The mitigation of microbial carbon and nitrogen limitations by shrub encroachment: extracellular enzyme stoichiometry of the alpine grassland on the Qinghai-Tibetan Plateau

Shrub encroachment changes the patterns of nutrition allocation in the below- and aboveground soil. However, influence of shrub encroachment on microbial carbon (C) and nitrogen (N) limitations remains unclear. Using the extracellular enzyme stoichiometry model, microbial nutrition limitations in bulk and rhizosphere soils at various soil layers were investigated at non-shrub alpine grasslands (GL) and shrub-encroached alpine grasslands including Spiraea alpina lands (SA), Caragana microphylla lands (CM) and Potentilla fruticosa lands (PF) on the Qinghai-Tibetan Plateau. We determined C-acquisition (β-1,4-glucosidase (BG); β-D-fibrinosidase (CBH)), N-acquisition (β-1,4-N-acetylglucosaminidase (NAG); leucine aminopeptidase (LAP)) and phosphorus (P)-acquisition (acid phosphatase (AP)) enzyme activities. The contents of soil organic carbon (SOC) in top- and subsoils significantly increased following shrub encroachment. Interestingly, (LAP + NAG) activities in subsoil increased following shrub encroachment. EC:N in subsoil decreased following shrub encroachment. Microbial C and N limitations were found in shrub-encroached and non-shrub alpine grasslands. Furthermore, microbial C and N limitations in bulk topsoil layers decreased following shrub encroachment. Microbial N limitations in subsoil decreased following shrub encroachment. This result indicates that shrub encroachment mitigated microbial C and N limitations. The limitations were gradually mitigated following shrub encroachment, which led to the decrease of the decomposition rate of organic carbon by microorganisms, indicating shrub encroachment might potentially contribute to SOC storage. In addition, the structural equation modeling (SEM) showed that increases of SOC and NH4+–N in top- and subsoils under shrub encroachment could mitigate microbial C and N limitations, respectively. This study provides available information on the environmental variables affecting the stoichiometry of extracellular enzymes following shrub encroachment, and the theoretical basis for the study of C and N cycling in alpine grasslands.


Introduction
The grassland ecosystem is a vital part in terrestrial ecosystems (Zhang et al. 2016).Many shrub encroachment phenomena have occurred in arid and semi-arid grassland ecosystem due to global climate change (Eldridge et al. 2011).The alpine grassland on the Qinghai-Tibetan Plateau is a significant C sink area and plays a crucial role in C emissions in China, and even in the world (Yan et al. 2015).Shrub encroachment is notably spreading in alpine grasslands on the eastern of the Qinghai-Tibetan Plateau, with a 39% raise in size from 1990 to 2009 (Brandt et al. 2013).Shrub encroachment seriously changes plant community, affecting the partitioning pattern of soil nutrient in above-and belowground, which in turn alters soil C input in alpine grasslands (Fayiah et al. 2019).The influences of shrub encroachment on C storage remains unclear at the ecosystem level, with some studies showing an increase of C storage in arid ecosystems (Liao et al. 2006b), while others demonstrated a decrease instead in moist ecosystems (Mureva et al. 2018).Meanwhile, the increase of C storage at deep layers following shrub encroachment was only found in a few investigations.For example, Biederman and Boutton (2009) showed that the well-developed root system of shrubs could distribute soil surface organic C and organic N to deeper soils, thus increasing the contents of C and N in deeper soils.It means that, the changes in C storage at different soil layers following shrub encroachment have not been measured and evaluated thoroughly, and the study of C storage in top-and subsoils following shrub encroachment is crucial for understanding the C dynamics in alpine grasslands.
Soil enzymes are mainly derived from plant roots, soil animals and microorganisms, and their activity levels reflect the soil nutrient and energy conversion and microbial metabolism status at a specific time (Dick 1994).Extracellular enzymes mainly decompose organic matter to acquire nutrients associated Vol.: (0123456789) with C, N and P (Soong et al. 2020).Specifically, β-1,4-glucosidase (BG) and β-D-fibrinosidase (CBH) hydrolyze cellulose and hemicellulose to acquire C sources for microorganisms (Sinsabaugh & Follstad Shah 2012).β-1,4-N-acetaminoglycosidase (NAG) and leucine aminopeptidase (LAP) hydrolyze chitin, protein and urea to acquire N source for microorganisms (Jian et al. 2016;Moorhead & Sinsabaugh 2006).Acid phosphatase (AP) hydrolyses phosphate monomers under acidic conditions to acquire P source for microorganisms (Tischer et al. 2015).Sinsabaugh et al. (2008) used soil extracellular enzyme stoichiometry to express microbial demand for soil nutrients.Further studies showed that the stoichiometry ratios of C:N:P acquisition enzymes are approximately 1:1:1, indicating that the nutrient requirements of microorganisms are in balance (Sinsabaugh et al. 2009).However, due to the great variety of soil extracellular enzyme stoichiometry among different ecosystems, the ratio was not followed in all regions (Sinsabaugh & Follstad Shah 2012).Especially, the ratio deviates from 1:1:1, suggesting microbial nutrient limitations.Moorhead et al. (2016) proposed an extracellular enzyme stoichiometric model, calculating vector lengths and angles by the ratio of soil C-, N-and P-acquisition enzyme activities for characterizing microbial nutrient metabolism limitations, which has been extensively used.In contrast, a recent study pointed out that when cellulose is the predominant C source in soils rather than chitin, peptidoglycan and protein, BG:(NAG + LAP) can represent microbial C and N limitations (Mori 2020).In the shrub-encroached alpine grasslands of the eastern Qinghai-Tibetan Plateau, soil litter enriches cellulose, which is the main C source for soil microbial growth (He et al. 2015).Especially in alpine grasslands on Qinghai-Tibetan Plateau, most genes involved in soil C cycling have high signal intensities.Numerouscellobiose genes involved in cellulose degradation exist in the soil, indicating soil cellulose is the main C source in alpine grasslands on the Qinghai-Tibetan Plateau (Zhang et al. 2013).Therefore, the application of soil extracellular enzyme stoichiometry is critical to reveal the characteristics of microbial metabolism in alpine grasslands.
Rhizosphere is a hotspot for soil extracellular enzyme-microbial interactions (Zheng et al. 2019).Rhizosphere soils are usually relatively nutrient-rich and have a higher microbial metabolic rate than bulk soils (Ai et al. 2012).Root exudates and rhizosphere metabolism change the composition of soil microbial communities (Bi et al. 2021;Ma et al. 2022).The variations in microbial community structure are most probably alter the metabolic process, thus resulting in the variations in the amount of microbial acquisition nutrients (Paterson 2003).Many researchers have used extracellular enzyme stoichiometry to disclose nutrient limitations in bulk and rhizosphere environments.Cui et al. (2021) used extracellular enzyme stoichiometry to find microbial P limitations in the high-altitude, which was significantly greater in rhizosphere than in bulk soils in alpine ecosystems.Consequently, determining metabolic characteristics of microbial communities in rhizosphere environments is of vital importance for understanding the function of root systems in C cycling during shrub encroachment in alpine grassland ecosystems.
Shrub encroachment is a primary factor affecting the C pool balance in alpine grasslands (Chen et al. 2022).Recent studies have found that climate warming on the Qinghai-Tibetan Plateau directly mitigates the temperature limitation on plant growth, which promotes plant growth and increases the biomass C allocated by plants to the ground, thus mitigating soil microbial C limitations (Chen et al. 2020).The activities of soil microorganisms increased with increasing temperature, and elevated SOC decomposition (Li et al. 2019), which may mitigate microbial N limitations to some extent.However, it is not clear whether the influence of shrub encroachment on soil C pool will change microbial nutrient limitations in bulk and rhizosphere soils in alpine grasslands.Therefore, in the study, we investigated the changes of extracellular enzyme activities and the stoichiometric characteristics in bulk and rhizosphere soils at different depths in non-shrub and three shrub-encroached grasslands (Spiraea alpina, Caragana microphylla and Potentilla fruticosa), as well as the effects of extracellular enzyme stoichiometric characteristics on soil C cycling in alpine grasslands.Understanding the characteristics of soil extracellular enzyme stoichiometry following shrub encroachment is beneficial to further determining nutrient cycling (eg., C, N, P) on the Qinghai-Tibetan Plateau.Therefore, we proposed the following hypotheses: ( 1

Study sites
This study is conducted in the Qinghai-Tibetan Plateau Research Base of Southwest Minzu University in western Sichuan Province, China (32°49′38″N, 102°34′21″E, 3485 m elevation), which belongs to the eastern margin of the Qinghai-Tibetan Plateau region.Th study area has a cold-temperate monsoon climate of the continental plateau.The temperature reaches its peak in July, with 25.6 ℃ as the highest record and 10.9 ℃ as the average.The annual relative humidity is 60-70%, and the moisture coefficient is 1.26.The mean annual precipitation ranged from 650 to 800 mm and mainly concentrated between May and August each year.Alpine grassland soil is the main soil type, and the soil parent material is mostly residual slope deposits and slope deposits.In the study area, shrub encroachment can be deemed serious (Brandt et al. 2013).The vegetation is typical alpine meadow and is dominated by Spiraea alpina, Caragana microphylla and Potentilla fruticosa.The main herbaceous plants are Kobresia filifolia and Elymus nutans.

Sampling design
Field sampling for this experiment was carried out in July 2021.A 10 m × 10 m plot of shrub-encroached alpine grassland including Spiraea alpina lands (SA), Caragana microphylla lands (CM), Potentilla fruticosa lands (PF) was randomly selected from the sampling area, and three 1 m × 1 m small plots heavily encroached by more than 75% were set in the large plot.All aboveground biomass in the plot was collected with scissors.Due to the special climatic conditions and geographical environment of the Qinghai-Tibetan Plateau, we found in the sampling process that the non-shrub alpine grassland basically had no living roots below 40 cm soil layer.Although the average root depth of Spiraea alpina, Caragana microphylla and Potentilla fruticosa was 0-80 cm, there were almost no roots at soil depths greater than 40 cm in Potentilla fruticose (Zhang & Li 2018), and most of the living roots of Spiraea alpina and Potentilla fruticose were found to be concentrated at 30-40 cm during sampling.To better collect rhizosphere soils in non-shrub and shrubencroached grasslands, 0-10 cm and 30-40 cm soils were selected as topsoil and subsoil.Rhizosphere soils were sampled using the shaking-off method (Fujii et al. 2005).Bulk soils were sampled at least 5 mm away from roots.Soil samples (physicochemical properties and enzyme samples) were put into ziplock bags.Soil enzyme samples were kept in zip-lock bags, and then immediately transferred into the ice box with ice packs.A 10 m × 10 m plot in non-shrub grassland (GL) was selected near where shrubs grew.Soil samples in non-encroached grassland plots were sampled in the same way as those from encroached grassland plots.Aboveground biomass and soil samples from each plot were brought back to the laboratory.Aboveground biomass was classified and measured.Soil physicochemical properties samples were divided into two parts.Some samples were baked at 105 °C for 24 h to calculate the soil water content and then dried naturally.The other part samples and soil enzyme samples were stored at −4 °C.

Soil properties measurements
Soil water contents (SWC) are determined by the natural drying method.A pH meter was used to measure soil pH, and the ratio of the soil and water was 1.0:2.5.The contents of SOC, total nitrogen (TN) and dissolved organic C (DOC) were determined by the Elementar Variomax CNS analyzer (Germany) (Jones & Willett 2006).The TP and available P (Olsen-P) contents were determined by NaOH alkali fusion-Molybdenum-antimony resistance colorimetry.The NO 3 − -N and NH 4 + -N contents were determined by a Flow Analyzer.

Assays of soil extracellular enzyme activity
Using the Saiya-cork fluorimetric test method, C-acquisition enzymes (BG and CBH), N-acquisition enzymes (NAG and LAP) and P-acquisition enzymes (AP) were identified (Saiya-Cork et al. 2002).To measure the enzyme activity, the 125 ml of deionized water and 1 g of fresh soil were homogenized for 2 h at a speed of 180 rotations per minute on a rotary shaker.Then, 200 μL of each sample was put into a 96-well microplate.We added 50 μL of the standard substrate (40 μmol L −1 4-MUB and 10 μmol L −1 7-amino-4-methylcoumarin) and 200 μL of test suspension to the extinguished standard microplates.We added 50 μL standard substrate and 200 μL deionized water to the negative control microplates.The 96-well microplates were brooded in obscurity at 25 °C for 4 h (German et al. 2011).Soil extracellular enzyme stoichiometry was computed using formulas on ln(BG + CBH):ln(NAG + LAP), ln(BG + CBH):ln(AP) and ln(NAG + LAP):ln(AP), respectively.

Quantification of soil microbial metabolic limitations
Based on the relative proportion of untransformed soil extracellular enzyme activities, such as (BG + CBH):(BG + CBH + AP), vector lengths and angles were used to quantify microbial nutrition limitations (Moorhead et al. 2013(Moorhead et al. , 2016)).
Equations 1 and 2 were used to compute vector lengths and angles, with x and y representing for, (BG + CBH):(BG + CBH + AP) and (BG + CBH):-(BG + CBH + NAG + LAP), respectively.The microbial C limitation is represented by vector lengths.Microbial C limitations would be enhanced with vector lengths.The microbial N or P limitation is represented by vector angles.Vector angles lower than 45° indicate microbial N limitations.Microbial N limitations increased with the decrease of vector angles.Vector angles greater than 45° imply microbial P limitations.Microbial P limitations would be enhanced with vector angles.

Statistical analysis
The experimental data were collected by Microsoft Excel 2019 and analyzed by IBM SPSS Statistics 26.0.Data conforming to normal distribution were analyzed by one-way variance analysis and an independent-sample t-test.The effects of different vegetation types on soil physicochemical properties, (1) extracellular enzyme activities, extracellular enzyme stoichiometry, and microbial nutrient limitations were examined using one-way variance analysis (Tukey's multiple comparison test) (P < 0.05).The effects of soil depths and locations (bulk and rhizosphere) on soil physicochemical properties, extracellular enzyme activities, extracellular enzyme stoichiometry, and microbial metabolic limitations were examined using an independent-sample t-test (P < 0.05).Canoco 5.0 was used to conduct the Redundancy analysis (RDA) to evaluate the relationship between soil physicochemical properties vs extracellular enzyme activities and soil physicochemical properties vs extracellular enzyme stoichiometry, respectively.A generalized linear regression analysis was used to determine the relationship between environmental factors and microbial nutrient limitations.Origin 2021 was used to conduct a generalized linear regression analysis.non-encroached grasslands, Spiraea alpina lands, Caragana microphylla lands and Potentilla fruticosa lands were ranked by continuous variable (1, 2, 3 and 4), respectively.Structural equation modeling (SEM) is used to identify the potential trend of the effects of shrub encroachment on microbial nutrient limitations.Amos v.23 was used to conduct the model.

Soil physical property and available nutrients
Soil physical properties and available nutrients varied among vegetation types.Generally, in top-and subsoils, the bulk and rhizosphere SWC, pH, NH 4 + -N, NO 3 − -N and DOC in CM were significantly higher than in GL (Table 1; P < 0.05).Soil physical properties and available nutrients varies between soil layers.In bulk and rhizosphere soils, NO 3 − -N and DOC in topsoil were significantly higher than in subsoil in GL, SA and CM (Table 1, P < 0.05).Soil physical properties and available nutrients varies between bulk and rhizosphere.In top-and subsoils, NH 4 + -N, DOC and Olsen-P in rhizosphere was significantly higher than in bulk in CM (P < 0.05).

Soil total nutrients and nutrient ratios
Soil total nutrients and nutrient ratios varied among vegetation types.In top-and subsoils, the bulk and rhizosphere SOC, C:P and N:P in CM were significantly higher than in GL (P < 0.05, Table 2).Soil total nutrients and nutrient ratios varied between soil layers.In bulk and rhizosphere soils, the contents of SOC and TN in topsoil were significantly higher than in subsoil in SA and CM (P < 0.05).Soil total nutrients and nutrient ratios varied between bulk and rhizosphere.In top-and subsoils, SOC and TP contents in rhizosphere were significantly higher than in bulk in PF (P < 0.05).

Soil extracellular enzyme activities (EEAs)
EEAs varied among vegetation types (Table 3).In top-and subsoils, rhizosphere (BG + CBH) activities in CM were significantly lower than in GL (P < 0.05); bulk (LAP + NAG) activities in CM were significantly higher than in GL (P < 0.05).EEAs differed between soil depths.Rhizosphere (BG + CBH) activities in topsoil were significantly higher than in subsoil in SA (P < 0.05).Rhizosphere (LAP + NAG) and AP activities in topsoil were significantly higher than in subsoil in GL (P < 0.05).EEAs differed between bulk and rhizosphere soils.In top-and subsoils, (BG + CBH) activities in rhizosphere were significantly higher than in bulk in GL and SA (P < 0.05).
In the redundancy analysis of soil physicochemical properties and extracellular enzyme activities, in topsoil, the interpretation rates of RDA1 and 2 axes is a total of 90.08% (Fig. 1a).TP, DOC and NH 4 + -N were the three factors with the highest explanatory degree, explaining 27.7, 25.6 and 13.9% respectively.(BG + CBH) activities were correlated with TP positively, and with NH 4 + -N and DOC negatively (P < 0.05).(LAP + NAG) activities were correlated with NH 4 + -N, DOC and TP positively (P < 0.05).
Table 1 The effects of vegetation types, locations (bulk and rhizosphere), and soil depths on physical property and available nutrients on the Qinghai-Tibetan Plateau Values are given as mean ± standard error (n = 3).The different letters in columns represent significant differences among four vegetation types and soil depths (P < 0.05).GL, non-shrub alpine grasslands; SA, Spiraea alpina lands; CM, Caragana microphylla lands; PF, Potentilla fruticose lands.Different lowercase letters represent significant differences (P < 0.05) among four vegetation lands in topsoil or subsoil.Different uppercase letters represent significant differences (P < 0.05) between top-and subsoils among four vegetation lands.* indicates the differences between bulk and rhizosphere soils within four vegetation lands and soil depths (*, P < 0.05; **, P < 0.01; ***, P < 0.001) AP activities were correlated with DOC and TP positively, and with NH 4 + -N negatively (P < 0.05).Additionally, in subsoil, the interpretation rates of RDA1 and 2 axes is a total of 86.07%(Fig. 1b).NH 4 + -N, Olsen-P and SOC were the three factors with the highest explanatory power, explaining for 38.6, 23.6 and 8.9% respectively.(BG + CBH) activities were correlated with Olsen-P contents positively, and with NH 4 + -N and SOC negatively (P < 0.05).(LAP + NAG) and AP activities were correlated with SOC and NH 4 + -N positively, and with Olsen-P negatively (P < 0.05).

Soil extracellular enzymatic stoichiometry (EES)
EES varied among different vegetation types (Table 4).In top-and subsoils, E C:N in SA were significantly lower than in GL (P < 0.05).EES differed between soil depths.The stoichiometric ratios of C-, N-and P-acquisition enzymes in non-shrub and encroached grasslands showed inconsistent trends in different soil layers.EES differed between bulk and rhizosphere soils.In top-and subsoils, E C:P in rhizosphere were significantly higher than in bulk in CM (P < 0.05).
In the redundancy analysis of soil extracellular enzyme stoichiometry and environmental factors, in topsoil, the interpretation rates of RDA1 and 2 axes is a total of 86.34% (Fig. 2a).TP, SWC and NH 4 + -N were the three factors with the highest explanatory degree, explaining for 36.2,10.2 and 8.0% respectively.Soil E C:N and E C:P were correlated with TP positively, and with the contents of SWC and NH 4 + -N negatively (P < 0.05).Soil E N:P were correlated with TP positively, and with SWC and NH 4 + -N negatively (P < 0.05).additionally, in subsoil, the interpretation rates of RDA1 and 2 axes is a total of 84.39% (Fig. 2b).C:N and N:P are the two factors with the highest explanatory degree, explaining 26.0, 20.1 and 11.8%, respectively.Soil E C:N and E C:P were correlated with C:N and N:P negatively (P < 0.05).Soil E N:P were correlated with C:N and N:P positively (P < 0.05).
Table 2 The effects of vegetation types, locations (bulk and rhizosphere), and soil depths on total nutrients and nutrient ratios on the Qinghai-Tibetan Plateau Values are given as mean ± standard error (n = 3).The different letters in columns represent significant differences among four vegetation types and soil depths (P < 0.05).GL, non-shrub alpine grasslands; SA, Spiraea alpina lands; CM, Caragana microphylla lands; PF, Potentilla fruticose lands.Different lowercase letters represent significant differences (P < 0.05) among four vegetation lands in topsoil or subsoil.Different uppercase letters represent significant differences (P < 0.05) between top-and subsoils among four vegetation lands.* represents the differences between bulk and rhizosphere soils within four vegetation lands and soil depths (*, P < 0.05; **, P < 0.01; ***, P < 0.001)

The characteristics of microbial C and N limitations
There were microbial C and N limitations in shrubencroached and non-shrub grasslands, as shown by the fact that all data points fell below the 1:1 line (P < 0.01; Fig. 3a, b).In addition, in top-and subsoil, there was a greatly positive correlation between microbial C and N limitations (P < 0.001; Fig. 3c, b).Microbial C and N limitations differed in vegetation types (Figs. 4,5).In top-and subsoil, vector lengths and angles in CM were significantly lower than in GL.Microbial C and N limitations differed between soil depths.Microbial C and N limitations in nonshrub and encroached grasslands showed inconsistent trends in different soil layers.Microbial C and N limitations differed between bulk and rhizosphere soils.In top-and subsoil, vector lengths in rhizosphere were significantly higher than in bulk in SA (P < 0.05).
By linear regression analysis, in topsoil, microbial C limitations were correlated with SWC, pH, NH 4 + -N, NO 3 − -N, DOC, SOC, TN, C:P and N:P negatively, and with TP and Olsen-P positively (P < 0.05; Fig. 6a).In subsoil, microbial C limitations were significantly correlated with pH, NH 4 + -N, NO 3 − -N, DOC, SOC, C:N and C:P negatively, and with TP and Olsen-P positively (P < 0.05; Fig. 6b).Meanwhile, linear regression analysis further showed that microbial N limitations were highly correlated with pH, DOC, TN and N:P negatively, and with TP and Olsen-P in topsoil positively (P < 0.05; Fig. 7a).In subsoil, microbial N limitations were significantly correlated with SWC, pH, NH 4 + -N, NO 3 − -N, DOC, SOC, TN, C:N, C:P and N:P negatively, and with TP and Olsen-P positively (P < 0.05; Fig. 7b).
The SEM was conducted to analyze the relationships between microbial C and N limitations and soil physiochemical properties following shrub encroachment, respectively (Fig. 8).In topsoil, SWC had a significantly positive correlation with SOC, and a significantly negative correlation with C:N; SOC had a significantly negative correlation with microbial C limitations; C:N had a significantly positive correlation Table 3 The effects of vegetation types, locations (bulk and rhizosphere), and soil depths on extracellular enzyme activities on the Qinghai-Tibetan Plateau Values are given as mean ± standard error (n = 3).The different letters in columns represent significant differences among four vegetation types and soil depths (P < 0.05).GL, non-shrub alpine grasslands; SA, Spiraea alpina lands; CM, Caragana microphylla lands; PF, Potentilla fruticose lands.C-acquisition enzymes, BG + CBH; N-acquisition enzymes, NAG + LAP; P-acquisition enzyme, AP.Different lowercase letters represent significant differences (P < 0.05) among four vegetation lands in topsoil or subsoil.Different uppercase letters represent significant differences (P < 0.05) between top-and subsoils among four vegetation lands.* represent the differences between bulk and rhizosphere soils within four vegetation lands and soil depths (*, P < 0.05; **, P < 0.01; ***, P < 0.001)

Layer (cm)
Vegetation type C-acquisition enzymes (nmol g −1 h −1 ) N-acquisition enzymes (nmol g −1 h with microbial C limitations (Fig. 8a).In subsoil, pH had significantly positive correlations with Olsen-P and SOC; Olsen-P had a significantly positive correlation with microbial C limitations; SOC had a significantly negative correlation with microbial C limitations (Fig. 8b).In addition, In topsoil, SWC had significantly positive correlations with Olsen-P and NH 4 + -N; Olsen-P had a significantly positive correlation with microbial N limitations; NH 4 + -N had a significantly negative correlation with microbial N limitations (Fig. 8c).In subsoil, SWC had significantly positive correlations with NH 4 + -N; pH had a significantly positive correlation with Olsen-P; Olsen-P had a significantly positive correlation with microbial N limitations; NH 4 + -N had a significantly negative correlation with microbial N limitations (Fig. 8d).

Discussion
The effect of shrub encroachment on soil nutritions Our study showed that shrub encroachment greatly affected physical properties.We found that pH in subsoil in shrub grasslands was significantly higher than in non-shrub grasslands (P < 0.05).The pH increased following shrub encroachment may be due to the fixation of shrubland root system, and little leaching loss occurs in shrubland after precipitation, which weakens nitrification (Pellegrini et al. 2021).In an alkaline environment, plant residues can reduce soil pH through nitrification (Binkley & Richter 1987).However, shrub encroachment weakens nitrification and does not reduce soil pH, resulting in the increase of pH following shrub encroachment.In the study, shrub encroachment affected soil available nutrients.4 The effects of vegetation types, locations (bulk and rhizosphere), and soil depths on extracellular enzyme stoichiometry on the Qinghai-Tibetan Plateau Values are given as mean ± standard error (n = 3).The different letters in columns represent significant differences among four vegetation types and soil depths (P < 0.05).GL, non-shrub alpine grasslands; SA, Spiraea alpina lands; CM, Caragana microphylla lands; PF, Potentilla fruticose lands.Different lowercase letters represent significant differences (P < 0.05) among four vegetation lands in topsoil or subsoil.Different uppercase letters represent significant differences (P < 0.05) between top-and subsoils among four vegetation lands.* represents the differences between bulk and rhizosphere soils within four vegetation lands and soil depths (*, P < 0.05; **, P < 0.01; ***, P < 0.001)   (Ding et al. 2021).Meanwhile, our results showed that the contents of NH 4 + -N and NO 3 − -N in rhizosphere soils in Caragana microphylla were the highest.To a certain extent, the phenomenon of nutrient enrichment in rhizosphere soils was explained (Dijkstra et al. 2013).This may be due to the fact that the input of shrub roots provides an effective N source for rhizosphere microorganisms, which are relatively active, leading to the secretion of compounds with the high physiological activity (Pinton et al. 2007).Our results showed that the rhizosphere DOC contents in Caragana microphylla was the highest compared with that in nonshrub grasslands.The loss of DOC depends on the concentration of DOC in soil pore water and hydrological conditions.Under low moisture conditions, DOC concentrations may decrease (Dieleman et al. 2016).However, our study showed that soil moisture in Caragana microphylla was generally higher than in non-shrub grasslands (Table 1).Therefore, high soil moisture conditions may lead to the increase of DOC concentrations.At the same time, we found that SOC contents increased significantly following shrub encroachment, which supports our first hypothesis.This is consistent with the conclusions by Gómez-Rey et al. ( 2013) and Liao et al (2006b) that shrub encroachment leads to the increase of soil C storage.The reason may be that shrub litter is not easy to be decomposed by soil microorganisms, leading to the accumulation of complex organic matter in shrubencroached grasslands (Liao et al. 2006a).In addition, the root growth of perennial shrubs often needs to consume more C to resist long-term environmental stress, so there will be more root sediments and nutrient enrichment in shrub-encroached grasslands (Bardgett et al. 2014).Moreover, Our results showed that TN contents in Caragana microphylla grasslands was significantly higher than in non-shrub grasslands (P < 0.05).This may be because Caragana microphylla is a leguminous shrub with a large Fig. 4 The variation of vector lengths affected vegetation types, soil depths, and sampling locations (bulk and rhizosphere).Figures (a) and (b) represent the variation trend of vector length between four vegetation types and locations in top-and subsoil layers, respectively.Figures (c) and (d) represent the variation trend of vector length of different soil layers in bulk and rhizosphere, respectively.Note: Different letters in columns denote significant differences among four vegetation types within sampling locations (bulk and rhizosphere) and soil depths (P < 0.05).GL, non-encroached alpine grasslands; SA, Spiraea alpina lands; CM, Caragana microphylla lands; PF, Potentilla fruticose lands.Different lowercase letters indicate significant differences (P < 0.05) among four vegetation lands in topsoil or subsoil.Different uppercase letters indicate significant differences (P < 0.05) between top-and subsoils among four vegetation lands.* indicates the differences between bulk and rhizosphere soils within four vegetation lands and soil depths (*, P < 0.05; **, P < 0.01; ***, P < 0.001) number of rhizobia in its root systems (Li et al. 2015), which may acquire N through symbiotic fixation and increase the accumulation of soil N (Cao et al. 2011).In addition, our study found that the contents and soil depths (P < 0.05).GL, non-encroached alpine grasslands; SA, Spiraea alpina lands; CM, Caragana microphylla lands; PF, Potentilla fruticose lands.Different lowercase letters indicate significant differences (P < 0.05) among four vegetation lands in topsoil or subsoil.Different uppercase letters indicate significant differences (P < 0.05) between top-and subsoils among four vegetation lands.* indicates the differences between bulk and rhizosphere soils within four vegetation lands and soil depths (*, P < 0.05; **, P < 0.01; ***, P < 0.001) of TP was much lower than that of SOC and TN in shrub-encroached and non-shrub grasslands, therefore the study area is the soil P limitation state.Shrub encroachment can change soil nutrient stoichiometry.In the study, we found that the C:N in subsoil in Caragana microphylla and Potentilla fruticosa Fig. 6 The effects of soil physicochemical properties and nutrient ratios on vector lengths in top-(a) and subsoils (b) Fig. 7 The effects of soil physicochemical properties and nutrient ratios on vector angles in top-(a) and subsoils (b) grasslands was significantly higher than in non-shrub grasslands (P < 0.05).This indicates that the rate of mineralization of organic matter is reduced following shrub encroachment.This may be due to shrubs producing large amounts of difficult-to-decompose leaf and root litter, which accumulate large amounts of SOC in soils, thus increasing C:N (Liao et al. 2006a).It may also be due to the fact that the age of shrub encroachment in the study area is 50 years.Although shrub encroachment in grasslands will immediately reduce soil C:N ratios, the ratio will gradually increase with the development of shrub encroachment (Feng & Bao 2018).
The effect of shrub encroachment on soil extracellular enzymes Our results showed that bulk (BG + CBH) activities increased following shrub encroachment, while rhizosphere bulk (BG + CBH) activities decreased.This may be due to the fact that our sampling time was August.In fact, there was no litter mass loss in the deep roots of shrubs between 6 and 12 months, and the catalytic decomposition rate of (BG + CBH) on soluble organic matter and sugars decreased, resulting in the decrease of the activity compared to previous months.This is consistent with previous findings by Hewins et al. (2017).We found that (LAP + NAG) activities mainly showed an increasing trend following shrub encroachment.This is consistent with the research results of Akinyemi et al. (2020).The reason may be that the "fertile island" formed under the shrub canopy following shrub encroachment contains abundant decomposable N-containing litters, resulting soils in shrubs may contain more substrates that stimulate (LAP + NAG) activities, which improved the rate of microbial N acquisition, therefore increasing (LAP + NAG) activities (Sardans et al. 2017).In addition, the primary production of plants and the proliferation of soil microorganisms are often limited by excess nutrients and lack of water, leading to differences in microbial nutrient acquisition and also affecting (LAP + NAG) activities (Wang et al. 2012).At the same time, our study further found that the rhizosphere (LAP + NAG) activities in Caragana microphylla grasslands was significantly higher than in non-shrub grasslands (P < 0.05).This may be due to the presence of N fixation-associated bacteria in the root of Caragana microphylla, which accelerates the secretion of (LAP + NAG) and thus increases (LAP + NAG) activities around the root of the shrub (Yu et al. 2020).Our study also showed that AP activities in subsoil in Caragana microphylla was significantly higher than in non-shrub grasslands.This may be due to the fact that the deep root in shrubs may be colonized by more arbuscular mycorrhiza compared to the root of gramineous plants (Yu-Ying & Wei 2004).Arbuscular mycorrhizal mycelium increases soil volume, plants can acquire P from the soil and produce AP in the rhizosphere (Javot et al. 2007).RDA analysis of soil extracellular enzyme activities and environmental factors in shrub encroached and non-shrub grasslands showed that TP, DOC and NH 4 + -N in topsoil were the three factors with the highest interpretation degree.NH 4 + -N, Olsen-P and SOC were the three factors with the highest explanatory power in subsoil.This result suggests that microorganisms can acquire relatively scarce resources by optimizing the distribution pattern of C, N and P during the synthesis of extracellular enzymes (Zhang et al. 2022).RDA analysis showed that there was a significant positive correlation between AP activities and TP contents in topsoil.This is consistent with the results of Jing et al. (2020).In an ecosystem with limited P, AP activities may increase as the demand of soil microorganisms for P (Spohn & Kuzyakov 2013).In addition, RDA analysis found that (LAP + NAG) activities in top-and subsoils were significantly positively correlated with NH 4 + -N.This may be because plants allocate more resources for the growth of their above-and underground biomass, increasing inputs from available N sources (Root exudates), which leads to accelerated metabolism of soil microorganisms and the release of more (LAP + NAG) into soils (Zuo et al. 2018).
The effect of shrub encroachment on soil extracellular enzyme stoichiometry and microbial nutrient limitations Soil extracellular enzyme stoichiometry can highlight factors limiting C and N cycling and predict the sustainability of alpine grassland ecosystems (Hu et al. 2021).In our study, the stoichiometry ratios of C-, Nand P-acquisition enzymes ranged from 1.27:1.48:1to 1.54:1.44:1,and the average was 1.36:1.44:1,which deviates from the 1:1:1 ratio at the global average level (Cui et al. 2021;Sinsabaugh et al. 2008).Moreover, E C:N were much lower than the average values of major terrestrial enzymes in the world, with E C:N being 1.41.E C:P and E N:P were much higher than the average values of major terrestrial enzymes in the world, that is, E C:P was 0.62, and E N:P was 0.44 (Sinsabaugh et al. 2009).Our finding is consistent with Luster et al. (2009) that soil extracellular enzyme stoichiometry can greatly change in different ecosystems, especially alpine ecosystems.RDA analysis showed that there was a significant positive correlation between E N:P and TP in topsoil.This may be due to the change of nutrient distribution pattern following shrub encroachment, stimulating the growth of plants and microorganisms, and intensifying the competition for available P between plants and microorganisms.When the N-acquisition enzyme activity become higher, microorganisms need to produce more AP to stimulate the acquisition of P (Kuzyakov & Xu 2013).We found that RDA analysis showed a significant negative correlation between E C:N and E C:P vs C:N and N:P in subsoil.This is consistent with the study of Cui et al. (2018).It may be because the stoichiometric ratio of soil nutrients strongly affects the composition and metabolic activities of microbial community, leading to the shift of soil extracellular enzyme stoichiometric ratios (Cui et al. 2018).
Microbial nutrient limitations play a crucial role in soil C cycling, especially in the decomposition of SOC (Sinsabaugh et al. 2009).In our study, the enzyme vector angle was less than 45°.At the same time, the characteristic figure on microbial nutrient limitations showed that all data points were below the 1:1 line in top-and subsoils (Fig. 3a, b), which indicated that there were microbial C and N limitations in shrub-encroached and non-shrub grasslands.This supports our second hypothesis.These microbial nutrient limitations indicated that C-and N-acquisition enzymes were more active than P-acquisition enzymes in shrub-encroached and non-shrub grasslands.This may be due to the fact that more C-and N-acquisition enzymes are produced during microbial metabolism to acquire C and N resources to sustain their respiration and growth under high water conditions (Wang et al. 2020).In particular, high SWC contents in non-shrub grasslands can accelerate plant growth and increase plant demand for soil C and N elements, and plants compete with microorganisms for C and N, resulting in soil microbial C and N limitations (Manzoni et al. 2012).At the same time, soil nutrient imbalance between shrub-encroached and non-shrub grasslands affects the utilization of nutrients by microorganisms, leading to microbial C and N limitations (Cui et al. 2018).Our results showed that soil microbial C and N limitations in shrubencroached grasslands were significantly lower than in non-shrub grasslands.This indicates that shrub encroachment mitigates soil microbial C and N limitations in alpine grasslands.This supports our third hypothesis.This means that vegetation types indirectly affect microbial nutrient limitations by shifting microbial community structure and soil nutrient supply.This may be because herbaceous plants have higher levels of carbohydrates and cellulose, while lignin contents is lower, therefore litters from herbaceous plants is more easily decomposed by microorganisms than from shrubs (Domenach et al. 1994).However, due to the N fixation in shrubs and arbuscular mycorrhiza, the C and N contents in shrubs are sufficient to maintain the growth of microorganisms (Carrasco et al. 2011).When microbial C and N limitations increased, the metabolism of microbial community changed from growth to maintenance of respiration.It can accelerate the decomposition of SOC by extracellular enzymes, thus promoting the release of soil C (Manzoni et al. 2012).Therefore, microbial C and N limitations may be detrimental to microbial assimilation of SOC and soil C retention.However, shrub encroachment mitigates microbial C and N limitations in alpine grasslands, which may be beneficial to soil C retention in alpine grasslands.The rhizosphere is a key region where plant roots interact with soil microorganisms (Luster et al. 2009).We found that the soil microbial C and N limitations in rhizosphere were greater than in bulk in shrubencroached and non-shrub grasslands.This may be due to the nutrient competition caused by the interaction of plants and microorganisms in the rhizosphere environment (Zhu et al. 2020).The nutrient competition among roots reduces soil nutrient availability and hinders the nutrient acquisition by microorganisms, leading to the increase of microbial C and N limitations in rhizosphere soils (Kuzyakov & Xu 2013).At the same time, our results showed that microbial C limitations in rhizosphere was significantly higher than in bulk in Spiraea alpina grasslands (P < 0.05).It may be due to the higher rhizosphere microbial activity in Spiraea alpina, the faster nutrient turnover rate, which requires more C input from root exudates (Grandy et al. 2007), accelerating the microbial secretion on C-acquisition enzyme, thus enhancing soil microbial C limitations.
Extracellular enzyme stoichiometry is usually affected by soil nutrient status (Sinsabaugh et al. 2009).Using the linear regression analysis and structural equation modeling, we found that soil nutrients and stoichiometry are important factors affecting shifts in microbial C and N limitations following shrub encroachment.This supports our fourth hypothesis.Linear regression analysis showed that the microbial C limitations in topsoil was negatively correlated with SWC contents.This may be due to the hypoxia environment caused by high soil moisture, which inhibits microbial metabolic activities (Grandy et al. 2007), and then affecting microbial demand for soil C. Linear regression analysis further showed that the microbial C limitation was significantly negatively correlated with NH 4 + -N, NO 3 − -N, DOC and SOC in top-and subsoils.This suggests that shifts in microbial C limitations may be driven by changes in soil nutrient levels.The results are consistent with the previous results of Yang et al. (2020).Through generalized linear correlation analysis, we also found that soil C:N and N:P were significantly negatively correlated with microbial C and N limitations, indicating the influence of nutrient stoichiometry on microbial nutrient acquisition.That is, soil nutrient stoichiometric imbalance, especially when soil P contents are much smaller than soil C and N contents, microorganisms will increase their investment in P, and then affecting their acquisition rates of C and N (Gao et al. 2021).In addition, the SEM also showed that soil nutrient changes could explain the shifts in microbial nutrient limitations, especially the increase of soil SOC and NH 4 + -N contents following shrub encroachment mitigated the degree of microbial C and N limitations, respectively.According to the theory of microbial resource allocation.This may be because the increased content of soil SOC and available N following shrub encroachment is enough to maintain the growth of shrubs, and soil microorganisms do not need to secrete more C-and N-acquisition enzymes to acquire C and N elements, thus mitigating microbial C and N limitations (Sun et al. 2021;Zechmeister-Boltenstern et al. 2015).In summary, shifts in microbial nutrient limitations may depend on soil nutrients and nutrient stoichiometry.The soil microbial C and N limitations in alpine grasslands were mitigated by shrub encroachment, which may be beneficial to the retention of SOC in alpine grasslands on the Qinghai-Tibetan Plateau.

Conclusions
The results demonstrated the ratio of log-transformed C-, N-, and P-acquisition enzyme activities ranged from 1.27:1.48:1to 1.54:1.44:1,and the average was 1.36:1.44:1,which deviates from the 1:1:1 at the global average level.Our study revealed that soil microbial C and N limitations were found in shrubencroached and non-encroached alpine grasslands on the Qinghai-Tibetan Plateau.Most microbial C and N limitations were higher in non-shrub grasslands than in shrub-encroached grasslands.It indicates that shrub encroachment mitigated microbial C and N limitations in non-shrub grasslands.At the same time, linear regression further revealed that microbial C and N limitations were directly mediated by soil nutrients and nutrient stoichiometry.Meanwhile, our results showed that microbial C and N limitations were different between bulk and rhizosphere soils in shrub-encroached and non-shrub grasslands.It suggests the influences of vegetation roots on microbial C and N metabolisms.Meanwhile, the SEM show that the increasing contents of SOC and NH 4 + -N following shrub encroachment directly mitigated microbial C and N limitations, respectively.Our results emphasized the change of microbial C and N limitations and their response to environmental factors following shrub encroachment.
) Shrub encroachment massively increased the contents of SOC; (2) Microbial C and N limitations are found in shrub-encroached and non-shrub grasslands; (3) Shrub encroachment Vol:.(1234567890) mitigated microbial C and N limitations; (4) Soil nutrients and nutrient stoichiometry are key factors affecting microbial C and N limitations following shrub encroachment.

Fig. 1
Fig. 1 Redundancy analysis of soil extracellular enzyme activity and environmental factors in top-(a) and subsoils (b).GLB, bulk soils in non-encroached alpine grasslands; GLR, rhizosphere soils in non-encroached alpine grasslands; SAB, bulk soils in Spiraea alpina lands; SAR, rhizosphere soils in

Fig. 3
Fig. 3 Extracellular enzyme stoichiometry of the relative proportions of C to N acquisition versus C to P acquisition (a and b) and the relationships of vector lengths and angles (c and d)

Fig. 5
Fig. 5 The variation of vector angles affected vegetation types, soil depths, and sampling locations (bulk and rhizosphere).Figures (a) and (b) represent the variation trend of vector angle between four vegetation types and locations in top-and subsoil layers, respectively.Figures (c) and (d) represent the variation trend of vector angle of different soil layers in bulk and rhizosphere, respectively.Note: The different letters in columns denote significant differences among four vegetation types within sampling locations (bulk and rhizosphere)

Fig. 8
Fig.8The effect relationships of nutrition limitations of microorganisms with soil physicochemical properties and nutrient ratios.the model of the influences of soil physicochemical properties and nutrient ratios on N limitations of microorganisms (represented by vector angles a and c), and C limitations of microorganisms (represented by vector lengths,