Arbuscular Mycorrhizal Fungus Improves Rhizobium –Glycyrrhiza Seedling Symbiosis under Drought Stress

: Rhizobia and arbuscular mycorrhizal (AM) fungi can potentially alleviate the abiotic stress on the legume Glycyrrhiza (licorice), while the potential beneﬁts these symbiotic microbes o ﬀ er to their host plant are strongly inﬂuenced by environmental factors. A greenhouse pot experiment was conducted to investigate the e ﬀ ects of single and combined inoculation with a rhizobium Mesorhizobium tianshanense Chen and an AM fungus Rhizophagus irregularis Walker & Schuessler on Glycyrrhiza uralensis Fisch. seedling performance under di ﬀ erent water regimes. Drought stress inhibited rhizobium nodulation but increased mycorrhizal colonization. Furthermore, co-inoculation of rhizobium and AM fungus favored nodulation under both well-watered and drought stress conditions. Glycyrrhiza seedling growth showed a high mycorrhizal dependency. The seedlings showed a negative growth dependency to rhizobium under well-watered conditions but showed a positive response under drought stress. R. irregularis -inoculated plants showed a much higher stress tolerance index (STI) value than M. tianshanense -inoculated plants. STI value was more pronounced when plants were co-inoculated with R. irregularis and M. tianshanense compared with single-inoculated plants. Plant nitrogen concentration and contents were signiﬁcantly inﬂuenced by inoculation treatments and water regimes. R. irregularis inoculation signiﬁcantly increased plant shoot and root phosphorus contents. AM fungus inoculation could improve Glycyrrhiza plant–rhizobium symbiosis under drought stress, thereby suggesting that tripartite symbiotic relationships were more e ﬀ ective for promoting plant growth and enhancing drought tolerance.


Introduction
The legume Glycyrrhiza (licorice) is widely used as a medicinal herb and as an industry material due to the large amount of glycyrrhizin, an important bioactive triterpenoid saponin derived from licorice roots and stolons [1,2]. Glycyrrhiza plants are also cultivated to restore degraded ecosystems, particularly in arid and semi-arid regions [3]. However, drought stress and nutrient deficiency are the two main factors limiting the growth and production of Glycyrrhiza seedling cultivation [4]. Nevertheless, plants have evolved a series of mechanisms to cope with these stresses. For example,

Soil Characteristics
The soil used for the experiment was a calcareous loamy soil collected from Duolun Restoration Ecology Research Station (116 • 17 29 E, 42 • 02 20 N). The soil is classified as Haplic Calcisol (FAO (Food and Agriculture Organization of the United Nations) classification) with a pH of 7.02 (1:2.5 soil to water ratio), organic matter content of 30.4 g/kg, total N of 2.4 mg/kg and extractable P (0.5 M NaHCO 3 , pH 8.5) content of 6.73 mg/kg. The soil was air-dried, ground to pass a 2-mm sieve, mixed with quartz sand (<1 mm, 1:1 soil to sand ratio (v/v)) as growth medium and sterilized by γ-irradiation (25 kGy, Institute of Atomic Energy, Chinese Academy of Agricultural Sciences).

Plant Materials
Seeds of G. uralensis Fisch. were provided by Chinese Materia Medica Resources Center, China Academy of Chinese Medical Sciences. The seeds were immersed in H 2 SO 4 (50%) for 30 min to break the thick seed coat, washed thoroughly (>3 times) with sterilized distilled water, and then surface sterilized with 10% H 2 O 2 for 10 min. The selected homogeneous seeds were germinated on filter paper soaked with distilled water in Petri dishes at 27 • C in the dark for 2 days in a growth chamber.

Microbial Inoculation
AM fungus Rhizophagus irregularis Walker & Schuessler (syn. Glomus intraradices Schenck & Smith BGC AH01) was provided by Beijing Academy of Agriculture and Forestry Sciences, China. Mycorrhizal inoculum consisting of spores (78 spores per gram), hyphae, and infected root fragments were used at a 1:10 (v/v) inoculum to growth medium ratio. The inoculum was added to the top 3 cm of the growth medium (i.e., the mycorrhizal inoculum layer) at sowing time just below the seeds for mycorrhizal treatments. Non-inoculated treatment received the same amount of sterilized inoculum together with a 10 mL aliquot of a filtrate (<20 µm) of the inoculum to provide a similar microbial population that was free from AM propagules.
A moderately growing rhizobium M. tianshanense Chen (strain CCBAU3306), originally isolated from the Glycyrrhiza plant's rhizosphere soil [7], was provided by the Culture Collection of Chinese Agricultural University, Beijing, China. The strain was grown at 28 • C in TY medium (5 g/L tryptone, 3 g/L yeast extract, 9 mmol/L CaCl 2 ) and added close to tap roots (10 mL containing 10 7 cells per pot) 3 weeks after sowing to establish the bacterial treatment. Non-inoculated treatment received 10 mL autoclaved (121 • C, 45 min) microbial suspension.

Experimental Design and Growth Conditions
Germinated seeds were sown in 400 mL of growth medium mixed with mycorrhizal inoculum in 0.5 L pots. Before water treatment, all the pots were maintained under well-watered conditions (well-watered, 12% relative water content equating to 55% field water capacity). At 14 weeks after sowing, the water regime was unified for two groups, namely, well-watered and drought stress (drought, 8% relative water content equating to 40% field water capacity). During the first 3 days of the drought period, the drought stress treatments received only 20% of the water consumption to avoid water heterogeneity in soil. The water loss, determined based on container and plant weight, was daily supplemented with deionized water to maintain the desired moisture content. The plants were harvested after 4 weeks of drought stress. All the plants were in vegetative growth during the whole experimental period, and no difference in phenological stages among the treatments was observed.
The experiment was set up in a three-factor randomized complete block design. Experimental treatments included the following: (1) non-AM fungus/rhizobium inoculation as control; (2) inoculated with AM fungus; (3) inoculated with rhizobium; and (4) co-inoculated with AM fungus and rhizobium. All the treatments were cultivated under well-watered and drought stress conditions. There were eight treatments (full combinations of inoculation status and water regimes) with four replicates, giving a total of 32 pots. The experiment was carried out under greenhouse conditions (15-25 • C, 16 h photoperiod with supplementary lighting of 700 µmol m −2 s −1 ).

Parameter Measurements
Plant roots and shoots were separately harvested. Roots were carefully washed with deionized water, and the nodules were separated. Fresh weight of shoots, roots and total nodules were recorded. The number of nodules were estimated by directly counting. Subsamples of fresh roots (0.5 g) were stored in a 4 • C refrigerator for measuring mycorrhizal colonization. The rest of the roots, shoots, and nodules were dried at 80 • C for 48 h for dry biomass and nutrient content analyses.
The percentage of mycorrhizal colonization was estimated by visual observation after clearing with 10% KOH and staining with 0.05% trypan blue in lactic acid (v/v) [28]. Microscopic quantification was performed using MYCOCALC program [29].
Dry samples of shoots and roots were milled to a fine powder, and N concentration was directly determined by using an element analyzer (Vario EL, Elementar, Hanau, Germany). To determine P concentrations in shoots and roots, plant subsamples (0.2 g) were digested by 10 mL HNO 3 using a microwave accelerated reduction system (Mars 5, CEM Co., Ltd., Matthews, NC, USA) and analyzed by Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES, Vista axial, Varian). Plant nutrient contents refer to total P or N contained within the shoot or the root.
The microbial dependency of inoculation treatment was calculated according to van der Heijden [30] as follows. If biomass of n 1 a n > bn, then microbial dependency = (1 − (bn/ n 1 a n ) × 100. If biomass of n 1 a n < bn, then microbial dependency = (−1 + ( n 1 a n / bn) × 100, where a is the plant dry weight of a treatment inoculated with microbe, n is the number of treatments where plants were inoculated with microbe and "b" is the plant dry mass of the non-inoculated treatments. These equations ensured that positive and negative values for microbial dependency were comparable and symmetrical.
Drought stress index as stress tolerance index (STI) was calculated using the following formula: STI=B c ×B s /M c 2 , where B c and B s were the plant biomass under control and stress conditions, respectively, and M c was the mean biomass over all plants under control condition [31].

Statistical Analysis
Data were shown as the mean ± standard error (SE) of independent replicates except for microbial dependency and STI, which were calculated from the mean value of the different treatments. The data were checked using the Shapiro-Wilk test for normality and homogeneity of variance before performing the analysis of variance (ANOVA) using SPSS (version 16.0, SPSS Inc. Chicago, IL, USA). For the percentage values of mycorrhizal colonization and arbuscule abundance, the data were transformed into arcsine square before statistical analysis. One-way ANOVA followed by Duncan's multiple-range test was first performed with treatment as the main factor. Then, a three-way ANOVA was performed to examine the significance of treatment effects and their interactions on the observed parameters using SPSS (version 16.0, SPSS Inc., Chicago, IL, USA).

Mycorrhizal Colonization and Nodulation
Non-inoculated plants showed no AM fungal colonization or nodulation on their root systems (Table 1). R. irregularis inoculated plants showed mycorrhizal colonization, which was significantly increased by 8.7% and 33.2% under drought stress compared with the well-watered conditions in plants inoculated without and with M. tianshanense, respectively. M. tianshanense inoculation showed a limited effect on root mycorrhizal development because of the reduction of the arbuscular abundance observed in the plants inoculated with M. tianshanense compared with that of the non-inoculated control under the well-watered conditions.
A three-way interaction was observed between water regimes, AM fungus, and rhizobium inoculation for nodule numbers and nodule dry weight. The co-inoculation of M. tianshanense with mycorrhizal fungus favored the nodulation with a nodule number that is 6.1-fold and nodule dry weight that is 6.4-fold higher compared with single inoculation with M. tianshanense under well-watered conditions. Although M. tianshanense nodulation was significantly decreased by drought stress, the co-inoculation with R. irregularis increased nodule number by 12.7-fold and nodule dry weight by 11.6-fold compared with the M. tianshanense single inoculation under drought stress. Table 1. Mycorrhizal colonization and nodulation of plants inoculated with/without arbuscular mycorrhizal (AM) fungus Rhizophagus irregularis and rhizobium Mesorhizobium tianshanense under well-watered and drought stress conditions (mean ± SE, n = 4). C represents non-inoculation control, F represents inoculation with R. irregularis, B represents inoculation with M. tianshanense treatments respectively, and FB represents the co-inoculation treatments. Data of columns indexed by the same letter are not significantly different at p < 0.05. ns, not significant; * p < 0.05; *** p < 0.001.

Plant Biomass
As expected, drought stress decreased plant growth ( Figure 1). In general, R. irregularis inoculation promoted a plant's shoot and root growth irrespective of water regimes. No difference was observed on dry weight of plants inoculated with M. tianshanense compared with non-inoculated control under well-watered conditions, whereas the promotion effects of M. tianshanense inoculation on shoot and root growth were significant under drought stress conditions. Significant synergetic effects of R. irregularis and M. tianshanense inoculation on plant shoot and root growth were recorded under both water regimes. A three-way interaction was observed between water regimes, AM fungus, and rhizobium inoculation for nodule numbers and nodule dry weight. The co-inoculation of M. tianshanense with mycorrhizal fungus favored the nodulation with a nodule number that is 6.1-fold and nodule dry weight that is 6.4-fold higher compared with single inoculation with M. tianshanense under wellwatered conditions. Although M. tianshanense nodulation was significantly decreased by drought stress, the co-inoculation with R. irregularis increased nodule number by 12.7-fold and nodule dry weight by 11.6-fold compared with the M. tianshanense single inoculation under drought stress.

Plant Biomass
As expected, drought stress decreased plant growth ( Figure 1). In general, R. irregularis inoculation promoted a plant's shoot and root growth irrespective of water regimes. No difference was observed on dry weight of plants inoculated with M. tianshanense compared with non-inoculated control under well-watered conditions, whereas the promotion effects of M. tianshanense inoculation on shoot and root growth were significant under drought stress conditions. Significant synergetic effects of R. irregularis and M. tianshanense inoculation on plant shoot and root growth were recorded under both water regimes. conditions (mean ± SE, n = 4). C represents non-inoculation control, F represents inoculation with R. irregularis, B represents inoculation with M. tianshanense treatments respectively, and FB represents the co-inoculation treatments. W represents water regime. The same letter above the error bars indicates no significant difference at P < 0.05. ns, not significant; * p < 0.05; ** p < 0.01; *** p < 0.001.

Microbial Dependency
Based on plant shoot and root dry weight, microbial dependency was calculated to evaluate contribution of mycorrhizal fungus and rhizobium inoculations to plant growth. A positive  Rhizophagus irregularis and rhizobium Mesorhizobium tianshanense under well-watered and drought stress conditions (mean ± SE, n = 4). C represents non-inoculation control, F represents inoculation with R. irregularis, B represents inoculation with M. tianshanense treatments respectively, and FB represents the co-inoculation treatments. W represents water regime. The same letter above the error bars indicates no significant difference at p < 0.05. ns, not significant; * p < 0.05; ** p < 0.01; *** p < 0.001.

Microbial Dependency
Based on plant shoot and root dry weight, microbial dependency was calculated to evaluate contribution of mycorrhizal fungus and rhizobium inoculations to plant growth. A positive contribution of R. irregularis and a negative contribution of M. tianshanense inoculation were observed on shoot and root growth under well-watered condition (Figure 2). A significant increase of microbial dependency was observed when plants were co-inoculated with R. irregularis and M. tianshanense as compared with a single inoculation. Microbial dependencies of single or co-inoculation of R. irregularis and M. tianshanense were significantly increased under drought stress, especially for M. tianshanense inoculation treatment, which shifted from negative to positive response. The shoots had a higher microbial dependency compared with roots.

Drought Tolerance Indices
Plants inoculated with R. irregularis or M. tianshanense alone had a higher STI value compared with non-mycorrhizal control, with R. irregularis and M. tianshanense showing approximately 8.5-fold and 41.2% increase, respectively ( Figure 3). Co-inoculation with R. irregularis and M. tianshanense showed the highest STI value, indicating a positive synergetic effect on plant drought tolerance.

Drought Tolerance Indices
Plants inoculated with R. irregularis or M. tianshanense alone had a higher STI value compared with non-mycorrhizal control, with R. irregularis and M. tianshanense showing approximately 8.5-fold and 41.2% increase, respectively ( Figure 3). Co-inoculation with R. irregularis and M. tianshanense showed the highest STI value, indicating a positive synergetic effect on plant drought tolerance.

Drought Tolerance Indices
Plants inoculated with R. irregularis or M. tianshanense alone had a higher STI value compared with non-mycorrhizal control, with R. irregularis and M. tianshanense showing approximately 8.5-fold and 41.2% increase, respectively ( Figure 3). Co-inoculation with R. irregularis and M. tianshanense showed the highest STI value, indicating a positive synergetic effect on plant drought tolerance.  weights. C represents non-inoculation control, F represents inoculation with R. irregularis, B represents inoculation with M. tianshanense treatments respectively, and FB represents the co-inoculation treatments.

Plant Nutrient Uptake
R. irregularis single inoculation decreased root N concentration under well-watered condition, and increased shoot and root N contents under drought stress (Figure 4). The promotion effects induced by M. tianshanense inoculation were observed both in shoot and root N contents, compared with non-inoculated plants under drought stress. Co-inoculation with R. irregularis and M. tianshanense showed a positive synergistic effect on shoot and root N content, irrespective of water regimes.  . C represents non-inoculation control, F represents inoculation with R. irregularis, B represents inoculation with M. tianshanense treatments respectively, and FB represents the co-inoculation treatments. W represents water regime. The same letter above the error bars indicates no significant difference at P < 0.05. ns, not significant; * p < 0.05; ** p < 0.01 ; *** p < 0.001.
R. irregularis inoculation significantly increased plant shoot and root P concentration and contents. These promotion effects were independent from M. tianshanense inoculation and water regimes ( Figure 5). . C represents non-inoculation control, F represents inoculation with R. irregularis, B represents inoculation with M. tianshanense treatments respectively, and FB represents the co-inoculation treatments. W represents water regime. The same letter above the error bars indicates no significant difference at p < 0.05. ns, not significant; * p < 0.05; ** p < 0.01; *** p < 0.001.
R. irregularis inoculation significantly increased plant shoot and root P concentration and contents. These promotion effects were independent from M. tianshanense inoculation and water regimes ( Figure 5).

Figure 5. Shoot (a) and root (b) P concentrations; shoot (c) and root (d) P contents of plants inoculated with/without AM fungus Rhizophagus irregularis and rhizobium
Mesorhizobium tianshanense under well-watered and drought stress conditions (mean ± SE, n = 4). C represents non-inoculation control, F represents inoculation with R. irregularis, B represents inoculation with M. tianshanense treatments respectively, and FB represents the co-inoculation treatments. W represents water regime. The same letter above the error bars indicates no significant difference at P < 0.05. ns, not significant; * p < 0.05; ** p < 0.01 ; *** p < 0.001.

Discussion
Legumes can form symbiosis both with rhizobia and AM fungi. However, the presence of these microbial symbionts in roots is not always beneficial for plant growth [20,22,23]. In the present study, single inoculation with M. tianshanense promoted seedling growth and N content under drought stress conditions, whereas no effect was observed under well-watered conditions. This result confirmed that symbiotic microbes are more effective with a substantially improved plant root and shoot growth under drought stress conditions, which has also been previously observed both in meta-analytical and empirical studies [6,32,33]. Furthermore, R. irregularis promoted Glycyrrhiza plant growth and increased N and P compensating for the growth suppression induced by rhizobium under well-watered conditions.
Glycyrrhiza plants could effectively form symbiosis with both rhizobia and AM fungi, and rhizobia or AM fungi inoculation could substantially improve Glycyrrhiza growth [3,9,34]. Unsurprisingly, our results showed that R. irregularis stimulated Glycyrrhiza seedling growth, which was more pronounced under drought stress conditions that increased AM fungal colonization. Such pronounced plant growth response to mycorrhizal colonization under water-deficit conditions has been observed previously [33] and was likely attributed to the improvement of phosphorus nutrition content and the combination of physical and cellular effects induced by AM inoculation [35,36]. Our results indicated that Glycyrrhiza seedling P acquisition were dependent on mycorrhizal inoculation especially under drought stress conditions. A plant's photosynthetic products were portioned out to support plant growth and reproduction of symbiotic microbes. Approximately 15-20% carbon is supporting the growth and respiration of   Figure 5. Shoot (a) and root (b) P concentrations; shoot (c) and root (d) P contents of plants inoculated with/without AM fungus Rhizophagus irregularis and rhizobium Mesorhizobium tianshanense under well-watered and drought stress conditions (mean ± SE, n = 4). C represents non-inoculation control, F represents inoculation with R. irregularis, B represents inoculation with M. tianshanense treatments respectively, and FB represents the co-inoculation treatments. W represents water regime. The same letter above the error bars indicates no significant difference at p < 0.05. ns, not significant; * p < 0.05; ** p < 0.01; *** p < 0.001.

Discussion
Legumes can form symbiosis both with rhizobia and AM fungi. However, the presence of these microbial symbionts in roots is not always beneficial for plant growth [20,22,23]. In the present study, single inoculation with M. tianshanense promoted seedling growth and N content under drought stress conditions, whereas no effect was observed under well-watered conditions. This result confirmed that symbiotic microbes are more effective with a substantially improved plant root and shoot growth under drought stress conditions, which has also been previously observed both in meta-analytical and empirical studies [6,32,33]. Furthermore, R. irregularis promoted Glycyrrhiza plant growth and increased N and P compensating for the growth suppression induced by rhizobium under well-watered conditions.
Glycyrrhiza plants could effectively form symbiosis with both rhizobia and AM fungi, and rhizobia or AM fungi inoculation could substantially improve Glycyrrhiza growth [3,9,34]. Unsurprisingly, our results showed that R. irregularis stimulated Glycyrrhiza seedling growth, which was more pronounced under drought stress conditions that increased AM fungal colonization. Such pronounced plant growth response to mycorrhizal colonization under water-deficit conditions has been observed previously [33] and was likely attributed to the improvement of phosphorus nutrition content and the combination of physical and cellular effects induced by AM inoculation [35,36]. Our results indicated that Glycyrrhiza seedling P acquisition were dependent on mycorrhizal inoculation especially under drought stress conditions. A plant's photosynthetic products were portioned out to support plant growth and reproduction of symbiotic microbes. Approximately 15-20% carbon is supporting the growth and respiration of nodules [37], whereas approximately 4-20% carbon goes to mycorrhizal fungi [38]. In our study, under well-watered conditions, there was a negative rhizobial growth dependency, which might be due to the net cost of the symbiosis exceeding the net benefits for the plant [23]. The plant's allocation of more resources to root nodules could reduce the benefits of rhizobia inoculation. However, a positive rhizobial dependency was marked for M. tianshanense-inoculated plants under drought stress, which confirmed the Glycyrrhiza plant-M. tianshanense relationships may vary from a positive to negative microbial dependency. In relation to underlying carbon and nutrient trade, the attention was mainly paid to carbon costs as causes of negative microbial growth responses [24]. However, the range of responses of plants to microbial colonization should be explored as well as the fitness of both plants and their microbial partners [39]. The rhizobial function can be determined by several factors, and our results showed plants had a higher rhizobial dependency to acquire nitrogen resources under drought stress.
The possibility of extending the beneficial effects of plant-rhizobia symbiosis by introducing AM fungus into the symbiont under stress conditions is becoming a popular research topic [40,41]. Our results showed AM fungus enhanced the nodulation and significantly improved plant growth compared to single rhizobial inoculation. Synergistic effects of mycorrhizal and rhizobial symbioses could be reflected by increased rhizobial infection rate, plant nutrient content, and plant growth [42]. Several common signals and processes were involved in the establishment of these two symbiotic associations [13]. An increase in P content by mycorrhizal symbiosis could increase nitrogenase enzyme activity, leading to a higher N 2 fixation of rhizobial symbiont and in return a better mycorrhizal development [43]. In our study, rhizobia infection rate in plants co-inoculated with R. irregularis was much higher compared with single inoculated plants under drought stress conditions, which indicated that AM colonization stimulated nodule formation under stress conditions [44]. AM fungi could also decrease oxidative stress occurring in the nodules [11,45] and improve carbon metabolism of nodules [46]. Our results also demonstrated that AM fungus could increase plant growth and root N content dependency on M. tianshanense under well-watered condition. These results indicated that AM inoculation might be a promising driver to regulate the tripartite symbiosis establishment.
The STI has been widely used to identify stress-sensitive and -resistant cultivars in plant cultivation [31,47]. Based on differences of plant growth decrease under drought stress with respect to favorable conditions, STI could identify plant drought tolerance among different treatments. In this study, R. irregularis symbiosis in combination with M. tianshanense inoculation resulted in synergistic effects on plant drought tolerance, thereby indicating that the tripartite symbiosis could adapt to a wide range of environmental stresses based on STI. AM fungi and rhizobia could help plants tolerate various environmental stresses [46]. Improved nutrient acquisition induced by symbiotic microbes has been postulated as a primary mechanism for enhancing plant drought tolerance. Several mechanisms have also been proposed to explain the mycorrhizal effects, including direct water uptake by mycorrhizal aquaporin, enhanced plant osmotic adjustment, and reduced oxidative damage [35,36]. Our results suggested that co-inoculation with rhizobia and AM fungi could be a useful strategy to enhance the drought tolerance of legume plants.

Conclusions
This study suggested that the combined inoculation of R. irregularis and M. tianshanense was more effective than the single inoculation in promoting plant growth and enhancing plant tolerance to drought stress. Further confirmation of plant stress tolerance induced by AM fungi and rhizobia inoculation is needed under field conditions. For sustainable Glycyrrhiza uralensis production systems, the interactive effects of Rhizobium and AM fungal co-inoculation should be elucidated in detail in the future so that the optimized combinations of microorganisms can be applied as effective soil inoculants for plant growth promotion and fitness.