The materials for this study were collected from Zhaodong city (46°2’5’’-46°2’51’’N, 125°54’6’’-125°42’31’’E) in Heilongjiang Province, China, which is located in the middle of the Songnen Plain and has a temperate continental monsoon climate. The landscape is flat, rainfall is mainly distributed in summer, and winter is dry and cold. The soil can be roughly sorted into saline soil, alkaline soil and meadow soil.

Five I. lactea samples were selected from each site according to the “multipoint parallel sampling method” and the “five sampling method”, and the roots and rhizosphere soils of samples from 10 sites were collected on July 6, 2016. The rhizosphere soil samples (approximately 1.5 Kg each) were collected from a soil depth of 0 cm–30 cm after the superficial rubbish was removed, and then the soil was packed in sealed bags marked with sample numbers, latitude and longitude. The samples were placed in a ventilated and dark place and air-dried until the roots could be separated from the soil completely. The roots were cut into 0.5-1-cm-long segments and soaked in FAA fixative solution (formaldehyde-glacial acetic acid) after they were washed with distilled water several times, and the water on their surface was sucked dry using filter paper. The air-dried soils were sieved through a 20-mesh sieve, numbered and then stored: the soils and root segments were stored at 4°C.

Fifty grams of soil was weighed from each single sample and mixed into suspension liquid (water:soil = 2:1). After the solution precipitated for 0.5 h, its pH value was measured using a pH meter (METTLER TOLEDO FE20). The average of three measurements, determined after the pH meter reading stabilized, was taken as the value of each sample.

The root segments were rinsed with clean water after they were removed from the FAA fixed solution and then stained with Alkaline Trypan Blue [17]. The cleaned root segments were placed in 10% KOH solution and bathed for 1 h to 1.5 h until they became soft and transparent. In addition, the root segments were dipped into alkaline hydrogen peroxide to soften, after which the roots were neutralized with 2% hydrochloric acid for 5 min and then washed before they were stained using a 0.05% Trypan Blue reagent and incubated for 0.5 h. The structural characteristics of the roots, including the characteristics of vesicles, hyphae, and arbuscules, were observed under a microscope (OLYMPUS-DSX500). The number of each structure was counted and recorded according to the method of Trouvelot et al. [18]. The AM fungal colonization rate in I. lactea roots and other indexes were analyzed using MYCOCALC, and the measurement results of the colonization intensities of the root segments were divided into five levels: (1) 0%–1% as Level I, (2) 1%–10% as Level II, (3) 11%–50% as Level III, (4) 51%–90% as Level IV, and (5) 91%–100% as Level V. The colonization assessment of each site was described using the colonization intensity, which was calculated based on the mycorrhizal colonization grade. Similarly, the arbuscular abundance was calculated based on the level of vesicular abundance: A₁ indicated few arbuscles, at <5%, A₂ indicated frequent arbuscles, at 5%–50%, and A₃ indicated abundant arbuscles, at >50%).

The colonization rate (F%) was calculated by dividing the number of colonized segments by the total number of root segments and multiplying the result by 100%.

The colonization intensity (M%) was calculated as follows: M% = (95 × N_V + 70 × N_IV + 30 × N_III + 5 × N_II + N_I)/(total number of root segments × 100) × 100%, where N is the number of roots at the same colonization level.

The arbuscular abundance (A%) was calculated as follows: A% = (100 × m × N_A3 + 50 × m× N_A2 + 10 × m × N_A1)/100, where m% = M × total number of root segments/amounts of colonized root segments and N is the number of roots at the same grade of vesicular abundance.

The quantity of spores per sampling site was counted according to the amount of spores in 50 g soil. In this experiment, we separated AM fungal spores in the rhizosphere soil of I. lactea using wet screening and sucrose density-gradient centrifugation. The isolated AM fungal spores were placed in a petri dish before the spores were enumerated under a double-tube solid-state dissecting microscope. Single spores were picked out with sterile tips; subsequently, the diameter, color, superficial decoration and thickness of each spore were observed under a microscope after a PVLG (a mixture of polyvinyl alcohol, lactic acid and glycerin) floating carrier had been added to the glass slide. The spores were identified and classified on the basis of the specific response of the spores to Melzer’s reagent. The morphological descriptions, pictures and newly published literature that were used to identify spore species in this experiment were provided by the INVAM International website (http://invam.wvu.edu/the-fungi/species-descriptions).

Diversity indexes were calculated according to the method of Yang et al. [19], including indexes of dominance grade classification, spore density, relative abundance, importance value and separation frequency. The equations used to calculate these indexes are as follows:

Spore Density (SD) = Spore number of all AM fungal species/soil sample number

Separation Frequency (F) = Occurrence frequency of certain pecies/total sample number × 100%

Relative Abundance (RA) = Spore number of certain species/total quantity of AM fungal spores × 100%

Importance Value (IV) = (F + RA)/2 × 100%

All symbiotic structures observed in the roots of I. lactea in this study were Arum-type (Fig. 1). Mycelia were abundant in roots, including septa hyphae (Figs. 1b and 1g) and nonsepta hyphae (Fig. 1a). The hyphae colonized the roots from outside (Fig. 1c) and formed large numbers of intercellular hyphae in the cortical cells. The lateral dichotomous branches directly penetrated the cell walls of the cortexes into the cells (Fig. 1d) and formed dendritic arbuscular structures (Fig. 1e); hyphae circles could be observed in the cells (Fig. 1f). The intercellular hyphae expanded and shaped vesicles that mostly presented as circles (Figs. 1e and 1i), ovals (Figs. 1g, 1h and 1l) or other irregular shapes (Fig. 1i). colonial structures formed by large numbers of vesicles bound together tightly were also seen (Fig. 1j), and internal spores were visible occasionally (Fig. 1k).

The Arum-type mycorrhizal structures formed after I. lactea was colonized by AM fungus could be seen through observations of mycorrhizae from 10 different sampling sites. As shown in Tab. 1, both the colonization rate and the colonization intensity reached maximum values at pH 8.2, and these maximum values were 73.33% and 20.8%, respectively. The maximal vesicular abundance (8.79%) and arbuscular abundance (4.07%) were observed when the soil pH reached 8.20 and 8.17, respectively. Overall, the colonization rate showed a fluctuating trend with increasing pH values, the colonization intensities were of level II or III, and the vesicle abundances were of grade A₂ or A₃ at the ten sampling sites. The colonization characteristics were dissimilar when the symbionts were under different pH conditions. Spearman analysis results showed that there was a significant and strong correlation between arbuscular abundance and spore density (P = 0.013, R = 0.748) as well as between vesicular abundance and colonization intensity (P = 0.008, R = 0.782). The spore density of AM fungi in the rhizosphere was 13–19/g at lower pH values, while it was 22/g at a pH value of 8.92 and up to 26/g at a pH value of 9.1. A significant strong correlation existed between pH and spore density (P = 0.010, R = 0.767) according to Pearson correlation analysis. However, there was no significant correlation observed between spore density and mycorrhizal colonization rate in this experiment, although both may be affected by pH.

AM fungi of 39 species in 15 genera were isolated from the samples taken from the Songnen saline-alkaline grassland (Fig. 2), among which 2 species were not identified by morphological identification. The 37 species of AM fungi that were identified included 10 species of Glomus, 5 species of Acaulospora, 4 species of Funneliformis, 3 species of Rhizophagus, 2 species each of Sclerocystis, Pacispora, Gigaspora and Claroideoglomus, and 1 species of Entropnospora, Scutellospora, Ambispora, Diversispora, Dominikia, Septoglomus and Dentiscutata.

Thirty-seven species of 15 AM fungal genera were isolated and identified from the soil samples (Tab. 2). Among these species, the dominant genera were Funneliformis and Glomus, which accounted for 21.97% and 20.58% of the total number of identified genera, respectively. Claroideoglomus and Rhizophagus were isolated from the rhizosphere with proportions of 13.56% and 13.08%, respectively, ranking second in dominance. Funneliformis mosseae and Rhizophagus intraradices were observed in all soil samples with importance values of 58.62 and 51.19, respectively, and these were the dominant species observed in the samples. Gigaspora etunicatum, Entrophospora infrequens, Glomus convolutum, Claroideoglomus lamellosum, and Sclerocystis sinuosa were common species, with spore proportions of 8.98%, 4.85%, 6.55%, 4.13% and 7.52%, respectively. Pacispora scintillans, Gigaspora gigantean, Sclerocystis rubiformis, and Rhizophagus diaphanus were occasional AM fungal species detected in the rhizospheres of I. lactea sampled in the Songnen saline-alkaline grassland.

AM fungi colonize host plants by recognizing signal substances from roots [20]. According to the colonization rate statistics and the identification of AM fungal species observed in this study, it was found that the species showed complex variation with changing soil pH, spore density was significantly and positively correlated with pH, and the colonization rate had no correlation with pH value. The colonization rate reached as low as 60% at the lowest soil pH (7.5) because the growth of I. lactea might not be inhibited and the secretion of signal substances would decrease under low pH circumstances, which would weaken the colonization of AM fungi and thus reduced the species number [21]. The experimental results showed that there was no significant correlation between spore density and mycorrhizal colonization rate, which was similar to the research results of Yang et al. [22]. However, many scholars have pointed out that there is a significant positive correlation between these two factors [23]. This inconsistency may be because the irregular spatial distributions of pores and the complex structures of roots are the most important factors affecting spore density [24], and these factors also lead to variations in the colonization rate among different conditions [25]. Inconsistent research conclusions appeared when symbionts were studied under different experimental conditions. In addition, it might be possible that there is no obvious correlation between the spore-producing ability of AM fungi and their colonization ability in host plants, and a strong spore- producing ability does not correlate to a strong colonization ability [22]. Colonization and spore production are two stages of the life cycle of AM fungi, the conditions required for these two stages are different, and there is no inevitable connection between them. Finally, the tolerance of different AM fungal spores to the same pH value varies, the AM fungal effectiveness could be affected by pH [26], and the mycelial growth and colonization capacity of AM fungi may be influenced under high pH conditions [27]. All these explanations are possible reasons for the inconsistencies between our results and the results of previous studies.

According to the statistics, the dominant observed species were F. mosseae and R. intraradices, and these species were isolated in all sampling sites, while the distributions of other AM fungi and spore densities showed some variance among different sampling sites, these variations not only reflected the diversity of AM fungi in Songnen saline-alkaline grassland but also showed the uneven distributions and varying colonization intensities of AM fungi. Some selectivity and adaptability were shown in the process of AM fungal recognition and colonization, and the mycorrhizal characteristics were determined jointly by the present AM fungi, host plants and environmental conditions [28]. The main ecological problem faced by Songnen saline-alkaline grassland is salinization, which directly manifests as increased pH values, and the total salinity of soil fluctuates with seasonal changes [29]. Subsequently, the content and availability of inorganic ions and heavy metal ions in soil are affected. The N application experiment showed that high N levels can increase the value of the N:P ratio and decrease the value of the C:N ratio, subsequently reducing the relative abundance of mycorrhizae. In addition, the colonization rate is enhanced at lower P levels [30]. The colonization indexes of AM fungi showed a trend of increased-decreased with the worsened heavy metal stress, such as stress due to the presence of cadmium or zinc [31]. These secondary effects of salinization may be factors affecting AM fungal diversity and spore density among different sites. However, different mycorrhizal characteristics have direct impacts on the physiological and ecological effects of AM fungi-host plant symbionts [32]. Therefore, it is of great significance for the application of AM fungal symbionts to select efficient AM fungi suitable for host plants.

It has been confirmed that many AM fungal species, such as F. mosseae [33], S. constrictum [34] and G. etunicatum [35], have strong salt-alkali tolerance in previous studies. Furthermore, these species could significantly improve the vegetal salt-alkali tolerances by enhancing the antioxidant enzymatic activities, increasing the proline, soluble sugar, soluble protein, and chlorophyll contents, reducing the malondialdehyde (MDA) contents, and regulating the palladone, dionine, quercetin 3’ –methyl ether and apigenin 7-O-neohesperidin concentrations in leaves [36]. In addition, these fungal species can also influence the expression of proteins related to active oxygen metabolism, phenylpropane biosynthesis, carbohydrate and energy metabolism, translation, sulfur metabolism, photosynthesis, nitrogen metabolism, and amino acid metabolism in plants [37] to improve the salt-alkali tolerances of host plants. The abovementioned AM fungal species were isolated in this research, and F. mosseae was distributed in all sampling sites, which indicates that AM fungal colonization improves the saline-alkaline tolerance of I. lactea growing in the Songnen saline grassland to a certain extent. Further verification of the roles of the other AM fungal species identified in this research in ecological restoration is needed.