Divergent responses of rhizosphere priming effect to N-P fertilization and plant traits in mandarin orange and alfalfa

1 Introduction

As the most substantial organic carbon reservoir in terrestrial ecosystems, soil harbored over 1500 Pg of organic carbon within the top 1-meter layer and nearly all the nutrients necessary for plant growth (1–4). Given its substantial carbon stocks, even a minor increase in soil organic matter (SOM) decomposition could exert positive feedback on atmospheric CO₂ concentrations, thereby intensifying global warming caused by fossil fuel combustion and land-use change (5). Beyond abiotic factors such as temperature and water availability, rhizosphere processes are being increasingly acknowledged as crucial in regulating SOM decomposition and nutrient cycling (6–8). Plant roots and associated rhizosphere microbes can either stimulate or inhibit native SOM decomposition, a phenomenon known as the rhizosphere priming effect (RPE), which constitutes a vital part of rhizosphere processes (9, 10). Previous investigations carried out in growth chambers or greenhouses have demonstrated that the magnitude of the RPE could vary from a 50% reduction to a 380% enhancement of SOM decomposition when compared with root-free soil (11). These magnitudes indicate that the RPE significantly influences SOM turnover and nutrient cycling. Hence, it is critical to comprehend the potential influencing factors and underlying mechanisms of the RPE for forecasting the global carbon cycle in response to climate change.

Previous research has established that both plant species identity and edaphic factors exert significant control over the direction and magnitude of the rhizosphere priming effect (RPE) (6, 8, 9, 11–13). Among soil properties, nitrogen (N) and phosphorus (P) availability—including that modified by fertilization—have been intimately linked to RPE dynamics (14–16). Compared with N, although the critical demands of plants and microbes for phosphorus (P), and soil P availability is hypothesized to directly or indirectly modulate RPE (15–17), this factor has received comparatively less attention (15, 18). Emerging evidence underscores its potential role: for instance, elevated P availability enhanced RPE in two near-isogenic wheat lines, likely through stimulating biomass production, root exudation, and microbial biomass (19–21). In a contrasting system, P fertilization amplified RPE in root-hair-deficient barley mutants, potentially by exacerbating plant N limitation, which in turn intensified microbial SOM decomposition to scavenge N (15, 22). Collectively, these studies not only establish P availability as a plausible driver of RPE but also suggest that P fertilization may interact synergistically or antagonistically with soil N to shape RPE outcomes.

Beyond abiotic factors, RPE magnitude is strongly contingent on plant species traits, including root morphology (23, 24), root exudation patterns (25), and symbiotic associations with mycorrhizal fungi (26, 27) or N₂-fixing bacteria (28). Notably, legumes, which fix atmospheric N via root-nodule symbioses with rhizobia, may exhibit reduced reliance on soil N and heightened dependence on soil P compared with non-leguminous species. In contrast, grasses depend entirely on soil-derived N and P, acquiring these resources through root exudation of labile substrates or extensive exploration of soil volumes via fibrous root systems. Despite these divergent nutrient acquisition strategies, the interactive effects of P availability and plant functional type (i.e., legume vs. shrub) on RPE—particularly whether legumes and shrubs differ in their sensitivity to N-P-driven RPE modulation—remain poorly understood.

Several hypotheses have been proposed to explain the mechanisms driving the rhizosphere priming effect (RPE) in response to nutrient availability, yet underlying processes remain incompletely understood (10, 18). The plant−microbe competition hypothesis posits that, under extreme nutrient limitation, intensified plant nutrient uptake intensifies competition with soil microbes, thereby suppressing microbial decomposition of soil organic matter (SOM) and resulting in negative RPEs (29, 30). Correspondingly, alleviating nutrient limitation via fertilization tends to increase the RPE magnitude. Contrary to the first hypothesis, the microbial nutrient mining hypothesis explains positive RPEs when soil nutrient availability is moderately low (31). Under this condition, root exudate dynamics can fuel microbial nutrient mining from recalcitrant SOM, leading to positive RPEs under moderate nutrient limitation, as documented in experimental systems (15, 18, 32, 33). As nutrient availability increases further, microbes may preferentially utilize labile exudates rather than catabolizing SOM, consistent with a preferential substrate utilization scenario often associated with reduced RPE responses, a phenomenon also described by the preferential substrate utilization hypothesis (8, 34, 35). Recent meta−analyses also underscore the role of microbial stoichiometry and N limitation in modulating RPE magnitude and direction (36). Despite this progress, the extent to which these nutrient−centered mechanisms apply to phosphorus (P) availability, especially in the context of interactive effects with soil nitrogen status, remains less well characterized and merits further empirical evaluation (37).

Here, we investigated the RPE of mandarin orange (Citrus reticulata Blanco, C₃ plant) and alfalfa (Medicago sativa L., C₃ plant) under three N-P fertilization regimes (two gradient N-P levels and a N-P-free control) at two sampling time points (45 and 90 days after planting) in an environmentally controlled growth chamber. A ¹³C natural abundance tracer method was employed to partition soil-derived CO₂ from root-derived CO₂, with RPE quantified as the difference in soil-derived CO₂ flux between planted treatments and unplanted controls. We measured a suite of leaf and root traits encompassing morphological, chemical, and physiological characteristics. Additionally, key soil properties (pH, dissolved organic C, mineral N, available P, and macroaggregate content) and microbial variables (microbial biomass C, N, and P, and activities of four extracellular enzymes) were measured to identify the factors driving RPE variation. We formulated three core hypotheses (1): RPE is more strongly driven by root traits (particularly chemical and physiological traits) than leaf traits, as roots directly govern rhizodeposition patterns; (2) Fertilization enhances RPE by stimulating plant biomass and functional traits, but this positive effect is partially counterbalanced by its negative impacts on soil and microbial properties; and (3) RPE-induced changes in soil organic carbon (SOC) decomposition are coupled with plant-mediated increases in net N mineralization, but decoupled from net P mobilization.

2 Materials and methods

2.1 Experimental design

The ¹³C natural abundance tracer method was employed to partition soil-derived CO₂ from root-derived CO₂, utilizing the distinct δ¹³C signatures between C₃ plants and C₄-derived soil organic matter (6). The experimental soil was collected from the top 0–20 cm layer of a long-term (>25 years) C4 crop field, rotated with sugarcane and maize, located in Taizhou City, Zhejiang Province, China (E: 121°15′42.79″, N: 28°18′20.62″). The soil was sieved through a 4 mm mesh to homogenize it and remove most roots and large stones. This soil, classified as yellow soil, had a pH of 7.48, with a particle size distribution of 11.7% sand, 77.87% silt, and 10.43% clay. Its basic nutrient properties were as follows: organic C (9.56 mg g^-1 soil), total N (1.55 mg g^-1 soil), total P (0.61 mg g^-1 soil), initial mineral N (NH⁺ + NO₃^-, 93.9 mg kg^-1 soil), and initial extractable P (45.0 mg kg^-1 soil). The initial δ¹³C value of C₄ soil was -18.66 ± 0.30‰ (mean ± SE, n = 5) and remained stable throughout the experiment.

The experiment included two perennial plant species with contrasting growth forms: one-year-old seedlings of mandarin orange (Citrus reticulata Blanco), a perennial woody C₃ species, and alfalfa (Medicago sativa L.), a perennial herbaceous C₃ legume. We selected the two species as our model species primarily to establish a robust contrast in plant functional traits and nutrient acquisition strategies. First, they represent distinct plant functional types with fundamentally different biomass allocation patterns and root architectures, providing an ideal gradient to evaluate trait-mediated RPE. Second, their nitrogen acquisition strategies differ significantly; alfalfa possesses a symbiotic nitrogen-fixing capacity, whereas mandarin orange relies entirely on soil-available nitrogen. Finally, in our specific study region, alfalfa is frequently cultivated as a green manure cover crop within mandarin orange orchards to supply organic nutrients and enhance soil fertility. Therefore, investigating the RPE of these two specific species not only provides theoretical insights into trait-driven carbon dynamics but also reflects a highly relevant local agroecological practice, offering practical guidance for soil carbon and nutrient management in regional intercropping systems. The δ¹³C values of mandarin orange and alfalfa were consistently below -27.00 ‰ and -26.40‰, respectively. The δ¹³C difference (>10‰) between root and soil respiration was sufficiently large to ensure reliable partitioning of the two CO₂ sources.

2.2 Growing system

The 90 polyvinyl chloride (PVC) pots (10 cm diameter × 40 cm height) were used as the experimental containers, with a circulation tube integrated at the bottom to serve dual functions of aeration and CO₂ trapping (3 treatments, two species and one unplanted soil × 3 N-P fertilization gradients × 2 sampling times × 5 replicates). To prevent clogging by fine soil particles, a nylon bag loaded with 50 g acid-washed sand was affixed inside the circulation tube. Before potting, the collected soil was sieved through a 4-mm mesh to homogenize its texture and remove large aggregates, stones, and residual plant materials. Each pot was filled with an equivalent of 3.54 kg oven-dried soil, adjusted to a bulk density of 1.16 g cm^-3 to mimic natural field conditions. All pots were pre-watered to 70% of the field water-holding capacity, which corresponded to a gravimetric soil moisture content of 21%. Soil moisture content was maintained at the target level (70% water-holding capacity) by weighing each pot every two days and replenishing the lost moisture with deionized water. To avoid anaerobic conditions in the rhizosphere, which could alter microbial activity and CO₂emissions, each pot was aerated for 15 minutes every 6 hours by connecting the bottom circulation tube to an aquarium air pump. All pots were arranged in a completely randomized design and re-randomized weekly to eliminate potential positional effects of light, temperature, and air flow within the shelter.

These species were grown under three N-P fertilization regimes and sampled at two time points, 45 and 90 days after planting, in an environmentally controlled growth chamber. Nitrogen (N, supplied as KNO₃) and phosphorus (P, supplied as KH₂PO₄) fertilizers were applied at two gradient rates: low N-P fertilization (L, P4N10: 10 g N m^-2 combined with 4 g P m^-2) and high N-P fertilization (H, P8N20: 20 g N m^-2 combined with 8 g P m^-2). The unfertilized treatment received no N or P addition and was designated as the P0N0 treatment. These N-P addition rates are consistent with typical agricultural fertilization practices in the study region and have been validated as ecologically relevant for simulating projected P enrichment scenarios in agroecosystems (15). To minimize nutrient pulse effects, the total fertilizer dose for each fertilized pot was split into four equal aliquots and applied on days 1, 7, 14, and 28 after planting. At each fertilizer application event, an equivalent volume of deionized water was added to the unfertilized pots to ensure uniform soil moisture across all treatments. No visual symptoms of nutrient deficiency were observed in either mandarin orange seedlings or alfalfa by the end of the experiment, indicating that soil nutrient levels, including both background nutrients and applied fertilizers, were sufficient to support plant growth. This also suggested that N and P additions did not induce co-limitation by other nutrients. Although the alfalfa seeds were not artificially inoculated with rhizobia prior to planting, well-developed root nodules were consistently observed during destructive sampling, confirming the presence of active symbiosis and nitrogen-fixing activity of native rhizobia in the experimental soil.

The experiment was conducted from 1 May 2024 to 1 August 2024 (a total of 90 days) under a rainout shelter (24 m length × 8 m width × 1.8 m height) at the Experimental Station of Taizhou University, Zhejiang Province, China (121.38° E, 28.66° N). The shelter roof was covered with a high-transparency plastic film that excluded precipitation while transmitting over 90% of incident sunlight. The side walls of the shelter were kept open to maintain natural air circulation and ambient temperature fluctuations, ensuring the experimental conditions were representative of the local agroclimatic environment.

2.3 Measurements and calculations

2.3.1 Soil CO₂ efflux

Total belowground CO₂ emissions (encompassing soil and root respiration) from each pot were quantified at 45 and 90 days after planting using a closed-circulation CO₂ trapping system, modified from methods described by Cheng et al. (6) and Lu et al. (15). Briefly, before each measurement, an inlet tube fitted with a porous sponge was inserted 3 cm below the soil surface to ensure uniform air sampling. The soil surface of each pot was then sealed with two-component non-toxic silicone rubber to create a gas-tight environment, preventing atmospheric CO₂ contamination. Gas tightness of the sealed pots was verified by applying negative pressure via the inlet tube while submerging the bottom circulation tube (serving as the outlet) in water; the absence of air bubbles confirmed an airtight seal. To eliminate initial CO₂ residues within the pot headspace, the isolated air was circulated through a soda lime column for 2 hours until no detectable CO₂ remained. Subsequently, CO₂ produced over 48 hours was trapped in 300 mL of 0.5 M NaOH solution, with air circulation activated for 15 minutes every 6 hours to ensure efficient CO₂ absorption. The trapping efficiency of this system has been validated at 99.9% in previous studies (6). Four blank traps (containing NaOH solution but no soil or plant) were run concurrently with each measurement batch to correct for background inorganic carbon (C) contamination from the NaOH solution and handling processes. Following the 48-hour trapping period, the NaOH solution was analyzed for total inorganic C concentration using a multi N/C^® 2000 TOC analyzer (Analytik Jena, Germany). The δ¹³C values of the trapped inorganic C were determined via cavity ring-down spectroscopy (CRDS) equipped with an Automate Module (Picarro G2131-i Analyzer, Picarro Inc., Santa Clara, CA, USA), enabling precise isotopic partitioning of CO₂ sources.

Total soil CO₂ efflux (C_Total) was separated into soil-derived CO₂ (C_Soil) and root-derived CO₂ (C_Root) in planted treatments using a linear two-source mixing model (15, Equations 1 and 2):

(1) CSoil = CTotal (δ13CRoot− δ13CTotal) / (δ13CRoot− δ13CSoil)

(2) CRoot=CTotal − CSoil

where δ¹³C_Total, δ¹³C_Soil, and δ¹³C_Root are the δ¹³C values of total soil CO₂, soil-derived CO₂, and root-derived CO₂ in planted treatments, respectively. The mean δ¹³C value of soil respiration in the unplanted control was used for δ¹³C_Soil in planted treatments. The δ¹³C fractionation between root-derived CO₂ and root bulk tissue was calculated as the difference between their δ¹³C values measured in three additional pots filled with acid-washed sand and C-free Hoagland solution for each species. The mean fractionation factors were -1.43‰ for mandarin orange and -2.12‰ for alfalfa. The fractionation factor of each species was only measured under P0N0 treatment and it was assumed that fertilization had little effect on the δ¹³C fractionation (6, 15). We then determined δ¹³C_Root as the δ¹³C value of root bulk tissue plus the fractionation factor of each species.

The rhizosphere priming effect (RPE) was calculated as the difference in soil-derived CO₂ (C_Soil) between planted and unplanted treatments using Equation 3:

(3) RPE=CSoil (planted)− CSoil (unplanted)

2.3.2 Root-derived MBC and soil-derived MBC

Total MBC (MBC_Total) in each planted pot was separated into soil derived MBC (MBC_Soil) and root-derived MBC (MBC_Root) using a two source mixing model (38; Equations 4 and 5):

(4) MBCSoil =MBCTotal × (δ13CRoot− δ13CTotal) / (δ13CRoot− δ13CSoil)

(5) MBCRoot =MBCTotal − MBCSoil

where δ¹³C_Root is the δ¹³C value of root tissue in planted pots, δ¹³C_Soil is the mean δ¹³C value of soil in unplanted control pots, and δ¹³C_Total is the δ¹³C value of total MBC in planted pots.

The latter was calculated using the following equation (Equation 6):

(6) δ13CTotal =(Cf ×δ13Cf− Cnf ×δ13Cnf)/(Cf − Cnf )

where C_f and C_nf are the total organic C concentrations of fumigated and non-fumigated extracts in planted pots, respectively, and δ¹³C_f and δ¹³C_nf are the δ¹³C values of C_f and C_nf, respectively.

2.3.3 Plant traits

After soil CO₂ efflux measurements, leaf traits analyses were conducted by selecting ten fully expanded, healthy leaves from each planted pot. The remaining aboveground plant tissues were harvested by cutting shoots at the soil surface, then separated into leaf and stem components. All aboveground samples were oven-dried at 65 °C for 48 h to a constant weight for biomass determination. The soil cores were collected from each pot, and visible roots were carefully manually removed via hand-picking. A homogenized fresh soil subsample (500 g) was subsequently taken from each pot for the determination of key soil properties, including soil moisture content, pH value, dissolved organic carbon (DOC), mineral nitrogen (N) concentrations, aggregate stability, microbial biomass carbon (MBC) and nitrogen (MBN), as well as extracellular enzyme activities.

Leaf traits were determined following the method described by Lu et al. (15), with minor modifications as detailed below. Specifically, ten leaves were sampled from five individuals per planted pot, with two fully expanded, healthy leaves (typically the 3rd and 4th leaves from the apex) selected per individual for trait measurements. It should be noted that the limited number of leaf samples might have introduced a certain degree of variability into the results. Leaf area was quantified using a portable scanner (Epson Perfection V33, Seiko Epson Corporation, Tokyo, Japan), and the scanned images were subsequently analyzed with WinFolia software (Regent Instruments Inc., Quebec City, QC, Canada).

Fresh leaf weight and dry leaf weight were measured using an electronic balance with a sensitivity of ±0.1 mg. Leaf carbon (C) and nitrogen (N) concentrations were determined with an elemental analyzer (Elementar III, Elementar Analysensysteme GmbH, Langenselbold, Germany). Based on these measurements, we calculated specific leaf area (SLA, cm² g^-1), leaf dry matter content (LDMC, mg g^-1), leaf N concentration (LNC, mg g^-1), and leaf C/N ratio. Root traits were quantified following the method described by Lu et al. (15), using a root subsample. The root subsample was scanned with a MICROTEK scanner (ScanMaker i800 Plus, Microtek International Inc., Hsinchu, Taiwan, China), and root morphological parameters including average diameter, total length, total surface area, total volume, total root tips, and total root forks were analyzed via WinRhizo software (Regent Instruments Inc., Quebec City, QC, Canada). After scanning, the scanned roots were oven-dried at 65 °C for 48 h and weighed. The remaining root fraction was also oven-dried, weighed, and ground into fine powder for the determination of root C and N concentrations using the same elemental analyzer mentioned above. Root functional traits were further calculated, including mean root diameter (MD, mm), specific root length (SRL, m g^-1), specific root surface area (SRA, cm² g^-1), root tissue density (RTD, g cm^-3), root branching intensity (RBr, n cm^-1), root tip density (R_tips, n g^-1), root N concentration (RNC, mg g^-1), and root C/N ratio.

2.3.4 Soil variables

Soil dissolved organic C (DOC) was measured by extracting 30 g fresh soil with 60 mL 0.05 M K₂SO₄ solution and analyzing total organic C using a multi N/C^® 2000 TOC analyzer (Analytik Jena, Germany). Soil mineral N (NH₄⁺and NO₃^-) was measured by extracting 30 g fresh soil with 60 mL 2 M KCl solution and analyzing the concentrations of NH₄⁺and NO₃^- using a flow injection analyzer (AA3, Bran-Luebbe, Germany). Soil moisture content was determined by oven-drying at 105 ^◦C for 24 h. Soil pH was measured in deionized water (1:2.5 soil: water ratio, shaking for 1 h, and standing for 30 min).

Aggregate size separation was conducted with wet sieving by a series of sieves (2 mm, 0.25 mm, and 0.053 mm) (39). Briefly, 50 g of air-dried soil was capillary-rewetted to make the soil moisture content reach 1.05 times field capacity and was then submerged in water for 5 min on the top sieve (2 mm). After manually moving a series of sieves up and down 3 cm for 50 repetitions within 2 min, the water-stable aggregates on each sieve and soil particles smaller than 0.053 mm remaining in the bucket were separately transferred to glass beakers, oven-dried at 60 ◦C, and weighed. The aggregates on each sieve were defined as large macroaggregates (>2 mm), small macroaggregates (0.25–2 mm), and microaggregates (0.053–0.25 mm), respectively, while the soil particles smaller than 0.053 mm were defined as silt and clay (<0.053 mm). The proportion of water-stable macroaggregates was calculated as the sum of large macroaggregates and small macroaggregates divided by the amount of sieved soil.

2.3.5 Microbial variables

Soil microbial biomass C (MBC), N (MBN), and P (MBP) were measured using the chloroform fumigation-extraction method. Briefly, one subsample (30 g) was extracted with 60 mL 0.05 M K₂SO₄ solution (40–42). Another subsample (30 g) was fumigated with ethanol-free chloroform in the dark at 25 ◦C for 24 h and was then extracted as for the non-fumigated soil. The extracts were analyzed for total organic C and N using a multi N/C^® 2000 TOC analyzer (Analytik Jena, Germany). Microbial biomass C, N, and P were calculated as the difference in total organic C, N and P between fumigated and non-fumigated soil. The conversion factors of 0.45 and 0.54 were used for MBC and MBN, respectively.

Four potential extracellular enzymes involved in the degradation of labile and recalcitrant SOM were assayed according to a fluorimetric method modified from Saiya-Cork et al. (43). β-glucosidase (BG) and β-cellobiohydrolase (CB) are hydrolytic enzymes commonly involved in the degradation of cellulose, which is considered to be labile soil organic C (15). N-acetyl-β-glucosaminidase (NAG) is a hydrolytic enzyme associated with the decomposition of chitin that is considered to be a moderate to fast cycling pool of N (40). In contrast, acid phosphate (ACP) is an enzyme involved in the degradation of soil organic phosphorus (40). Thus, the four potential enzymes could indicate the capacity of microbes to decompose labile and recalcitrant soil organic C and N. Briefly, BG, CB, NAG, and ACP were determined by using 4-methylumbelliferyl β-D-glucopyranoside, 4-methylumbelliferyl β-D-cellobioside, 4-methylumbelliferyl N-acetyl-β-D-glucosaminide, and Disodium p-nitrobenzene phosphate as substrates, respectively. The soil suspensions were prepared by adding 1.5 g fresh soil to 125 mL of buffer (50 mM Tris, pH =7). After adding suspensions, standards, and substrates, the 96-well microplates were incubated in the dark at 25 ◦C for 4 h (BG, CB, and NAG) or 24 h (ACP). Then, sample fluorescence for BG, CB, and NAG was measured with 360 nm excitation and 460 nm emission filters, and sample absorbance for ACP was measured with a 405 nm filter on a microplate reader (Biotek Synergy H1). Enzyme activities were expressed in nmol g^-1 soil h^-1.

2.4 Statistical analyses

A two-way ANOVA was conducted to assess the main and interactive effects of sampling time (Time) and fertilization gradients (F) on plant traits, plant biomass, and RPE. A two-way ANOVA was conducted to assess the main and interactive effects of plant presence (Planting) and fertilization (F) on soil pH, DOC, mineral N, mineral P, macroaggregates, MBC, MBN, MBP, BG, CB, NAG, and APC. Significant differences among means were identified using post hoc Tukey’s HSD test. Pearson correlation analyses and simple linear regressions were used to examine the bivariate relationships between the RPE and the measured variables at each fertilization gradient. To identify the relative importance of these variables in influencing the RPE, a random-forest analysis was conducted at each N level. Differences at P < 0.05 were considered significant, and differences between P > 0.05 and P < 0.1 were considered marginally significant. All statistical analyses were performed with SPSS 20.0 (IBM SPSS Statistics 20, Armonk, USA), except for random-forest analysis, which was conducted in R 3.6.1 (R Core Team, 2019) using the vegan package.

3 Results

3.1 Root-derived CO₂, Soil-derived CO₂, and rhizosphere priming effect

Root-derived CO₂ (i.e., rhizosphere respiration) was significantly affected by plant species, N-P fertilization, sampling time, and their interactions (Figure 1A). Mandarin orange exhibited larger root-derived CO₂ than alfalfa by Day 45, while alfalfa exhibited larger root-derived CO₂ than mandarin orange by Day 90, similar to their shoot biomass (Table 1; Figure 1A). Both low and high N-P fertilization significantly increased root-derived CO₂ of both species by Day 45 and Day 90. Moreover, the lower N-P had no significant influence on the root-derived CO₂ either for mandarin orange or alfalfa by Day 90. However, the high N-P fertilization significantly decreased root-derived CO₂ of alfalfa by Day 90 (Table 1; Figure 1A), which aligned well with shoot biomass and shoot N content in response to P fertilization (Tables 1, 2).

Table 1. Leaf traits and biomass of two species with different N-P fertilization.

Mandarin orange, Citrus reticulata Blanco.; Alfalfa, Medicago sativa L.; SLA, specific leaf area; LDMC, leaf dry matter content; LNC, leaf nitrogen concentration; LCN, leaf C, N ratio; F, Fertilization; S, Species; T, Time.Data are shown as mean ± standard error (n = 5). Different letters represent significant differences among means (Tukey’s test, P < 0.05).

Bold P values indicate statistically significant effects at P < 0.05.

Table 2. Root traits and biomass of two species with different N-P fertilization.

Mandarin orange, Citrus reticulata Blanco; Alfalfa, Medicago sativa L.; MD, root mean diameter; RTD, root tissue density; SRL, specific root length; SRA, specific root area; RBr, root branching intensity; Rtips, root tip density; RNC, root nitrogen concentration; RCN, root C, N ratio; SRR, specific rhizosphere respiration; RB, root biomass; F, Fertilization; S, Species; T, Time.Data are shown as mean ± standard error (n = 5). Different letters represent significant differences among means (Tukey’s test, P < 0.05).

Bold P values indicate statistically significant effects at P < 0.05.

RPE was significantly affected by plant species, sampling time, N-P fertilization, and their interactions (Figure 1B). Both species caused larger soil-derived CO₂ than the unplanted control, indicating positive RPEs, except for the alfalfa treatment with high N-P fertilization by Day 45 (Figure 1B). In general, mandarin orange exhibited a larger RPE than alfalfa by Day 45 (0.076 vs. 0.042 mg C kg^-1 Soil h^-1), while alfalfa exhibited a larger RPE than mandarin orange by Day 90 (0.23 vs. 0.18 mg C kg^-1 Soil h^-1 for low N-P fertilization, 0.13 vs. 0.12 mg C kg^-1 Soil h^-1 for high N-P fertilization (Figure 1B). Otherwise, the N-P fertilization had no significant effect on the RPE of mandarin orange but showed a negative effect on the RPE of alfalfa by Day 45. Meanwhile, the high N-P fertilization significantly decreased the RPE both for mandarin orange and alfalfa by Day 90 (Figure 1B).

3.2 Interspecific biomass and traits

Leaf and root biomass were significantly influenced by plant species, nitrogen-phosphorus (N-P) fertilization, and sampling time. On average, mandarin orange had larger leaf biomass than alfalfa by Day 45, while alfalfa had larger leaf biomass than mandarin orange by Day 90 (Table 1). By day 45, low N-P fertilization generally facilitated leaf biomass with an average increase of 6.7%, 14.2% for mandarin orange and alfalfa, respectively. High N-P fertilization generally facilitated leaf biomass accumulation with an average increase of 12.8%, 19.8% for mandarin orange and alfalfa, respectively. Meanwhile, low N-P fertilization increased the leaf biomass of mandarin orange and alfalfa with an average of 2.6%, 14.0%, high N-P fertilization with the average of 7.3%, 9.8% for mandarin orange and alfalfa by Day 90, respectively. Furthermore, the impact of fertilization on root biomass was species-dependent: nitrogen addition increased leaf biomass by an average of 19% but reduced root biomass in alfalfa by 14%, while slightly increasing root biomass in mandarin orange by 12% (Table 2).

Significant interspecific divergence was also observed across all measured morphological, chemical, and physiological traits (P < 0.001; Tables 2–4). Mandarin orange consistently exhibited significantly higher leaf and root biomass compared to the alfalfa. Specifically, alfalfa displayed significantly higher specific leaf area (SLA) and leaf nitrogen concentration (LNC), but markedly lower leaf dry matter content (LDMC) and leaf C: N ratio (LCN) than mandarin orange across all treatments.

Table 3. Summary of the leaf and root traits under different N-P fertilization treatments after 90 days.

SD, standard deviation.

CV%, the coefficient of variation.

Mandarin orange, Citrus reticulata Blanco; Alfalfa, Medicago sativa L.; SLA, specific leaf area; LDMC, leaf dry matter content; LNC, leaf nitrogen concentration; LCN, leaf C, N ratio. MD, root mean diameter; RTD, root tissue density; SRL, specific root length; SRA, specific root area; RBr, root branching intensity; Rtips, root tip density; RNC, root nitrogen concentration; RCN, root C, N ratio; SRR, specific rhizosphere respiration.

Table 4. Pearson correlation coefficients between the rhizosphere priming effect (RPE) and leaf, root, microbial, and soil properties using all data, and the data under different fertilization treatments, respectively.

LDMC, leaf dry matter content; MD, root mean diameter; Rtips, root tip density; RNC, root nitrogen concentration; RCN, root C, N ratio; SRR, specific rhizosphere respiration; MBC, microbial biomass C; MBN, microbial biomass N; DOC, dissolved organic C.*, 0.01 < P < 0.05; **, P < 0.01.

In addition, sampling time exerted a highly significant influence on all leaf traits and biomass (P < 0.001; Table 3). Comparing Day 45 to Day 90, there was a substantial increase in SLA and leaf biomass for both species. For instance, the SLA of alfalfa increased from an average of 187–198 cm² g ^-1 at Day 45 to 287–307 cm cm² g ^-1 at Day 90, representing a significant shift toward a more acquisitive leaf strategy over the growing season. Otherwise, N-P fertilization significantly affected leaf chemical traits, specifically LDMC (P = 0.002), LNC (P = 0.036), and LCN (P = 0.027), while its effect on SLA and leaf biomass was generally non-significant (P>0.05; Tables 1-3). Increasing fertilization levels led to a reduction in leaf C/N and an increase in LNC across species. A significant three-way interaction between fertilization, species, and time (F×S×T) was observed only for leaf biomass (P = 0.04), highlighting complex combined effects on productivity.

3.3 Soil pH, DOC, mineral N, and macroaggregates

Soil physicochemical properties and aggregate structure were significantly influenced by both N-P fertilization and plant species (Figure 2). N-P fertilization significantly reduced soil pH across all treatments, with a decrease of 0.1–0.2 units observed from the P0N0 to P8N20 levels (P < 0.05) (Figure 2A). Planted treatments exhibited significantly higher soil pH (0.2–0.7 units increase) but markedly lower mineral N concentrations compared to the unplanted control (CK). Among species, mandarin orange maintained a significantly higher soil pH and mineral N level than alfalfa across the fertilization gradient.

Planted treatments exhibited markedly higher DOC concentrations compared to the unplanted control (CK) (Figure 2B). Among species, alfalfa showed the highest soil DOC levels, particularly under the zero-fertilization (P0N0) treatment, where it increased DOC by 113% relative to the CK. Mandarin orange generally maintained lower DOC concentrations than alfalfa across all fertilization levels. Moreover, the response of DOC to N-P fertilization was highly species-specific. For the unplanted control and mandarin orange, DOC concentrations showed an incremental trend with increasing fertilization levels (P8N20 > P4N10 > P0N0). In contrast, N-P addition resulted in a significant reduction in soil DOC for alfalfa, suggesting that fertilization may alter the carbon exudation patterns or stimulate microbial consumption of labile carbon in the alfalfa rhizosphere.

Compared to the unplanted control (CK), all planted treatments significantly reduced soil mineral N levels (P < 0.05), reflecting active nutrient uptake by the plants (Figure 2C). Among the studied species, mandarin orange maintained significantly higher soil mineral N concentrations than alfalfa across all fertilization gradients. Fertilization Effects: As expected, N-P fertilization significantly increased soil mineral N concentrations in a dose-dependent manner (P8N20 > P4N10 > P0N0). This increasing trend was most pronounced in the unplanted control. However, in planted treatments, while mineral N also increased with fertilization, the magnitude of the increase was substantially dampened by plant sequestration. Notably, the effect of fertilization on mineral N was consistent across species, though the absolute concentrations remained lowest in the alfalfa rhizosphere.

The presence of plants generally decreased the proportion of macroaggregates compared to the CK (Figure 2D). Mandarin orange retained a significantly higher macroaggregate fraction than that of alfalfa. Furthermore, heavy fertilization (P8N20) resulted in a significant reduction (average ~6%) in macroaggregate proportions across most species, although a slight compensatory increase was observed specifically in alfalfa.

3.4 Soil microbial biomass C and N, and enzyme activities

Soil microbial biomass and extracellular enzyme activities exhibited significant but divergent responses to plant species and N-P fertilization (Figure 3). Microbial biomass was generally higher in planted soils than in the unplanted control (CK), with a significant effect of planting observed for both MBC and MBN (P < 0.001; Figures 3A, B). Specifically, the presence of mandarin orange led to the highest MBC levels, whereas alfalfa treatments maintained significantly higher MBN compared to the CK and mandarin orange treatments under zero-fertilization (P0N0). N-P fertilization did not significantly affect MBC or MBN, except for a significant N × Planting interaction observed for MBN (P = 0.04), where N addition tended to increase MBN in mandarin orange but showed a neutral or slight decreasing trend in alfalfa.

Planting and nitrogen fertilization had significant interactive effects on the activities of β-glucosidase (BG) and β-cellobiohydrolase (CB) (P < 0.05; Figures 3C, D). In the absence of fertilization (P0N0), both mandarin orange and alfalfa significantly stimulated BG and CB activities compared to the CK. However, N-P fertilization significantly reduced the potential activities of BG, CB, and N-acetyl-β-glucosaminidase (NAG) in both planted treatments in a dose-dependent manner (P0N0 > P4N10 > P8N20; Figures 3C–E).

Phenol oxidase (POX) activity was significantly influenced by the interaction between fertilization and planting (P = 0.014; Figure 3F). While the CK showed no response to fertilization, POX activity in planted treatments was significantly stimulated by N-P addition, especially in mandarin orange and alfalfa soils, where P4N10 and P8N20 treatments maintained markedly higher POX levels compared to their respective P0N0 controls.

3.5 Correlations between the RPE and measured variables

Pearson correlation analysis identified key drivers of interspecific variation in RPE (Table 4). RPE was significantly and positively correlated with specific rhizosphere respiration (SRR), leaf biomass, root nitrogen concentration (RNC), and microbial biomass C (MBC) across treatments (P < 0.01). Otherwise, under N-P limited conditions (N0P0), RPE was negatively related to soil mineral N. Under N-P supplemented conditions (N10P4 and N20P8), RPE showed a significant positive relationship with root mean diameter and a negative relationship with root tip density.

In addition, the linear regression models revealed that leaf biomass, SRR, and root biomass could collectively explain a substantial portion of the interspecific variation in RPE. Specifically, leaf biomass was the dominant predictor under N0 treatments (41%), while RNC became the most influential variable under N10 treatments (44%) (Figure 4).

4 Discussion

In this study, we systematically investigated how nitrogen-phosphorus (N-P) fertilization levels and specific plant functional traits interactively regulate the rhizosphere priming effect (RPE) in mandarin orange and alfalfa. Overall, this study provides three major insights into how plant functional traits and N-P fertilization interactively regulate rhizosphere priming effects (RPE). First, the dominant plant traits controlling RPE shifted with nutrient availability rather than remaining constant. Under low fertilization, leaf biomass emerged as the strongest predictor of RPE, suggesting that plant carbon supply was the primary constraint on rhizosphere microbial activity. Greater leaf biomass likely promoted photosynthate production and belowground carbon allocation, thereby stimulating microbial decomposition of soil organic matter. In contrast, under high fertilization, root nitrogen concentration became the dominant predictor, indicating that substrate quality and root metabolic activity became more important than carbon quantity in regulating microbial responses. Second, N-P fertilization exerted contrasting effects on RPE by simultaneously enhancing plant carbon inputs through improved plant growth and rhizosphere respiration while suppressing microbial decomposition through soil acidification and reduced extracellular enzyme activities. Consequently, the net priming response reflected the balance between plant-mediated stimulation and soil-mediated microbial inhibition. Third, RPE-induced soil organic matter decomposition was more closely associated with plant nitrogen acquisition than phosphorus mobilization, suggesting that microbial nutrient mining in this system remained primarily driven by nitrogen demand despite simultaneous N-P fertilization.

4.1 Species effects on RPE

Consistent with previous studies on plant–soil interactions (15, 38, 44), both mandarin orange and alfalfa induced predominantly positive rhizosphere priming effects (RPEs) throughout the experiment, although alfalfa exhibited a negative RPE under high N-P fertilization at Day 45. Despite this overall similarity, the magnitude and temporal dynamics of RPE differed markedly between the two species, indicating that plant functional type strongly influenced rhizosphere carbon cycling. These interspecific differences are most likely attributable to contrasting growth strategies, carbon allocation patterns, and nutrient acquisition mechanisms.

During the early growth stage (Day 45), the two species adopted distinctly different carbon allocation strategies. Alfalfa, as a fast-growing herbaceous legume, exhibited high specific rhizosphere respiration (SRR), indicating elevated root metabolic activity (Tables 1, 2; Figure 1). Such metabolically active roots are generally associated with greater release of labile carbon compounds into the rhizosphere, thereby stimulating microbial activity and accelerating soil organic matter (SOM) decomposition (44). In contrast, mandarin orange, a perennial woody species, allocated a larger proportion of assimilated carbon to structural biomass accumulation rather than rapid metabolic turnover. This conservative growth strategy likely resulted in a slower but more sustained priming response during the early developmental stage.

As plant growth progressed to Day 90, the mechanisms regulating RPE appeared to shift. The contrasting nutrient acquisition strategies of the two species became increasingly important in determining rhizosphere processes. Mandarin orange maintained continuous biomass accumulation throughout the experiment, which likely increased its demand for soil-derived nutrients (Tables 1, 2). To satisfy this increasing nutrient requirement, greater allocation of photosynthetically derived carbon belowground may have stimulated microbial decomposition of native SOM, consistent with the microbial nutrient mining hypothesis (18, 44, 45). This mechanism suggests that plants invest carbon into the rhizosphere to enhance microbial mineralization and thereby improve nutrient acquisition under increasing nutritional demand.

In contrast, alfalfa exhibited a more complex response because of its symbiotic nitrogen-fixing capacity. Although biological nitrogen fixation reduces dependence on soil mineral nitrogen, it is an energetically expensive process that requires substantial carbon investment to maintain rhizobial activity. Consequently, the combination of high carbon expenditure for nitrogen fixation and the relatively high root nitrogen concentration observed in alfalfa (Table 2) likely created rhizodeposits with high substrate quality. Previous studies have demonstrated that nitrogen-rich rhizodeposits can stimulate microbial activity and induce stronger priming than carbon-rich but nitrogen-poor substrates (15, 44, 46). Therefore, despite its potentially lower reliance on soil nitrogen uptake, alfalfa was still capable of maintaining considerable RPE through the production of metabolically active, nitrogen-rich root tissues and exudates.

Together, these findings demonstrate that species differences in RPE cannot be explained solely by plant functional type or nitrogen-fixing capacity. Instead, they arise from dynamic interactions among plant growth strategy, biomass allocation, root physiological traits, and nutrient availability. This also explains why the simplistic assumption that nitrogen-fixing species necessarily exhibit lower RPE is not universally valid. Rather, both aboveground carbon production and belowground root quality jointly determine microbial responses, and their relative importance changes with nutrient availability.

4.2 The relationships between the RPE and plant traits

The distinct RPEs observed between mandarin orange and alfalfa can be largely explained by their divergent leaf and root traits, which regulate carbon allocation and rhizodeposition strategies. Among the measured traits, leaf biomass emerged as one of the strongest predictors of RPE. Mandarin orange consistently maintained higher leaf biomass than alfalfa during the first 45 days, providing a larger source of photosynthates for belowground carbon allocation. This abundant carbon supply likely stimulated microbial activity and subsequent soil organic matter (SOM) decomposition through enhanced rhizodeposition (27, 47). Although alfalfa exhibited higher specific leaf area (SLA) and leaf nitrogen concentration (LNC), traits generally associated with greater photosynthetic capacity (15), its smaller total leaf biomass appeared to limit the absolute amount of carbon available for rhizosphere inputs. These results suggest that the quantity of photosynthetic tissue may outweigh tissue quality in determining the magnitude of RPE because total biomass ultimately determines the amount of carbon supplied to soil microorganisms (48). Thus, plants with greater biomass can deliver larger absolute carbon inputs belowground, even when their leaf functional traits indicate slightly lower photosynthetic efficiency. Specifically, even when tissue quality indicators like SLA and LNC are relatively high, a limited total plant biomass may act as a bottleneck, constraining the overall carbon allocation downward to the rhizosphere. Conversely, plants characterized by a larger biomass—even if exhibiting slightly lower tissue quality—possess a greater capacity to deliver larger absolute amounts of total carbon inputs belowground, thereby exerting a more pronounced influence on soil microbial activities and associated priming processes.

Beyond aboveground traits, root physiological characteristics provided further insight into the interspecific differences in RPE. Consistent with our hypothesis, specific rhizosphere respiration (SRR) and root nitrogen concentration (RNC) were strong predictors of RPE. Alfalfa exhibited a significantly higher RNC than mandarin orange (Tables 2–4). Roots with high nitrogen concentrations generally exhibit greater metabolic activity (49) and release nitrogen-rich rhizodeposits that are readily utilized by soil microorganisms (28). These high-quality substrates can stimulate microbial decomposition more effectively than lower-quality exudates (46), potentially explaining why alfalfa induced substantial RPE despite its lower total biomass. This interpretation is consistent with previous studies reporting that legumes frequently generate stronger priming effects through the release of nitrogen-rich rhizodeposits (6, 44). Therefore, while leaf biomass primarily reflects the quantity of carbon entering the rhizosphere, root physiological traits determine the quality of substrates available for microbial metabolism.

However, root morphology told a different story. Although mandarin orange generally possessed thicker roots (greater mean root diameter) than the finer root system of alfalfa, these morphological traits were much less predictive of RPE than physiological traits. This agrees with previous studies showing that root architecture alone, including specific root length and root tip density, does not necessarily correspond to rhizodeposition rates because carbon release is concentrated in metabolically active regions such as root tips and young roots rather than being uniformly distributed throughout the root system (27, 45, 50, 51). Consequently, root morphology appears to influence rhizosphere priming mainly through its indirect effects on carbon allocation and root physiology rather than acting as a direct driver of microbial activity.

Overall, our results suggest that variation in RPE between mandarin orange and alfalfa reflects a trade-off between “source strength” and “substrate quality”. Mandarin orange primarily stimulated RPE through greater carbon inputs associated with its larger leaf biomass, whereas alfalfa compensated for its smaller biomass by supplying metabolically active, nitrogen-rich root tissues and rhizodeposits that provided high-quality substrates for microbial growth. Importantly, these two mechanisms should be regarded as complementary rather than mutually exclusive. The relative importance of leaf biomass and root nitrogen concentration varied under different environmental conditions (Figure 4), indicating that plant trait regulation of RPE is dynamic rather than fixed. This finding highlights that physiological traits generally provide stronger mechanistic insight into rhizosphere priming than morphological traits alone and provides a conceptual basis for understanding why the dominant drivers of RPE shift under different nutrient regimes, a mechanism that is discussed further in the following section.

4.3 The interactive effects of N-P fertilization and plant traits on RPE

N-P fertilization also differentially affected the RPE of mandarin orange and alfalfa demonstrating that nutrient availability interacts with species-specific traits to regulate soil carbon dynamics, and these differences became more evident across the fertilization gradient. Under low N-P availability, RPE was mainly governed by plant carbon acquisition and microbial nutrient demand, whereas increasing N-P inputs progressively shifted the rhizosphere environment toward greater nutrient availability. Consequently, both plant carbon allocation and microbial decomposition processes varied with fertilization intensity, resulting in species-specific priming responses.

For mandarin orange, N-P fertilization generally increased leaf biomass and leaf nitrogen concentration (LNC) (Tables 1, 2), suggesting enhanced photosynthetic capacity and greater potential carbon assimilation. Under low fertilization, mandarin orange likely relied more strongly on rhizosphere priming to mobilize nutrients because plant nutrient demand exceeded the available inorganic nutrient supply. In this case, greater photosynthate production and belowground carbon allocation could stimulate microbial activity and promote soil organic matter (SOM) decomposition. However, as fertilization intensity increased, the additional carbon assimilated by mandarin orange did not lead to a proportional increase in RPE. This was likely because the positive effect of increased plant carbon input was progressively offset by fertilization-induced changes in soil properties.

Specifically, N-P fertilization significantly decreased soil pH while increasing mineral nitrogen availability (Figure 2). Higher inorganic nutrient availability can reduce microbial dependence on SOM-derived nutrients, thereby weakening microbial nutrient mining. Consistent with this interpretation, the activities of β-glucosidase and β-cellobiohydrolase, two enzymes involved in carbon acquisition and SOM decomposition, declined markedly following fertilization (Figure 3). Reduced extracellular enzyme production suggests that microorganisms invested less energy in decomposing native SOM once mineral nutrients became more readily available (10, 16, 52). Therefore, in mandarin orange, fertilization enhanced the potential for plant-derived carbon inputs but simultaneously constrained microbial decomposition capacity, resulting in a weakened or non-proportional RPE response under higher fertilization. Moreover, N-P fertilization-induced soil acidification may act as a strong environmental filter that shapes microbial community structure and functional traits. Prolonged acidification typically shifts the soil microbiome from bacteria-dominated to fungi-dominated or specifically suppresses copiotrophic microbial groups (such as Proteobacteria and Bacteroidetes) that rely on high pH and labile carbon, while favoring acid-tolerant, oligotrophic groups (such as Acidobacteria) (53). This structural shift directly alters the functional guild composition, downregulating the expression of key extracellular enzymes responsible for recalcitrant organic matter decomposition (12). Consequently, this microbial trade-off constrains the overall capacity of the soil microbiota to respond to enhanced rhizodeposition, ultimately dampening the magnitude of the rhizosphere priming effect (12, 54).

In contrast, alfalfa exhibited a different response pattern along the N-P fertilization gradient because of its legume-specific nutrient acquisition strategy. Under low N-P availability, its symbiotic nitrogen-fixing capacity and metabolically active root system likely enabled alfalfa to maintain substantial rhizosphere activity despite limited soil nutrient supply. With increasing fertilization, alfalfa likely shifted carbon allocation away from root growth and toward aboveground tissue maintenance or reduced its reliance on costly root systems when mineral N became abundant. The correlation analysis also revealed that under high N-P conditions (N20P8), root N concentration (RNC) became a critical predictor of RPE (Figure 4B). Alfalfa consistently maintained a higher RNC compared to Mandarin orange (Table 2). This N-rich root litter promotes microbial growth and turnover, but coupled with the direct inhibition of N-mining enzymes by fertilization, the net effect on SOM decomposition remains complex. Under low N-P conditions (N10P4), leaf biomass-a proxy for carbon supply-was the dominant driver of RPE (Figure 4A). Mandarin orange, with its consistently larger leaf biomass, exerted a strong priming effect through massive carbon input. However, under high N-P (N20P8) fertilized conditions, the importance of leaf biomass declined while root traits like RNC gained prominence (Figure 4B). This aligns with the “allocation shift” hypothesis (55), where plants reduce the proportion of photosynthetic carbon allocated to root exudates when soil nutrients are sufficient. Consequently, the rhizosphere microbial community likely switches from relying on fresh, labile exudates to utilizing N-rich root litter. This explains why alfalfa, despite its smaller biomass compared to mandarin orange, maintained a distinct influence on RPE through its high-quality, nitrogen-rich root tissues.

Overall, the influence of N-P fertilization on RPE depended on the balance between plant-mediated stimulation and soil-mediated microbial inhibition, and this balance shifted along the fertilization gradient. Under low fertilization, microbial nutrient demand remained relatively high, allowing plant-derived carbon inputs to be more closely coupled with SOM decomposition. Under high fertilization, increased nutrient availability, soil acidification, and suppressed extracellular enzyme activities weakened microbial investment in SOM decomposition, thereby reducing the efficiency with which additional plant carbon inputs were translated into stronger RPE. Thus, although mandarin orange and alfalfa adopted contrasting strategies for carbon acquisition and nutrient use, their RPE responses were jointly regulated by plant functional traits and fertilization-induced changes in soil microbial processes. Ultimately, while N-P fertilization improved specific plant traits—such as increasing LNC in both species and enhancing specific rhizosphere respiration (SRR) in mandarin orange - these indirect benefits were seemingly outweighed by the direct suppression of microbial decomposition. The interaction between species identity and N availability thus determines RPE magnitude: Mandarin orange drives priming through sheer biomass and carbon supply, while Alfalfa influences it through tissue quality and morphological plasticity, with both pathways being modulated by the suppressive effects of inorganic N on microbial enzyme activity.

4.4 Implications and future perspectives

The present study provides important implications for improving predictions of soil carbon dynamics under changing nutrient conditions. Our results demonstrate that the plant functional traits regulating rhizosphere priming are context dependent rather than fixed. The dominant predictors of RPE shifted from leaf biomass under low N-P fertilization to root nitrogen concentration under high fertilization, indicating that the relative importance of carbon source strength and substrate quality changes with nutrient availability. This finding suggests that trait-based models of soil carbon cycling should incorporate dynamic plant trait responses to nutrient availability rather than relying on static trait values. Furthermore, our results show that plant traits alone cannot fully explain variation in RPE because fertilization simultaneously modifies soil physicochemical conditions and microbial functioning. Consequently, accurate prediction of soil carbon turnover requires integrating plant carbon allocation, soil nutrient availability, and microbial processes within a unified framework.

Several limitations should also be acknowledged. First, although plant functional traits explained a substantial proportion of the observed variation in RPE, belowground carbon allocation was inferred indirectly rather than quantified through direct measurements of root exudate quantity and chemical composition. Second, microbial regulation was evaluated primarily through extracellular enzyme activities, whereas microbial community composition and functional characteristics were not directly determined. Finally, this study was conducted under controlled conditions using two representative plant species, which allowed the mechanisms underlying trait-mediated RPE to be isolated but may not fully capture the complexity of natural ecosystems.

Future research should integrate direct measurements of root exudate chemistry with microbial community analyses, such as metagenomics or metatranscriptomics, to better resolve the mechanistic links between plant functional traits and microbial decomposition processes. Extending trait-based studies across broader plant functional groups and validating these mechanisms under long-term field fertilization experiments will further improve our understanding of rhizosphere priming under realistic environmental conditions. Such efforts will provide a stronger mechanistic basis for predicting soil organic carbon dynamics under future nutrient enrichment and global environmental change.

Correction note

A correction has been made to this article. Details can be found at: 10.3389/fsoil.2026.1959770.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Dener Oliveira, Universidade Federal de Viçosa, Brazil

Reviewed by: Carolin Córdova, Laboratorio de Fisiología Vegetal Aplicada. Universidad de Aysén., Chile

Tianfu Han, Zhengzhou University, China