Removal of Soil Microbes Alters Interspecific Competitiveness of Epichloë Endophyte-Infected over Endophyte-Free Leymus chinensis

Epichloë endophytes may not only affect the growth and resistances of host grasses, but may also affect soil environment including soil microbes. Can Epichloë endophyte-mediated modification of soil microbes affect the competitive ability of host grasses? In this study, we tested whether Epichloë endophytes and soil microbes alter intraspecific competition between Epichloë endophyte-colonized (EI) and endophyte-free (EF) Leymus chinensis and interspecific competition between L. chinensis and Stipa krylovii. The results demonstrated that Epichloë endophyte colonization significantly enhanced the intraspecific competitive ability of L. chinensis and that this beneficial effect was not affected by soil microbes. Under interspecific competition, however, significant interactions between Epichloë endophytes and soil microbes were observed. The effect of Epichloë endophytes on interspecific competitiveness of the host changed from positive to neutral with soil microbe removal. Here higher mycorrhizal colonization rates probably contributed to interspecific competitive advantages of EI over EF L. chinensis. Our result suggests that Epichloë endophytes can influence the competitive ability of the host through plant soil feedbacks from the currently competing plant species.


Introduction
Epichloë endophytes are a group of fungi characterized by their ability to infect the aerial tissues of several cool-season grasses without causing obvious disease [1]. Symbioses between grasses and Epichloë endophytes occur in both natural and agricultural grassland communities and are generally considered to be mutualistic [2][3][4]. The host plant provides food, shelter and a mode of reproduction for the fungus, while the fungus can increase host growth, reproduction, and resistance and tolerance to abiotic and biotic stress, such as drought [5][6][7], low nutrients [8][9][10], herbivory [3,11,12] and pathogen attack [13,14].
Microorganisms 2020, 8 Epichloë endophytes may not only affect the growth and resistances of host grasses, but may also affect soil microbes, such as nitrogen-fixing bacteria, phosphorus-solubilizing rhizospheric fungi, arbuscular mycorrhizal fungi (AMF) and other microbial communities [30][31][32][33]. Can Epichloë endophyte-mediated modification of soil microbes affect the competitive ability of host grasses indirectly? Limited studies focused on AMF. Omacini et al. [15] found that Epichloë endophyte colonization reduced AMF colonization of endophyte-infected (EI) Lolium multiflorum, while increased AMF colonization of neighbouring endophyte-free (EF) ryegrass. As AMF was neutral to both EI and EF plants, the modified colonization did not change the completive advantage of EI over EF plants. In a recent study, we found that Epichloë endophyte colonization significantly inhibited AMF colonization of the host grass Achnatherum sibiricum, and the effects of AMF on host competition were variable and depended on the identity of the AMF species [18]. As AMF diversity is high in natural grasslands, and Epichloë endophyte colonization can affect AMF as well as other microbes [32][33][34], we hypothesized that Epichloë endophyte colonization could not only affect the competitive ability of the host directly by changing its growth and resistance, but also indirectly by changing soil microbes.
Leymus chinensis and Stipa krylovii are two dominant species of the Inner Mongolia steppe. In the present study, we investigated the effects of Epichloë endophytes and soil microbes on the intraspecific competitive interactions between EI and EF L. chinensis and interspecific competitive interactions between L. chinensis and S. krylovii. Specifically, we focused on two primary questions: (1) Does Epichloë endophyte colonization affect the growth and intra-and interspecific competitive ability of L. chinensis? (2) Is the effect of Epichloë endophyte colonization on the growth and competitive ability of L. chinensis influenced by soil microbes?

Plant and Fungal Material
L. chinensis, a rhizomatous perennial grass, was originally sampled from their natural populations at Abaga Banner in Inner Mongolia (43.90 • N, 115.34 • E). It can be colonized by the leaf endophyte Epichloë bromicola, and the infection rate was about 63% [35]. In this area, L. chinensis seldom produces seeds, and stromata have not ever been observed, thus the endophyte is highly likely transmitted via vegetative propagation. EI and EF plants used in this experiment were originally collected from their natural population in summer in 2015, multiplied and selected for uniformity in 2016 and 2017. During this period, we clipped the plants repeatedly and kept them growing vegetatively aiming to avoid possible horizontal transmission of Epichloë endophyte, but only vertical transmission between different tillers. The Epichloë endophyte status of each L. chinensis plant was not only checked microscopically by examining the leaf sheaths for the presence of fungal hyphae after staining with lactophenol aniline blue [36] ( Figure 1A) but also isolated ( Figure 1B) and identified before conducting experiments to confirm endophyte status.
S. krylovii, a perennial bunchgrass, cannot be infected by Epichloë endophyte [37]. S. krylovii seeds were sampled from their natural populations at Abaga Banner in Inner Mongolia. After nine months of growth in the greenhouse at Nankai University, the plants that were almost the same size as L. chinensis were selected.
Soil was collected from the natural grassland hosting the two plant species at Abaga Banner in Inner Mongolia. We sieved the soil samples to remove roots and then pooled them for the experiment. Sterilized soils (which were autoclaved for 90 min at 121 • C) served as a microbe-free (MF) treatment. Non-sterilized soils were used as microbe-inclusive (MI) treatment.

Experimental Design
The experiment had a randomized block design with three factors. One factor, Epichloë endophyte colonization status of L. chinensis, contained two treatments: EI and EF L. chinensis. The second factor, soil microbe status, contained two treatments: MI and MF. These two factors were the same for both intra-and interspecific competition. The difference in the test of the two competition types were the third factor, plant mixture type: monoculture and mixture. The test of intraspecific competition included three groups: monoculture of EI L. chinensis (EI), monoculture of EF L. chinensis (EF) and mixture of EI and EF L. chinensis. In contrast, the test for interspecific competition involved five groups: EI, EF, monoculture of S. krylovii (S), mixture of EI L. chinensis and S. krylovii (EIS) and mixture of EF L. chinensis and S. krylovii (EFS) ( Table 1) A de Wit-type replacement series [38] with equal plant densities was used to assess the intraspecific competitive interaction between EI and EF L. chinensis and the interspecific competitive interaction between L. chinensis and S. krylovii. Six plant individuals per plant species for the monocultures and three + three individuals of each plant species for the mixtures were transplanted into each plastic pot (22-cm diameter and 17-cm depth) containing 1.5 kg sterilized or non-sterilized soil, resulting in 12 combinations. Each combination was replicated five times, yielding a total of 60 pots. The experiment lasted 4 months, from 12 April 2018 to 12 August 2018, and was conducted in the campus experimental field at Nankai University, Tianjin, China. During the experiment, all pots were watered two or three times a week, and nutrients were supplied by the addition of Hoagland nutrient solution once per week to ensure the normal growth of plants.

Biomass and Relative Yield (RY)
At harvest, the shoots of each species were cut at the soil level, dried at 80 °C for 24 h, and then weighed. Roots were collected by gentle washing, separated according to species and divided into two groups. One group was dried at 80 °C for 24 h and then weighed; the other group was kept at

Experimental Design
The experiment had a randomized block design with three factors. One factor, Epichloë endophyte colonization status of L. chinensis, contained two treatments: EI and EF L. chinensis. The second factor, soil microbe status, contained two treatments: MI and MF. These two factors were the same for both intra-and interspecific competition. The difference in the test of the two competition types were the third factor, plant mixture type: monoculture and mixture. The test of intraspecific competition included three groups: monoculture of EI L. chinensis (EI), monoculture of EF L. chinensis (EF) and mixture of EI and EF L. chinensis. In contrast, the test for interspecific competition involved five groups: EI, EF, monoculture of S. krylovii (S), mixture of EI L. chinensis and S. krylovii (EIS) and mixture of EF L. chinensis and S. krylovii (EFS) ( Table 1) A de Wit-type replacement series [38] with equal plant densities was used to assess the intraspecific competitive interaction between EI and EF L. chinensis and the interspecific competitive interaction between L. chinensis and S. krylovii. Six plant individuals per plant species for the monocultures and three + three individuals of each plant species for the mixtures were transplanted into each plastic pot (22-cm diameter and 17-cm depth) containing 1.5 kg sterilized or non-sterilized soil, resulting in 12 combinations. Each combination was replicated five times, yielding a total of 60 pots. The experiment lasted 4 months, from 12 April 2018 to 12 August 2018, and was conducted in the campus experimental field at Nankai University, Tianjin, China. During the experiment, all pots were watered two or three times a week, and nutrients were supplied by the addition of Hoagland nutrient solution once per week to ensure the normal growth of plants.

Biomass and Relative Yield (RY)
At harvest, the shoots of each species were cut at the soil level, dried at 80 • C for 24 h, and then weighed. Roots were collected by gentle washing, separated according to species and divided into two groups. One group was dried at 80 • C for 24 h and then weighed; the other group was kept at −20 • C for the determination of the AMF colonization rate. Relative yield (RY) was calculated as the ratio of the weight of the species in the mixture to the weight of the species in monoculture.

AMF Colonization Rate
Colonization of the roots by AMF was microscopically assessed. A subsample (ca. 2 g) of roots was cleared with 10% (w/v) KOH before being stained using 0.05% trypan blue in lactic acid (v/v) using a modified version of the procedure described in Phillips and Hayman [39]. Mycorrhizal colonization was determined by examining 15-25 1-cm root segments at 200× magnification. The AMF colonization rate was recorded using the cross-hair eyepiece method [40], with a minimum of 200 intersections per replicate.

Competitiveness
The intraspecific competitive ability of EI against EF L. chinensis or interspecific competitive ability of L. chinensis against S. krylovii in the mixture pots was assessed using the aggressivity index (AGR) and the relative interaction intensity index (RII), both of which were calculated using the total biomass of the plant species. The AGR of species x relative to species y was measured according to McGilchrist and Trenbath [41]: where RY is the relative yield of species x or y, defined as the dry matter yield of a species grown in mixture (DM xy or DM yx ) relative to the dry matter in the respective monoculture (DM xx or DM yy ).
If the AGR xy value is zero, then x and y have the same competitive ability. Species x had a higher competitive ability than species y if the AGR xy value was greater than zero. A negative AGR xy value indicated the opposite. The RII allows for simple comparisons of interaction strength across species and treatments [42]. The RII equation is as follows: The RII is a measure of the strength of interactions between species and is centred on zero, with negative interactions (competition) indicated by values between 0 and −1 and positive interactions (facilitation) indicated by values between 0 and +1.

Statistical Analyses
We performed all statistical analyses in the SPSS software (version 20.0, SPSS, Chicago, IL, USA). A two-way ANOVA was conducted to determine the effects of Epichloë endophyte colonization (E) and soil microbes (M) on the relative yield, AGR of L. chinensis and RII of both L. chinensis and S. krylovii and to analyse the effects of E and plant mixture type (C) on the mycorrhizal colonization rate of L. chinensis and to examine the effects of M and C on the relative yield of S. krylovii. We compared the competitiveness of EI relative to EF L. chinensis in intraspecific competition and mycorrhizal colonization rate among S, EIS and EFS in interspecific competition, so we only used one-way ANOVA examining the effect of M on the AGR of L. chinensis and determining the effect of C on the mycorrhizal colonization rate of S. krylovii.

Plant Growth Performance
Growth of L. chinensis, in terms of RY was significantly increased by Epichloë endophyte colonization under both intra-and interspecific competition, but the extent of this effect was higher for intraspecific competition (26% versus 18%, respectively) (Table 2, Figure 2A,C). Microbe removal significantly reduced the RY of L. chinensis under inter-but not intraspecific competition (Table 2, Figure 2B,D). There was no significant interaction between Epichloë endophytes and soil microbes on the RY of L. chinensis under either intra-or interspecific competition (Table 2).  Figure 2B,D). There was no significant interaction between Epichloë endophytes and soil microbes on the RY of L. chinensis under either intra-or interspecific competition ( Table 2). The RY of S. krylovii was significantly affected by the interaction between soil microbes and plant mixture types (Table 2). In addition, there was no significant effect of Epichloë endophytes in L. chinensis on the RY of S. krylovii in the MF treatment, but Epichloë endophyte presence significantly suppressed the RY of S. krylovii in the MI treatment ( Figure 2E). Thus, S. krylovii showed a higher RY when grown with EF than EI L. chinensis in the MI treatment ( Figure 2E).  Values are means ± SE (n = 5). The different letters above bars denote means that are significantly different among treatments (p < 0.05). EI, Epichloë endophyte-colonized L. chinensis; EF, endophyte-free L. chinensis; EIS, S. krylovii grown with EI L. chinensis; EFS, S. krylovii grown with EF L.chinensis; MI, soil microbe inclusive; MF, soil microbe free. Symbol '×', horizontal line in the box, top and bottom of box edges and bars indicate the average value, median, 25% and 75% quartiles, and 5% and 95% quantiles, respectively.
The RY of S. krylovii was significantly affected by the interaction between soil microbes and plant mixture types (Table 2). In addition, there was no significant effect of Epichloë endophytes in L. chinensis on the RY of S. krylovii in the MF treatment, but Epichloë endophyte presence significantly suppressed the RY of S. krylovii in the MI treatment ( Figure 2E). Thus, S. krylovii showed a higher RY when grown with EF than EI L. chinensis in the MI treatment ( Figure 2E).

AMF Colonization Rate
Epichloë endophyte colonization significantly increased the mycorrhizal colonization rate of L. chinensis under both intra-and interspecific competition, and this increase was more pronounced under interspecific competition (Table 2, Figure 3A). The mycorrhizal colonization rate of S. krylovii was similar in mixtures with EF L. chinensis to those in monocultures, which suggests that L. chinensis did not affect the mycorrhizal colonization rate of its neighbouring plants. However, the mycorrhizal colonization rate of S. krylovii was higher when grown with EF L. chinensis plants than when grown with EI plants, which suggests that Epichloë endophyte colonization could inhibit the mycorrhizal colonization rate of neighbouring plants (Table 2, Figure 3B). chinensis under both intra-and interspecific competition, and this increase was more pronounced under interspecific competition (Table 2, Figure 3A). The mycorrhizal colonization rate of S. krylovii was similar in mixtures with EF L. chinensis to those in monocultures, which suggests that L. chinensis did not affect the mycorrhizal colonization rate of its neighbouring plants. However, the mycorrhizal colonization rate of S. krylovii was higher when grown with EF L. chinensis plants than when grown with EI plants, which suggests that Epichloë endophyte colonization could inhibit the mycorrhizal colonization rate of neighbouring plants (Table 2, Figure 3B).

Competitiveness
The AGR was used to measure the competitive ability of EI L. chinensis relative to EF L. chinensis. Under intraspecific competition, the AGR of L. chinensis was positive in the MI treatment, indicating a greater competitive ability of EI L. chinensis relative to that of EF L. chinensis ( Figure 4A), but no significant main effect of soil microbes was observed ( Table 2). Under interspecific competition, the AGR was significantly affected by Epichloë endophyte colonization, soil microbes and their interaction ( Table 2). L. chinensis had a greater competitive ability than S. krylovii (AGR > 0). In addition, there was no significant effect of Epichloë endophytes on the interspecific competitive ability of L. chinensis in the MF treatment, but a higher competitive ability of EI L. chinensis compared to EF L. chinensis was observed in the MI treatment ( Figure 4B).

Competitiveness
The AGR was used to measure the competitive ability of EI L. chinensis relative to EF L. chinensis. Under intraspecific competition, the AGR of L. chinensis was positive in the MI treatment, indicating a greater competitive ability of EI L. chinensis relative to that of EF L. chinensis ( Figure 4A), but no significant main effect of soil microbes was observed ( Table 2). Under interspecific competition, the AGR was significantly affected by Epichloë endophyte colonization, soil microbes and their interaction ( Table 2). L. chinensis had a greater competitive ability than S. krylovii (AGR > 0). In addition, there was no significant effect of Epichloë endophytes on the interspecific competitive ability of L. chinensis in the MF treatment, but a higher competitive ability of EI L. chinensis compared to EF L. chinensis was observed in the MI treatment ( Figure 4B).  The RII provides a simple comparison of interaction strength across species and treatments. Under intraspecific competition, the RII values of both EI and EF L. chinensis were negative in both the MI and MF treatments, and only the RII of EI L. chinensis was less negative compared to that of EF L. chinensis (Table 2, Figure 5A,B). The results indicate that EI and EF L. chinensis suppressed each other, but the magnitude of the negative effect of EI on EF L. chinensis was larger. The presence of soil The RII provides a simple comparison of interaction strength across species and treatments. Under intraspecific competition, the RII values of both EI and EF L. chinensis were negative in both the MI and MF treatments, and only the RII of EI L. chinensis was less negative compared to that of EF L. chinensis (Table 2, Figure 5A,B). The results indicate that EI and EF L. chinensis suppressed each other, but the magnitude of the negative effect of EI on EF L. chinensis was larger. The presence of soil microbes did not have a significant effect on the RII of L. chinensis (Table 2, Figure 5B). In terms of the response of L. chinensis to competition with S. krylovii, the presence of S. krylovii facilitated the growth of EI L. chinensis (RII > 0) but had no influence on EF L. chinensis, and a significant main effect of Epichloë endophytes was observed (Table 2, Figure 5C). In addition, the presence of soil microbes significantly enhanced the promoting effect of S. krylovii on L. chinensis under interspecific competition (Table 2, Figure 5D). However, no significant interactive effect between Epichloë endophytes and soil microbes was observed ( Table 2). The RII of S. krylovii was significantly affected by Epichloë endophyte colonization, soil microbes and their interaction (Table 2). Furthermore, L. chinensis had a negative effect on S. krylovii (RII < 0). Epichloë endophyte colonization had no significant effect on the RII in the MF treatment, but a stronger negative effect of EI L. chinensis on S. krylovii was observed in the MI treatment ( Figure 5E).

Discussion
With respect to the effects of Epichloë endophyte colonization on plant competitiveness, published studies on the intraspecific or interspecific competitive ability of hosts have focused on different grass-endophyte symbionts. Epichloë endophytes can enhance the competitive ability of host plants either by promoting self-growth or by inhibiting the growth of companion plants. For example, Brem and Leuchtman [22] conducted an intraspecific competition experiment and found that Epichloë endophytes enhanced the competitive ability of Br. benekenii by increasing its aboveground dry matter yield. In relation to the interspecific competition experiment, Craig et al. [25] showed that Epichloë endophytes increased the biomass of host P. alsodes. Saikkonen et al. [24] also demonstrated that Epichloë endophytes conferred a competitive advantage to the host, F. pratensis, thereby reducing weed invasions. However, Vázquez-de-Aldana et al. [16] evaluated the effect of Epichloë endophyte Figure 5. Effects of Epichloë endophyte colonization and soil microbes on the relative interaction intensity index (RII) between EI and EF Leymus chinensis (A) and (B) in intraspecific competition and between L. chinensis and Stipa krylovii for L. chinensis in interspecific competition (C) and (D) and effect of the interaction between Epichloë endophyte colonization and soil microbes on the RII between L. chinensis and S. krylovii for S. krylovii in interspecific competition (E). The different letters above bars denote means that are significantly different among treatments (p < 0.05). EI, Epichloë endophyte-colonized L. chinensis; EF, endophyte-free L. chinensis; MI, soil microbe inclusive; MF, soil microbe free. Symbol '×', horizontal line in the box, top and bottom of box edges and bars indicate the average value, median, 25% and 75% quartiles, and 5% and 95% quantiles, respectively.

Discussion
With respect to the effects of Epichloë endophyte colonization on plant competitiveness, published studies on the intraspecific or interspecific competitive ability of hosts have focused on different grass-endophyte symbionts. Epichloë endophytes can enhance the competitive ability of host plants either by promoting self-growth or by inhibiting the growth of companion plants. For example, Brem and Leuchtman [22] conducted an intraspecific competition experiment and found that Epichloë endophytes enhanced the competitive ability of Br. benekenii by increasing its aboveground dry matter yield. In relation to the interspecific competition experiment, Craig et al. [25] showed that Epichloë endophytes increased the biomass of host P. alsodes. Saikkonen et al. [24] also demonstrated that Epichloë endophytes conferred a competitive advantage to the host, F. pratensis, thereby reducing weed invasions. However, Vázquez-de-Aldana et al. [16] evaluated the effect of Epichloë endophyte colonization on the competitive ability of F. rubra against five other grassland species in binary mixtures and found that EI plants had better competitive ability than EF plants, as indicated by the lower RY of companion plants when growing in mixture with EI plants. Certainly, neutral [28,29] and negative [26,27] effects of symbiosis when infected plants grow in competition with other species have also been reported. Are the effects of Epichloë endophyte colonization on intraspecific and interspecific competition similar for a single grass-endophyte symbiont? The only study that has addressed this question was conducted by Marks et al. [20], who found that Epichloë endophyte colonization enhanced both the intra-and interspecific competitive ability of tall fescue but decreased the intra-and interspecific competitive ability of perennial ryegrass. In our study, we found that Epichloë endophyte colonization enhanced both the intra-and interspecific competitive ability of L. chinensis, only the beneficial effects of Epichloë endophyte colonization on the interspecific competitiveness of the host were affected by soil microbes.
In natural systems, plants often experience direct competition with neighbouring plants and simultaneous complex interactions with soil biota [43]. Epichloë endophyte colonization can shape communities of soil biota [44][45][46][47][48] and influence subsequent plant growth and survival [49][50][51][52]. Can simultaneous occurrence of foliar Epichloë endophytes and soil microbes affect the competitivity of the host grasses? Up to now, only the effects of simultaneous occurrence of foliar Epichloë endophytes and AMF were reported and the results were varied. Omacini et al. [15], in binary mixtures of EI and EF Lolium multiflorum, found that the competitive ability of plants was increased by the presence of Epichloë endophytes, while AMF did not affect host performance in the presence or absence of Epichloë endophytes. Vignale et al. [53], in binary mixtures of EI and EF Bromus auleticus, showed that neither Epichloë endophytes nor AMF affected the intraspecific competitive ability of the plants. In our previor study [18], we found that Epichloë endophyte colonization significantly enhanced the interspecific competitive ability of A. sibiricum, but there was no interaction between AMF and Epichloë endophytes. In the present study, we found that the effect of Epichloë endophyte colonization on the intra-and interspecific competitive ability of L. chinensis was differentially mediated by soil microbes. Soil microbe presence did not affect the intraspecific competitive ability of EI relative to EF L. chinensis. However, with respect to interspecific competition, we did find significant interactions between Epichloë endophytes and soil microbes on the competitiveness of L. chinensis. The effect of Epichloë endophytes on interspecific competitiveness of the host changed from positive to neutral with soil microbe removal.
Why did soil microbes not affect the intraspecific competitive ability but change the interspecific competitive ability of EI over EF L. chinensis? In the present study, we investigated AMF colonization rates of L. chinensis and S. krylovii, and found that AMF colonization rates of EF L. chinensis was significantly lower in interspecific competitive studies than those in monoculture and intraspecific competitive studies, while the AMF colonization rates of EI L. chinensis did not change significantly. The advantage of EI over EF L. chinensis in the mycorrhizal colonization was more evident in interspecific than intraspecific studies. Higher mycorrhizal colonization rates probably contributed to higher interspecific competitive advantages of EI over EF L. chinensis.
Plants can shape communities of soil biota and alter soil structure and chemistry in ways that influence subsequent plant growth and survival. The effects of Epichloë endophyte colonization on these plant-soil feedbacks (PSFs) have been found both in host grasses as well as in neighbouring plants in the community. Matthew and Clay [50] found that EI Festuca had reduced growth in soil that previously supported EI Festuca, but was unaffected by soil that once supported EF Festuca. Rudgers and Orr [48] reported that soil conditioning by EI tall fescue reduced the biomass of three native tree species via altered soil microbes. In the present study, we found that the effect of Epichloë endophytes on interspecific competitiveness changed from positive to neutral with microbe removal. The result suggests that the soil feedback effect of the EI plants can build up quickly during competition. Thus, Epichloë endophytes can influence the competitive ability of the host through plant-soil feedbacks from the currently competing plant species.