2.1. Study site and sample collection

The experimental materials consisted of fresh C. songaricum Rupr. Plants at the “outcrop stage” were collected from 33 sampling sites in the Alxa region of Inner Mongolia Autonomous Region between July and August 2019. The study was conducted in the Alxa region, Inner Mongolia, China. A total of 33 sampling sites were selected to cover the main distribution area of C. songaricum. No specific permits were required for the collection of C. songaricum samples in the Alxa region, because at the time of collection, this species was not protected under local wildlife protection regulations in that area. All collection work complied with the relevant laws and regulations of China.

At each site, three healthy individuals with consistent growth phenotypes and no visible disease spots or insect damage were selected. Species identification was performed by Professor Zhou Yubi from the Northwest Institute of Plateau Biology, Chinese Academy of Sciences. On the day of sampling, three types of samples were collected simultaneously: the fleshy stems of C. songaricum, rhizosphere soil, and surrounding water samples. The specific procedures were as follows:

The fleshy stems of C. songaricum were cleaned of surface impurities, such as sediment, and then partitioned. Rhizosphere soil samples were collected using a sterile soil auger at a depth of 20–40 cm after removing non-soil impurities such as fallen leaves and stones. A 500 mL groundwater sample was collected from local herder wells near each sampling site, sealed, and preserved.

The collected C. songaricum stems were peeled with a sterile knife to separate the epidermis from the internal fleshy tissue. The epidermis and internal fleshy stems were then packaged separately, transported to the laboratory on the same day, and immediately frozen at −80°C for subsequent determination of amino acid composition and content. The soil and water samples were transported to the laboratory on the same day and stored at −4°C for physicochemical property analysis.

All samples were assigned unique identification numbers (S1–S33) corresponding to the 33 sampling sites. Detailed geographic coordinates, elevation, and host plant information are provided in Supplementary S1 Table.

2.2. Soil and water analysis

2.3. Amino acid quantification of C. songaricum samples

Following portioning, 100 g of the stem epidermis and internal fleshy stem of C. songaricum were randomly selected from each sample and mechanically homogenized for further use. A 0.5 g portion of the powdered sample was weighed into a hydrolysis tube, followed by the addition of 10 mL of 6 mol/L hydrochloric acid solution and 2 drops of phenol. The tube was sealed under a nitrogen atmosphere and placed in an electric thermostatic blast drying oven at 103 °C for 22 hours. After hydrolysis, the tube was removed and cooled to room temperature. The hydrolysate was filtered into a 50 mL volumetric flask. The hydrolysis tube was rinsed with pure water, and the rinsate was combined into the same volumetric flask. The solution was then brought to volume and mixed thoroughly by shaking. Exactly 1 mL of the filtrate was pipetted into a test tube and dried under reduced pressure at 40 °C. The residue was dissolved in 1 mL of pure water and subjected to further drying under reduced pressure. Finally, the residue in the tube was dissolved in 1 mL of sodium citrate buffer solution (pH = 2.2), vortex-mixed, and filtered through a 0.22 μm membrane to obtain the test solution for instrumental analysis.

A sulfonic acid-based cation-exchange chromatography column was used for the analysis. The detection wavelengths were set at 570 nm and 440 nm. A sodium citrate-citric acid buffer was used as the mobile phase with a flow rate of 0.25 mL/min. The separation column (15 cm in length) was maintained at 53 °C, with a column pressure ranging from 9.7 to 10.1 Pa. The ninhydrin solution was delivered at a flow rate of 0.3 mL/min, with a pump pressure maintained between 80 and 90 kgf/cm². The reaction coil temperature was set at 98 °C. An injection volume of 20 μL was used, and the total analysis time was 45 minutes.

2.4. Statistical analysis

The data analysis was performed using the following software tools: Microsoft Excel 2021 was used for organizing and performing basic statistical calculations on the soil property and amino acid content data. Origin 2021 was employed to generate the box plots illustrating the amino acid contents of C. songaricum and the principal component analysis (PCA) score plots for the soil physicochemical properties. SPSS Statistics 26.0 was utilized to calculate the contribution rates of the principal components for the soil physicochemical factors. Based on the R language environment, Z-score normalized heatmaps for the water and soil environmental factors, as well as a correlation network analysis diagram between environmental factors and amino acids, were constructed.

The climatic factors used in this study were obtained from the National Tibetan Plateau Data Center (https://data.tpdc.ac.cn/home). The map boundaries of sampling sites were derived from the National Geomatics Center of China – Tianditu (https://www.tianditu.gov.cn/). The elevation data were acquired from the Geospatial Data Cloud (https://www.gscloud.cn/search). All environmental and spatial data were preprocessed uniformly before being used in correlation network analysis and spatial heterogeneity tests of samples, so as to reveal the spatial correlation characteristics among different geographical populations.

3.1. Distribution of amino acids in different tissues of C. songaricum

To investigate the tissue-specific differences in amino acid content between the stem and stem epidermis of C. songaricum (Supplementary S4–S5 Tables), we performed a box plot analysis of 17 free amino acids and total amino acids in both tissues. Overall, all measured amino acids except proline (Pro) and cysteine (Cys) (which were significantly higher in the stem) and aspartic acid (Asp) (which showed no significant difference) were significantly or highly significantly more abundant in the stem epidermis than in the fleshy stem (Fig 1). The detailed comparisons are as follows.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Box plots of 17 free amino acids and total amino acid contents in the fleshy stem and the stem epidermis of Cynomorium songaricum.

Note: Stem denotes fleshy stem, and Epidermis denotes stem epidermis; *, **, and *** indicate significant differences at P < 0.05, P < 0.01 and P < 0.001, respectively; (a) shows the distribution of the 33 sampling sites, and (b, c) show representative C. songaricum specimens from the Alxa region.

https://doi.org/10.1371/journal.pone.0354162.g001

The total amino acid content in the stem epidermis was significantly higher than that in the stem. Specifically, the contents of threonine (Thr), serine (Ser), glycine (Gly), alanine (Ala), valine (Val), methionine (Met), isoleucine (Ile), leucine (Leu), tyrosine (Tyr), phenylalanine (Phe), histidine (His), lysine (Lys), and arginine (Arg) were all significantly higher in the epidermis. The content of glutamic acid (Glu) was also significantly higher in the epidermis. In contrast, the stem exhibited a significantly higher content of proline (Pro) than the epidermis, and a significantly higher content of cysteine (Cys). Only aspartic acid (Asp) showed no significant difference between the stem and the stem epidermis.

In summary, the stem and stem epidermis of C. songaricum exhibited distinct tissue specificity in their amino acid composition and content. The stem epidermis serves as an enrichment site for most amino acids, particularly essential and functional amino acids, while the stem possesses a higher abundance of proline and cysteine. This tissue-specific distribution pattern may be related to their physiological functional differentiation and adaptation strategies to the desert environment.

3.2. Principal component analysis of soil and water environmental factors

Principal component analysis (PCA) was performed on soil and water environmental factors to investigate their spatial variability (Fig 2; Supplementary S6–S7 Tables). For the soil factors (Fig 2a), four principal components (PC1–PC4) were extracted, collectively explaining 70.55% of the total variance. To identify the main driving factor for each principal component, we selected the environmental variable with the highest absolute loading value within that component (i.e., the maximum loading criterion). Among these, PC1 (32.83% of variance) was primarily driven by water-soluble salts; PC2 (18.60%) was dominated by Soil_Fe; while Soil_Al and Soil_Co were the highly loaded variables for PC3 (11.64%) and PC4 (7.47%), respectively (Table 1). The principal component score plot illustrated the distribution pattern of the 33 geographic populations in the PC1–PC2 space, reflecting the spatial heterogeneity in soil factor combinations. Populations S1 and S19 were clearly separated from others in the score plot, suggesting that the combined characteristics of soil water-soluble salts and iron content in these regions differ significantly from those in other areas.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Variance contribution rates of principal components for soil and water environmental factors.

https://doi.org/10.1371/journal.pone.0354162.t001

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Principal component analysis (PCA) of soil and water environmental factors.

Note: (a) PCA score plot of soil environmental factors; (b) PCA score plot of water environmental factors. S1–S33 represent rhizosphere samples of Cynomorium songaricum from 33 different geographic populations.

https://doi.org/10.1371/journal.pone.0354162.g002

For the water environmental factors, four principal components (PC1–PC4) were also extracted, with a cumulative variance contribution rate of 76.06%. Using the same maximum loading criterion, EC (electrical conductivity) was the key driving factor for PC1 (39.59% of variance), As (arsenic content) drove PC2 (14.86%), CO₃²⁻ (carbonate ion) drove PC3 (12.02%), and pH drove PC4 (9.59%) (Table 1). The principal component score plot (Fig 2b) displayed the distribution of the 33 geographic populations in the PC1–PC2 space, indicating spatial differentiation in water environmental factor combinations. For instance, populations S27, S28, and S32 clustered together in the score plot, while S29 was distinctly separated, reflecting grouping differences in hydrochemical properties—such as ionic strength and arsenic content—among different regions.

In summary, these principal components captured most of the variability in the soil and water environments, laying a foundation for subsequent analysis of the driving mechanisms of environmental factors on the amino acid distribution in C. songaricum.

3.3. Correlation analysis between environmental factors and amino acids

Using heatmaps and correlation network analysis, this study systematically investigated the associations among soil, water, environmental, and climate factors and the amino acid contents in the stem and rind of C. songaricum. The Z-score heatmaps of soil factors (Fig 3a) and water environmental factors (Fig 3b) visually displayed the spatial variation patterns of each environmental factor across the 33 geographic populations. Both soil and water environmental factors exhibited significant spatial heterogeneity across populations, providing the context for environmental variation in the subsequent correlation analysis.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Environmental factor heatmaps and correlation network analysis of Cynomorium songaricum.

Note: Correlations of C. songaricum amino acids with environmental factors. (a) Z-score heatmap of soil environmental factors across the 33 geographic populations; (b) Z-score heatmap of water environmental factors across the 33 geographic populations; (c) Correlation network showing significant associations (|r| ≥ 0.5, P < 0.05). Purple edges indicate positive correlations, green edges indicate negative correlations. Abbreviations: Soil indicators: As arsenic, Cd cadmium, Cr chromium, Al aluminum, Mg magnesium, Na sodium, total potassium, total nitrogen, total phosphorus, water-soluble calcium, water-soluble potassium, water-soluble sodium, water-soluble salts, available potassium, Fe iron, organic matter, available phosphorus, Co cobalt, Ti titanium, Hg mercury, ammonium nitrogen, chloride ion, nitrate nitrogen; Water indicators: electrical conductivity, Ca calcium, K potassium, Na sodium, Mg magnesium, Cr chromium, As arsenic, total dissolved solids, carbonate ion, bicarbonate ion, fluoride ion, phosphate ion, sulfate ion, chloride ion, nitrate ion; Climate factors: T average temperature, Pre annual precipitation.

https://doi.org/10.1371/journal.pone.0354162.g003

Pearson correlation coefficients were calculated between all environmental factors (soil, water, and climate) and amino acids in the stem epidermis and fleshy stem. For network visualization (Fig 3c), only correlation pairs with |r| ≥ 0.5 and P < 0.05 were retained. The correlation network analysis (Fig 3c) further revealed differential association patterns between environmental factors and tissue-specific amino acids. Soil factors showed significant negative correlations with most amino acids in the fleshy stem, indicating that high salinity and high ionic strength inhibit amino acid accumulation in the fleshy stem. This may represent an adaptive strategy of C. songaricum to cope with salt stress, reduce intracellular osmotic pressure, and alleviate ion toxicity. In contrast, water factors exhibited significant positive correlations with various amino acids in the stem epidermis, suggesting that the epidermis, as the outermost interface tissue, is highly sensitive to hydrochemical characteristics and can achieve environmental perception and defense responses by enriching functional amino acids. Meanwhile, climatic factors were also significantly correlated with multiple amino acids in both tissues, implying that large-scale climatic conditions jointly regulate amino acid metabolism together with soil and water factors.

In summary, these distinct correlation patterns clearly reveal the tissue-specific environmental response strategy of C. songaricum: amino acids in the fleshy stem are mainly regulated by soil salinity to maintain osmotic balance and redox homeostasis, whereas amino acids in the epidermis are primarily driven by water conditions to support environmental perception and stress defense. The physiological functional differentiation between the two tissues further elucidates the core mechanism by which C. songaricum adapts to heterogeneous desert habitats through amino acid metabolic division of labor.

4.1. Amino acid metabolism and ecological adaptation

This study reveals that C. songaricum has developed a sophisticated internal strategy at the organ level to cope with the extreme desert environment. Data indicate that the contents of 13 amino acids—including Thr, Ser, and Phe—in the stem epidermis are significantly higher than those in the fleshy stem, with an average increase of 42.7%. This suggests that the stem epidermis is far from being a simple protective tissue but rather serves as an active “frontier for defense and signal perception.” The enrichment of functional amino acids such as threonine and serine may provide the molecular basis for constructing complex defense metabolic networks, including glycosylation modifications and stress signal transduction [22]. Meanwhile, aromatic amino acids like Tyr and Phe, which act as precursors for UV-screening compounds and alkaloids, directly enhance the physical barrier function against intense radiation and biotic stress [23,24]. This epidermal specialization represents a frontline sensing and defensive module that enables rapid detection and response to fluctuating desert microenvironments.

Notably, Asp showed no tissue difference but was strongly regulated by environmental factors in the epidermis. As a central precursor for multiple functional amino acids, Asp maintains stable content while its metabolism is highly sensitive to soil–water cues [25,26]. In contrast, the fleshy stem has evolved into an efficient internal homeostasis maintenance system. The proline content in the fleshy stem is 28.3% higher than that in the stem epidermis, which not only contributes to classical osmotic regulation to maintain cell turgor pressure [11], but also plays a critical role in stabilizing protein structures as a molecular chaperone. This function is essential for the water-storage organ to cope with drastic fluctuations in internal and external water conditions [27]. More importantly, the significant accumulation of cysteine in the fleshy stem highlights its role as a strategic hub in the antioxidant defense system of C. songaricum [28]. As a precursor for glutathione synthesis, high cysteine levels ensure a robust redox buffering capacity, enabling the plant to mitigate oxidative damage induced by intense light and drought [29].

In summary, the highly specific differentiation in amino acid metabolism between the stem epidermis and fleshy stem of C. songaricum is not coincidental but represents an evolutionary adaptation strategy to multiple stressors in desert habitats. This strategy involves internal division of labor and synergistic adaptation: the stem epidermis specializes in environmental perception and immediate defense, while the fleshy stem focuses on osmotic regulation and long-term maintenance of redox homeostasis. Such functional complementarity and optimized resource allocation between organs significantly enhance the overall adaptability and survival probability of C. songaricum in harsh environments.

4.2. Environmental factors and their linkages to amino acids

This study further elucidates the significant functional divergence between the stem epidermis and fleshy stem of C. songaricum in response to rhizosphere environmental factors [30]. Principal component analysis indicated that soil and water environmental factors collectively form the key environmental context influencing the metabolism of C. songaricum, with soil water-soluble salts and iron content, as well as water electrical conductivity and carbonate ions, being the dominant factors driving the variation in soil and water environments, respectively.

Correlation network analysis revealed that amino acids in the stem epidermis showed significant positive correlations with water electrical conductivity and carbonate ions, suggesting that the stem epidermis, as an environmental interface tissue, may rapidly respond to hydrochemical changes through osmotic or epidermal signaling pathways [31,32]. In contrast, amino acids in the fleshy stem were mainly negatively regulated by soil water-soluble salts, reflecting their physiological role as a storage organ in mitigating soil osmotic stress through the accumulation of proline and cysteine [33]. This differential response mechanism reflects a synergistic adaptation strategy developed by C. songaricum at the organ level: “stem epidermis perceives water environment, while fleshy stem buffers soil stress.” This dual modular response allows the holoparasite to maintain metabolic autonomy despite relying on hosts for nutrients, thereby finely tuning its physiology to heterogeneous soil–water conditions. Future studies could integrate transcriptomic and enzymatic activity analyses to further unravel the underlying molecular regulatory network and construct a trinity ecological adaptation model of “perception-storage-regulation,” providing a theoretical framework for understanding resource allocation and metabolic division of labor in plants within heterogeneous habitats [34–36].

4.3. Research significance and conservation implications

The innovation of this study lies in the analysis of plant ecological adaptability at the internal substructural level of plant organs. It preliminarily explores that the “spatial metabolic division of labor of amino acids” may be an important mechanism for C. songaricum to respond to environmental heterogeneity [37,38]. Based on large-sample data from 33 geographic populations and multivariate statistical analysis, we have demonstrated the significant value of intra-organ specificity studies in elucidating plant environmental adaptability. This understanding not only holds theoretical importance for plant stress ecology but also provides clear guidance for the conservation of C. songaricum resources [3]. It is worth noting that this study mainly focuses on amino acid metabolism without systematically considering other metabolites closely related to plant water status, such as carbohydrates and organic acids. Therefore, future research can further integrate various metabolites, including sugars and inorganic ions, to construct a tissue-specific global metabolic regulatory network, so as to more comprehensively and systematically clarify the adaptation strategies of C. songaricum to extreme environments. In terms of resource conservation practice, fine regulation of the rhizosphere microenvironment can be carried out and differentiated management can be implemented according to the metabolic characteristics of different plant tissues, which is expected to further improve the applicability and long-term effectiveness of conservation strategies for C. songaricum [5]. Beyond species-specific insights, this work establishes a generalizable framework for studying metabolic specialization and environmental adaptation in other desert-adapted parasitic or succulent plants. By revealing how tissue-specific metabolism translates environmental signals into adaptive phenotypes, this study opens new avenues for integrative research linking metabolomics, ecophysiology, and conservation biology.