Located in Xinjiang Uygur Autonomous Region in northwestern China, Ebinur Lake Wetland National Nature Reserve (EWNR) is the lowest depression and water-salt aggregation center of Ebinur Lake Wetland on the western edge of the Gurbantunggut Desert (44°30′-45°09′N, 82°36′-83°50′E), with a typical desert ecosystem. The annual maximum temperature is 44°C and the minimum temperature is -34°C. The average annual precipitation is less than 100 mm, whereas the annual evaporation is 16 times the average annual precipitation (Li et al., 2023). Ebinur Lake is the largest saltwater lake in Xinjiang. The soils in this basin are mostly grey desert soil, grey brown desert soil, and aeolian sandy soil with severe salinization (Chen et al., 2023). The dominant plant communities are the salt and drought stress tolerant plants such as P. euphratica, H. ammodendron, and Tamarix chinensis (Wang et al., 2022).

The soil was collected in an empty aluminum box that had been weighed beforehand, weighed in the field, brought back to the laboratory, dried. The dry soil was weighed to calculate the soil moisture content (SWC). The shaken soil was dried naturally and then sieved for analysis of soil physical and chemical properties, including soil organic carbon (SOC), electrical conductivity (EC), soil pH (pH), total phosphorus (TP), total nitrogen (TN), available phosphorus (AP), ammonium nitrogen (NH₄-N), nitrate nitrogen (NO₃-N), urease (UA), and alkaline phosphatase (ALP). The specific methods were as follows (Bao, 2000): organic carbon content was determined by high temperature exothermic potassium dichromate oxidizing capacity method; soil conductivity was measured by precision conductivity meter; pH value was measured by a pH meter in the solution with a water soil ratio of 2.5:1; soil total phosphorus content was determined by HClO₄-H₂SO₄ Mo-Sb colorimetric method; the content of soil total nitrogen was determined by H₂SO₄ mixed accelerant digestion; and available nitrogen content was determined by NaHCO₃ leaching Mo-Sb colorimetric method; KCl leaching-indophenol blue colorimetric method for the determination of soil ammonium nitrogen content; colorimetry method with phenol disulfonic acid for the determination of soil nitrate nitrogen content; phenol-sodium hypochlorite colorimetric method for the determination of soil urease content; and alkaline phosphatase content in soil was determined by phenol-sodium hypochlorite colorimetric method.

Sufficient fine roots were randomly selected from the root samples. The cells were stained using the Phillips and Hayman staining method (Phillips and Hayman, 1970). The structures of AMF and DSE were observed under a microscope (Nikon ECLIPSE Ti2, Japan), and photographs of the arbuscules, hyphae, vesicles of AMF and hyphae, and microsclerotia of DSE were obtained. The root sample observation method and mycorrhizal colonization grading criteria were used to calculate the root colonization rate (F%), colonization density (M%), and arbuscule abundance(A%) (Trouvelot et al., 1986). Microsclerotia colonization rate (MS%) was measured and calculated using the cross method under a 200x microscope (McGONIGLE et al., 1990). F% = number of infected root segments/total number of microscopic examination root segments* 100%. M% = (0.95 * N5 + 0.7 * N4 + 0.3 * N3 + 0.05 * N2 + 0.01 * N1)/total number of root segments examined by microscopy* 100%, where 0.95, 0.7… respectively represent the weight of each level. N5=sum of fifth-level root segments, N4, N3, N2, and N1 have the same meaning. A% = sum of the intersection points of arbuscule colonization/sum of cross colonization points* 100%. MS% = number of cross-colonization points/sum of cross-colonization points * 100%.

Fungal DNA extraction and detection were completed, followed by PCR amplification using ITS1F (CTTGGTCATTTAGAGGAAGTAA) and ITS2R (GCTGCGTTCTTCATCGATGC) primers (Gardes and Bruns, 1993; Lekberg et al., 2018; Wang et al., 2024), and the PCR products were analysed using QuantiFluor ^™ - ST blue fluorescence quantitative system (Promega Corporation), followed by the construction of an Illumina library and Illumina sequencing. DNA extraction and sequencing were performed by Shanghai Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China) to rapidly detect target strains. Then we used UPARSE platform version 7.1 to analyze the obtained gene sequences. To improve the accuracy and stability of the results, we clustered the data sequences and eliminated all chimeric sequences, ensuring that those with 97% similarity were grouped into the same operational taxonomic unit (OTU) to optimize the features of these datasets. An in-depth comparative analysis of the SILVA rRNA database (Release 138 https://www.arb-silva.de/) with the Unite eukaryotic ITS region database (Release 8.0 http://unite.ut.ee/index.php) was performed. The Majorbio cloud computing platform (http://cloud.Majorbio.com) was used to analyze the microbial communities in the soil. The initial data was stored in NCBI Sequence Read Archive database with the accession number PRJNA1046269. The data were initially organized and statistically analysed. By integrating the AMF and DSE taxonomic information and references, we created a species abundance table using the Majorbio platform to screen and download biodiversity information and facilitate subsequent data analysis.

SPSS 27 (IBM Corporation, Armonk, NY, USA) was used to perform normality and homogeneity of variance tests on the colonization rates of AMF and DSE, and the colonization rates under different water gradients were compared using multiple group comparisons (Kruskal-Wallis test). Bar charts were plotted using Excel to represent the status of the rhizosphere soil physical and chemical factors in P. euphratica and H. ammodendron. The ggcor package for R (version 4.3.1) was used to analyze the correlation between fungal colonization in rhizosphere and roots of P. euphratica and H. ammodendron and soil physical and chemical factors. Heat maps were drawn using the ggpubr, corrplot, and ggplot2 packages to visualize the correlation between the AMF and DSE colonization status. AMF and DSE with different gradients in rhizosphere and roots were classified and visualized by genus using circular and statnet packages for R. Redundancy analysis (RDA) between AMF and DSE colonization status and soil physical and chemical factors in the roots of the two plants was performed using CANOCO 5. Prior to RDA analysis, a species-sample analysis was done using CANOCO 5 to determine the choice of RDA analysis. The collinearity degree between soil physical and chemical factors was assessed by variance inflation factor (VIF) analysis using the R package (version 4.3.0) to ensure that the respective VIF values for all factors were less than 10.

It can be seen that the AMF colonization rate increased significantly from 13.89% to 71.83% in the roots of H. ammodendron from high to low water gradients ( Table 1 ). The arbuscule abundance of P. euphratica decreased with a decrease in soil moisture, whereas that of H. ammodendron increased with a decrease in soil moisture. From high to low water gradients, the DSE colonization rate increased significantly from 53.24% to 92.41% in the roots of P. euphratica, and decreased significantly from 67.10% to 45.06% in the roots of H. ammodendron ( Table 2 ). Similar to P. euphratica, the DSE colonization rate and density of H. ammodendron in the medium water gradient were the lowest among the three gradients, and the number of microsclerotia was the highest among the three gradients. In most cases, both AMF and DSE had higher colonization rates under low water gradient conditions, indicating that a certain degree of drought promotes the formation of mycorrhizal structures between AMF, DSE and plant roots.

Under the same water gradient, the colonization status of AMF and DSE in P. euphratica and H. ammodendron differed. Under a high water gradient, the AMF colonization rate of P. euphratica was significantly higher than that of H. ammodendron. AMF were not detected in the roots of P. euphratica under a medium water gradient, but were detected in all water gradients of H. ammodendron. The DSE colonization rate of P. euphratica in the medium water gradient was significantly higher than that of H. ammodendron. Under a low water gradient, the DSE colonization rate, colonization density, and number of microsclerotia in the roots of P. euphratica were significantly higher than in H. ammodendron.

Different interactions were observed between DSE and AMF in the roots of P. euphratica and H. ammodendron ( Figure 5 ). For example, in the roots of P. euphratica, AMF colonization (AM) was significantly positively correlated with the DSE colonization rate (DF) (R=0.74), and DF was significantly positively correlated with the colonization rate of AMF (AF) roots (R=0.72).

Among all the 27 samples from the rhizosphere soil and roots of P. euphratica, 18 samples were detected in seven genera of AMF, with samples belonging to high and low water gradients in the rhizosphere and roots ( Figure 6 ). More than one AMF genus coexisted in each of the 18 samples. At low water gradients, the dominant AMF genera from the rhizosphere to the root interior changed from Kamienskia to unclassified_f__glomerraceae. Under high water gradients, the dominant genus changed from unclassified_p__Glomeromycota in the rhizosphere to unclassified_f__Glomerraceae in the roots. Overall, Kamienskia had the highest abundance, followed by unclassified_f__Glomerraceae.

Among the 27 samples in the rhizosphere soil and within the roots of H. ammodendron, 18 samples were detected in three genera of AMF, belonging to high water gradients in the rhizosphere and high and low water gradients in the roots, all of which had AMF coexisting in one or more genera. Only two types of AMF were detected in the roots of H. ammodendron under a high water gradient, with unclassified_p__Glomeromycota being the dominant genus. Under a low water gradient, the dominant genus in the rhizosphere was Kamienskia, and the dominant genus in the root was unclassified_p__Glomeromycota.

The diversity of AMF species inside and outside the roots of P. euphratica was greater than that in H. ammodendron. Only three of the seven AMF genera were detected inside and outside the roots of H. ammodendron, while P. euphratica detected all seven genera including unclassified_o__Glomerales, Rhizophagus, Glomus, and Dominikia, with a total number greater than that of H. ammodendron. Among all AMF-detected samples, the dominant genus was unclassified_f__Glomerraceae. The AMF in the rhizosphere soil of P. euphratica was approximately six times the amount of H. ammodendron, whereas the amount of AMF in H. ammodendron is 2 times more than that of P. euphratica.

In total, 25 DSE genera were detected in the rhizosphere soil and roots of P. euphratica, and more than one DSE genus coexisted in all 27 samples ( Supplementary Table S2 ). The dominant genera were Cyphellophora, followed by Fusarium ( Figure 7 ). The composition of DSE genera of P. euphratica varied among the different water gradients. Cyphellophora and Fusarium have been widely detected in low water gradients. However, Cadophora, the dominant genus in the high-water gradients were rarely detected in the low water gradients.

In total, 25 DSE genera were detected in the rhizosphere soil and roots of H. ammodendron, and more than one DSE genus coexisted in all 27 samples. The three most abundant genera were Neocamrosporium, Cyphellophora, and Preussia. Neochamrosporium was widely distributed in high water gradients in roots, whereas Scytalidium was mainly detected in the low water gradient rhizosphere soil. Other genera of DSE were detected in samples from multiple water gradients. The abundance of DSE in the roots of each water gradient was lower than that in the rhizosphere. But the quantities of DSE in the rhizosphere were greater than that in the rhizosphere.

The TOP 10 DSE genera inside and outside the roots of the two plant species were also different: Scytalidium and Microascus were not dominant in the roots of P. euphratica, and Cadophora, Phialophora, and Embellisia were not dominant in the collected samples of H. ammodendron roots.

A redundancy analysis was conducted on the relationship between AMF and DSE colonization status of the roots of P. euphratica and H. ammodendron and soil physical and chemical factors at the plots. The results showed that the eigenvalues of the first and second sorting axes were 0.5842 and 0.1294, respectively. The two axes explained 71.36% of the changes in endophytic fungi, indicating that the first and second sorting axes could better reflect the changes between endophytic fungi and soil physical and chemical factors ( Figure 8 ). The first axis was positively correlated with soil physical and chemical factors TP, negatively correlated with other soil physical and chemical factors, such as pH, AP, NO₃-N, ALP, UA, SOC, SWC, NH₄-N, TN and EC. It was positively correlated with AA, AM, AF, DF, and DM and negatively correlated with MS. The second axis was negatively correlated with MS, DF, DM, and all environmental factors, and positively correlated with AM, AF, and AA.

The results indicated ( Supplementary Table S1 ) that SWC, AP, TP, and TN were significantly and positively correlated with AM (P<0.01). UA levels were significantly positively correlated with AM (P<0.01), and significantly negatively correlated with AF (P<0.05). ALP levels were significantly positively correlated with AM (P<0.01). There was a significant negative correlation between NH4-N and AF (P<0.05). There was a highly significant negative correlation between NO3-N and AF (P<0.01) and a significant positive correlation with DM (P<0.05). PH was significantly and positively correlated with DF (P<0.05) and DM (P<0.05).

Soil moisture can directly affect the growth and development of mycorrhizal fungi and the formation of mycorrhizal symbioses (Liu et al., 2018). Under low water gradient, for example, endophytic fungi can more effectively connect the roots of H. ammodendron to the surrounding soil microenvironment, thus exhibiting more significant functions in absorbing and transporting water and nutrients. In the present study, the DSE colonization rate of H. ammodendron decreased with decreasing water content. We speculate that this may be due to the drought stress exceeding a certain limit, which may have resulted in the production of certain substances by fungal cells (Pegler and Allen, 1993), leading to changes in the cell membrane permeability of H. ammodendron (Dick et al., 1996) and affecting the growth of fungi and plants. The colonization rate of AMF in the roots of P. euphratica and H. ammodendron was negatively correlated with soil water content, consistent with previous findings (Shen and Zhu, 2021). It has been speculated that appropriate level of humidity is beneficial for AMF colonization and spore germination (Frew, 2023). In this study, it was also found that under the same drought stress conditions, host plant species have an impact on the biomass and colonization rate of mycorrhizal fungi (Ilyas et al., 2024). For example, in contrast to the changes in the DSE colonization rate of the roots of H. ammodendron mentioned earlier, the DSE colonization rate of P. euphratica was negatively correlated with soil moisture content. This may be because the soil moisture content of the W1 plot in this study was more suitable for DSE colonization of the roots of P. euphratica. Under these conditions, DSE can have a relatively tight colonization of the roots of H. ammodendron (Ge et al., 2018). In our study, we also found that the total colonization rate of DSE under a low water gradient was higher than that of AMF, which not only indicates that DSE has a stronger colonization ability than AMF, but also indicates that DSE has greater colonization advantages under extreme environments (Alberton et al., 2010; Postma et al., 2007). The changes in the colonization structure that occurred in AMF and DSE with increasing drought severity may be an adaptation of endophytic fungi to environmental influences (Xie et al., 2017).

Zhao et al. (2021) found a positive correlation between the colonization rate of DSE microsclerotia and the content of AMF groups, which is consistent with the results of this study. In the present study, we also found a correlation between DSE colonization and AMF in the roots of the two plants. The colonization rate and density of AMF in P. euphratica were significantly positively correlated with the colonization rate of DSE, whereas the colonization rate and density of AMF were negatively correlated with the DSE colonization rate and density in H. ammodendron. This indicates that AMF and DSE can form different colonization relationships in the roots of different species, and that the interaction between the two mycorrhizal fungi is complex and greatly influenced by the host plant and environmental conditions. This is consistent with previous research results. The study of Shang et al. (2021) study showed that the colonization of AMF and DSE in the roots of cultivated Vaccinium uliginosum in Guizhou is highly significantly negatively correlated. Seerangan and Thangavelu, (2014) found that there was no clear relationship between AMF colonization and DSE colonization when studying plants growing in aquatic and wetland habitats, and the co-occurrence of AMF and DSE is more common in terrestrial habitats.

In this study, we found that the majority of AMF and DSE were not unique to P. euphratica or H. ammodendron but were shared by both, indicating that AMF and DSE strains do not exhibit strong host specificity in arid areas (Liu et al., 2017; Bi and Xie, 2021). For example, unclassified_f:Glomeraceae, an unclassified member of the Glomeraceae family, was found in the roots of both H. ammodendron and P. euphratica. Glomeraceae is a widely distributed community in the AMF population currently discovered, and was found in both soil and plant roots (Stürmer et al., 2018). Unlike AMF, DSE exhibits host preference (Stroheker et al., 2021). For example, Cyphellophora, the dominant genus of DSE, had a significantly greater amount of colonization in the rhizosphere soil and roots of P. euphratica than H. ammodendron, whereas another genus, Neocamrosporium, had a higher number in the rhizosphere soil and roots of H. ammodendron, which also reflects the varying degrees of host preference for different mycorrhizal fungi (Joshee et al., 2009). Under the same drought conditions, there was a significant difference in the colonization status of AMF between the roots and rhizospheres of P. euphratica. The diversity and quantity of AMF in the rhizosphere soil were higher than in the roots, which is consistent with the results of Tian et al (2018). on AMF in the rhizosphere of potato roots. Differences in AMF communities between the soil and roots may be determined by the heterogeneity of the rhizosphere soil (Liu et al., 2009). Shi et al. (2022) and Wang et al. (2010) conducted a diversity survey of AMF in the Xinjiang region and found that Glomus was the dominant genus. In this study, Rhizophagus and Kamienskia were more dominant than Glomus, which may be due to differences in hosts.

Many of the AMF and DSE genera detected in this study are highly beneficial to the host. For example, Glomus genera are used for Se biofortification (Ye et al., 2020). The Phialocephala genus can promote the biological adaptability of Saussurea involucrata and Rhodiola rosea to extremely cold and strong radiation environments (Chen et al., 2018), and improve the gold tolerance properties of host plants. Thielavia in the soil provides plants with highly active endoglucanases, increases the accumulation of plant glucose, and promotes the absorption of soil nutrients by plants, thereby increasing the yields of Medicago sativa (Medicago sativa Linn) and Zea mays (Badhan et al., 2014; Shi et al., 2019). These findings demonstrated that endophytic fungi play an important role in improving plant adaptability to the environment.

Studies have shown that the distribution of endophytic fungi that colonize plant roots is related to soil physical and chemical factors (Wang et al., 2015). Both redundancy and correlation analyses in this study indicated that soil physical and chemical factors have different effects on fungal colonization within roots. Nitrogen deposition is one of the most important global change factors (Tang et al., 2023), and nitrogen content plays a role in AMF colonization. When the distribution ratio of nitrogen, phosphorus, and potassium in the soil is unbalanced, the infectivity of AMF is improved, helping plant roots absorb more nutrients from the soil to cope with adversity (Zhou et al., 2021). Increasing nitrogen fertilizer application can promote the proliferation of bacteria and actinomycetes in the rhizosphere soil of Oryza sativa, and the nitrogen application level is positively correlated with the number of bacteria and actinomycetes. TP and ALP were highly significantly positively correlated with AMF colonization density, which is consistent with the results of Li et al. (2021). Soil phosphatase can promote the hydrolysis of organic phosphorus compounds into inorganic phosphorus, thereby increasing the available phosphorus content in the soil (Luo et al., 2022). However, under high phosphorus levels, the colonization rate of AMF significantly decreases (Salmeron-Santiago et al., 2023), which is not conducive to plant growth. Therefore, controlling and managing soil phosphorus levels are beneficial for improving the richness and diversity of soil AMF, thereby facilitating the establishment of sustainable ecosystems (Qin et al., 2020). It is obvious that the soil moisture content was not the reason for the lack of AMF colonization in the roots of P. euphratica under a moderate water gradient, and the soil moisture content did not cause AMF colonization in the other two gradients. We speculate that this may be due to the aforementioned soil physical and chemical factors (Yan et al., 2023). Under nitrate-nitrogen conditions, endophytic fungal inoculation can promote seedling growth (Sun et al., 2022). Nitrate-nitrogen can reflect soil nitrogen levels, and nitrogen can accelerate the growth metabolism of soil microorganisms (Qu et al., 2022), strongly supporting the results of our study. In this study, there was a significant positive correlation between pH and the colonization status of the DSE communities, which is consistent with previous studies. Hou (2020) found that the colonization rate of DSE was positively correlated with pH in his research on various desert plants. This may be because the pH affects the soluble salt content in the soil and the material exchange capacity of DSE, thus affecting its growth status of DSE (Zhang et al., 2015).