Multistrain Microbial Inoculant Enhances Yield and Medicinal Quality of Glycyrrhiza uralensis in Arid Saline–Alkali Soil and Modulate Root Nutrients and Microbial Diversity

Abstract

Glycyrrhiza uralensis (G. uralensis), a leguminous plant, is an important medicinal and economic plant in saline–alkaline soils of arid regions in China. Its main bioactive components include liquiritin, glycyrrhizic acid, and flavonoids, which play significant roles in maintaining human health and preventing and adjuvantly treating related diseases. However, the cultivation of G. uralensis is easily restricted by adverse soil conditions in these regions, characterized by high salinity, high alkalinity, and nutrient deficiency. This study investigated the impacts of four multistrain microbial inoculants (Pa, Pb, Pc, Pd) on the growth performance and bioactive compound accumulation of G. uralensis in moderately saline–sodic soil. The aim was to screen the most beneficial inoculant from these strains, which were isolated from the rhizosphere of plants in moderately saline–alkaline soils of the Hexi Corridor and possess native advantages with excellent adaptability to arid environments. The results showed that inoculant Pc, comprising Pseudomonas silesiensis, Arthrobacter sp. GCG3, and Rhizobium sp. DG1, exhibited superior performance: it induced a 0.86-unit reduction in lateral root number relative to the control, while promoting significant increases in single-plant dry weight (101.70%), single-plant liquiritin (177.93%), single-plant glycyrrhizic acid (106.10%), and single-plant total flavonoids (107.64%). Application of the composite microbial inoculant Pc induced no significant changes in the pH and soluble salt content of G. uralensis rhizospheric soils. However, it promoted root utilization of soil organic matter and nitrate, while significantly increasing the contents of available potassium and available phosphorus in the rhizosphere. High-throughput sequencing revealed that Pc reorganized the rhizospheric microbial communities of G. uralensis, inducing pronounced shifts in the relative abundances of rhizospheric bacteria and fungi, leading to significant enrichment of target bacterial genera (Arthrobacter, Pseudomonas, Rhizobium), concomitant suppression of pathogenic fungi, and proliferation of beneficial fungi (Mortierella, Cladosporium). Correlation analyses showed that these microbial shifts were linked to improved plant nutrition and secondary metabolite biosynthesis. This study highlights Pc as a sustainable strategy to enhance G. uralensis yield and medicinal quality in saline–alkali ecosystems by mediating microbe–plant–nutrient interactions.

1. Introduction

Saline–sodic soils, also known as salt-affected soils, represent a critical land resource and encompass various soil types including salinized soils, salt soils, solonetzic soils, and alkali soils. According to data from the United Nations Food and Agriculture Organization (FAO, 21 October 2021), global saline–alkali lands cover over 833 million hectares (8.7% of the Earth’s surface), primarily distributed in arid or semi-arid regions of Africa, Asia, and Latin America [1]. Statistically, the area of salt-affected soils is expanding at a rate of approximately 1.0–2.0 million hectares per year [2]. About 70% of soil salinization occurs in the semi-arid and arid regions of Northwest in China [3], while approximately 1.02 × 10⁶ hectares of cropland are affected in Gansu Province [4], primarily distributed in the arid and semi-arid regions of northern Hexi Corridor and Jingtai Basin [5]. Soil saline–alkalization of arable land will be increasing in the future if effective solutions are not applied [6,7]. Constrained by water resource scarcity and high economic costs, physical measures such as salt leaching by irrigation are difficult to implement in these regions [8], while applying chemical amendments proves prohibitively expensive [9]. Although these approaches can effectively reduce soil salinity in the short term [10], long-term reliance on them leads to salt rebound [11], soil structure degradation, or environmental pollution [12,13]. Against this backdrop, utilization-oriented bioremediation technologies have gradually emerged as a crucial approach for the amelioration and restoration of arid inland saline–alkali soils [14,15,16].

In the field of integrated saline–alkali land utilization, certain medicinal plants exhibit unique eco-economic dual value [17,18,19,20,21]. These species can improve site conditions by reducing rhizospheric soil pH and increasing organic matter content, leveraging their excellent salt-tolerant physiological mechanisms and salt accumulation capacity. Compared with food crops or forages, they also show significant advantages in economic output per unit area [22]. However, even salt-tolerant plants have limited salinity–alkalinity tolerance. When the salt tolerance threshold is exceeded, the growth of salt-tolerant plants becomes restricted [23,24,25]. Moreover, when such stress persists, single phytoremediation models often fall into a dilemma where stress resistance metabolism takes precedence over growth metabolism [26,27,28].

Take Glycyrrhiza uralensis (G. uralensis), a leguminous medicinal plant widely cultivated in the inland saline–alkali areas of the Hexi Corridor, as an example. It holds an important position in both traditional medical systems (such as Traditional Chinese Medicine, TCM) and modern pharmacology. Its roots and rhizomes are rich in bioactive compounds, including liquiritin, glycyrrhizic acid, and flavonoids. These compounds, as key indicators for evaluating the quality of licorice [29], are widely used in the treatment of inflammatory diseases [30], liver diseases [31], and viral infections [32], as well as in immunomodulation [33] and antioxidation [34,35]. Due to its irreplaceable medicinal properties, G. uralensis has been included in the Pharmacopoeia of the People’s Republic of China. Its ability to grow in salt-stressed environments is a crucial mechanism underlying its status as an “authentic medicinal material” [36].

Numerous studies have shown that mild soil salinization can promote the growth and accumulation of bioactive components in G. uralensis [37,38,39]. However, as salinization intensifies to moderate and severe levels, the dual stress of high-concentration Na⁺/Cl⁻ or Na⁺/SO₄²⁻ toxicity and osmotic pressure imposes significant limitations. These stresses hinder root development, reduce photosynthetic efficiency [40], and cause a 40–60% reduction in biomass accumulation compared to mild saline–alkali conditions, thereby becoming a key bottleneck restricting the species’ application in saline–alkali ecological restoration [41].

Recent research highlights that integrating microbial intervention with plant cultivation not only preserves the saline–alkali land improvement capacity and economic value of medicinal plants but also enhances soil desalination and alkalinity reduction. Through microbial community regulation, the growth inhibition dilemma of single phytoremediation models is overcome, representing a promising synergistic effect in saline–alkali land restoration [42,43,44]. To address this challenge, leveraging the theory of rhizospheric microbe–plant interactions, this study evaluates four liquid microbial inoculants developed by our research group for cultivating G. uralensis in arid saline environments.

The strains in these four inoculants were isolated from the rhizosphere of plants growing in moderately saline–alkaline soils of the Hexi Corridor. Endowed with the advantage of their native origin, they exhibit excellent adaptability to arid environments. By applying these indigenous strains, this study aims to achieve three objectives: (1) screen the optimal inoculant for promoting G. uralensis growth and bioactive component accumulation through comparative analysis of growth indices, yield, and component content; (2) characterize the effects of the optimal inoculant on rhizospheric soil nutrients and microbial community structure; and (3) elucidate the inter-relationships between microbial communities, soil nutrient factors, growth indices, and nutrient contents. By enhancing the rhizospheric microenvironment, the selected inoculant aims to overcome the stress of moderate-to-severe salinization, achieving synergistic improvements in yield and bioactive component enrichment. This research seeks to provide a sustainable solution for biological measures of arid saline–alkali soils, integrating ecological benefits with economic viability.

2. Materials and Methods

2.1. Field Experiment Site and Design

The experimental field is situated in Xicha Town, Gaolan County, Lanzhou City, Gansu Province, China (36°42′ N,104°19′ E), characterized by a temperate semi-arid continental climate. Climatic conditions include an average annual temperature of 6.9 °C, annual precipitation of 300–350 mm, annual evaporation of 1880 mm, a 139-day frost-free period, annual sunshine duration of 1744–2659 h (60% sunshine percentage), and prevailing northwest winds (average annual wind speed: 2.3 m·s⁻¹), with southeast winds dominant in summer–autumn and northwest winds in winter-spring. The soil is classified as Huangmian soil, a subtype of loessial soil.

Soil properties in the 0–20 cm topsoil layer are as follows: total nitrogen 0.22 g·kg⁻¹, available phosphorus 41.80 mg·kg⁻¹, available potassium 227.22 mg·kg⁻¹, organic matter 6.84 g·kg⁻¹, pH 7.80 (slightly alkaline), and soil soluble salt content 4.08 g·kg⁻¹, indicating moderate salinization [45].

One month prior to planting, the experimental site was prepared by uniformly applying fully decomposed cow manure at a rate of 30,000 kg·ha⁻¹ to the soil surface, followed by plowing to a depth of 25–30 cm. Concurrently, a mixture fertilizer of superphosphate (300 kg·ha⁻¹), diamine phosphate (225 kg·ha⁻¹), and potassium fulvate (75 kg·ha⁻¹) (Yunnan Ruilinfeng Chemical Industry Co., Ltd., Kunming, China)was evenly broadcasted over the soil and incorporated through a second plowing to ensure thorough homogenization with the native soil. The field was subsequently leveled to facilitate uniform planting.

The experiment was officially initiated on 25 April 2023. A furrow planting approach was adopted. One-year-old G. uralensis seedlings (sourced from Minqin Gendegen Chinese Medicinal Materials Co., Ltd., Wuwei, China) with uniform growth (root length 40–45 cm, diameter 0.8–1.0 cm) and free of pests/diseases were selected. Seedlings were disinfected by immersion in an 800–1000-fold diluted solution of 50% carbendazim wettable powder (Shandong Chengxin Biotechnology Co., Ltd., Weifang, China) for 10–15 min, air-dried to remove surface moisture, and then coated with inoculant slurries or a control slurry. Each slurry was prepared by diluting microbial inoculant solutions 100-fold with irrigation water and mixing with 2 L of sterile soil to form a uniform mud coating.

A randomized block design was employed, with each plot measuring 9 m × 1.5 m (13.5 m²) and five replicates per treatment. Buffer rows (1.5 m wide) were established around each plot to minimize edge effects. Two planting furrows (15–20 cm deep) were dug per plot, with 30 seedlings placed horizontally and in linear rows with 30 cm spacing between root collar (60 seedlings per plot, 300 seedlings total across five replicates). Seedlings were covered with soil, gently compacted, and irrigated with tap water to settle the roots.

Post-germination, liquid microbial inoculants were applied three times during the growing season at a rate of 2 L·ha⁻¹ (diluted 100-fold), targeting the root zone; the control group received only irrigation water.

Four liquid multistrain inoculants (Pa, Pb, Pc, Pd) were tested, each containing distinct dominant strains (determined based on careful evaluation of functional complementarity and antagonistic interactions):

Pa: Bacillus subtilis and Paenibacillus peoriae;

Pb: Pseudomonas silesiensis, Arthrobacter globiformis, and Sinorhizobium meliloti;

Pc: Pseudomonas silesiensis, Arthrobacter sp. GCG3, and Rhizobium sp. DG1;

Pd: Bacillus subtilis, Pseudomonas silesiensis, and Arthrobacter globiformis.

The four inoculants described above collectively comprise seven strains, which were isolated from the rhizosphere of plants growing in moderately saline–alkaline soils of the Hexi Corridor, Gansu Province, China. The specific plant species are listed in

Table 1. Source of the seven strains.

Strains	Source
Bacillus subtilis	Tamarix chinensis
Paenibacillus peoriae	Lycium ruthenicum
Pseudomonas silesiensis	Elaeagnus angustifolia
Arthrobacter globiformis	Glycyrrhiza uralensis
Sinorhizobium meliloti	Medicago sativa
Arthrobacter sp. GCG3	Glycyrrhiza uralensis
Rhizobium sp. DG1	Nitraria tangutorum

.

All of them can tolerate more than 5% NaCl and a pH of 8 or higher in LB medium. Detailed information regarding the specific host plants and functional characteristics is provided in

Table 2. Functional characteristics of the seven strains.

Strains	Nitrogenase Activity (IU·L−1)	IAA Increment (μg·mL−1)	Organic Phosphorus Increment (μg·mL−1)	Inorganic Phosphorus Increment (μg·mL−1)	ESP Increment (mg·L−1)	Siderophore Halo Diameter (mm)	Antifungal Rate (%)
							Bc	Fs	Fo	Rs	Ss
Bacillus subtilis	197.77 ± 1.69 a	6.10 ± 0.78 c	0.00 ± 0.07 c	1.44 ± 0.43 b	234.38 ± 16.44 a	2.67 ± 1.20 c	-	-	-	-	-
Paenibacillus peoriae	186.49 ± 1.95 b	4.44 ± 0.70 d	7.27 ± 1.56 a	0.00 ± 0.56 c	195.49 ± 9.63 bc	5.50 ± 0.29 b	-	-	-	-	-
Pseudomonas silesiensis	170.88 ± 0.99 c	8.28 ± 1.33 b	4.41 ± 1.77 b	3.38 ± 0.44 a	171.45 ± 2.59 c	8.17 ± 0.44 a	-	-	-	-	-
Arthrobacter globiformis	168.44 ± 2.28 c	8.82 ± 0.69 b	0.00 ± 0.05 c	0.00 ± 0.06 c	177.14 ± 4.86 c	4.33 ± 0.72 bc	39.09 ± 0.89 b	15.69 ± 0.91 b	15.89 ± 0.54 b	0.00 ± 0.11 c	8.25 ± 0.23 b
Sinorhizobium meliloti	184.46 ± 2.21 b	15.94 ± 1.35 a	0.00 ± 0.17 c	0.00 ± 0.04 c	171.82 ± 5.04 c	5.83 ± 0.44 b	-	-	-	-	-
Arthrobacter sp. GCG3	169.64 ± 2.00 c	17.70 ± 2.89 a	0.00 ± 0.04 c	0.00 ± 0.30 c	200.26 ± 3.88 bc	0.00 ± 0.00 d	76.36 ± 1.05 a	53.92 ± 0.76 a	52.34 ± 0.57 a	63.27 ± 1.65 a	80.00 ± 0.88 a
Rhizobium sp. DG1	154.58 ± 1.42 d	8.82 ± 0.29 b	4.74 ± 0.89 b	0.00 ± 0.02 c	211.08 ± 5.28 ab	6.17 ± 0.44 ab	73.82 ± 0.56 a	50.98 ± 1.38 a	49.72 ± 2.06 a	21.43 ± 0.06 b	76.67 ± 2.33 a

Lowercase letters indicate significant differences between treatments, ANOVA, p < 0.05. Nitrogenase activity was determined following the manufacturer’s instructions for the Microbial Nitrogenase (NITS) ELISA Kit (Jiangsu Enzyme-linked Biotechnology Co., Ltd., Yancheng, China); phosphorus solubilization capacity was measured using the molybdenum-antimony anti-colorimetric method; the increment of IAA in fermentation broth was determined by the Salkowski colorimetric method; the increment of extracellular polysaccharides (ESP) was determined by the phenol-sulfuric acid method; siderophore detection was conducted using the CAS medium plate inoculation method; the inhibitory effects of strains against plant pathogenic fungi were assessed using the dual culture assay; plant pathogenic fungi abbreviations: Bc (Botrytis cinerea), Fs (Fusarium solani), Fo (Fusarium oxysporum), Rs (Rhizoctonia solani), Ss (Sclerotinia sclerotiorum); “-” indicates that the strain exhibits no inhibitory activity against the pathogenic fungus.

.

Each strain was individually cultured in LB liquid medium with constant shaking at 180 r·min⁻¹ and 28 °C. Upon reaching the logarithmic phase, the strains were mixed in equal proportions according to the predefined combinations, with the total viable count adjusted to 2 × 10¹⁰ colony-forming units per milliliter (CFU·mL⁻¹).

2.2. Sample Preparation and Analysis

Plants were harvested in their entirety in September of the following year. Fifteen plants were randomly selected from each plot, yielding 75 plants per treatment. Rhizospheric soil was collected using the root-shaking method: one subsample was immediately stored in sterile self-sealing bags at −80 °C for subsequent high-throughput sequencing, while another subsample was air-dried at room temperature, sieved through a 2 mm mesh, and used for soil nutrient analysis.

Root morphological traits—including main root length (MRL), main root diameter (MRD), and number of lateral roots (LRN)—were recorded in the field. Roots were then transported to the laboratory and gently rinsed with deionized water to remove adhering soil, air-dried, sliced into 0.5 cm segments, and oven-dried at 40 °C to constant weight for biomass determination (single-plant dry weight, SPDW).

Liquiritin and glycyrrhizic acid are the bioactive compounds measured in the chemical analysis. Specifically, their contents (%) were quantified following the protocols outlined in the Chinese Pharmacopoeia (2020 Edition, Volume I) [46]. Total flavonoids were extracted using ultrasonic-assisted extraction [47] and assayed via the AlCl₃ colorimetric method [48], with absorbance measured at 510 nm using a spectrophotometer (TU-1901, Beijing Purkinje General Instrument Co., Ltd., Beijing, China). The contents of liquiritin, glycyrrhizic acid, and total flavonoids per plant (SPLQT, SPGA, SPTF) were calculated based on SPDW and their respective mass percentages.

Soil pH, soil soluble salt (SSS) and soil nutrient including soil organic matter (SOM), total nitrogen (TN), available phosphorus (AP), available potassium (AK), and nitrate nitrogen (NO₃⁻-N) were determined for the content, as mentioned in the book by Bao [49].

2.3. High-Throughput Sequencing of Rhizospheric Microbiota and Data Processing

Soil microbial DNA extraction, PCR amplification, library preparation, and Illumina MiSeq sequencing were performed by Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China). Bacterial and fungal community structures were analyzed by targeting the V3–V4 hypervariable regions of the bacterial 16S rRNA gene and the ITS2 region of the fungal ITS rRNA gene, respectively. Bacterial amplification used universal primers 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′), while fungal amplification employed primers ITS1 (5′-TCCGTAGGTGAACCTGCGG-3′) and ITS4 (5′-TCCTCCGCTTATTGATATGC-3′). Quality control of paired-end raw sequencing reads was performed using the fastp v0.19.6 [50], followed by read assembly with the FLASH v1.2.11 [51]. The UPARSE v7.1 [52] was used to cluster the quality-controlled and assembled sequences into Operational Taxonomic Units (OTUs) based on 97% sequence similarity, with chimeras removed simultaneously. To minimize the impact of sequencing depth on subsequent Alpha diversity (α diversity) and Beta diversity (β diversity) analyses, all samples were rarefied to an equal number of sequences. After rarefaction, the average sequence coverage of each sample remained at 97.00%. For species taxonomic annotation, appropriate databases were selected according to the research objects: bacterial OTU sequences were aligned against the Silva 16S rRNA gene database (v138) using the RDP classifier v2.11 [53], while fungal OTU sequences were aligned against the UNITE fungal ITS database (v9.0). A confidence threshold of 70% was set for both, and the community composition of each sample was statistically analyzed at different taxonomic levels.

2.4. Data Analysis

Initial data processing (e.g., mean calculation, error propagation) was performed using Microsoft Excel 2019, while graphical representations were generated using Origin 2021. Statistical significance (p < 0.05) was determined via independent samples t-tests in SPSS 26.0.

Microbial community analyses were conducted on the Majorbio platform (https://www.majorbio.com/, accessed on 12 December 2024). Specifically, α diversity and community structure were analyzed using a combination of mothur v1.46.1 [54]. Bray–Curtis dissimilarity matrices were calculated using the vegan package (v2.4-6), followed by principal coordinate analysis (PCoA) implemented via phyloseq (v1.16.2). The resulting PCoA plots were visualized using ggplot2 (v2.2.1). Differential abundance analysis between treatment groups was performed using Wilcoxon rank-sum tests implemented in the base stats package. Taxonomic composition bar charts were generated using phyloseq (v1.16.2), while Spearman correlation heatmaps depicting relationships among microbial taxa, soil nutrient factors, G. uralensis growth parameters, and nutrient contents were constructed using the pheatmap package (v1.0.10). All statistical analyses and visualizations were conducted using R software (v3.3.1; R Foundation for Statistical Computing, Vienna, Austria).

3. Results

3.1. Effects of Different Microbial Inoculants on Growth Indices, Yield, and Component Content of G. uralensis

Different microbial inoculants exhibited differential effects on MRL, MRD, LRNSPDW, SPLQT, SPGA and SPTF in G. uralensis (Figure 1 and Figure 2). No significant differences in MRL were observed among treatments. However, LRN was significantly lower in Pb and Pc treatments compared to the control (p < 0.05). The Pc treatment exhibited the highest SPDW, significantly exceeding Pb, Pd, and control groups. Additionally, Pc showed significant differences in bioactive component contents compared to other treatments and the control (p < 0.05). These results indicate that, compared to other treatments and the control, inoculant Pc exhibited more stable overall growth-promoting performance on G. uralensis. Therefore, inoculant Pc was identified as the optimal microbial compound inoculant for promoting licorice growth and improving quality.

3.2. Effect of Inoculant Pc Application on Rhizospheric Soil Physicochemical Properties of G. uralensis in Saline–Alkali Soil

The effects of inoculant Pc application on rhizospheric soil physicochemical properties of G. uralensis in saline–alkali soil are presented in

Table 3. Effect of inoculant Pc application on rhizospheric soil physicochemical properties of G. uralensis in saline–alkali soil.

Treatments	pH	SSS (g kg−1)	TN (mg kg−1)	AK (mg kg−1)	AP (mg·kg−1)	NO3−-N (μg·g−1)	SOM (mg·g−1)
Pc	8.05 ± 0.08 a	7.38 ± 0.02 a	291.65 ± 1.61 b	269.37 ± 0.40 a	45.91 ± 0.39 a	13.56 ± 0.54 c	4.30 ± 0.07 c
CK	7.91 ± 0.05 a	7.75 ± 0.18 a	356.64 ± 2.57 a	97.42 ± 2.20 c	36.64 ± 0.60 c	15.14 ± 0.34 b	6.48 ± 0.08 b
Basic soil data.	7.80 ± 0.13 ab	7.38 ± 0.13 a	216.11 ± 1.97 c	227.22 ± 0.52 b	41.80 ± 0.68 b	48.02 ± 0.48 a	6.84 ± 0.06 a

Lowercase letters indicate significant differences between treatments, ANOVA, p < 0.05. The abbreviations are indicated as follows: SSS (soil soluble salt); SOM (soil organic matter); TN (total nitrogen); AP (available phosphorus); AK (available potassium); NO₃⁻-N (ni-trate nitrogen).

. Application of inoculant Pc caused no significant changes in soil pH or soluble salt content, indicating that one year of seedling cultivation did not significantly alter the salinity–alkalinity of moderately saline–alkali soil. However, other indicators showed that SOM and NO₃⁻-N in the rhizosphere were significantly lower when compared with both pre-planting soil and CK. AK and AP contents decreased in the CK treatment relative to pre-planting soil but exhibited an increasing trend in treatment Pc.

3.3. Effect of Inoculant Pc Application on Rhizospheric Soil Microbial Diversity of G. Uralensis in Saline–Alkali Soil

3.3.1. Sequencing Depth and α Diversity of the Sampled Soils

With the increase in sequencing depth, the rarefaction curves of rhizospheric soil bacterial communities gradually plateaued. The results (Figure 3a) showed that the sequencing data for each treatment were sufficiently large to accurately reflect the majority of rhizospheric bacterial diversity characteristics. Consistently, coverage indices for all samples exceeded 97.5% (Figure 3b), corroborating the rarefaction curve results and indicating that the sequencing data reliably represented the community composition of rhizospheric soil bacteria in G. uralensis.

Following the application of compound microbial agent Pc, the Chao index (species richness), Sobs index (possibly referring to the number of species in specific groups), and Shannon index (comprehensive diversity index) of rhizospheric soil bacterial communities were significantly lower than the control, while the Simpson index (dominance index) increased significantly (Figure 3d–f). This may be attributed to the target strains in the agent forming a competitive advantage in the rhizospheric environment, suppressing indigenous flora through resource preemption or metabolic products. This leads to a decrease in community species richness and evenness, with dominant species taking the leading role, thereby shifting the community structure toward a direction of low diversity and high prevalence of dominant species.

Similarly, combining the results of rarefaction curves and coverage indices (Figure 4a,b), the sequencing data can also truly reflect the community composition of rhizospheric soil fungi in G. uralensis.

Application of the compound microbial agent Pc significantly increased the species richness of rhizosphere fungi (as indicated by elevated Chao and Sobs indices) and reduced the community dominance (reflected by a decreased Simpson index, i.e., improved evenness). However, the magnitude of the increase in richness and the improvement in evenness offset each other in the calculation of the Shannon index, resulting in no significant difference in the Shannon index (Figure 4c,d). These results indicate that Pc reshapes the rhizosphere fungal community structure primarily through “increasing species number” and “breaking the monopoly of dominant species” but does not significantly alter the overall diversity level of the fungal community.

3.3.2. β Diversity of Microbial Communities in the Sampled Soils

A Bray–Curtis distance-based PCoA analysis of the rhizospheric soil microbial communities in G. uralensis revealed significant statistical differences (p < 0.05) in the community structures between the Pc treatment group and the control group at the OTU level (Figure 5), for both bacterial and fungal communities. This indicates that the application of compound microbial agent Pc can significantly alter the composition of the rhizospheric microbial community in G. uralensis.

3.3.3. Relative Abundance and Differential Analysis of Dominant Rhizospheric Genera of Microbial Communities in the Sampled Soils

The genus-level relative abundances of rhizospheric soil bacteria are depicted in Figure 6a. Among the top 10 dominant taxa, Arthrobacter and Pseudomonas—the keystone constituents of compound microbial agent Pc—ranked first and second, comprising 25.67% and 6.82% of the community, respectively. In the control group, their abundances were 18.68% and 0.96%, corresponding to increases of 37.41% and 608.33% in the Pc treatment. Rhizobium also showed marked enrichment (2.40% vs. 0.66%, +263.63%). Non-inoculant genera exhibited divergent trends: Novosphingobium increased by 26.41%, whereas Sphingomonas decreased by 38.06% (1.53% vs. 2.47%).

Wilcoxon rank-sum tests (Figure 6b) confirmed significant abundance shifts in target genera (Pseudomonas, Rhizobium), as well as Novosphingobium, Sphingomonas, and norank_o_Vicinamibacterales between Pc and control treatments.

Genus-level relative abundances of rhizospheric fungi in G. uralensis are depicted in Figure 6c, illustrating pronounced shifts in fungal community structure following Pc treatment. Noteworthy, the plant pathogen Fusarium exhibited a 26.27% reduction in relative abundance within the Pc-treated rhizosphere (8.56%) compared to the control (11.61%), a suppression effect consistent with our prior findings [55]. Relative abundances of weak pathogenic genera Pseudogymnoascus and Pseudombrophila decreased by 42.44% and 99.55% in the Pc treatment (15.10%, 0.03%) compared to the control (26.29%, 6.68%). Concurrently, beneficial fungi such as Mortierella and Cladosporium were promoted, with Cladosporium exhibiting a 33.39-fold increase (9.35% vs. 0.28%). Filobasidium and Vishniacozyma also showed marked enrichment.

Wilcoxon rank-sum tests (Figure 6d) revealed significant abundance shifts (p < 0.05) between the Pc treatment and control groups for Fusarium, Filobasidium, Vishniaco-zyma, Cladosporium, Pseudombrophila, and Alternaria.

3.4. Correlations Between Dominant Rhizospheric Microbes in G. Uralensis and Soil Nutrients, Plant Growth, and Bioactive Compounds

Correlation analysis between dominant rhizospheric microorganisms and soil properties, plant growth indices, and bioactive compounds (Figure 7) revealed that after colonization by Pc strains, Arthrobacter exhibited a significant positive correlation with single-plant glycyrrhizic acid content (SPGA). Pseudomonas showed positive correlations with single-plant total flavonoids (SPTFs), single-plant total flavonoids (SPLQT), SPGA, single-plant dry weight (SPDW), and available potassium (AK). Similarly, Rhizobium correlated positively with SPTF, SPLQT, SPDW and AK. Conversely, Sphingomonas abundance decreased significantly post-Pc treatment and negatively correlated with SPGA, SPLQT, SPTF, AK, and available potassium (AP).

Similarly, despite lacking fungal components, Pc application significantly increased the abundances of Filobasidium, Chaetomium, Vishniacozyma, and Alternaria. Among these genera, Filobasidium positively correlated with SPLQT, SPGA, SPDW, and AK and AP. Alternaria exhibited positive correlations with SPTF, SPLQT, SPGA, SPDW, and AK and AP. Both Vishniacozyma and Chaetomium were positively associated with SPTF, SPLQT, SPGA, AK and AP; moreover, they were negatively correlated with soil organic matter (SOM) and total nitrogen (TN). In contrast, Pseudombrophila abundance decreased significantly under Pc treatment and negatively correlated with SPTF, SPLQT, SPGA, SPDW, and AK and AP.

4. Discussion

4.1. Microbial Inoculant-Mediated Enhancement of G. uralensis Performance

Microbial inoculants have emerged as a pivotal strategy to promote plant growth and secondary metabolism, with their impacts on medicinal plant quality garnering increasing research attention [56,57,58,59]. This study investigated the effects of four microbial inoculants (Pa, Pb, Pc, Pd) on the growth and bioactive components of G. uralensis, aiming to provide a theoretical basis for their application in licorice cultivation. While the four inoculants did not significantly affect main root length or diameter, Pb and Pc treatments notably reduced lateral root number, with Pc uniquely increasing single-plant dry weight and contents of liquiritin, glycyrrhizic acid, and total flavonoids (p < 0.05). This phenomenon is likely linked to the functional specificity of microbial consortia and root resource allocation strategies.

The inoculants Pc and Pb exhibited similar microbial compositions. Previous studies have shown that Pseudomonas can restructure rhizosphere architecture through plant hormone regulation, volatile compound release, and other pathways, and its positive regulatory effect on lateral root development has been confirmed in most studies [60,61,62]. However, this study found that the number of lateral roots in G. uralensis treated with Pc and Pb consortia containing Pseudomonas was significantly lower than that in the control group. It is speculated that the synergistic interaction between Pseudomonas and Arthrobacter as well as nitrogen-fixing bacteria (such as Sinorhizobium meliloti and Rhizobia sp. DG1) may drive changes in carbon–nitrogen allocation patterns in the rhizosphere by improving nitrogen supply efficiency. Arthrobacter species are mostly involved in nitrogen fixation [55,63,64], while Sinorhizobium meliloti and Rhizobia sp. DG1 through nodule formation and nitrogen fixation maintain the carbon–nitrogen dynamic balance in the rhizosphere [65,66]. This synergistic process may reduce carbon source allocation to lateral roots, prompting carbon skeletons to preferentially transfer to dry matter accumulation and secondary metabolite synthesis in the main roots of G. uralensis.

Notably, the Arthrobacter strains present in both Pc and Pb were isolated from the rhizosphere of G. uralensis. The key difference between these two inoculants lies in their rhizobial components: the Rhizobium sp. DG1 in Pc was derived from the rhizosphere of Nitraria tangutorum, while the Sinorhizobium meliloti in Pb was isolated from the rhizosphere of Medicago sativa. Under the same resource utilization pattern, Pc exhibited significantly superior performance compared to Pb in promoting host rhizosphere growth and accumulation of functional components. This discrepancy may result from functional differentiation among the strain combinations, or differences in signal transduction pathways and nitrogen-fixing efficiency. Although G. uralensis and Medicago sativa both belong to the Fabaceae family, and numerous studies have demonstrated that G. uralensis can form symbiotic relationships with certain Rhizobium species (included in treatment Pc) and Sinorhizobium meliloti (included in treatment Pb) [66,67], the combination of Sinorhizobium meliloti (from Medicago sativa) and Arthrobacter in this study failed to show a synergistic advantage. This phenomenon is likely associated with the specific binding patterns between each leguminous plant and certain rhizobia. These findings provide a new perspective for deciphering the differences in molecular ecological mechanisms through which similar microbial communities regulate host root development.

From the perspective of root functional trade-offs, although reduced lateral root number may decrease root absorption surface area, the rhizospheric mycelial networks formed by Pc strains may compensatorily enhance the availability of soil phosphorus, potassium (Table 3), and other elements, offsetting the absorption disadvantages of fewer lateral roots. This hypothesis warrants further validation in subsequent studies.

4.2. Impact on Rhizospheric Soil Properties

The inoculant Pc induced no significant changes in soil pH or soluble salt content, a phenomenon closely associated with the short-term (1-year) growth cycle. The physicochemical properties of saline–sodic soils typically exhibit strong buffering capacity [68], whereby the synergistic effects of plant roots and microbes may be insufficient to alter baseline soil salinity/alkalinity in the short term [69]. Moreover, most current studies have reported that the combined action of plants and microorganisms indirectly ameliorates saline–alkali soils by enhancing plant adaptability and microecological functions [70,71,72,73], rather than directly altering soil physical and chemical properties. Thus, applying microbial inoculants in moderately saline–sodic soils may require long-term positioning observations for soil property changes (e.g., 2–3-year cycles) or integration with other amelioration measures (e.g., organic fertilization, salt-tolerant crop rotation) to overcome the “threshold effect” of soil salinity [74,75,76].

The significant decrease in SOM and NO₃⁻-N in the rhizosphere of G. uralensis under Pc treatment indicates a synergistic interaction between functional microbes and roots. Specifically, this synergy may be reflected in the enhanced decomposition of SOM by functional microbes and the promotion of efficient uptake of NO₃⁻-N by roots. The smaller decline in SOM and NO₃⁻-N in CK indirectly indicates that the decomposition of SOM by indigenous microbes is less active and the efficiency of the roots in acquiring nitrogen nutrients is relatively low. The increase in available potassium (AK) and available phosphorus (AP) in the rhizosphere treated with the composite microbial inoculant (Pc) can be attributed to multiple microbially mediated mechanisms. For potassium, microorganisms may secrete organic acids, which can enhance the solubility of potassium-bearing minerals in the soil [77]. In addition to solubilization, microorganisms may facilitate the mobilization and release of potassium into the soil through ion exchange processes [78,79]. Furthermore, extracellular polysaccharides or siderophores secreted by them may further promote potassium release [80,81]. For phosphorus, microorganisms secrete phosphatases (to hydrolyze organic phosphorus) and organic acids (to solubilize fixed inorganic phosphorus), converting them into plant-available forms [82,83]. Pseudomonas and Rhizobium sp. DG1 (unpublished results from preliminary experiments) have been confirmed to possess phosphorus-solubilizing capabilities [83,84,85]. Additionally, Arthrobacter sp. GCG3 in the PC inoculant can secrete a large amount of extracellular polysaccharides [55]. Previous research has demonstrated that when EPS-producing salt-tolerant bacteria are inoculated into plants, they enhance the plants’ ability to absorb sodium, calcium, and potassium ions from the soil or increase the content of AP [86,87]. The decline in AK and AP under CK indicates that G. uralensis uptakes available nutrients at a rate exceeding the natural replenishment capacity of the soil. By contrast, the Pc treatment mitigates this nutrient deficit through microbial-mediated processes.

4.3. Microbial Community Restructuring by Pc Inoculation

α and β diversity revealed that Pc significantly restructured the rhizospheric microbial community, likely driven by the competitive dominance of target strains in the inoculant. This is further corroborated by the relative abundances of the top 10 dominant bacterial genera. Core components of Pc—Arthrobacter, Pseudomonas, and Rhizobium—exhibited remarkable rhizospheric colonization, with abundances increasing by 37.41%, 608.33%, and 263.63% compared to the control, respectively. As typical plant growth-promoting rhizobacteria (PGPR), the surge in Pseudomonas abundance may suppress pathogens by secreting siderophores and antibiotics, while establishing a favorable microenvironment for symbiotic microbes [88,89,90]. The enrichment of Rhizobium suggests that Pc enhances host nutrition via improved nitrogen fixation. Existing studies have demonstrated that Novosphingobium plays a pivotal role in environmental pollutant degradation, with strains capable of metabolizing diverse organic contaminants [91], whereas Sphingomonas contributes to plant growth promotion, stress resilience enhancement, soil structure amelioration, and biocontrol [92]. The increase in Novosphingobium and decrease in Sphingomonas reflect that Pc’s functional microbes may possess a competitive edge in nutrient utilization and spatial occupation, leading to reduced abundances of certain beneficial taxa, while others remain less affected or gradually dominate through competition, exhibiting increased relative abundances.

In terms of fungal community structure, Pc treatment significantly inhibited the abundance of pathogenic fungi (such as Fusarium) and markedly reduced the abundance of certain saprophytic fungi, such as Pseudombrophila (a 99.55% reduction). The mechanism may involve antimicrobial secretion and nutritional competition. Meanwhile, beneficial fungi (Mortierella, Cladosporium) also showed obvious proliferation, especially the relative abundance of Cladosporium increased by 33.39-fold, which is consistent with the documented role of Cladosporium in biological control of plant diseases [93]. This indicates that Pc enhances the host’s disease resistance by establishing a beneficial microbial barrier. The enrichment of Filobasidium and Vishniacozyma—also reported to suppress plant diseases [94,95,96]—further enhances rhizosphere stress resistance.

These results provide strong corroboration for the “niche theory” of microbiome regulation [97,98]. Through the introduction of functional strains, Pc systematically reorganizes rhizospheric resource allocation patterns, promoting microbial community succession in a direction conducive to host health. This targeted regulatory effect is not only manifested in the dominant colonization of inoculated target strains but also extends to the remodeling of the entire microbial interaction network via cascade effects—evident in the concurrent suppression of pathogenic fungi and proliferation of beneficial microbial taxa. These findings establish a theoretical foundation for the development of high-efficiency microbial inoculants. Future investigations should prioritize exploring the signal transduction mechanisms between Pc and host plants, as well as evaluating the influence of soil physicochemical properties on field application efficacy, to facilitate the practical translation of microbiome technologies in agricultural ecosystems.

4.4. Mechanistic Links Between Microbes, Nutrients, and Plant Traits

The structural alterations in rhizospheric microbial communities induced by Pc inoculant colonization exhibited a significant association with the functional remodeling of the plant–soil system [99,100]. Positive correlations between the abundance dynamics of plant growth-promoting bacteria (Arthrobacter, Pseudomonas, Rhizobium) and single-plant dry weight (SPDW)/bioactive compound accumulation (single-plant dry weight, SPTF; single-plant dry weight, SPLQT and single-plant glycyrrhizic acid SPGA) suggested their potential role in promoting host growth and bioactive component synthesis by regulating secondary metabolic pathways or activating soil AK/AP. The significant decrease in abundances of Sphingomonas and Pseudombrophila following Pc treatment, along with their negative correlations with bioactive compound content (eg. SPGA and SPLQT), suggests their potential role as competitive microbes in the rhizospheric ecosystem. Their negative correlations with AK and AP may reflect their inhibitory effect on soil nutrient activation, which was alleviated by Pc treatment—thus indirectly promoting plant nutrient uptake. While Pc treatment resulted in significantly increased relative abundances of Vishniacozyma and Chaetomium, which showed significant positive correlations with bioactive components (SPTF and SPGA), SPDW and available nutrients (AK/AP), these genera exhibited negative correlations with soil organic matter and total nitrogen. These fungi may release small carbon molecules (e.g., simple sugars, organic acids) when decomposing rhizosphere organic matter. Absorbed by plants, these participate in major metabolic pathways (TCA cycle, glycolysis) to form intermediate metabolites, which act as carbon skeletons for secondary metabolites [101,102]. For instance, acetyl-CoA, derived from organic carbon, is key for synthesizing terpenoids like glycyrrhizic acid [103,104]. Therefore, these fungi may be involved in promoting the increase in active components in G. uralensis in this study. This finding provides a novel perspective for deciphering the multi-kingdom regulatory mechanisms of microbial inoculants.

4.5. Limitations and Future Directions

While this study demonstrates short-term efficacy, long-term field trials are needed to assess Pc’s durability in dynamic soil environments. Mechanistic studies should explore the specific metabolites mediating microbe–plant interactions. Additionally, combining Pc with organic amendments could optimize nutrient cycling in severely salinized soils. Genomic analyses of key strains (e.g., Arthrobacter sp. GCG3) may reveal functional genes underlying stress tolerance and secondary metabolite promotion.

5. Conclusions

In this study, four multistrain microbial inoculant formulations were developed using rhizospheric microbial resources isolated from plants in saline–alkali soils of the Hexi Corridor, an arid region in northwestern China. The primary objective was to evaluate their growth-promoting effects on the indigenous economic plant G. uralensis and identify the most beneficial inoculant.

Systematic evaluations demonstrated that inoculant Pc—comprising Pseudomonas silesiensis, Arthrobacter sp. GCG3, and Rhizobium sp. DG1—significantly enhanced the yield and medicinal quality of G. uralensis in arid saline–alkali soils. It exerts its benefits through synergistic regulation of plant–microbe–soil interactions, offering a sustainable solution to the bottlenecks in cultivating medicinal plants in saline–alkali ecosystems. Specifically, Pc increased single-plant dry weight by 101.70% and markedly elevated the contents of key bioactive compounds: liquiritin (177.93%), glycyrrhizic acid (106.10%), and total flavonoids (107.64%) by reshaping the rhizospheric microenvironment.

Mechanistically, Pc promoted root utilization of soil organic matter and nitrate while enhancing the availability of soil potassium and phosphorus. Furthermore, colonization of Pc strains in the rhizosphere of G. uralensis restructured the rhizospheric microbial community, enriching beneficial fungi (e.g., Mortierella, Cladosporium) and suppressing pathogenic taxa (e.g., Fusarium), thereby establishing a microecological balance conducive to plant growth and secondary metabolism.

These findings not only validate the effectiveness of utilizing native salt-tolerant microorganisms to overcome growth limitations of medicinal plants in saline–alkali environments but also provide a mechanistic framework for understanding microbe-mediated nutrient activation and microbial community regulation. Practically, Pc exhibits dual ecological and economic value: it enhances the ecological restoration potential of G. uralensis in saline–alkali lands while improving the yield of high-quality medicinal materials, aligning with goals of sustainable agriculture and ecological remediation.

Looking forward, this research lays a foundation for developing region-specific microbial inoculants tailored to arid saline–alkali ecosystems. It highlights the promise of integrating indigenous microbial resources into agricultural practices to mitigate abiotic stresses, offering a scalable model for improving global medicinal plant productivity and restoring degraded lands.