Effects of drought stress and Morchella inoculation on the physicochemical properties, enzymatic activities, and bacterial community of Poa pratensis L. rhizosphere soil

Experimental design

The soil for this experiment was provided by the Academy of Agriculture and Forestry Sciences experimental base at Qinghai University, which is located at N36°43.2034′ and E101°45.1752′ in the North District of Xining, Qinghai Province, China. Specifically, the soil was collected from the top 0–20 cm in a greenhouse used previously for Brassica chinensis L. cultivation. The soil had a calcareous chestnut texture, with a pH of 8.2, and contained 1.98 g/kg total nitrogen, 1.79 g/kg total phosphorus, and 17.99 g/kg total potassium. The indoor pot experiments employed three replications and a two-factor randomized block design. The experiment consisted of two Morchella inoculation treatments (inoculated (X) and uninoculated (K)) and three drought stress treatments (30%, 50%, and 70% soil moisture content).

Pot experiment

Each 12 × 15 cm diameter plastic pot was filled with 300 g of pre-prepared soil. Morchella-inoculated pots were established by adding 10 g of Morchella inoculant to each pot and using sterile wooden sticks to mix the soil. The Morchella sextelata variety QJ-2 was utilized for inoculation purposes. To prepare the inoculum, the stock culture maintained on potato dextrose agar (PDA) medium was transferred to sterile substrate for mycelium propagation. The substrate was composed of the following ingredients: 72.80% wood chips, 19.40% bran, 4.80% humus soil, 1% sucrose, 1% calcium superphosphate, 1% gypsum, and a water content ranging from 60% to 65%. The inoculated substrate was maintained at 22 °C for 30 days until the mycelium fully filled the cultivation bag. Uninoculated pots were established by adding 10 g of sterile substrate to each pot and using sterile wooden sticks to mix the soil. Thirty surface-sterilized P. pratensis seeds were distributed evenly across the soil surface in each pot. The pots were transferred to a light incubator maintained at 23 °C, 80% humidity, 3,000 Lx light intensity, and 16 h of light per day (Su, Qi & Yin, 2024). The humidity was adjusted to 60%, and the temperature was kept at 18 °C at night. Pots were watered every 2 days with distilled water using the constant weight method according to their assigned drought stress treatment.

Determination of plant growth performance

Following emergence, the seedlings in the pots were randomly thinned, with only 10 plants ultimately being retained. To evaluate the growth performance of P. pratensis, the plant height and leaf number per plant were measured 100 days after sowing.

Determination of soil chemical properties

Rhizosphere samples were collected 100 days after sowing. To ensure that each sample contained at least 30 g of soil, all rhizosphere soil from each of the three pots used for each treatment was removed. Soil samples were frozen in liquid nitrogen for cryopreservation. In all, three lyophilization tubes were collected for each of the six treatments, for a total of eighteen sample tubes. Soil alkaline nitrogen (AN) content was determined according to the method of Prasad (1965). The determination was carried out using a Kjeldahl nitrogen analyzer, following a brief procedure as follows: 10 mL of 4 M NaOH and 0.1 g of FeSO₄ were added to 1 g of air-dried soil. After distillation, the ammonium is absorbed by boric acid solution, and the AN content was calculated after titrating it with 0.005 M sulfuric acid. Soil available phosphorus (AP) content was determined according to the method of Olsen et al. (1954). A total of 50 mL of 0.5 M NaHCO₃ (pH = 8.5) was added to 2.5 g of air-dried soil, and the suspension was shaken for 30 min in an oscillator for extraction. 10 mL of filtrate was mixed with 5 mL of molybdenum-antimony solution and incubated at room temperature for 30 min. The AP content was quantified with a spectrophotometer. Soil available potassium (AK) content was determined according to the method of Leaf (1958). 50 mL of 1 M NH₄OAc (pH = 7) was added to 5 g of air-dried soil. After shaking for 30 min, the filtrate was collected and subjected to measurement of excitation intensity using a flame photometer. The AK content was then calculated.

Determination of soil enzymatic activity

Soil catalase (CAT) activity was determined according to the method of Minczewski & Marczenko (1973). Briefly, 40 mL of distilled water and 5 mL of 3% hydrogen peroxide solution were added to 2 g of fresh soil. The mixture was shaken for 30 min and subsequently filtered. A volume of 25 mL of the filtrate was titrated with 0.1 M potassium permanganate until the solution turned pink. Soil alkaline phosphatase (APA) activity was determined according to the method of Tabatabai & Bremner (1969). Briefly, 2 g fresh soil was gently mixed with 5 mL CaCl₂ and 1 mL p-nitrophenyl phosphate (pNPP). After incubation at 37 °C for 1 h, the reaction was terminated with 0.5 M NaOH. The mixture was then filtered, and pNP content was determined with a spectrophotometer. Soil protease (PA) activity was determined according to the method of Speir & Ross (1975). A total of 5 g of fresh soil was thoroughly mixed with 10 mL of 2% casein. After incubation at 50 °C for 2 h, the reaction was terminated by adding 17.5% trichloroacetic acid. The protein content in the filtrate was determined spectrophotometrically, and the PA activity was subsequently calculated.

Soil DNA extraction and PCR amplification

Total microbial DNA was recovered from soil samples using the Soil Microbial DNA Quick Extraction Kit (Mo Bio Laboratories Inc., Carlsbad, CA, USA). Following PCR amplification of the complete soil microbial DNA sample, a sequencing library was created. The V3-V4 variable region, specific to the bacterial 16S rRNA, was amplified using the upstream primer 341F 5-CCTACGGGNGGCWGCAG-3 and the downstream primer 806R 5-GGACTACHVGGGTATCTAAT-3 (Klindworth et al., 2013). The following thermal conditions were applied for amplification: 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 45 s for a total of 35 cycles; and a final extension step at 72 °C for 10 min. After passing quality control, the PCR amplification products were cut, recovered, and quantified using a Quanti Fluor fluorometer. The sequencing junctions were ligated and the sequencing libraries were created. Finally, an Illumina sequencer PE250 was used to sequence a mixture of equal amounts of the purified amplification products.

Bioinformatics analysis

Sequencing was followed by data filtering, quality control, and single-base precision amplicon sequence variant (ASV) clustering. ASV clustering is equivalent to clustering operational taxonomic units (OTUs) with 100% similarity (Bokulich et al., 2013; Callahan et al., 2016). Chimera filtering was used to splice the bipartite data using overlap and quality control to obtain high-quality, clean data. Following the acquisition of the ASVs/OTUs (DeSantis et al., 2006; Pruesse et al., 2007; Wang et al., 2007), α-diversity analysis (Caporaso et al., 2010), β-diversity analysis, and community function prediction were carried out. The species annotation method was applied sequentially following the analytical process (Asshauer et al., 2015; Louca, Parfrey & Doebeli, 2016).

Statistical analysis

Microsoft Excel was used to calculate percentages and SPSS 23.0 was used for statistical analysis. Statistically significant differences among groups were determined using one-way ANOVA. Histograms were created with Origin 2021. The R packages psych, igraph, labdsv, Vegan, ggplot2, and pheatmap were used to calculate inter-species correlation coefficients, create inter-species correlation network and ternary diagrams, perform redundancy and correlation analyses, and graph the outputs.

Plant growth performance

As illustrated in Fig. 1, the X30 treatment group exhibited significantly greater plant height and leaf number per plant compared to the K30 treatment group, showing enhanced growth performance of Morchella-inoculated P. pratensis plants under severe drought conditions. However, the growth-promoting effect was not readily apparent under 50% and 70% moisture conditions. These findings indicate that the growth promotion resulting from Morchella inoculation is contingent upon the moisture content of the soil.

Rhizosphere soil physicochemical properties and enzymatic activities

Morchella inoculation considerably decreased the AK content, increased the AN and AP content, enhanced the PA and APA activity, and diminished the CAT activity, in Poa pratensis rhizosphere soil (Figs. 2, 3). In Morchella-inoculated samples, soil AN increased with soil water content, and was significantly higher in the 70% treatment than in the 30% treatment (Fig. 2A). In uninoculated samples, soil AP increased with soil water content, and was significantly higher in the 50% and 70% treatments than in the 30% treatment (Fig. 2B). Furthermore, soil APA activity was notably higher in the 50% treatment as compared to the 30% and 70% treatments, whereas the PA activity was considerably lower in the 70% treatment than in both the 30% and 50% treatments (Figs. 3A, 3C).

High-throughput sequencing results

After sequence quality control, denoising, splicing, and chimera removal, a total of 944,833 reads were recovered from Morchella-inoculated samples, with an average of 104,981 reads per sample. In addition, a total of 44,533 ASVs, with an average of 4,948 ASVs per sample, were identified in Morchella-inoculated samples. A total of 914,435 reads were recovered from uninoculated samples, with an average of 101,603 reads per sample. A total of 43,018 ASVs, with an average of 4,779 ASVs per sample, were identified in uninoculated samples (Table S1).

Microbiological community analysis

After excluding the “unclassified” and “other genera” categories, the top-ten bacterial genera across all samples were Sphingomonas, RB41, AKYG587, Bacillus, Flavobacterium, Pseudomonas, Pir4_lineage, Pirellula, JGI_0001001-H03, and Stenotrophobacter (Fig. 4). In comparison to the uninoculated treatments (K30, K50, and K70), the abundance of Bacillus decreased by 0.99%, 1.68%, and 1.38% in X30, X50, and X70, respectively. In contrast, the abundance of Pseudomonas exhibited an increase of 1.46%, 2.53%, and 2.18% in X30, X50, and X70, respectively. For JGI_0001001-H03, its abundance increased by 1.28%, 1.49%, and 0.76% in X30, X50, and X70, respectively. Specifically concerning RB41 and AKYG587, their respective abundance levels decreased by 1.64% and 0.85% in X30. Sphingomonas abundance decreased with increasing water stress, falling by 0.73% in X30 and 1.01% in K30 compared to X70 and K70, respectively. Overall, Morchella inoculation appears to significantly alter the bacterial community structure of P. pratensis rhizosphere soil under various levels of drought stress.

We observed substantial variation in the abundances of 19 different bacterial genera between X30 and K30 (Fig. 5), with the abundances of 15 bacterial genera being higher in K30 than in X30. In contrast, the abundances of Pseudomonas, Luteolibacter, Pseudarthrobacter, and Arenimonas were significantly higher in X30 compared to K30. These results suggest that Morchella inoculation under severe drought stress may lead to decreased rhizosphere community diversity and abundance.

There were significant differences (P < 0.05) in the abundances of 14 different bacterial genera between X50 and K50 (Fig. 6), including JGI 0001001-H03, Stenotrophomonas, Pseudarthrobacter, and Phyllobacterium, among others. Among these, the abundances of Pir4 lineage, PAUC26f, and Planctopirus were significantly lower in X50 than in K50, while the remaining 11 genera were significantly higher in X50. These results suggest that Morchella inoculation under moderate drought stress may lead to increased rhizosphere community diversity and abundance.

There were significant differences (P < 0.05) in the abundances of 16 different bacterial genera between X70 and K70 (Fig. 7). The abundances of 10 genera, including JGI_0001001-H03, Ohtaekwangia, Gemmata, and Pseudarthrobacter, were significantly higher in X70 than in K70. The abundances of Aquicella, PAUC26f, Nitrosospira, Planctopirus, Verrucomicrobia_bacterium_LX191, and Urania-1B-19_marine_sediment_group were significantly lower in X70 than in K70. The findings suggest that under adequately watered conditions, the introduction of Morchella results in an increase in the diversity of the rhizosphere bacterial community.

There were notable differences among bacterial genera in uninoculated P. pratensis rhizosphere soil under different water conditions (Fig. 8), including for RB41, Nitrospira, and Pseudomonas. The abundance of RB41 was significantly higher in K30 than in K70, and the abundance of Nitrospira was significantly higher in K30 than in K50 and K70. The abundance of Pseudomonas was significantly higher in K70 than in K30 and K50. Fewer differences were observed in Morchella-inoculated P. pratensis rhizosphere soil under different water conditions (Fig. 9). Specifically, the only notable difference was that of AKYG587 in X30, the abundance of which was significantly lower in X30 than in X50 and X70. Taken together, these results suggest that Morchella inoculation may help to stabilize the rhizosphere bacterial community under various levels of soil moisture.

Alpha diversity

Alpha diversity is primarily utilized to evaluate bacterial diversity among different ecological zones. Considerable differences in species richness (Chao1 index) were observed between moisture conditions, with 70% > 50% > 30% across both Morchella-inoculated and uninoculated soils (Fig. 10). In addition, both the Shannon and Pielou indices were significantly higher under 70% moisture than under 30% moisture (Figs. 11 and 12). In this experiment, we observed that soil moisture had a greater impact on alpha diversity than did inoculation.

Beta diversity

The genus-level principal co-ordinates analysis (PCoA) analysis based on Bray distances showed considerable separation between certain groups (Fig. 13A), particularly the 30% soil moisture treatments. Differences in bacterial community structure were explained by PCo1 in 36.57% of cases and PCo2 in 17.31%. In general, uninoculated soil samples were concentrated between −0.1–0.2 along the PCo1 axis, but more scattered across the PCo2 axis. The Morchella-inoculated soil samples tended to be distributed more randomly along both the PCo1 and PCo2 axes. The PCo2 axis was found to be extended by the overall tendency of progressive decline with decreasing water content.

PLS-DA analysis revealed that t1 explained 33.4% of the variation in bacterial community structure between treatments, while t2 explained 14.5% of the variation (Fig. 13B). The Morchella-inoculated soil samples were primarily found near the 0 attachment and the negative end of the t1 axis, while the absolute value of the t2 axis was within 2.5. The uninoculated samples were more dispersed along the t2 axis, primarily between the t1 axis 0 and 2.5. The t2 axis was found to be extended by the propensity to progressively rise with diminishing water.

Correlation between soil factors and soil community structure

To evaluate any correlations between soil physicochemical properties and enzymatic activities and rhizosphere bacterial community structure, we performed a redundancy analysis (RDA) (Fig. 14A). Overall, RDA1 and RDA2 contributed 60.78% and 22.61% of the explanatory factors, respectively, while the two axes combined contributed 83.39%. The redundancy analysis further indicated that Morchella-inoculated and uninoculated P. pratensis rhizosphere soil had different bacterial community structures. Four soil variables (AN, AP, APA, and PA) were strongly associated with the Morchella inoculation treatment group. Conversely, AK and CAT were more strongly associated with the uninoculated treatment group.

Bacterial community structure was significantly influenced by soil factors (Fig. 14B). The relative abundance of Bacillus showed a significant positive correlation with both AK content and CAT activity, whereas it exhibited a significant negative correlation with the levels of AN, APA, and PA activity. The relationship between the relative abundance of Pseudomonas and soil factors was exactly the opposite of that of Bacillus. Specifically, AN and AP content, as well as APA and PA activity, were positively correlated with the relative abundance of Pseudomonas; whereas AK content and CAT activity were negatively correlated. The correlation between the relative abundance of JGI_0001001-H03 and soil factors displayed a similar pattern to that observed in Pseudomonas as a whole, despite differences in correlation and significance. The relative abundance of Adhaeribacter also demonstrated significant correlations with multiple soil factors. Notably, a negative correlation was observed with the content of AN, AP, APA, and PA activity, whereas a positive correlation was found with CAT activity.

The distribution of soil bacteria was most significantly influenced by PA activity (Fig. 14C), which contributed 14.45%, followed by AN content. The distribution of soil bacteria was least significantly influenced by AP content, which contributed 4.9%.