Environmental Effects on Atractylodes macrocephala Rhizome Growth and Compounds

Abstract

The rhizome of Atractylodes macrocephala, a perennial herb in the Asteraceae family, is valued for its bioactive atractylenolides, but achieving consistent quality in cultivation is challenging. This study aimed to decipher how environmental factors differentially regulate its biomass and atractylenolide content. We sampled from 22 Korean cultivation sites and performed correlation analyses, rigorously controlled by a False Discovery Rate (FDR) correction. Our analysis revealed that the environmental networks governing quantitative growth and qualitative composition are largely independent. While growth was weakly correlated with environmental factors, likely due to suboptimal temperatures at our sites, atractylenolide content was robustly associated with soil properties and climate. Specifically, soil texture was a dominant factor, with sand content showing a strong negative correlation (−0.717 ***) with total atractylenolides, whereas silt (0.675 ***) and clay (0.622 ***) had strong positive correlations. Additionally, cation exchange capacity (0.517 *) and temperature were positively correlated, while relative humidity showed a negative correlation (−0.553 **). This decoupling suggests that optimizing yield and phytochemical quality requires distinct cultivation strategies, providing a foundational framework for developing site-specific practices and quality control for this high-value medicinal herb.

1. Introduction

Atractylodes (Asteraceae) is a perennial herb found in Korea, China, and Japan [1]. Atractylodes macrocephala Koidz., grown in temperate and subtropical areas for over seven centuries [2], is a major medicinal herb in East Asia, particularly in China, where it has been industrialized with an annual production of approximately 12,450 tons between 2020 and 2021 [3]. The dried rhizomes of Atractylodes are categorized into two primary herbal medicines, ‘Baekchul’ (Chin.: Baizhu, Jap.: Byakujutsu) and ‘Changchul’ (Chin.: Cangzhu, Jap.: Soujutsu) [1], with the pharmacopoeias of Korea, Japan, and China designating the rhizomes of A. japonica Koidz. and A. macrocephala Koidz. as ‘Baekchul,’ and those of A. lancea DC. and A. chinensis Koidz. as ‘Changchul’ [4,5,6]. In Korea and Japan, A. japonica and A. macrocephala are sources of the herbal medicine Baekchul [4,6], whereas in China, only A. macrocephala is recognized as a source [5]. It has been traditionally used in Korea, as documented in Donguibogam, for the effects of A. macrocephala rhizomes on strengthening the spleen and to promote digestion, regulate body fluids, treat diarrhea, improve blood circulation, and inhibit sweating [7], as well as in China, with records of its use dating back to the 5th century AD in Bencaojingjizhu [2]. Furthermore, the rhizomes of A. macrocephala exhibit a wide range of pharmacological activities, most notably for improving gastrointestinal function, as well as anti-inflammatory and anti-tumor effects [8]. In modern times, phytochemical studies on A. macrocephala have revealed various phytochemicals, including sesquiterpenoids, triterpenoids, polyacetylenes, coumarins, phenylpropanoids, flavonoids, flavonoid glycosides, steroids, and benzoquinones [2,9]. Sesquiterpenes, including atractylenolides, have been extensively studied, revealing the presence of various compounds with diverse biological activities [10].

With the growing interest in sustainable agriculture [11] and natural remedies, interest in medicinal herbs has increased. However, the growing importance of A. macrocephala rhizomes, which possess diverse physiological properties, continues to be challenged by issues such as falsified origin, as their quality and active compound content can vary significantly by region [12]. Thus, understanding the influence of environmental factors on the secondary metabolites of A. macrocephala is crucial. Abiotic stresses, such as soil property variations and climate gradients, are known to substantially impact the biosynthesis of phytochemicals in plants [13,14], often reflecting a growth–defense trade-off. For instance, our own previous work on Cudrania tricuspidata [15] demonstrated this principle, where smaller fruits produced under the composite stresses of higher altitudes exhibited significantly increased concentrations of their major chemical constituents. This established pattern, where stressful conditions can enhance chemical potency at the expense of size or biomass, is rooted in the plant’s strategy to synthesize secondary metabolites for protection [16].

Hwang et al. (2022) [17] reported that artificial lighting can enhance the biological activity of A. macrocephala, implying changes in its active compounds. However, this finding is a limited result from a specific condition. However, comprehensive studies on how complex environmental variables, such as soil and climate, affect the growth and accumulation of key active compounds under actual cultivation conditions remain remarkably limited. Moreover, a critical knowledge gap persists regarding the relationship between developmental conditions and metabolic changes in medicinal plants.

Previous studies on A. macrocephala have often focused on single environmental factors or have analyzed growth and phytochemical content separately. Consequently, a comprehensive understanding of how the entire suite of environmental factors coordinately regulates both aspects under real-world cultivation conditions is still lacking. Therefore, the primary aim of this multi-site study was to elucidate these complex relationships, specifically to determine whether the environmental networks governing quantitative growth and qualitative traits are distinct. The results of this study will provide a foundational framework for developing site-specific cultivation strategies to enhance the quality and industrial value of A. macrocephala.

2. Materials and Methods

2.1. Plant Samples and Analytical Chemicals

A total of 198 rhizome samples of A. macrocephala (sample numbers FMRC-B2310001–B2310198) were collected from 22 different cultivation sites across 8 regions of South Korea in September 2023 (Figure 1 and Table S1). Within the cultivation site, we defined three distinct zones. From each of these zones, three root rhizomes and one soil sample were collected. The samples were identified by Dr. Hyun-Jun Kim of the National Institute of Forest Science (NIFoS), Korea, deposited in the Forest Medicinal Resources Research Center (FMRC) Herbarium, and assigned to NIFoS, Korea (specimen numbers FMRC-A2310001–A2310088).

The collected A. macrocephala rhizomes had an average moisture content of 65.47 ± 1.96%. Four morphological characteristics (length of rhizome, LR; diameter of rhizome, DR; number of lateral roots, NLR; fresh weight of underground parts, FWU) were evaluated to determine the growth characteristics of A. macrocephala. LR and DR were measured using a digital Vernier caliper (DC200-1; CAS Korea, Yangju, Republic of Korea), the NLR was measured using a 4-digit mechanical hand tally counter, and the FWU was measured using an electronic weighing scale (HS3200S; HANSUNG Instrument Co., Gwangmyeong, Republic of Korea). High-performance liquid chromatography-grade acetonitrile, methanol, and water were purchased from J. T. Baker (Avantor Inc., Radnor, PA, USA). Atractylenolide I (≥98%, CAS No. 73069-13-3) and III (≥98%, CAS No. 73030-71-4) reference standards were purchased from ChemFaces (Wuhan, Hubei, China).

2.2. Preparation of Analytical Samples

The collected rhizome specimens were cleansed of soil debris with a water rinse, sliced, and then dried via a hot-air dryer at 60 °C for 48 h (DY-330H, Lassele, Ansan, Republic of Korea). The dried rhizomes were ground into a powder using a pulverizer (KSP-35, Korea Medi Co., Ltd., Daegu, Republic of Korea). Precisely 500 mg of this powder was mixed with 10 mL of 100% methanol in a 50 mL tube, vortexed, and sonicated at 30 °C for 1 h. Extraction was performed in an ultrasonic bath (JAC-5020, KODO Technical Research Co., Ltd., Hwaseong, Republic of Korea) at 350 W and 40 kHz. The crude extract was centrifuged (UM-1248, Labogene, Seoul, Republic of Korea) at a rcf of 1763× g for 10 min, and the supernatant was passed through qualitative filter paper (ADVENTEC^® No. 2, Toyo Roshi Kaisha, Ltd., Tokyo, Japan) and stored at 4 °C before analysis. Before ultra-high-performance liquid chromatography (UHPLC) analysis, the extract was filtered using a 0.2 µm PTFE syringe filter (Cat. No. 6784-1302; Whatman Co., Maidstone, UK). Reference standard solutions of atractylenolides I and III (1–400 µg/mL) were prepared via serial dilution in methanol and stored at 4 °C.

2.3. UHPLC Analysis and Validation

The samples were analyzed using a Shimadzu Nexera X2 UHPLC system (Shimadzu Co., Kyoto, Japan). The system comprised two LC-30AD pumps, a DGU-20A_5R degasser, an SIL-30AC autosampler, a CTO-20A column oven, a CBM-20A system controller, and an SPD-M20A PDA detector. The analysis utilized an Acquity UPLC HSS T3 column (part number 186003539, C18, 2.1 mm × 100 mm, 1.8 µm, 100Å, Waters Co., Milford, MA, USA) within a column oven at 30 °C. The flow rate was maintained at 0.3 mL/min, and the injection volume was adjusted to 1 µL. The mobile phase consisted of water with 0.1% formic acid (A) and ACN with 0.1% formic acid (B), with the following gradient conditions: initial to 6 min, 50% B; 6 to 25 min, 55% B; 25 to 25.1 min, 65% B; and 25.1 to 28 min, 100% B. The extract samples were analyzed at UV 235 nm. The data were processed using LabSolutions software (version 5.82, Shimadzu, Kyoto, Japan), and the samples were analyzed in triplicate and expressed as mean values.

Validation of the analytical methods, including linearity, detection limit (DL), quantitation limit (QL), and precision, adhered to the ICH Guideline Q2 [18]. Calibration curves for major compounds were constructed using standard solutions at five concentrations (atractylenolides I and III, 1–400 µg/mL). DL and QL were determined based on signal-to-noise ratios (3.3:1 for DL and 10:1 for QL). Precision was assessed using the relative standard deviation (RSD, %) for repeatability and reproducibility. The validation samples were tested three times at three different concentrations, and the results were expressed as mean RSD values.

2.4. Soil Properties and Meteorological Data

In the field, soil samples were collected from a depth of 20 cm following the removal of the surface layer, air-dried until a consistent weight was reached, and homogenized by passing them through a 2 mm mesh sieve. According to the USDA soil taxonomy [19], the soil texture was categorized into 12 types. The soil physicochemical property analysis followed the standard manual outlined by the Rural Development Administration of Korea [20]. To determine soil pH and electrical conductivity (EC), a 1:5 (w/v) soil-to-water suspension was prepared by mixing 10 g of dried soil with 50 mL of distilled water, and measurements were taken with respective meters. Soil organic matter (OM) content was quantified by the Tyurin method, while total nitrogen (TN) was assessed using the Kjeldahl sulfuric acid distillation method. Available phosphorus (AP, as P₂O₅) was determined colorimetrically with the molybdenum blue method, employing a 1-amino-2-naphthol-4-sulfonic acid solution. For the analysis of exchangeable cations (K⁺, Ca²⁺, Mg²⁺, Na⁺) and cation exchange capacity (CEC), the soil samples were leached with 1 N ammonium acetate (NH₄OAc, pH 7.0). The cation concentrations in the leachate were measured by inductively coupled plasma optical emission spectrometry (ICP-OES). The CEC was determined from the amount of exchanged ammonium using the Kjeldahl distillation method. Base saturation (BS) was calculated as the ratio of the sum of exchangeable cations to the CEC, expressed as a percentage. All analyses were performed in triplicate, with the results presented as mean ± standard error (S.E.). Meteorological data (annual average temperature, AAT; annual average maximum temperature, AAMT; annual average minimum temperature, AAmT; annual maximum temperature, AMT; annual minimum temperature, AmT; total precipitation, TP; relative humidity, RH; total sunshine hours, SSH) were obtained from the Korea Meteorological Administration (KMA) open data platform (data.kma.go.kr, accessed on 7 August 2024).

2.5. Statistical Analysis

Statistical analyses were performed using SPSS software (version 26.0, IBM Corp., Chicago, IL, USA). Data are expressed as the mean ± standard error (SE). To explore the relationships between the major compound content, cultivation environment, and rhizome growth, multivariate analysis of variance (MANOVA) was employed. Tukey’s honest significant difference (HSD) test was used for post hoc analysis, with statistical significance set at p < 0.05. Pearson’s correlation coefficients were calculated to assess the relationships among the contents of major compounds, rhizome growth, and the environment. To account for multiple comparisons, the resulting p-values were adjusted using the Benjamini–Hochberg procedure to control the False Discovery Rate (FDR). Correlation plots were visualized using R Studio, version 4.2.3 (Boston, MA, USA), with the ‘ggcorrplot2’ and ‘qgraph’ packages.

3. Results

3.1. Soil and Meteorological Data

The physicochemical properties of the soils from the 22 A. macrocephala cultivation sites were analyzed (

Table 1. Soil physicochemical data of 22 different cultivation sites of Atractylodes macrocephala.

Cultivation Sites (n = 3)	Soil Texture	pH [1:5]	EC	OM	TN	AP	Exchangeable Cations				CEC	BS
							K+	Ca2+	Mg2+	Na+
			(dS/m)	(%)	(%)	(mg/kg)	(cmol+/kg)				(cmol+/kg)	(%)
1	SCL	6.17 ± 0.14 a,b,c,d,e,f	0.80 ± 0.07 a,b	4.99 ± 0.13 a,b,c	0.31 ± 0.00 a	2267.91 ± 29.24 b	1.96 ± 0.18 a	9.15 ± 0.64 b,c,d	2.25 ± 0.24 a,b,c,d,e,f	0.10 ± 0.03 a	22.67 ± 1.02 a	59.88 ± 5.52 a,b,c,d,e
2	SCL	5.44 ± 0.22 g	1.19 ± 0.73 a,b	5.38 ± 0.07 a,b	0.28 ± 0.01 a,b,c	1338.13 ± 176.91 b,c,d,e	0.89 ± 0.22 a,b,c	5.67 ± 0.84 d,e,f,g,h	1.48 ± 0.26 c,d,e,f	0.04 ± 0.01 a	21.99 ± 0.41 a	36.96 ± 6.39 e
3	SL	6.26 ± 0.30 a,b,c,d,e,f	0.42 ± 0.05 a,b	2.24 ± 0.76 d,e,f,g,h	0.13 ± 0.02 d,e,f	466.95 ± 186.82 d,e,f	0.99 ± 0.16 a,b,c	6.47 ± 1.10 b,c,d,e,f,g,h	1.90 ± 0.44 b,c,d,e,f	0.42 ± 0.39 a	14.80 ± 1.33 b,c,d,e	65.28 ± 9.02 a,b,c,d,e
4	SL	6.84 ± 0.03 a	0.47 ± 0.07 a,b	2.31 ± 0.71 d,e,f,g,h	0.13 ± 0.04 d,e,f	472.32 ± 116.75 d,e,f	1.06 ± 0.21 a,b,c	7.04 ± 1.40 b,c,d,e,f,g,h	2.69 ± 0.34 a,b,c,d	0.04 ± 0.01 a	13.28 ± 1.54 c,d,e	83.17 ± 9.19 a,b
5	SCL	5.79 ± 0.16 e,f,g	0.46 ± 0.18 a,b	1.91 ± 0.29 e,f,g,h	0.12 ± 0.01 d,e,f	468.67 ± 124.67 d,e,f	0.34 ± 0.16 c	4.67 ± 0.38 f,g,h	1.51 ± 0.23 c,d,e,f	0.06 ± 0.01 a	13.24 ± 0.74 c,d,e	49.98 ± 4.31 b,c,d,e
6	SL	6.85 ± 0.09 a	1.27 ± 0.86 a,b	3.80 ± 0.48 a,b,c,d,e	0.23 ± 0.03 a,b,c,d	1483.90 ± 392.23 b,c,d	1.20 ± 0.59 a,b,c	10.46 ± 0.97 b	3.55 ± 0.68 a,b	0.16 ± 0.12 a	19.92 ± 3.57 a,b,c	78.15 ± 4.59 a,b,c
7	SL	6.44 ± 0.09 a,b,c,d,e	0.49 ± 0.05 a,b	2.57 ± 0.08 d,e,f,g,h	0.15 ± 0.01 d,e,f	694.48 ± 10.73 d,e,f	0.93 ± 0.07 a,b,c	8.02 ± 0.25 b,c,d,e,f,g	2.42 ± 0.11 a,b,c,d,e	0.48 ± 0.22 a	14.73 ± 0.21 b,c,d,e	80.45 ± 2.07 a,b,c
8	SL	6.61 ± 0.09 a,b,c	0.34 ± 0.07 b	1.67 ± 0.44 e,f,g,h	0.12 ± 0.03 d,e,f	372.13 ± 120.98 d,e,f	0.39 ± 0.11 c	8.00 ± 0.63 b,c,d,e,f,g	1.65 ± 0.58 c,d,e,f	0.14 ± 0.04 a	12.52 ± 1.16 d,e	81.13 ± 5.57 a,b,c
9	SCL	5.54 ± 0.15 f,g	2.44 ± 0.49 a	5.64 ± 0.46 a	0.30 ± 0.02 a,b	1979.81 ± 104.40 b,c	1.04 ± 0.13 a,b,c	10.16 ± 0.18 b,c	2.32 ± 0.09 a,b,c,d,e,f	0.23 ± 0.13 a	21.30 ± 2.11 a,b	65.76 ± 6.01 a,b,c,d,e
10	SL	6.41 ± 0.13 abcde	1.21 ± 0.07 ab	3.36 ± 0.77 bcdefg	0.19 ± 0.05 bcdef	879.94 ± 205.24 cdef	0.63 ± 0.18 bc	9.23 ± 0.75 bcd	1.78 ± 0.38 cdef	0.24 ± 0.08 a	13.93 ± 0.57 cde	84.87 ± 5.36 a
11	SL	6.02 ± 0.11 b,c,d,e,f,g	0.22 ± 0.03 b	1.25 ± 0.14 g,h	0.09 ± 0.01 e,f	208.83 ± 15.55 f	0.23 ± 0.03 c	3.74 ± 0.26 h	0.62 ± 0.04 f	0.02 ± 0.00 a	10.47 ± 0.18 e	43.87 ± 2.28 d,e
12	SL	5.95 ± 0.09 b,c,d,e,f,g	0.21 ± 0.04 b	0.84 ± 0.25 h	0.07 ± 0.02 f	285.17 ± 69.56 e,f	0.30 ± 0.06 c	8.16 ± 0.50 b,c,d,e,f,g	1.88 ± 0.17 b,c,d,e,f	0.03 ± 0.00 a	13.05 ± 0.30 d,e	79.40 ± 2.61 a,b,c
13	SL	5.92 ± 0.09 c,d,e,f,g	0.72 ± 0.06 a,b	1.91 ± 0.17 e,f,g,h	0.14 ± 0.02 d,e,f	861.59 ± 37.71 d,e,f	0.85 ± 0.05 a,b,c	4.82 ± 0.49 f,g,h	1.73 ± 0.30 c,d,e,f	0.02 ± 0.00 a	12.19 ± 1.46 d,e	61.48 ± 4.15 a,b,c,d,e
14	SL	6.48 ± 0.07 a,b,c,d,e	0.73 ± 0.05 a,b	3.52 ± 0.32 a,b,c,d,e,f	0.20 ± 0.02 a,b,c,d,e	1371.78 ± 54.76 b,c,d,e	1.52 ± 0.35 a,b,c	6.37 ± 0.16 c,d,e,f,g,h	1.83 ± 0.01 b,c,d,e,f	0.09 ± 0.02 a	13.62 ± 0.73 c,d,e	72.05 ± 0.46 a,b,c,d
15	SCL	6.55 ± 0.04 a,b,c,d	1.75 ± 0.35 a,b	5.28 ± 0.31 a,b	0.30 ± 0.00 a,b	2390.30 ± 468.45 b	0.84 ± 0.06 a,b,c	15.33 ± 1.02 a	3.18 ± 0.16 a,b,c	0.04 ± 0.00 a	23.73 ± 1.08 a	81.66 ± 1.81 a,b,c
16	SL	6.79 ± 0.03 a	1.07 ± 0.09 a,b	4.11 ± 0.17 a,b,c,d	0.27 ± 0.00 a,b,c	3526.50 ± 94.64 a	1.86 ± 0.12 a,b	8.59 ± 0.22 b,c,d,e,f	3.84 ± 0.30 a	0.06 ± 0.01 a	17.38 ± 0.17 a,b,c,d	82.48 ± 2.75 a,b,c
17	SL	6.19 ± 0.10 a,b,c,d,e,f	0.82 ± 0.16 a,b	2.80 ± 0.26 d,e,f,g,h	0.18 ± 0.01 c,d,e,f	676.79 ± 130.59 d,e,f	0.99 ± 0.09 a,b,c	7.16 ± 0.53 b,c,d,e,f,g,h	1.64 ± 0.22 c,d,e,f	0.14 ± 0.02 a	14.70 ± 1.15 b,c,d,e	68.03 ± 5.3 a,b,c,d,e
18	SL	5.85 ± 0.03 d,e,f,g	0.59 ± 0.04 a,b	2.82 ± 0.34 c,d,e,f,g,h	0.16 ± 0.01 c,d,e,f	887.28 ± 113.63 c,d,e,f	0.42 ± 0.15 c	4.11 ± 0.43 g,h	1.29 ± 0.09 d,e,f	0.16 ± 0.09 a	15.16 ± 0.58 b,c,d,e	39.30 ± 2.96 d,e
19	SL	6.65 ± 0.16 a,b	0.32 ± 0.02 b	1.58 ± 0.07 f,g,h	0.11 ± 0.01 d,e,f	293.88 ± 7.09 e,f	0.29 ± 0.03 c	9.04 ± 0.15 b,c,d,e	1.62 ± 0.10 c,d,e,f	0.02 ± 0.00 a	13.23 ± 0.13 d,e	82.90 ± 0.87 a,b
20	SL	5.42 ± 0.06 g	1.08 ± 0.13 ab	2.78 ± 0.26 defgh	0.16 ± 0.01 cdef	1307.19 ± 194.97 bcdef	0.37 ± 0.11 c	3.74 ± 0.27 h	0.91 ± 0.07 ef	0.63 ± 0.31 a	13.36 ± 0.25 cde	42.37 ± 1.90 de
21	SL	6.56 ± 0.07 a,b,c,d	0.36 ± 0.04 b	2.05 ± 0.19 d,e,f,g,h	0.13 ± 0.00 d,e,f	599.28 ± 89.83 d,e,f	0.35 ± 0.03 c	5.01 ± 0.53 e,f,g,h	1.30 ± 0.06 d,e,f	0.05 ± 0.02 a	13.74 ± 0.61 c,d,e	48.96 ± 4.36 c,d,e
22	SL	6.30 ± 0.28 a,b,c,d,e	1.67 ± 1.15 a,b	1.70 ± 0.75 e,f,g,h	0.13 ± 0.05 d,e,f	986.87 ± 550.45 c,d,e,f	1.06 ± 0.69 a,b,c	6.26 ± 1.88 c,d,e,f,g,h	1.41 ± 0.76 d,e,f	0.35 ± 0.13 a	10.83 ± 1.68 d,e	80.21 ± 19.63 a,b,c

SCL, sandy clay loam; SL, sandy loam; EC, electrical conductivity; OM, organic matter; TN, total nitrogen; AP, available phosphate; CEC, cation exchange capacity; BS, base saturation. The values are expressed as mean ± standard error (S.E.). Tukey’s Honestly Significant Difference (HSD) multiple comparison tests showed that there were statistically significant differences in means between groups with different letters (p < 0.05).

). The soils investigated were sandy loam and sandy clay loam. The pH of the cultivation site soils ranged from 5.42 to 6.85 for A. macrocephala rhizome cultivation. EC was highest at site 9 (2.44 ± 0.49 dS/m) and lowest at site 12 (0.21 ± 0.04 dS/m). OM was highest at site 9 (5.64 ± 0.46%) and lowest at site 12 (0.84 ± 0.25%). TN was highest at site 1 (0.31 ± 0.00%) and lowest at site 12 (0.07 ± 0.02%). AP showed a large variation across the surveyed sites (208.83 ± 15.55 to 3526.50 ± 94.64 mg/kg), and significant variation was observed even within the same site. BS ranged from 36.96 ± 6.39% to 84.87 ± 5.36% across the cultivation sites.

Meteorological data were collected for the eight variables (Table S2). The altitude of the fields ranged from 128 to 438 m. AAT ranged from 11.4 to 15.0 °C, AAMT from 17.7 to 21.4 °C, and AAmT from 5.8 to 9.6 °C. The highest AMT was at site 15 (37.8 °C), and the lowest was at sites 19–22 (−20.2 °C). TP ranged from 1518.7 to 2216.1 mm, RH from 62.6 to 74.6%, and SSH from 2205.5 to 2358.9 h.

3.2. Growth Characteristics of A. macrocephala Rhizomes

Table 2. Growth characteristics of Atractylodes macrocephala rhizomes in 22 cultivation sites.

Cultivation Sites (n = 3)	LR	DR	NLR	FWU
	(mm)	(mm)		(g)
1	60.31 ± 4.95 d,e	62.75 ± 6.86 c	66.67 ± 7.51 c	106.18 ± 6.21 b
2	80.99 ± 11.43 a,b,c,d	50.31 ± 6.15 c,d	38.78 ± 13.36 c	63.67 ± 8.35 b
3	72.94 ± 7.85 b,c,d,e	48.41 ± 2.54 c,d	42.78 ± 11.52 c	53.64 ± 3.60 b
4	61.57 ± 6.39 b,c,d,e	47.28 ± 2.04 c,d	42.67 ± 7.45 c	48.51 ± 7.74 b
5	56.11 ± 4.94 d,e	38.23 ± 0.97 d	39.33 ± 7.43 c	30.46 ± 2.60 b
6	69.97 ± 0.17 b,c,d,e	40.85 ± 3.11 c,d	30.56 ± 8.42 c	37.20 ± 5.96 b
7	91.43 ± 12.87 a,b,c	51.99 ± 4.60 c,d	34.22 ± 11.05 c	63.30 ± 7.66 b
8	84.28 ± 4.71 a,b,c,d	56.99 ± 3.03 c,d	35.67 ± 9.60 c	64.15 ± 10.25 b
9	86.23 ± 1.76 a,b,c,d	93.62 ± 3.01 b	71.56 ± 6.25 c	115.63 ± 5.57 b
10	73.69 ± 1.72 a,b,c,d,e	51.95 ± 6.28 c,d	58.56 ± 14.29 c	48.87 ± 11.10 b
11	62.94 ± 1.04 b,c,d,e	48.13 ± 2.46 c,d	49.56 ± 3.30 c	40.97 ± 2.25 b
12	55.86 ± 3.43 d,e	53.55 ± 4.05 c,d	48.67 ± 7.17 c	41.67 ± 4.53 b
13	60.71 ± 3.66 c,d,e	47.00 ± 2.17 c,d	45.11 ± 1.28 c	34.17 ± 3.31 b
14	49.84 ± 2.53 e	54.12 ± 6.32 c,d	40.22 ± 9.61 c	39.54 ± 7.34 b
15	104.70 ± 4.21 a	141.57 ± 3.71 a	149.00 ± 7.35 b	312.00 ± 13.23 a
16	48.48 ± 1.80 e	49.59 ± 2.17 c,d	44.00 ± 4.53 c	32.83 ± 3.57 b
17	62.28 ± 2.82 b,c,d,e	53.27 ± 2.89 c,d	47.78 ± 3.95 c	46.22 ± 4.64 b
18	59.55 ± 7.14 d,e	57.73 ± 9.97 c,d	33.67 ± 5.05 c	59.03 ± 22.54 b
19	92.37 ± 5.37 a,b	138.53 ± 7.33 a	231.56 ± 33.61 a	401.58 ± 78.34 a
20	64.86 ± 8.47 b,c,d,e	54.31 ± 2.16 c,d	54.11 ± 6.34 c	64.24 ± 18.54 b
21	63.40 ± 0.52 b,c,d,e	56.22 ± 0.36 c,d	46.56 ± 0.79 c	66.09 ± 5.57 b
22	70.84 ± 5.19 b,c,d,e	57.60 ± 2.91 c,d	41.56 ± 0.42 c	64.99 ± 5.85 b

LR, length of rhizome; DR, diameter of rhizome; NLR, number of lateral roots; FWU, fresh weight of underground parts. Values are expressed as mean ± standard error (S.E.). Tukey’s HSD test revealed statistically significant differences in the group means, as indicated by different letters, at a significance level of p < 0.05.

presents the growth characteristics of A. macrocephala rhizomes across the 22 cultivation sites. Site 15 recorded the highest LR at 104.70 ± 4.21 mm, whereas site 16 had the lowest at 48.48 ± 1.80 mm. DR was also highest at site 15, at 141.57 ± 3.71, while the lowest value was recorded at cultivation site 5, at 38.23 ± 0.97. The NLR was highest at site 19, measuring 231.56 ± 33.61, and lowest at site 6, at 30.56 ± 8.42. Lastly, the FWU was most substantial at site 19, weighing 401.58 ± 78.34 g, and least substantial at site 5, with 30.46 ± 2.60 g.

3.3. Method Validation and Quantitation

The developed UHPLC quantitative method was validated, and the results are shown in Figure S1 and Tables S3 and S4. Two major compounds were identified by matching the retention times of the extract with those of the reference standard solution. These compounds exhibited excellent linearity with correlation coefficients exceeding 0.9998. The DLs for atractylenolides I and III were 0.03 and 0.02 µg/mL, respectively, with QLs of 0.10 and 0.08 µg/mL. The method’s precision was assessed by analyzing the major compounds at three concentration levels, with triplicate measurements performed for each level within a single day and across multiple independent days. The RSD, defined as the standard deviation normalized to the mean, was used to evaluate the accuracy, and the obtained RSD values were within the specified acceptance criteria, indicating high method accuracy. The intraday RSD values ranged from 0.05% to 0.30%, and the interday RSD values ranged from 0.04% to 0.19%. These results establish the reliability of the developed UHPLC method for the quantitative determination of the major compounds in A. macrocephala rhizomes.

Atractylenolide I and atractylenolide III, the two major compounds in A. macrocephala rhizomes, were analyzed at 22 different cultivation sites (

Table 3. Contents of two major compounds of Atractylodes macrocephala from 22 different cultivation sites.

Cultivation Sites (n = 3)	Atractylenolide I	Atractylenolide III	Total
		(μg/mL)
1	24.09 ± 5.23 b	72.12 ± 20.80 a	96.21 ± 24.79 a
2	13.85 ± 3.94 b,c,d,e	39.61 ± 10.27 b,c	53.46 ± 13.29 b,c,d
3	13.29 ± 3.73 b,c,d,e	30.91 ± 9.13 b,c,d	44.19 ± 11.71 c,d,e
4	9.75 ± 1.05 c,d,e	20.76 ± 6.65 c,d,e	30.51 ± 6.90 d,e,f,g
5	15.16 ± 4.30 b,c,d	48.57 ± 14.75 a,b	63.73 ± 18.60 b,c
6	5.79 ± 0.92 d,e	6.47 ± 2.09 d,e	12.26 ± 2.27 e,f,g
7	11.80 ± 2.76 b,c,d,e	6.34 ± 2.57 d,e	18.14 ± 5.01 e,f,g
8	6.21 ± 3.85 d,e	2.44 ± 1.01 e	8.65 ± 4.83 f,g
9	15.43 ± 2.32 b,c,d	5.62 ± 1.28 d,e	21.05 ± 1.24 e,f,g
10	1.98 ± 0.41 e	4.46 ± 1.21 e	6.44 ± 1.20 g
11	10.96 ± 3.85 c,d,e	11.21 ± 4.80 d,e	22.17 ± 7.71 d,e,f,g
12	8.96 ± 3.51 c,d,e	6.06 ± 2.75 d,e	15.01 ± 5.09 e,f,g
13	14.56 ± 2.42 b,c,d	6.83 ± 0.83 d,e	21.39 ± 2.70 e,f,g
14	7.31 ± 1.49 d,e	6.19 ± 1.47 d,e	13.50 ± 1.91 e,f,g
15	40.89 ± 2.23 a	39.67 ± 13.44 b,c	80.56 ± 15.08 a,b
16	10.56 ± 1.50 c,d,e	22.41 ± 8.53 c,d,e	32.97 ± 9.77 c,d,e,f,g
17	20.75 ± 5.13 b,c	17.88 ± 5.28 c,d,e	38.63 ± 8.09 c,d,e,f
18	4.73 ± 1.50 d,e	4.44 ± 2.47 e	9.17 ± 3.87 f,g
19	7.11 ± 0.88 d,e	8.60 ± 2.72 d,e	15.71 ± 3.49 e,f,g
20	6.40 ± 1.70 d,e	4.82 ± 1.14 e	11.22 ± 2.30 f,g
21	4.57 ± 1.59 d,e	3.22 ± 1.30 e	7.80 ± 2.42 f,g
22	9.19 ± 1.54 c,d,e	6.82 ± 1.41 d,e	16.02 ± 2.90 e,f,g

The values are expressed as mean ± standard error (S.E.). Tukey’s Honestly Significant Difference (HSD) multiple comparison tests showed that there were statistically significant differences in means between groups with different letters (p < 0.05).

). The atractylenolide I content ranged from 1.98 ± 0.41 μg/mL at site 10 to 40.89 ± 2.23 μg/mL at site 15, while the atractylenolide III content ranged from 2.44 ± 1.01 μg/mL at site 8 to 72.12 ± 20.80 μg/mL at site 1. The total content of major compounds was highest at site 1 (96.21 ± 24.79 μg/mL) and lowest at site 10 (6.44 ± 1.20 μg/mL).

3.4. Correlation Between Atractylenolides, Rhizome Growth, and Environments

We employed principal component analysis (PCA) and Pearson’s correlation analysis to investigate the interconnections between environmental factors (including soil physicochemical properties and meteorological data), the growth traits of A. macrocephala rhizomes, and the concentrations of key compounds (atractylenolides I and III) found in these rhizomes.

The PCA results (Figure 2) revealed relationships between the variables. Although the first two principal components accounted for only 50.9% of the total variance, they provided meaningful insights into the multivariate relationships between atractylenolide contents, growth traits, soil properties, and meteorological variables. Both atractylenolides and growth traits were located on the negative side of PC1, indicating a negative relationship between precipitation and relative humidity. Along PC2, the atractylenolides were positioned on the positive side and growth traits on the negative side, indicating associations with temperature-related factors and soil nutrients, respectively. This relatively low cumulative variance is characteristic of observational data collected from non-controlled field environments. It suggests the system’s complexity is driven by numerous interacting factors, including key unmeasured variables, like soil microbial communities, genetic variations among plants, and undocumented differences in cultivation practices. Therefore, while acknowledging these inherent complexities, the PCA serves as a robust exploratory tool to identify the most significant trends within our dataset.

To identify the key environmental factors influencing the rhizome growth of A. macrocephala, we performed a Pearson’s correlation analysis and applied a stringent FDR correction to all p-values to account for multiple comparisons. All correlations reported hereafter are statistically significant after this adjustment (p < 0.05). A correlation analysis between the environment and rhizome growth of A. macrocephala (Figure 3 and Table S5) was performed. Rhizome growth was only significantly correlated with exchangeable Mg²⁺ (DR, 0.575 ***; FWU, 0.495 **). After the FDR correction, the correlation with meteorological factors was no longer valid.

On the other hand, a clear correlation was still confirmed between the environments and the atractylenolides (Figure 3 and Table S6). Among the soil chemical properties, CEC showed a significant positive correlation, specifically with the total atractylenolide content (0.517 *). In contrast, soil texture, particularly the sand, silt, and clay contents, showed strong and significant correlations with all individual and total atractylenolides (sand, −0.717 **; silt, 0.675 **; clay, 0.622 **). Soil humidity positively correlated with total atractylenolides (sand, −0.717 **; silt, 0.675 **; clay, 0.622 **). However, RH had the opposite reaction (total, −0.553 **; Atr I, −0.535 **; Atr III, −0.490 *). All temperature factors were positively correlated with atractylenolides. Latitude negatively correlated with atractylenolides.

To visualize the overall structure of relationships among the variables, a network analysis was performed using only the correlations that remained significant after FDR correction (Figure 4). The network analysis revealed two distinct and largely independent clusters. The first, larger cluster consisted of major compound contents, climatic factors, and soil properties, indicating a complex interplay among these variables in determining the phytochemical profile of the rhizome. Within this cluster, soil texture (sand, silt, and clay) and temperature-related factors appeared to be central hubs. The second, smaller cluster was composed of rhizome growth parameters (NLR, DR, and FWU), which were strongly intercorrelated but showed only weak connections to the first cluster. This structural separation in the network strongly suggests that the factors regulating quantitative growth and those modulating qualitative chemical composition are largely independent in A. macrocephala under the studied conditions. Notably, some variables, such as Na⁺ and TP, were isolated from the main network, indicating their lack of significant direct influence on other measured variables.

4. Discussion

The results of this study provide mechanistic insights into how environmental factors differentially regulate the growth and chemical quality of A. macrocephala.

Analysis of the correlation between A. macrocephala rhizomes and soil physicochemical properties showed that exchangeable Ca²⁺ positively correlated with rhizome growth (DR and FWU). Burström (1968) [21] indicated that Ca²⁺ could bind to pectin in the cell wall to strengthen it and regulate the movement of substances inside and outside the cell, thereby contributing to maintaining normal cell function and promoting growth. Interestingly, after applying a rigorous statistical correction (FDR), most direct correlations between rhizome biomass and individual climate factors lost statistical significance. This result may be due to the fact that the cultivation sites in this study did not fall within the optimal temperature range for A. macrocephala growth. Previous studies have reported the optimal temperature for A. macrocephala growth to be 24–29 °C [22]. However, the average temperature range for the cultivation sites in this study was 11.4–15 °C, consistently falling below the optimal range and corresponding to suboptimal conditions. Therefore, the disappearance of growth–climate correlations after FDR correction does not imply that climate does not affect growth. Rather, it suggests that temperature conditions in the study sites were either insufficiently diverse or suboptimal to explain growth.

Regarding atractylenolides, the correlation analysis revealed a positive relationship between CEC and atractylenolide content. CEC is one of the most important chemical properties of soil, playing a key role in maintaining soil fertility by preventing nutrient leaching and regulating their availability to plants [23]. This stable nutrient status likely contributed to securing the biochemical resources necessary for plants to produce secondary metabolites. An investigation of the correlation between soil texture and atractylenolide content revealed a negative correlation between sand content, whereas silt and clay contents showed positive correlations. In this study, the soil texture types at the cultivation sites were sandy loam and sandy clay loam, suggesting that although excessive drainage is beneficial for A. macrocephala cultivation [22], it does not contribute to the high accumulation of atactylenolides. The correlation between meteorological factors and atractylenolide content showed that temperature positively correlated with atractylenolide content. An atractylenolide is a substance corresponding to a sesquiterpene lactone, and its accumulation in plants is mainly affected by abiotic factors, such as light intensity, temperature, and soil composition [24]. Since the temperatures in our study sites were below the optimal growth range, it is plausible that the reaction rate of terpene synthases involved in atractylenolide biosynthesis increased with rising temperatures. This heightened enzyme activity in areas with high average annual maximum temperatures may have promoted the accumulation of the final product. This temperature–metabolite relationship is consistent with the findings obtained by Xu et al. (2016), who demonstrated that temperature elevation within physiological ranges can increase the expression of jasmonic acid biosynthesis genes (LOX, AOS, AOC), thereby promoting sesquiterpene biosynthesis through enhanced metabolic activity rather than stress responses [25]. In addition, relative humidity negatively correlated with atractylenolides. Generally, a high relative humidity (>70%) reduces leaf transpiration. When the root moisture was sufficient and water stress was relieved, secondary metabolite synthesis decreased. Therefore, we hypothesize that the high relative humidity (70.50 ± 0.70%) at our sites contributed to lower atractylenolide accumulation. This may be because high humidity alleviates water stress by maintaining sufficient rhizome moisture, which, in turn, reduces the plant’s need to produce stress-induced secondary metabolites.

Plant secondary metabolites are typically synthesized as a defense mechanism against stressors [26,27], often leading to a well-documented trade-off between growth and defense. This occurs because the synthesis of these complex compounds competes for the same pool of primary metabolites required for biomass accumulation [28].

Based on this principle, an inverse relationship between rhizome growth and atractylenolide content would be expected. However, our correlation analysis, even before the stringent FDR correction, did not show a significant negative correlation. After the correction, any direct statistical link between growth and phytochemical content vanished, suggesting that a simple trade-off model does not apply in this system.

We hypothesize that this lack of a trade-off is due to the specific nature of sesquiterpene lactone (STL) accumulation, including atactylenolides [29], in underground storage organs like rhizomes [30,31]. For self-protection and long-term storage, STLs are often sequestered in specialized sites such as laticifers or vacuoles within the root tissue. This sequestration physically separates the defensive compounds from active metabolic sites, potentially minimizing their feedback inhibition on growth processes. More importantly, the development of these storage structures occurs concurrently with rhizome growth. As the rhizome expands and matures, it creates more physical space for STL accumulation. This concept is supported by Chen et al. (2017) [32], who reported that the activation of STL synthesis in A. lancea coincides with simultaneous root growth. Therefore, rather than competing, the processes of growth and accumulation can occur in parallel, neutralizing the statistical signal of any underlying metabolic trade-off.

This finding—that the apparent decoupling of growth and secondary metabolite production in A. macrocephala rhizomes may be explained by a specialized storage mechanism—offers a more refined understanding of the plant’s physiological strategy. It suggests that the relationship between primary and secondary metabolism is not always a direct competition, but it can be spatially and temporally coordinated within the plant’s overall development program.

Ultimately, this study underscores the necessity of moving beyond simple correlation analyses to understand the complex, multifaceted interactions between medicinal plants and their environments. While our FDR-corrected analysis revealed that distinct environmental factors regulate growth and phytochemical quality, the lack of a simple trade-off highlights the importance of considering the unique biology of the plant organ in question. For the sustainable cultivation of high-quality A. macrocephala, future strategies must therefore consider not only the external environmental conditions but also the internal physiological mechanisms, such as sink–source dynamics and storage capacity development. This integrated approach will be crucial for developing precision cultivation practices that can co-optimize both yield and phytochemical value in a changing world.

As our study is based on observational data, these inferred conditions should be considered strong hypotheses rather than definitive prescriptions.

5. Conclusions

This study aimed to elucidate the complex relationships between environmental factors, rhizome growth, and the content of major bioactive compounds in Atractylodes macrocephala. Through a comprehensive correlation analysis, rigorously controlled by an FDR correction, we found a pivotal result: the environmental networks that regulate quantitative growth and those that control qualitative phytochemical composition are largely independent. Rhizome growth showed weak correlations with specific climatic factors, likely because the suboptimal temperature conditions of our study sites acted as a general limiting factor rather than a discriminating variable. In contrast, atractylenolide content demonstrated robust correlations with soil properties (soil texture and CEC) and specific climatic factors (temperature and humidity), highlighting its sensitive response to the immediate physicochemical environment. Furthermore, the absence of a clear trade-off between growth and secondary metabolite accumulation may be due to the unique role of rhizomes as storage organs.

These findings have significant implications for the sustainable cultivation of high-quality A. macrocephala. Our results suggest that optimizing yield and phytochemical quality may require different, tailored cultivation strategies. Future research should focus on validating these distinct regulatory pathways under controlled experimental conditions and developing precision agriculture practices that can co-optimize both growth and the production of valuable secondary metabolites in a changing climate.