Effects of Epichloë Endophytes on The Litter Decomposition of Three Host Grass Species in The Qinghai-Tibetan Plateau

Grass fungal endophyte symbioses are widespread in the Qinghai-Tibetan plateau grasslands. It is not clear whether endophytes will inuence litter decomposition in alpine grassland. It is important to understand the role of fungal endophytes in the litter decomposition of host grasses in the grasslands of Qinghai-Tibetan Plateau. Method s This study utilized Festuca sinensis, Stipa purpurea and Achnatherum inebrians as objects and compared their litter with endophyte (E+) infection and without (E-) during the change in litter weight, total nitrogen, lignin and cellulose contents and their residual rate during the decomposition process. The microbial biomass carbon and nitrogen of soil under litters were also compared.


Introduction
For the past few years, many studies have been conducted to delineate the relationship between Epichloë endophytes, which were formerly classi ed in the genus Neotyphodium, and their hosts grasses in the subfamily Pooideae or to delineate the diversity of symbionts in different ecosystems (Chen et Siegel et al. 1987). Studies have shown that these endophytes confer the ability to tolerate stress tolerance to host species and play a signi cant role in the survival of some plants in high-stress environments, such as those subjected to drought, salt, cold, heat, waterlogging, insects and diseases ( . These mutualistic symbioses have been proven to in uence some ecosystem processes, such as soil respiration and litter decomposition among others (Gartner and Cardon 2004;Hector et al. 2000;Omacini et al. 2004).
Litter decomposition is a fundamental ecosystem process that drives nutrient and carbon cycling at the local, regional, and global scales (Aerts 1997;Bradford et al. 2017). The rates of decomposition can be affected by both abiotic and biotic factors (Couteaux et al. 1995;Zhou et al. 2008). These factors include climate (e.g., temperature and soil moisture), litter quality (e.g., the contents of N, lignin and cellulose) and the nature and abundance of decomposing organisms (e.g., bacteria, fungi and soil animals) (Couteaux et al. 1995;Porre et al. 2020). In the same environment, the original litter quality can affect the abundance, composition and activity of microorganism related to litter decomposition and then adjust the rate of decomposition and nutrient cycling (Austin et al. 2014;Berg et al. 1993;Chomel et al. 2016). Climatic condition is also a primary factor determining litter decomposition (Anderson 1991;Fierer et al. 2005). Alpine grasslands are very fragile ecosystems that are highly sensitive to climatic and ecological variations (Hope 2014;Tang et al. 2021).The Qinghai-Tibetan Plateau is with low temperatures, limited precipitation and low oxygen concentrations at high altitudes . Furthermore, the alpine grasslands of the Qinghai-Tibetan Plateau have been affected by anthropogenic factors and have degenerated seriously in recent decades, not only decreased the herbage yields but also altered the soil microorganism activity and the soil nutrient cycling (Dong and Sherman 2015; Li et al. 2018; Wang et al. 2015). Hence, the studies of nutrient dynamics (like litter decomposition) or factors that might in uence these processes in the Qinghai-Tibetan Plateau are essential.
Although endophytes only occur in aerial plant tissues, the effects on root-feeding herbivores and soil-dwelling organisms suggest that endophyte byproducts could be exuded by the roots of infected plants, which would affect the soil micro ora (Bernard et al. 1997;Latch 1993;Omacini et al. 2004). It has been proposed that foliar endophytes of grasses could in uence the decomposition of litter by altering its quality or the microenvironment for decomposition (Bernard et al. 1997;Clay 1997). Casas et al. (2011) found that fungal endophytes in Lolium multi orum increased the activity of soil fungi and affected the metabolic diversity of soil microbial community. However, an increase in microbial diversity does not always accelerate decomposition rates. For example, part of the fungal endophytes could change their role from endophytes to saprophytes to become one of the decomposers (Purahong and Hyde 2011). Thus, the rate of decomposition might be increased by the higher diversity of decomposers. However, studies have reported that there is a competitive relationship between fungal endophytes and secondary saprophytic decomposing fungi (Dowson et al. 1988). So the endophytes may prevent the colonization of these new saprotrophic invaders resulting in lower decomposition rates (Fukasawa et al. 2009; Purahong and Hyde 2011). In addition, Franzluebbers et al. (1999) also concluded that infection by fungal endophytes in Festuca arundinacea could decrease the soil microbial mass and soil respiration and also partly inhibit soil microbial activity. Therefore, the responses of the decomposing microorganism to fungal endophytes will vary with the difference of endophyte species, host species and growing environments. Fungal endophytes not only affect the decomposer community; they also affect the litter quality that is the part of abiotic component (Lemons et  Based on the evidence above, we hypothesized that 1) fungal endophytes will affect the quality of host grass litter by changing the degradation of its components to affect the rate of decomposition of the grass litter; 2) the microenvironment for decomposition will be altered by the presence of fungal endophytes; and 3) different grass-endophyte symbioses will have different rates of litter decomposition in the same environmental conditions. The main focus of this study was to determine how the decomposition of litter from the host plant was affected by fungal endophytes and the related potential mechanism. These results provide insight into the role of grass fungal endophytes in nutrient cycling and ecological protection in alpine grasslands.

Origin of plant litter
The study site was located at an alpine grassland in Haiyan County (N 37°04′, E 100°52′), Qinghai Province of China. Mean elevation of the area is 3,200 m. This area has a typical plateau continental climate, with a mean annual solar radiation of 2,580 h, mean annual temperature of 0.4 to 3.4°C, and annual precipitation of 277.8 to 499.5 mm (most of which falls between May and September). The vegetation is typical of an alpine grassland, with Kobresia and Elymus species serving as the dominant plants in our study area. Other companion species included F. sinensis, Poa pratensis, Melissitus ruthenica, K. humilis, Carex atrofusca and S. purpurea.
We collected natural F. sinensis, S. purpurea and A. inebrians plants with mature reproductive tillers from the study site in October 2017. The endophyte infection status of the plants was determined by the microscopic examination of stained sheaths (Li et al., 2004). The seeds of plants infected with endophytes (E+) were divided into two groups. One group was treated with a 100-fold dilution of the fungicide thiophanate-methyl for 1.5 h and then rinsed with distilled water to obtain seeds without endophyte (E-). Simultaneously, the other group was treated with distilled water (Chen et al., 2018). E + and E-seeds of the three species were planted separately in a greenhouse. The temperature and light cycle in the greenhouse was adjusted to 25:18°C and a 14:10 h light : dark cycle. The leaves of E + and E-plants were collected after four weeks of germination, and the status of their endophyte infection was con rmed using the leaf staining method (Li et al., 2004).

Litter decomposition experiment
We collected litter from newly senesced E + and E-F. sinensis, S. purpurea and A. inebrians at the end of growing season in October 2018. All the plant materials were air dried in the laboratory. To observe the longterm impact of endophytes on the rates of decomposition, litter bags (10 × 10 cm) composed of nylon mesh (1 × 1 mm openings) were used (Chuan et al. 2018). Air dried litter was cut into pieces of approximately 2 cm, and 5 g was inserted into each litter bag. A total of 18 litter bags of each type of litter (108 bags in total) with E + and E-F. sinensis, S. purpurea and A. inebrians were placed at a study plot in the alpine grassland study site in December 2018. The study plot was enclosed in 2016 to prevent disturbance by livestock. After the manual removal of aboveground vegetation, the litter bag sets were laid on the soil surface and xed using plastic sticks. Three randomly selected litter bags of each type of litter were retrieved after 0, 3, 6, 8, 10, and 12 months (December 2018, and March, June, August, October and December 2019). Soil samples that were 10 cm deep were collected under each litter bag at 6 months using a soil auger with a diameter of 5 cm. The soil samples were sieved through a 1 cm mesh to remove stones and stored at 4°C to determine the soil microbial biomass carbon and nitrogen.
The litter in bags was collected at each time point, gently washed in deionized water, and dried at 70°C for 48 h. Soil and little stones were manually removed, and the litter was weighed. Dried samples were ground to powder to measure the contents of total nitrogen (N), lignin and cellulose. The content of total N in litter was determined using a Nitrogen Analyzer System (Kjeltec 2300 Auto System II, Foss Tecator AB, Höganäs, Sweden). The contents of lignin and cellulose were determined after methanol-chloroform extractions and hydrolysis (Rowland and Roberts 1994). The soil microbial biomass carbon was estimated using the chloroform-fumigation-extraction method (FE) described by Vance et al. (1987). The soil microbial biomass nitrogen was estimated using the chloroform-fumigation-incubation method (FI) described by Horwath and Paul (1994).

Data analysis
We tted the data for loss of litter mass to a negative exponential model (Olson 1963): where X 0 is the initial litter mass; X t is the residual litter mass at time t expressed as a proportion of the initial dry mass; and k is the decay constant expressed in year − 1 .
The element residual rate (R) was calculated using the following formula: where C t is element content (mg·g − 1 ) in the residual litter mass at t time; X t is the residual litter mass (g) at t time; C 0 is the initial element content (mg·g − 1 ); and X 0 is the initial litter mass.
Data analyses were performed with SPSS 22.0 for Windows (IBM, Inc., Armonk, NY, USA). A repeated measures analysis of variance (ANOVA) with Fisher's LSD test was utilized to estimate the effect of species, endophyte infection and time on the plant litter weight, total N content and residual rate, lignin content and residual rate, cellulose content and residual rate in the litter samples. The determination of signi cance of the difference between E + and E-plants in all of the parameters was conducted using an independent t-test. Statistical signi cance was de ned at the 95% con dence level. The means are reported with their standard errors.

Results
The weights of the F. sinensis, S. purpurea and A. inebrians litters declined over time; the litters of E + F. sinensis and S. purpurea had a higher rate of decomposition than those of E-, but the speed of decomposition of E + A. inebrians litter was slower than that of the E- (Fig. 1, Table 1). The range of k-values was 0.613 ~ 0.836, and the time at which 95% of the tissue that decomposed was calculated from the k-value indicating that it was 3.67 ~ 4.98 years ( Table 1). The weight of litter was signi cantly lower in E + F. sinensis than that of the E-plants at 6, 8 and 10 months. A lower level was observed in the weight of E + S. purpurea litter from 6 to 12 months when compared with the E-plants. In contrast, E + A. inebrians had signi cantly higher weights of litter compared with the E-plant from 8 to 12 months ( Fig. 1, P < 0.05). There was a signi cant interaction effect between endophyte and species over time ( Table 2). After one year of decomposition in the eld, the total N content in the F. sinensis, S. purpurea and A. inebrians litters increased gradually over time (Fig. 2a-c, Table 2). The total N content of E + F. sinensis litter changed from increasing to stable during the period of decomposition. However, the E-litter had an increased N content throughout the time. Thus, the E + F. sinensis litter had a signi cantly higher content of N at 6 and 8 months but a signi cantly lower N content at 12 months compared with that of E-F. sinensis (Fig. 2a). The total N content in E + litter of S. purpurea was signi cantly higher than that in the E-litter at 6, 10 and 12 months (Fig. 2b). The total N content in E + A. inebrians was higher than that of E-between the decomposition time, and it was signi cantly different at 8 and 12 months (Fig. 2c, P < 0.05). There was a signi cant interaction effect on the content of N in the litter between endophyte and species over time ( Table 2).
The changes in content of lignin in F. sinensis, S. purpurea and A. inebrians litters declined from quick to slow during the decomposition process ( Fig. 2d-f). In addition, the contents of lignin in all that litters were lower than 1% from 8 months. There was no signi cant difference in the content of lignin in litter between E + and E-F. sinensis during the process of decomposition (Fig. d, P > 0.05). In contrast, the E + litter from S. purpurea and A. inebrians had a signi cantly higher content of lignin compared with the E-litters at 3 and 6 months (Fig. 2e, f). Species, endophyte and the time of decomposition had a signi cant effect on the contents of lignin in litters, but there was no signi cant interaction effect for these three factors ( Table 2). The contents of cellulose of F. sinensis, S. purpurea and A. inebrians litters clearly decreased over the period of decomposition (Fig. 2g-i). Similarly with the content of lignin, the change in speed of decomposition of cellulose in litter changed from fast to slow over the course of experiment. The content of cellulose in the E + F. sinensis litter was higher than that of E-at 12 months, and there was no signi cant difference between the E + and E-litters before 12 months (Fig. 2g). However, the E + S. purpurea and A. inebrians litters had lower contents of cellulose compared with those of E-during the period of decomposition, and the difference between E + and Ewas signi cant for S. purpurea during 3, 8 and 10 months and 8, 10 and 12 months for A. inebrians (Fig. 2h, i). There was no signi cant interaction effect on the content of cellulose in litter for species, endophyte and time ( Table 2).
The residual rate of N in F. sinensis litter decreased with the progression of time. There was no signi cant difference between the E + and E-F. sinensis litter during the whole decomposition process (Fig. 3a). The residual rate of N in S. purpurea and A. inebrians litters had no discernable trend during the decomposition (Fig. 3b, c). In addition, the difference between E + and E-litters was not signi cant during most of the decomposition stage. Species and decomposition time had a signi cant effect on the residual rate of lignin in the litters, respectively, but there was no signi cant interaction effect for species, endophyte and time (Table 3). The residual rate of lignin had a similar trend of variation with the content of lignin in F. sinensis, S. purpurea and A. inebrians litters, which decreased gradually with the progression of time (Fig. 3a-c). The litter from E + F. sinensis had a signi cantly lower residual rate of lignin at 6 months, and E + A. inebrians litters had a signi cantly higher residual rate of lignin at 3 months. With the exception of that, the difference between E + and E-litters of the three species was not signi cant during the time of decomposition (Fig. 3d-f). The time had a signi cant effect on the residual rate of lignin in litters, and the interaction effect induced by species, endophyte and time was signi cant (P = 0.006, Table 3). Alternatively, the residual rate of cellulose in the litters of F. sinensis, S. purpurea and A. inebrians also declined as the decomposition time was prolonged (Fig. 3g-i).
The residual rate of cellulose in E + F. sinensis litter was lower than that in the E-litter at 6 and 8 months ( Fig. 3g, P < 0.05). The E + S. purpurea litter had a lower residual rate of cellulose during the process of decomposition, and the difference was signi cant at 8 and 10 months (Fig. 3h). However, the residual rate of cellulose in A. inebrians litter was not signi cantly different among the whole period of decomposition (Fig. 3i). The residual rate of cellulose in litter was signi cantly in uenced by endophyte and time separately, but the interaction effect of species, endophyte and time was not signi cant (Table 3). The microbial biomass carbon of soil under the E + and E-F. sinensis litters did not differ signi cantly (P < 0.05, Fig. 4a). However, the microbial biomass carbon of soil under the E + S. purpurea and A. inebrians litters was signi cantly lower than that of the E-litters. The species did not result in a signi cant effect on microbial biomass carbon, but the effect of endophyte was signi cant (Table 4). There was no signi cant interaction effect for these two factors. The microbial biomass nitrogen of soil in the E + F. sinensis and S. purpurea litters was higher when compared with that of E-(P < 0.05, Fig. 4b), but the difference between E + and E-A. inebrians litters was not signi cant. Similarly, the interaction effect of species and endophyte was not signi cant (P > 0.05, Table 4).

Discussion
As hypothesized, we found that the fungal endophyte affects the decomposition of litter of their host grasses F. sinensis, S. purpurea and A. inebrians. Different species varied in their response to fungal endophyte infection in the rate of decomposition of litter and related factors, such as litter quality and the surrounding microbial mass.
Grasses infected with fungal endophytes can release allelochemicals and alter the area of soil in which they grow (Malinowski et al. 1999;Petroski et al. 1990). Most studies about the effect of endophytes on litter decomposition of their host showed that the endophyte played a negative role during the process of decomposition of litter. For example, Omacini et al. (2004) found that the rates of decomposition in E + litters were 18% lower on average than that in the E-litters in both a garden microcosm experiment and a greenhouse experiment. Lemons et al. (2005) also found that the rates of decomposition were 6% slower in E + litter than in E-litter in an agricultural eld experiment. In this study, the speed of decomposition in E + A. inebrians litter was slower than that in the E-litter (Fig. 1c), which is similar to previous results. Other studies demonstrated that the enhanced accumulation of soil carbon in Festuca pastures that were highly infected with endophytes compared with those that had low levels of infection could be attributed to the reduction in decomposition rates of E + litters (Franzluebbers et al. 1999;Osono 2006;Schomberg et al. 2000).
However, in contrast with A. inebrians, the rate of decomposition of litter in F. sinensis and S. purpurea with endophytes was quicker and the period of decomposition was shorter compared with the litters that lacked endophytes (Fig. 1a, b, Table 1). These results are consistent with a previous study that found that the E. uncinatum fungal endophyte increased the rate of decomposition of Schedonorus pratensis litter when incubated in a garden (Gundel et al. 2016). Grass endophytes can utilize simple sugars, such as glucose, sucrose, and xylose, as a sole carbon source in tissues that had recently died (White et al. 1991). Thus, endophytes have an advantage in their ability to utilize these readily available components before the fungi that colonize after tissue death. Such an ecological advantage suggests that grass infected by endophytes will decompose at a quicker rate (Gundel et al. 2016;Osono 2006).
Within individual ecosystems or biome types, litter quality becomes a more effective determinant of the rate of decomposition than the climate (Aerts 1997;Meentemeyer 1978). At the ecosystem scale, litter quality is most often related to the chemical characteristics of the litter, for example, carbon : nitrogen ratios and/or lignin content (Aber et al. 1990;Aerts 1997). In this study, the content of nitrogen in the A. inebrians, F. sinensis and S. purpurea litters increased with the progression of time (Fig. 2a). The content of nitrogen or phosphorus in litter would gradually increase during the decomposition of litter because of the lower speed of release of nutrient elements compared with the rate of loss of litter mass (Gallardo and Merino 1993). This could also be re ected in the change in residual rate of nitrogen, which decreased with the progression of time of decomposition (Fig. 3a). However, the residual rates of total nitrogen in the litters of three species had no signi cant difference between the E + and E-plants during most of the time of decomposition. Omacini et al. In addition to changes in the mineral content, endophytes have been associated with changes in plant structural parameters, such as the content of bers and lignin, that could also be linked with the decomposition of litter (Gundel et al. 2017;Rogers et al. 2011;Soto-Barajas et al. 2016). In this study, there was no visible effect from fungal endophytes on the original contents of lignin and cellulose of host grass.
However, the degradation of lignin and cellulose varied between E + and E-symbiont for different species.
First, S. purpurea had a higher content of lignin in the E + litter compared with that of E-during the process of decomposition, but the residual rates of lignin did not differ signi cantly (Figs. 2e, 3e). This may be owing to the higher speed of rate of decomposition of the E + litter. The content and residual rate of lignin in the E + litter from A. inebrians were higher than those of the E-litter to some degree during the process of decomposition (Figs. 2f, 3f). Thus, the slower rate of decomposition of the E + A. inebrians litter compared with that of E-may be partly attributed to the slower decomposition of lignin. Secondly, the content and residual rate of cellulose in the E + S. purpurea litter were lower than those in the E-litter to some degree during the year of decomposition (Figs. 2h, 3h). Thus, the lower concentration of cellulose in the lower mass weight of E + litter could result from the faster degradation of cellulose compared with that of the E-S. purpurea litter. The lower residual rate of cellulose in E + F. sinensis compared with that of E-could also induce the quicker decomposition of E + litter (Fig. 3g). However, the lower concentration of cellulose in E + A. inebrians litter was mostly caused by the slower decomposition of the E + litter based on the lack of a signi cant difference in the residual rate of cellulose between the E + and E-litters (Fig. 2i). These results suggest that the effect of fungal endophytes on the decomposition of litter could be partly attributed to the in uence of lignin degradation for A. inebrians and the in uence of cellulose degradation for F. sinensis and S. purpurea.
Soil microbial biomass is the main driving force in the decomposition of organic materials and is frequently used as an early indicator of changes in the chemical and physical properties of soil (Baaru et al. 2007;Brookes 1995). The quicker rate of decomposition of organic materials related to the higher values of soil microbial biomass carbon that resulted from the ready supply of nutrients that provide for microbial growth (Baaru et al. 2007). Grasses infected with fungal endophytes can release allelochemicals and alter the area of soil in which they grow (Malinowski et al. 1999;Petroski et al. 1990 Omacini et al. 2004). Our results suggest that the endophyte of A. inebrians could decrease the degradation of lignin in A. inebrians litter, reducing the microbial mass in soil, to slow the decomposition of host grass litter. We found that the endophyte of S. purpurea accelerated the degradation of cellulose in S. purpurea litter, enhanced the microbial activity of soil, and nally promoted the rate of decomposition of this litter. The fact that the promotion of rate of decomposition by the F. sinensis endophyte may be owing to the higher microbial biomass of nitrogen in the soil under E + litter, whose litter quality did not differ signi cantly between E + and E-during the process of decomposition. Accordingly, the endophytes in A. inebrians, F. sinensis and S. purpurea played different roles in the decomposition of litter from their host, resulting in varied litter quality and environment for decomposition. Our results on decomposition further demonstrate a role for endophyte-plant mutualisms in ecosystem processes under eld conditions. To fully understand the mechanism by which one endophyte affects the decomposition of litter, a more solid conclusion must await further examination on the effects of endophytes on litter quality or microbial community composition and the effect of environmental conditions in the future.

Conclusions
In conclusion, different fungal endophytes played different roles in the litter decomposition of their host grasses. The endophytes in F. sinensis and S. purpurea promoted the rate of decomposition of host and the degradation of cellulose in litter, and they also increased the microbial biomass nitrogen of soil under litter. While the endophyte in A. inebrians delayed the decomposition of host litter and the degradation of lignin degradation in the litter, it also decreased the microbial biomass carbon of soil under litter. Therefore, F. sinensis and S. purpurea endophyte symbioses are suggested to be applied in alpine grasslands in the Qinghai-Tibetan Plateau for accelerating the nutrient ow and improving the vegetation coverage of degraded grassland. However, the mechanism of different fungal endophytes modifying the decomposition of host grass and the role of endophytes playing in the microbial community needs additional study.
Declarations Figure 1 Change in litter weights of E+ and E-Festuca sinensis (a), Stipa purpurea (b) and Achnatherum inebrians (c) during the decomposition time. E+: with endophyte; E-: without endophyte. *P < 0.05 between the E+ and Elitters simultaneously Change in the residual rate of N, lignin and cellulose in E+ and E-Festuca sinensis (a, d, g), Stipa purpurea (b, e, h) and Achnatherum inebrians (c, f, i) during the decomposition time. E+: with endophyte; E-: without endophyte. *P < 0.05 between the E+ and E-litters simultaneously