element for nutrient cycling in subtropical pine forest ecosystems. Soil enzyme extracellular enzyme activity and microbial biomass C:N:P stoichiometry can be integrated into biogeochemical balance models, linking microbial growth efficiency with extracellular enzyme activity39. phosphorous availability is crucial for microbial enzyme allocation, biomass growth, and carbon mineralization. In situations where soil nitrogen and phosphorous utilization are limited, microorganisms secrete corresponding enzymes to fulfill their nutrient requirements. This experiment indicates that in subtropical forests, the phosphorous extracted from the soil may rapidly transform into refractory phosphorous through geochemical adsorption, reducing phosphorous availability. Our identification of Chloroflexi (notably class Ktedonobacteria) and Basidiomycota taxa (Tremellomycetes order) as positively associated with P limitation suggests these groups are pivotal responders to N enrichment in P-poor soils.

The phosphorous limitation in the region is primarily driven by several key factors. Firstly, the subtropical climate, characterized by its high temperatures and abundant rainfall, depends predominantly on rock minerals as the primary source of phosphorous. The intense weathering in this region results in the easy loss and limited availability of phosphorous40. Secondly, the area is subject to significant nitrogen deposition, as demonstrated by nitrogen application experiments that alter soil ecological stoichiometry. This alteration promotes plant growth and increases the demand for phosphorous by both plants and microbes, in accordance with ecological stoichiometry theory. Thirdly, prolonged elevated nitrogen levels contribute to reduced vegetation diversity beneath the forest canopy over the long term, potentially causing soil erosion and subsequent nutrient loss.

Key microbial taxa associated with nutrient limitation

Microbial communities in the soil may face various nutrient limitations in diverse environmental conditions36. These microorganisms acquire nutrients through the secretion of enzymes and the decomposition of organic matter41. The differential responses of bacterial and fungal communities to soil nutrient gradients following nitrogen enrichment not only reflect distinct nutrient niche preferences among soil microbial taxa but also underscore the prevalence of interspecific nutrient competition in the rhizosphere. From a functional adaptation perspective, the modulation of microbial community functional traits enables these taxa to acclimate to resource limitation (e.g., phosphorous scarcity) and environmental stressors (e.g., soil acidification induced by nitrogen deposition) in subtropical forest or agricultural ecosystems, thereby sustaining ecosystem functional stability and nutrient cycling homeostasis. In our study area, characterized by abundant precipitation and low phosphorous and high-N conditions, nitrogen application does not cause a loss of microbial biomass or a collapse in function; instead, the community adjusted in composition. Instead, adaptive changes in microbial community abundance and composition occur (Fig. 1). Therefore, microorganisms can adjust their physiological metabolism to obtain carbon, nitrogen, and phosphorous resources, and adapt to the habitat environment after nitrogen application.

N addition led to a markedly different profile of indicator taxa, reflecting significant shifts in community composition through LEfSe analysis (Fig. 6). In this study, the bacterial community structure was primarily influenced by carbon and phosphorous resources (Fig. 2, Table S4), as well as the nutrient availability dictated by soil type (Table S1). Given bacteria’s preference for utilizing available nutrients in the soil42, it is suggested that available nutrients play a more significant role in regulating the bacterial community structure in this context. Ktedonobacteria are a unique group of prokaryotes that morphologically resemble actinomycetes and are considered to participate in the carbon cycle. Cavaletti et al.43 first reported the class Ktedonobacteria and classified it within the Phylum Fibrobacteres44, a group of distinct bacteria. The fungal community was mainly affected by soil chemical properties (TN, AN) and soil microbial biomass (Fig. 2, Table S5). Zhang et al.15 highlighted the significant impact of vegetation on the fungal community, emphasizing the crucial role fungi play in regulating plant productivity and soil carbon storage in forest ecosystems. Fungi are key contributors to the initial stages of litter decomposition, a process greatly influenced by vegetation type and litter quantity5,13.

The structure of microbial communities determines their functions and metabolic patterns25. Biomarkers responsible for distinctions in microbial communities were identified across different sampling points, and some of these biomarkers exhibited significant correlations with microbial nutrient limitations (Tables 3 and 4). The sequencing analysis results of this study revealed that Tremellomycetes, Hydnum, Trimorphomycetaceae, and Tremellales were positively correlated with TERC:N. Notably, Trimorphomycetaceae belongs to a genus within Tremellomycetes identified as a low-N treatment biomarker, while Hydnum belongs to the genus Agaricomycetes identified as a control treatment biomarker (Fig. 6A). As these fungi are all part of the Basidiomycota phylum, and the control treatment exhibited higher TERC:N values than the other treatments (Fig. 3), it suggests that Trimorphomycetaceae and Hydnum significantly influence microbial nitrogen metabolism. Hence, these two microbial groups can serve as biological markers for microbial nitrogen limitation. Basidiomycete fungi like Hydnum are often ectomycorrhizal or involved in organic matter breakdown, which might give them an advantage in scavenging for N under N-poor conditions. Moreover, the bacterial group Chloroflexi and three fungal biomarkers (Tremellomycetes, Tremellales, and Trimorphomycetaceae) exhibited positive correlations with TERC:P (Table 3). The control treatment, on average, demonstrated higher TERC:P values than the LN and HN treatments (Fig. 3), indicating that microbial classes from Chloroflexi and the Tremellomycetes orders within Tremellomycetes strongly influence microbial phosphorous metabolism. Thus, these two microbial groups can be regarded as biological markers for microbial phosphorous limitation. Furthermore, Chloroflexi showed close associations with vector length (L) and vector angle. For Chloroflexi (Ktedonobacteria), their abundance under P limitation might be due to their ability to use recalcitrant carbon sources (hence thriving when P is limiting and easily available C is exhausted, perhaps). Specifically, in response to the observed soil C:P stoichiometric imbalance, soil microorganisms exhibit enhanced phosphorous acquisition strategies to meet their metabolic demands. A key ecological function of the phylum Chloroflexi lies in its robust capacity for organic phosphorous (Po) mineralization, primarily mediated by the secretion of alkaline phosphatase (ALP)—a signature phosphorous -acquiring enzyme in soil. This hydrolase catalyzes the cleavage of phosphate ester bonds in Po compounds (e.g., phytate, phospholipids, and nucleic acids), releasing bioavailable inorganic phosphate (Pi) that can be assimilated by both heterotrophic microorganisms and autotrophic plants. In addition, nitrogen addition induced shifts in the microbial community composition toward more efficient phosphorus acquisition, primarily by modulating the enzyme–substrate binding affinities of NAG, LAP, ACP, and ALP15 As typical slow-growing K-strategists, members of the phylum Chloroflexi produce enzymes with high substrate affinity (low Km)15. Significant negative correlations were observed between the Km values of phosphorus-degrading enzymes and the relative abundances of Acidobacteria and Chloroflexi15 Through this ALP-driven Po mineralization process, Chloroflexi establishes a specialized phosphorous-acquisition niche, which not only supports its competitive survival in low-P soil environments but also facilitates its dominance in the microbial community. Therefore, the results suggest that biomarkers influencing changes in microbial community structure, representing distinct species, may be key microbial taxa contributing to microbial phosphorous limitation. Furthermore, relative abundance of Chloroflexi possesses an intrinsic genomic trait of "low phosphorous requirement" and employs an active "resource allocation strategy toward phosphorous acquisition" —two adaptive traits that further enhance its tolerance to phosphorous limitation. The high relative abundance of Chloroflexi in nitrogen-amended soils thus not only serves as a biological indicator of soil phosphorous limitation but also plays a pivotal role in alleviating ecosystem-level phosphorous limitation by promoting P turnover. Although this study offers valuable insights into the immediate microbial responses to nitrogen deposition, longer-term experiments are essential to ascertain its chronic effects. Collectively, the results indicate that short-term nitrogen enrichment in subtropical forest soils exacerbates microbial phosphorous limitation and induces significant shifts in microbial community composition, with certain taxa serving as indicators of altered nutrient dynamics. These findings contribute to the optimization of forest soil and ecosystem management strategies, such as targeted phosphorous fertilization or microbial inoculation, and support the development of sustainable agricultural practices and improvement of forest productivity in subtropical regions.

Conclusion

This study demonstrates that in forest soils of P. Taiwanensis, short-term nitrogen enrichment primarily intensifies microbial phosphorous limitation rather than carbon limitation. In response to nitrogen, the microbial community structure undergoes significant shifts, with both bacteria and fungi exhibiting distinct adjustments aligned with their respective nutrient acquisition strategies. Notably, the microbial community composition shifted: P-sensitive taxa (like specific Basidiomycota fungi and Chloroflexi bacteria) increased in relative abundance, highlighting their role as potential bio-indicators of P limitation stress. These alterations subsequently impacted microbial nutrient status. Identifying five biomarkers linked to associated with our stoichiometric indicators (e.g., Tremellales) highlights pivotal microbial taxa influencing both community structure and function, particularly in the context of nutrient limitation. The divergent responses of bacterial and fungal groups to nitrogen application in soil imply distinct nutrient preferences and competitive resource utilization among soil microorganisms. Through regulating microbial community function, these microorganisms adapt to resource scarcity and environmental pressures in subtropical ecosystems, contributing to the preservation of ecological stability. Our findings imply that sustained N deposition in subtropical forests may require increased phosphorous availability to maintain microbial activity and nutrient cycling, an important consideration for forest management in the face of changing nutrient inputs. This study not only clarifies the connection between nitrogen application and microbial nutrient limitation but also provides insight into the responses of microbial communities to nutrient limitations in subtropical ecosystems.