Functional traits mediate the effects of soil property on the growth performance of Ardisia gigantifolia cuttings

Abstract

Background

Understanding how soil types influence functional traits and growth performance is essential for effective plant cultivation and management. Ardisia gigantifolia (A. gigantifolia) is a rare traditional medicinal plant in China; however, the mechanisms underlying its growth responses remain poorly understood. This study utilized a pot experiment to investigate the effects of different soil types (humus, loam, and sand) on the functional traits and growth performance of one-year-old A. gigantifolia cuttings and to characterize the interrelationships among these factors.

Results

Compared to sand, humus soil significantly increased the height, as well as root and stem biomass of A. gigantifolia cuttings. In humus, the specific leaf area, leaf water content, root average diameter, root volume, stem length, and stem average diameter were significantly higher, whereas leaf dry matter content and the tissue density of leaves, roots, and stems were lower. Additionally, loam significantly improved the photosynthetic and pigment traits of the cuttings. Results from RDA and PLS-SEM demonstrated that the growth performance of A. gigantifolia cuttings is primarily and directly governed by root, stem, and leaf morphological traits, which also mediate soil-driven regulation. Root morphological traits, in particular, showed the greatest direct and indirect impacts, highlighting them as key potential indicators for predicting the performance of A. gigantifolia cuttings.

Conclusion

This study demonstrates that the growth performance and functional traits of A. gigantifolia cuttings are modulated by soil properties, while organ morphological traits serve as reliable predictors of plant growth performance. Our findings provide a scientific foundation for the conservation of wild A. gigantifolia resources and the advancement of its sustainable industrial development.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12870-026-09323-x.

Background

Ardisia gigantifolia Stapf (A. gigantifolia), a small shrub in the genus Ardisia (Primulaceae), is primarily distributed in tropical and subtropical regions of southern China, Vietnam, Thailand, and Laos [1]. A. gigantifolia possesses high medicinal value; in China, it is commonly known as “Zoumatai” and is frequently used to treat gout and traumatic injuries, as well as in health foods [2, 3]. Notably, various saponins extracted from A. gigantifolia have been proven to significantly inhibit the in vitro proliferation of cancer cell lines, including breast [4], gastric [5], and colorectal cancers [6]. Increasing demand has led to the over-exploitation of wild populations, resulting in a sharp decline in their numbers. Therefore, it is urgent to improve artificial cultivation techniques and related research.

Soil is a primary source of water and nutrients required for plant development and is a key environmental factor influencing plant performance. The interaction between different parent materials and climates produces soils with distinct physicochemical properties (e.g., pH, moisture, and nutrient content), which further filter and shape the plants growing in those soils. It has been reported that the growth heights of Melissa officinalis, Taraxacum officinalis, and Ocimum basilicum are pH-dependent; neutral soils facilitate photosynthesis, whereas excessively high pH induces stomatal closure and reduces photosynthetic rates [7]. Additionally, soil water deficiency inhibits cell elongation in Ardisia gigantifolia, significantly reducing plant height and basal diameter, and leading to decreased biomass [8]. As shown in multiple global meta-analyses, soil fertility (especially nitrogen) exerts significant and widespread effects on leaf, root, and stem traits [9–12], making it a robust predictor of functional trait variation [13]. Generally, high-fertility soils tend to increase leaf nutrient content, leaf area, specific leaf area (SLA), water-use efficiency, and photosynthesis, while decreasing leaf dry matter content (LDMC), leaf thickness, and tissue density, thereby promoting growth; the opposite is observed under low-fertility conditions [10, 11, 14]. The effects of high fertility on roots primarily involve increased root biomass and nutrient content, but decreased specific root length (SRL) and root tissue density [9, 15]. However, due to the multidimensional nature of root traits and the complexity of the soil environment, the influence of soil on root traits is frequently overlooked in many contexts [16].

A. gigantifolia typically grows in shady, humid areas under dense forests, primarily colonizing sandy and humus-rich soils, and exhibits slow growth. Individuals suitable for medicinal use usually require more than three years of growth [17]. Current research on A. gigantifolia has focused on seedlings; however, their slow and uneven growth rates limit modern commercial cultivation. Cutting is a distinctive vegetative propagation method characterized by rapid growth and the reliable maintenance of superior parental traits. Research has demonstrated that, compared to seed-derived seedlings, the leaf and root traits of cuttings exhibit greater plasticity in response to nutrient and water availability [18]. The genetic stability of cuttings may enhance the consistency of their responses to soil conditions; furthermore, distinct rooting modes can lead to modifications in root architecture, which in turn determine water and nutrient acquisition strategies and ultimately influence overall growth performance. Nevertheless, there remains a lack of understanding regarding how different soils regulate the growth and functional traits of A. gigantifolia cuttings.

The plant economics spectrum (PES) theory describes variations in plant traits across multiple dimensions and serves as a comprehensive framework for assessing plant performance and environmental adaptation [19–22]. For specific populations, the plastic shifts in functional traits reflect adaptive strategies to local ecosystem changes [23, 24]. For instance, in response to water deficit, stomatal closure leads to an increase in leaf thickness (LT) [25]. Similarly, wood tissue density explains a substantial portion (> 80%) of the variation in the growth-mortality trade-off among tree seedlings [26]. Furthermore, studies have demonstrated that leaf traits (e.g., leaf area, specific leaf area, net photosynthetic rate, and chlorophyll content) and root traits (e.g., root nitrogen content and biomass) are highly sensitive to soil nitrogen availability [27–30]. On the other hand, utilizing functional traits to predict plant growth and productivity remains challenging, as environmental complexity often results in diverse trade-off patterns among traits [31, 32]. However, under specific conditions or within controlled environments, certain traits exert a profound influence on growth performance [33]. For example, a study conducted in the transition zone between the Loam Plateau and the desert in Northwest China found that the net assimilation rate and leaf area ratio of eight xerophytic shrub species drove variations in their growth rates [34]. Another study from the Eastern Amazon indicated that leaf photosynthesis and hydraulic traits were superior predictors of plant growth [35]. These findings suggest that the functional traits predictive of growth performance are not universal but are influenced by local environmental factors. Therefore, understanding the regulation of A. gigantifolia functional traits by complex soil characteristics and their influence on growth performance warrants further investigation. This will provide theoretical support for improving the cultivation practices of A. gigantifolia seedlings, with a specific focus on cuttings.

Based on this background, the present study focused on A. gigantifolia cuttings to quantify their growth performance and the traits of their leaves, roots and stems across different soil types, thereby exploring the effects of various cultivation substrates. We addressed two primary questions: (1) What are the effects of different soil types on the functional traits and growth of A. gigantifolia cuttings? (2) How do functional traits mediate the soil-driven regulation of growth in these cuttings? We hypothesized that: (a) differences in the physicochemical properties of various cultivation substrates determine the plasticity of A. gigantifolia functional traits; and (b) leaf and root traits are the primary mediators of the effects of soil on plant growth and biomass accumulation.

Materials and methods

Plant materials and experimental design

The experiment was conducted in June 2025 at germplasm resource nursery of the Guangxi Institute of Botany, Chinese Academy of Sciences (25° 04′ 18.1″ N, 110° 18′ 11.8″ E). Approximately 15 perennial A. gigantifolia plants, which were propagated via tissue culture (using stem segments as explants) and transplanted by our team in 2017, served as the donor plants. Due to the characteristic single-stem growth habit of A. gigantifolia, one-year-old branches were harvested from the middle section of these donor plants and partitioned into cuttings approximately 14 cm in length, each containing axillary buds. To mitigate potential maternal effects, cuttings from the same donor plant were evenly and systematically allocated to different experimental treatments. Before planting, the cuttings were placed in a shaded and well-ventilated area to allow the cut surfaces to air-dry.

The study utilized a completely randomized design with three substrate treatments: humus, Loam, and river sand. On clear days, the nursery maintained a light intensity of approximately 10,000 lx, an ambient temperature of 20–25 °C, and a relative humidity of about 65%. Rectangular pots (100 × 25 × 25 cm) with identical drainage capacities were used for planting. The humus was sourced from Zhaoqing Hengzhan Agricultural Science and Technology Co., Ltd., while the loam and sand were obtained from Yuxin Material Store (Guilin, China); none of the substrates underwent sterilization. Prior to planting, all substrates were sieved through a 10 mm mesh and filled into the pots to a level 5 cm below the rim, with each pot containing approximately 50 L of soil. Cuttings of uniform size were inserted into the substrates at an angle of approximately 60° to the ground, with a planting depth of about 3 cm. Each treatment consisted of three replicates, with five cuttings per replicate evenly spaced within a single pot, totaling 9 pots and 45 cuttings. Following planting, an equal volume of water was applied daily as required. One month later, all cuttings had successfully survived, with no root interference or mutual shading observed among the seedlings. The seedlings were harvested five months after planting. Prior to planting, soil samples were collected from each treatment to determine their physicochemical properties.

Soil physicochemical properties

Soil moisture (SM) was determined by drying the soil at 105 °C to a constant weight and was calculated as the ratio of the weight loss (fresh weight minus dry weight) to the initial fresh weight. Soil pH was measured using a pH meter in a 1:2.5 soil-to-distilled water suspension after a 30-minute equilibrium period. Soil organic matter (SOM) was determined using the potassium dichromate oxidation method followed by titration with ferrous sulfate. Total nitrogen (TN) was measured via sulfuric acid digestion and hydrochloric acid titration. Total phosphorus (TP) was determined using the molybdenum-antimony colorimetric method after digestion with sulfuric acid (H₂SO₄) and perchloric acid (HClO₄). Alkali-hydrolyzable nitrogen (AN) was determined by the alkaline-diffusion method, in which NH₃ released by NaOH hydrolysis was absorbed by boric acid (H₃BO₃) and then titrated with hydrochloric acid. Available phosphorus (AP) was extracted using NaHCO₃ (for soils with pH > 6.5) or NH₄F-HCl (for soils with pH < 6.5) and subsequently measured using the molybdenum-antimony colorimetric method. Total potassium (TK) was measured via flame photometry after the soil was digested with a mixture of nitric, perchloric, and hydrofluoric acids (HNO₃ : HClO₄ : HF = 3:1:5, v: v:v). Available potassium (AK) was extracted with ammonium acetate (NH₄OAc) and determined using flame photometry.

Seedling growth and biomass

Seedling height was measured as the vertical distance from the germination point to the apical growth point using a ruler. Stem diameter was measured at 1.0 cm above the base using a digital vernier caliper. Upon harvesting, the seedlings were separated into leaves, stems, and roots, and their fresh weights were recorded. All components were heat-inactivated at 90 °C for two hours, then oven-dried at 60 °C to a constant weight. Total seedling biomass was calculated as the sum of the weights of the leaves, stems, and roots.

Leaf traits

At harvest, all leaves from each seedling were collected. To minimize errors caused by transpirational water loss, leaf thickness (LT) and leaf fresh mass (LFM) were measured immediately following excision. LT was determined using a micrometer screw gauge (0.01 mm precision) by taking the average of three measurements along the primary vein, avoiding major lateral veins. Leaf area (LA), leaf length (LL), and leaf width (LW) were determined by scanning the leaves and analyzing the images with the ImageJ2 software image processing package (Fiji) (National Institutes of Health). Subsequently, leaf dry mass (LDM) was measured using the same electronic balance after the samples were oven-dried at 60 °C to a constant weight. Based on these primary measurements, several functional traits were calculated: specific leaf area (SLA) was calculated as the ratio of LA to LDM; leaf dry matter content (LDMC) was determined as the ratio of LDM to LFM; and leaf water content (LWC) was calculated as (LFM - LDM) / LFM. Leaf volume (LV) was estimated as the product of LA and LT, while leaf tissue density (LTD) was calculated by dividing LDM by LV. At the whole-plant level, the leaf number (LN) per plant was recorded during sampling [15].

Prior to harvest, gas exchange parameters were measured on clear days between 9:00 AM and 11:00 AM using a Li-6800 portable photosynthesis system (LI-COR, Lincoln, NE, USA). Measurements were performed on fully expanded mature leaves from the middle section of the seedlings. Based on light-response curves, A. gigantifolia cuttings exhibited peak photosynthetic performance at an irradiance of 200 µmol m⁻² s⁻¹. Therefore, the photosynthetic photon flux density (PPFD) was set at 200 µmol m⁻² s⁻¹, and the ambient CO₂ concentration was maintained at 400 µmol (CO₂) mol⁻¹. The parameters recorded included net photosynthetic rate (Pn), stomatal conductance (gs), transpiration rate (Tr), and intercellular CO₂ concentration (Ci). Leaf water use efficiency (WUE) and Rubisco carboxylation efficiency (CE) were calculated as the ratios of Pn: Tr and Pn: Ci, respectively [36]. Additionally, leaf chlorophyll a (Chl a), chlorophyll b (Chl b), total chlorophyll (Chl), and carotenoids (Car) were extracted using 80% (v/v) acetone under dark conditions. The samples were agitated on an orbital shaker for 24 h, after which the supernatant was collected. The absorbance of the supernatant was measured at 470, 646, and 663 nm using a Libra S22 UV/Visible spectrophotometer (Biochrom Ltd., Cambridge, UK) or a microplate reader. Pigment concentrations were calculated according to the equations described by Lichtenthaler [37]:

1

2

3

4

where A₄₇₀, A₆₄₆, and A₆₆₃ represent the absorbance values of the supernatant at 470, 646, and 663 nm, respectively.

Root traits

Following seedling harvest, the root systems were scanned, and morphological traits were analyzed using the Rhizo Vision Explorer software (v. 2.0.3). The quantified traits included the number of root tips (Tips), total root length (RL), root average diameter (RAD), root volume (RV), and root surface area (RSA). Derived traits were subsequently calculated: specific root length (SRL) and specific root surface area (SRSA) were determined as the ratios of RL and RSA to root dry mass, respectively. Root tissue density (RTD) was calculated as the ratio of root dry mass to RV.

Stem traits

Stem length (SL) was measured using a ruler. The diameters of the stems were measured at the top, middle, and base using a digital caliper, and the average of these three measurements was defined as the stem average diameter (SAD). Stem volume (SV) was then calculated based on their respective lengths and diameters. Finally, stem tissue density (STD) was determined as the ratio of dry mass to volume for each component.

Statistical analysis

Data were organized and pre-processed using Microsoft Excel. One-way analysis of variance (ANOVA) was performed in SPSS (v.26.0) to compare the effects of different substrate treatments on the growth and traits of A. gigantifolia cuttings. Significant differences between treatment means were determined using Duncan’s multiple range test at the p < 0.05 level. Redundancy analysis (RDA) and hierarchical partitioning (HP) were conducted using the “rdacca.hp” package in R (v.4.5.1) to investigate the influence of various traits on the growth performance of A. gigantifolia cuttings. HP is a global decomposition method based on all variable subsets; it quantifies the importance of each variable by calculating its independent and shared contributions across different subset combinations, effectively mitigating the impacts of multicollinearity [38]. A random forest model was constructed in R (v.4.5.1) using the ‘randomForest’ package to rank the importance of soil physicochemical properties in determining the growth and trait expression of the cuttings. Furthermore, a formative partial least squares structural equation model (PLS-SEM) was constructed using the ‘plspm’ package in R (v.4.5.1) to explore the potential pathways and mechanisms by which soil properties regulate the functional traits and growth of A. gigantifolia cuttings. Indicators were categorized into several latent variables: soil physicochemical properties (SM, pH, SOM, TN, AN, TP, AP, TK, and AK), leaf morphological traits (LN, LA, LL, LW, LFM, LDM, LT, SLA, LDMC, LWC, LV, and LTD), leaf physiological traits (Pn, gs, Ci, Tr, WUE, CE, Chl a, Chl b, Chl, Car, Chl a/Chl b, and Car/Chl), root morphological traits (Tips, RL, RAD, RV, RSA, SRL, SRSA, and RTD), stem morphological traits (SL, SAD, SV, and STD), and growth performance (height, diameter, leaf fresh mass, root fresh mass, stem fresh mass, seedling fresh mass, leaf dry mass, root dry mass, stem dry mass, seedling dry mass, and root-to-shoot ratio). Prior to model construction, multicollinearity was assessed, and dimension reduction was performed for the explanatory variables within each latent variable. To retain maximum original data characteristics, principal components (PCs) with eigenvalues > 0.5 were selected to construct the latent variables, ensuring a cumulative explained variance of > 90%. Subsequently, to prevent potential biases from multicollinearity on the overall model, collinearity analysis was conducted for all soil- and trait-related PCs. PCs with high collinearity were sequentially removed based on their Variance Inflation Factor (VIF) until all remaining PCs fell within a reasonable range. Finally, the overall model fit was evaluated using the coefficient of determination (R²) for the explained latent variables and the goodness-of-fit (GOF) index.

Results

Soil physicochemical properties

One-way ANOVA revealed significant variations among the different soil types. Humus soil exhibited the highest levels of SM, SOM, TN, AN, and AP, while sand showed the highest pH and TK values (p < 0.05). Furthermore, the SM, SOM, TN, AN, TP, and AK in sand were significantly lower than those in humus soil and Loam. The lowest AP content was recorded in Loam (p < 0.05; Fig. 1).

Fig. 1.

Differences in the physicochemical properties of various soils. Different lowercase letters indicate significant variations in the physicochemical properties of different substrates

Variations in growth and development of A. gigantifolia seedlings across different soil

The height of A. gigantifolia seedlings cultivated in humus soil (14.12 ± 0.66) and Loam (12.87 ± 0.49) was significantly higher than that of seedlings grown in sand (8.62 ± 0.99; Fig. 2b). Fresh root mass in the humus treatment was significantly greater than in Loam and sand (p < 0.05); similarly, fresh stem mass was significantly higher in humus than in sand (p < 0.05; Fig. 2c). Although there were no significant differences in total dry mass across the different substrates, the root-to-shoot ratio of seedlings grown in humus was significantly higher than that of seedlings in Loam and sand (p < 0.05; Fig. 2d).

Fig. 2.

Growth status and biomass of A. gigantifolia cuttings in different soils. (a) Growth status, (b) seedling height and diameter, (c) fresh biomass, and (d) dry biomass. Different lowercase letters indicate significant differences in growth indices among soil treatments. * indicate significant differences in seedling growth indices between humus soil and other treatments at the 0.05 level

Variations in leaf, root and stem traits across different soil type

Leaf SLA and LWC of A. gigantifolia in humus were significantly higher than those in Loam and sand, whereas LDMC and LTD were significantly lower (Fig. 3). Across the different substrates, leaf Pn, gs, Tr, WUE, and CE reached their highest values in Loam, followed by humus, and were lowest in sand (p < 0.05); in contrast, Ci exhibited the opposite trend (Fig. 3). Similarly, the concentrations of leaf pigments (including Chl a, Chl b, Chl, and Car) were highest in Loam, followed by humus, and lowest in sand (p < 0.05), aligning with the observed trends in photosynthetic parameters. Furthermore, the leaf Chl a/Chl b ratio in humus was significantly lower than in Loam and sand, while the Car/Chl ratio in sand was significantly higher than those in humus and Loam (Fig. 3). RL, RSA, and SRSA of A. gigantifolia seedlings in humus were significantly higher than those in sand, while their RAD and RV were significantly greater than those in Loam. In loam, SRL and SRSA were significantly higher than the values recorded in sand. Additionally, RTD in both humus and Loam was significantly lower than that in sand (p < 0.05; Fig. 3). SAD and SV of seedlings grown in Loam were significantly smaller than those in humus, whereas their SL was significantly higher than that of seedlings in sand. For seedlings in sand, SL and SV were significantly lower than those in humus, but STD was significantly higher than in the humus treatment (p < 0.05; Fig. 3).

Fig. 3.

Effects of different soil types on leaf, root and stem traits. LN: Leaf number, LA: Leaf area, LL: Leaf length, LW: Leaf width, LFM: Leaf fresh mass, LDM: Leaf dry mass, LT: Leaf thickness, SLA: Specific leaf area, LDMC: Leaf dry matter content, LWC: Leaf water content, LV: Leaf volume, LTD: Leaf tissue density, Pn: Net photosynthetic rate, gs: Stomatal conductance, Ci: Intercellular carbon dioxide concentration, Tr: Transpiration rate, WUE: Water use efficiency, CE: Carboxylation efficiency, Chl a: Chlorophyll a, Chl b: Chlorophyll b, Chl: Chlorophyll, Car: Carotenoid, Tips: Root tip number, RL: Root length, RAD, Root average diameter, RV: Root volume, RSA: Root surface area, SRL: Specific root length, SRSA: Specific root surface area, RTD: Root tissue density, SL: Stem length, SAD: Stem average diameter, SV: Stem volume, STD: Stem tissue density. Different lowercase letters indicate significant differences among substrates based on one-way ANOVA (p < 0.05)

Impact of different traits on the growth of A. gigantifolia cuttings

The first two axes of the RDA for leaf morphological, leaf physiological, root morphological, and stem morphological traits explained 91.95%, 85.50%, 92.82%, and 96.37% of the total variation in growth performance, respectively (Fig. 4). Regarding leaf morphological traits, most indicators were distributed along the first axis, whereas SLA, LDMC, and LTD primarily varied along the second axis. Specifically, SLA was positively associated with the humus treatment, while LTD and LDMC were more closely aligned with the river sand treatment (Fig. 4a). Leaf physiological indicators, including Chl a, Chl b, Car, and Ci, tended toward the humus treatment, whereas the other physiological traits were associated with loam and sand (Fig. 5c). Similarly, the majority of root morphological indicators clustered toward the humus direction, but SRL and RTD were primarily distributed toward the loam and sand treatments (Fig. 4e).

Fig. 4.

Redundancy analysis and explained variation of traits of leaf morphological traits (a, b), leaf physiological traits (c, d), root morphological traits (e, f), and stem morphological traits (g, h) with growth indicators

Fig. 5.

Functional traits mediate potential regulatory pathways governing plant growth. Solid and dashed lines indicate significant (p < 0.05) and non-significant regulatory pathways, respectively. Line width indicates the correlation coefficient between variables

Hierarchical partitioning (HP) analysis revealed that all leaf morphological traits significantly influenced the growth indicators, with the individual explanatory power of LN, LL, LDM, LA, LV, and LW each exceeding 10% (Fig. 4b). Among the leaf physiological traits, Car, Chl a, Chl, Chl b, and the Car/Chl ratio exerted significant effects on growth, collectively accounting for 54.72% of the total explanation (Fig. 4d). For root traits, RL, RSA, Tips, and RV significantly influenced the growth of A. gigantifolia, contributing 23.10%, 22.88%, 19.89%, and 19.50% to the total explanation, respectively (Fig. 4f). Additionally, stem traits, specifically STD, SV, and SL, also significantly impacted the growth indicators (Fig. 4h).

Potential regulatory pathways of functional traits mediating seedling growth

The PLS-SEM demonstrated a robust goodness-of-fit (GoF = 0.406). Specifically, root morphological traits were significantly influenced by soil physicochemical properties (β = 0.573, p < 0.05) and exerted significant positive effects on stem morphological traits (β = 0.584, p < 0.05), leaf morphological traits (β = 0.512, p < 0.05), leaf physiological traits (β = 0.400, p < 0.05), and seedling growth indicators (β = 0.484, p < 0.05). Soil physicochemical properties also significantly affected leaf morphological traits (β = 0.267, p < 0.05). In addition to root morphology, stem (β = 0.262, p < 0.05) and leaf morphological traits (β = 0.304, p < 0.05) both had significant impacts on growth indicators. Overall, functional traits mediated the regulatory effects of soil physicochemical properties on seedling growth to a substantial degree (R² = 0.866). Notably, root, stem, and leaf morphological traits exhibited high direct effects on seedling growth, whereas stem and root morphological traits, along with soil physicochemical properties, showed substantial indirect effects on growth performance (Fig. 5).

Discussion

Effects of soil types on above- and below-ground growth of A. gigantifolia cuttings

Plant growth and development are strongly influenced by soil physicochemical properties [39–41]. This study found that both humus and loam significantly enhanced the height of A. gigantifolia cuttings (Fig. 2b). Analysis of differences in soil physicochemical properties suggests that this enhancement may be attributed to higher soil moisture and total phosphorus (TP) content, as well as an optimal pH (Fig. 1). Phosphorus is a critical component of biological membranes, and an adequate supply of both water and phosphorus is essential for plant cell division and growth [42]. Furthermore, the availability of soil phosphorus is heavily influenced by pH. High pH levels not only increase the adsorption of phosphorus by soil particles but also inhibit phosphorus uptake by the root system [43]. Consequently, the pH range of humus and loam (5.0–6.0) likely facilitated phosphorus absorption, thereby promoting seedling height. These results are consistent with previous findings in Brassica campestris, Medicago sativa, and Oryza sativa [44]. On the other hand, compared to loam and sand, humus was more conducive to the root and stem growth of A. gigantifolia cuttings. Significant differences in the root-to-shoot ratio further suggest that humus exerts a more pronounced stimulatory effect on below-ground components (Fig. 2c, d). Humus is characterized by its high nutrient and water content. Previous research has shown that humus supplementation significantly enhances the biomass and dry matter content of Lactuca sativa [45]. Similarly, Gao et al. found that nutrient-rich fertile soil improved the ground diameter, fresh weight, and leaf size of A. gigantifolia seed-derived seedlings [46]. For cuttings, we propose that they respond with greater sensitivity and speed to high fertility, accelerating root development to absorb more water and nutrients to sustain stem growth. The prediction results from the Random Forest model support this hypothesis (Supplementary Fig. 1). However, the specific underlying mechanisms warrant further nutrient analysis for verification. Additionally, the height and ground diameter of the 5-month-old cuttings in this study were substantially higher than those of the 1-year-old seed-derived seedlings grown in identical substrates as reported by Gao et al. [46], further validating the superior growth characteristics of A. gigantifolia cuttings.

Responses of leaf, root and stem traits to soil type

The leaf economics spectrum (LES) theory posits that plants in high-fertility environments prioritize resource acquisition, whereas those in stressful environments allocate more resources to defense and structural persistence [19, 22, 24]. In this study, cuttings in the humus treatment—which featured the highest moisture and fertility levels—exhibited the highest SLA and LWC, alongside the lowest LDMC and LTD (Fig. 3). This aligns with typical leaf trait shifts along the LES [14].Random forest modeling further indicated that these variations were likely driven by factors such as soil TP, AN, and SM (Supplementary Fig. 2). Notably, the photosynthetic performance and pigment content of A. gigantifolia leaves were significantly higher in the loam treatment than in the humus treatment (Fig. 3). This finding deviates from the conventional view that higher soil nutrients promote photosynthesis, given that the moisture and nutrient availability in humus were substantially higher than those in loam. A plausible explanation is that the rapid leaf construction of A. gigantifolia in humus reduced the area-based gas exchange capacity and photosynthetic pigment concentrations. Such trade-offs between leaf construction and photosynthesis are widespread across plant species and represent a fundamental component of the LES [47]. This process can lead to reduced photosynthetic rates, as a smaller proportion of leaf nitrogen leaf nitrogen is allocated to the photosynthetic apparatus [48]. Additionally, high soil pH (7.0–7.2) is known to inhibit nutrient uptake and suppress photosynthesis in acidophilic species [49]. Furthermore, soil moisture regulates gas exchange and nutrient utilization via stomatal control [25, 50]. In the present study, the high pH (8.94 ± 0.19) and low nutrient/moisture availability in the river sand treatment likely hindered water and nutrient uptake, resulting in suboptimal growth and photosynthetic performance across the leaves, roots, and stems (Fig. 1). Random forest modeling also corroborated that leaf physiology was potentially regulated by soil pH, moisture, and multiple nutrients (Supplementary Fig. 3).

Root trait variation is multidimensional, reflecting adaptations to complex soil environments [10, 16, 21, 51]. Previous studies have shown that functional trait values at the individual level are closely coupled with growth rates, as root systems typically expand in tandem with overall plant growth [52]. This explains why the RL, RV, and RSA of A. gigantifolia cuttings peaked in humus and decreased progressively in other soil treatments (Fig. 3). In resource-rich environments, plants tend to increase root diameter and decrease root tissue density (RTD) to facilitate rapid resource acquisition; conversely, they adopt the opposite strategy in stressful environments to ensure survival [15, 53–55]. In this study, the variations in RAD and RTD of A. gigantifolia cuttings conformed to the expectations of the root economics spectrum (RES). However, the differences in SRL and SRSA were not predicted by the RES (Fig. 3), a result that more intuitively reflects the multidimensional nature of root traits. Research indicates that at the species level, root diameter and SRL are orthogonal to the economics spectrum, implying that plants can construct root systems with a wide range of SRL values regardless of their position on the RES [15]. Redundancy analysis also suggested that RAD and SRL occupy distinct dimensions (Fig. 5). Since root construction is governed by multiple soil resources, trait variations likely align with the most limiting soil properties [16]. Furthermore, because the root systems of cuttings consist primarily of adventitious roots, they may be more susceptible to environmental influences [56]. Our random forest model predicted that soil phosphorus and potassium might be critical elements influencing the root traits of these cuttings (Supplementary Fig. 4). Previous research on citrus rootstocks has demonstrated that variations in root diameter and RTD are direct drivers of differences in SRL, and larger-diameter roots in cuttings are thought to possess faster expansion rates [57]. Consequently, humus likely promoted larger root diameters and lower tissue density in A. gigantifolia cuttings, thereby enhancing their capacity for water and nutrient acquisition.

Regarding stem traits, humus-grown cuttings exhibited higher SL and SAD, leading to the maximum SV but the lowest STD (Fig. 3). High soil fertility is often associated with lower stem tissue density, reflecting a trade-off between rapid elongation and structural investment [14]. As stem traits are inherently linked to hydraulic conductivity and safety [58], A. gigantifolia likely adjusts these traits to balance moisture differences between substrates.

Functional traits as mediators of soil-regulated seedling growth

Functional ecology theory posits that functional traits play critical roles throughout a plant’s life history, serving as both primary drivers of growth and development and as conduits through which plants respond to environmental variations [59–62]. Although some researchers argue that individual-level functional traits are poor predictors of growth performance [32], an alternative perspective suggests that these traits can robustly predict seedling growth when competition and environmental noise are controlled [52, 63]. Our RDA revealed that leaf morphology, leaf pigment contents, specific root morphological traits, and stem morphological traits exerted significant impacts on A. gigantifolia cuttings (Fig. 5). Consistent with the RDA findings, PLS-SEM also demonstrated that leaf, root, and stem morphological traits—excluding leaf physiological traits—exerted significant direct effects on seedling growth (Fig. 5). Consequently, these morphological traits can serve as reliable reliable predictors of the growth and biomass of A. gigantifolia. This outcome is corroborated by related pot-based studies; for instance, Guo et al. found that the leaf size and thickness of Spartina alterniflora were positively correlated with growth performance in both field and greenhouse experiments [64]. Similarly, the responses of LA, SRL, and SRSA to elevated soil pH shifted potted Cunninghamia lanceolata seedlings toward an acquisitive strategy [65]. Furthermore, the PLS-SEM indicated that functional traits mediated the effects of soil properties on seedling performance. Most notably, root morphological traits played a pivotal role in the soil-driven regulation of A. gigantifolia cuttings, exhibiting the greatest direct and indirect effects on growth among all trait categories (Fig. 5). This suggests that root system development not only directly determines growth but also indirectly facilitates seedling development by promoting stem and leaf growth. We also observed that leaf morphological traits mediated the relationship between soil properties and seedling performance, supporting our second hypothesis. An unexpected finding was the non-significant influence of leaf physiological traits on seedling growth (Fig. 5). This aligns with research suggesting a trade-off between leaf construction and photosynthetic performance based on the optimal allocation principle to favor rapid plant growth [66]. In the present study, leaf construction rather than physiological performance, appeared to be the primary mediator of the soil-driven differences in growth performance.

Overall, our study demonstrates that functional traits largely mediate the regulatory effects of soil properties on the growth performance of A. gigantifolia cuttings. However, unlike seed-derived seedlings, the unique structural characteristics and growth patterns of cuttings may render them more sensitive to soil fertility and enable them to maintain rapid growth following establishment [18], necessitating caution when interpreting their trait-performance relationships. Additionally, pot size may have imposed potential constraints on seedling traits [67], and the five-month experimental duration might not have fully captured the long-term dynamics of the cuttings. Nevertheless, this study provides a valuable reference for the conservation and commercial cultivation of A. gigantifolia, while further research should encompass longer monitoring periods and additional developmental stages.

Conclusion

In summary, this study investigated the regulatory mechanisms by which different soil types influence the growth performance of A. gigantifolia cuttings. Our findings demonstrate that soil physicochemical properties govern growth primarily through direct effects on root and leaf morphological traits, while indirectly modulating performance via stem traits. Specifically, root morphology serves as a pivotal mediating hub that directly drives seedling development. These results reinforce the utility of functional traits in predicting A. gigantifolia performance and offer valuable scientific insights for cultivation and breeding practices. Future research should expand to investigate the associations between soil, traits, and plant performance across diverse propagation methods and developmental stages to further refine trait-based predictive capabilities. Moreover, to translate these promising results into sustainable field applications, validation through similar trials in natural settings or industrial plantations is warranted.

Abbreviations

SM

Soil Moisture

SOM

Soil Organic Matter

TN

Total Nitrogen

AN

Available Nitrogen

TP

Total Phosphorus

AP

Available Phosphorus

TK

Total Potassium

AK

Available Potassium

LN

Leaf Number

LA

Leaf Area

LL

Leaf Length

LW

leaf Width

LFM

Leaf Fresh Mass

LDM

Leaf Dry Mass

LT

Leaf Thickness

SLA

Specific Leaf Area

LDMC

Leaf Dry Matter Content

LWC

Leaf Water Content

LV

Leaf Volume

LTD

Leaf Tissue Density

Pn

NetPhotosynthetic Rate

Gs

Stomatal Conductance

Ci

Intercellular Carbon Dioxide Concentration

Tr

Transpiration Rate

WUE

Water-use Efficiency

CE

Carboxylation Efficiency

Chl a

Chlorophyll a Content

Chl b

Chlorophyll b Content

Chl

Chlorophyll Content

Car

Carotenoid Content

RL

Root Length

RAD

Root Average Diameter

RV

Root Volume

RSA

Root Surface Area

SRL

Specific Root length

SRSA

Specific Root Surface Area

RTD

Root Tissue Density

SL

Stem Length

SAD

Stem Average Diameter

SV

Stem Volume

STD

Stem Tissue Density

CAD

Cutting Average Diameter

CV

Cutting Volume

LFB

Leaf Fresh Biomass

RFB

Root Fresh Biomass

SFB

Stem Fresh Biomass

PFB

Plant Fresh Biomass

LDB

Leaf Dry Biomass

RDB

Root Dry Biomass

SDB

Stem Dry Biomass

PDB

Plant Dry Biomass

Authors’ contributions

H.Z. collected the data and drafted the manuscript. H.Z. and Y.L. analyzed the experimental data and prepared the figures. C.Z. verified and stored the data and reviewed the draft manuscript. X.W. supported the methodology and experimental sites for this study. F.T. provided the resources and projects required for this research and reviewed the manuscript.

Funding statement

This work was supported by the National Key R&D Program of China (2024YFC3506703, 2024YFC3506701, 2024YFC3506702), the Forestry Science and Technology Promotion Program of Guangxi ([2025]GT19), and the Basic Research Fund of Guangxi Institute of Botany (Guizhiye 26002).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.