Introduction

Endophyte communities are widely distributed and are an important component of grasslands (Rudgers et al. 2004; Sánchez Márquez et al. 2007, 2008). Endophytic fungi may increase nutrient uptake of hosts (Richardson et al. 1999; Omacini et al. 2006; Zabalgogeazcoa et al. 2006), enhance their tolerance to abiotic and biotic stresses such as drought (Hesse et al. 2005; Kannadan and Rudgers 2008; Jaber and Vidal 2010), insects (Clay et al. 1993; Emery et al. 2010) and herbivores (Hartley and Gange 2009). Endophytic fungi can therefore play important ecological roles in plant communities (Clay and Holah 1999; Rudgers and Clay 2007; Giordano et al. 2009; Wäli et al. 2009; Aly et al. 2010) and have become important candidates for novel compound discovery (Huang et al. 2008, 2009; Tejesvi et al. 2009; Aly et al. 2010).

There have been many studies of endophytic fungi focused on grass leaves (e.g., Saikkonen et al. 2000; Sánchez Márquez et al. 2007, 2008) and roots (Skipp and Christensen 1989; Schulz 2006; Porras-Alfaro et al. 2008), but with the exception of Sánchez Márquez et al. (2007, 2008), few studies have compared the endophytic fungi between leaves and roots.

The composition of fungal communities in plants in general may be affected by many factors, e.g., host species, tissues, age, geographically distant regions, and environment (Petrini 1996; Collado and Platas 1999; Fröhlich et al. 2000; Higgins et al. 2007; Wang and Guo 2007; White and Backhouse 2007). However, the relationship between plant diversity and endophytic fungal diversity has generally been neglected. It is not clear whether the diversity of endophytic fungi can be influenced by the presence of neighboring plants. It is therefore important to establish whether endophytic fungi in one host plant will change when surrounded by differing levels of plant diversity.

The Inner Mongolia steppe, distributed at the eastern end of Eurasian steppe zone, is the largest grassland in China and is an important natural resource in these arid and semiarid regions. The region contributes significantly to the ecology and economy of China. In this ecosystem, plants are divided into different functional groups, e.g., perennial rhizome grass, based on their traits, responses to environmental constraints, and effects on main ecosystem processes (Díaz and Cabido 1997; Lavorel et al. 1997). The plant functional groups are important in ecosystem functioning (Díaz et al. 2003; Bai et al. 2004), and are the main drivers of carbon, nutrient and water cycling in the ecosystem (Urcelay et al. 2009). Stipa grandis is one of the most common grass species and of highly important nutritional forage value for sheep and cattle in the Inner Mongolia steppe. Its endophytic fungi have an important ecological function in grassland ecosystems (Rudgers et al. 2004). The differences in communities of endophytes in leaves and roots of S. grandis in the steppe, however, have received limited attention. This difference is considered important because leaves and roots represent the most dynamic interfaces between plants and the environment. Fungi that inhabit these biologically active tissues may share characteristics that allow them to grow and persist in an ever-changing biochemical milieu, as host tissues grow and age, and play important ecological roles (Arnold 2007).

The aims of this study, carried out at the Inner Mongolia steppe, were to investigate the colonization rates and communities of endophytes in leaves and roots of S. grandis. The study also aimed to establish the effect of plant function group removal on the endophytic fungal communities in S. grandis.

Materials and methods

Site and sampling procedure

This study was conducted in the plant functional group removal experiment site in the Inner Mongolia Grassland Ecosystem Research Station, the Chinese Academy of Sciences (43°26′–44°08′N, 116°04′–117°05′E), which is located in the typical steppe zone of the Inner Mongolia Plateau. It has a semiarid continental temperate steppe climate with a dry spring and a moist summer. The annual mean temperature is 2ºC and an annual precipitation is 350 mm. The altitude is 1,180–1,250 m above sea level, and the soil is Chestnut (Chen 1988).

There are 32 plant species in the study site, divided into five plant functional groups, i.e., perennial rhizome grass, perennial bunchgrasses, perennial forbs, shrubs and semi-shrubs, and annuals and biennials. Of these, perennial rhizome grass comprises Leymus chinensis; perennial bunchgrasses comprise Achnatherum sibiricum, Agropyron cristatum, Cleistogenes squarrosa, Koeleria cristata, Poa subfastigiata, and Stipa grandis; perennial forbs comprise Allium anisopodium, A. bidentatum, A. ramosum, A. senescens, A. tenuissimum, Artemisia pubescens, Carex korshinskyi, Cymbaria dahurica, Haplophyllum dauricum, Heteropappus altaicus, Potentilla acaulis, P. bifurca, P. tanacetifolia, Pulsatilla turczaninovii, Saposhnikovia divaricata, Serratula centauroides, and Thalictrum petaloideum; annuals and biennials comprise Artemisia scoparia, Axyris amaranthoides, Chenopodium aristatum, C. glaucum, and Salsola collina; and shrubs and semi-shrubs comprise Artemisia frigida, Caragana microphylla, and Kochia prostrata (Bai et al. 2004). The plant functional group removal experiment site was established in 2005, and includes eight blocks (55 m × 85 m each block) as replicates. Each block contains 96 plots (6 m × 6 m each plot) with different treatments of plant functional group removal—the removal is carried out every year.

Four treatments were selected in our study. In treatment I no plant species were removed. In treatment II two plant functional groups (perennial forbs and perennial rhizome grass) were removed. In treatment III three plant functional groups (annuals and biennials, perennial forbs, and perennial rhizome grass) were removed. In treatment IV four plant functional groups (annuals and biennials, perennial forbs, perennial rhizome grass, and semi-shrubs) were removed. Three individuals of S. grandis were collected randomly from each plot. The plant samples with leaves and roots were immediately placed in plastic bags, labeled, put in an ice box and taken to the laboratory. Samples were stored at 4ºC and processed within 2 d of collection. A total of 96 plants of S. grandis were collected from the four treatments with 8 replicates in this study.

Isolation and culture of endophytic fungi

The leaves and roots of S. grandis were washed thoroughly under running tap water and cut into 5 mm-long segments. In total, 768 leaf or root segments (8 segments × 3 individuals × 8 replicates × 4 treatments) were used in this study. Surface sterilization of the plant tissue segments was carried out using the method of Guo et al. (2000). Samples were surface sterilized by consecutive immersions for 1 min in 75% ethanol, 3 min in 5% (roots) or 3% (leaves) sodium hypochlorite and 30 s in 75% ethanol.

The segments were surface dried with sterile paper towels. Sets of four segments were evenly placed in each 90 mm Petri dish containing malt extract agar (MEA, 2%) supplemented with Rose bengal (30 mg/L) to slow down fungal growth. Streptomycin sulfate (50 mg/L) was added to suppress bacterial growth. Petri dishes were sealed, incubated for 2 months at 25ºC, and examined periodically. When colonies developed, they were transferred to new Petri dishes with potato dextrose agar (PDA, 2%), and sterile leaf segments of S. grandis to promote sporulation (Guo et al. 1998). Isolates were incubated at 25ºC, with cool white fluorescent light (12 h light, 12 h dark) to induce sporulation.

Identification of endophytic fungi

Subcultures on PDA were examined periodically and identified based on morphological characteristics, if they sporulated. Cultures that fail to sporulate were designated “mycelia sterilia”, and were divided into morphotypes according to cultural characteristics (Lacap et al. 2003). All living cultures were deposited in the China General Microbiological Culture Collection Center (CGMCC) in Beijing, China.

DNA extraction, amplification and sequencing

DNA extraction was carried out for sterile isolates. Total DNA was extracted from fresh cultures following the protocol of Guo et al. (2000). Fresh fungal mycelia (c. 50 mg) was scraped from the surface of the agar plate and transferred into a 1.5 ml microcentrifuge tube with 700 μl of preheated (60ºC) 2× CTAB extraction buffer (2% (w/v) CTAB, 100 mM Tris–HCl, 1.4 M NaCl, 20 mM EDTA, pH 8.0), and c. 0.2 g sterilized quartz sand (Sigma). The mycelium was ground using a glass pestle for 5–10 min and then incubated in a 60ºC water bath for 30 min with occasional gentle swirling. 500 μl of phenol : chloroform (1:1) was added into each tube and mixed thoroughly to form an emulsion. The mixture was spun at 11 900 g for 15 min at room temperature in a microcentrifuge and the aqueous phase was removed into a fresh 1.5 ml tube. The aqueous phase containing DNA was re-extracted with chloroform : isoamyl (24:1) until no interface was visible. 50 μl of 5 M KOAc was added into the aqueous phase followed by 400 μl of isopropanol and inverted gently to mix. The genomic DNA was precipitated at 9200 g for 2 min in a microcentrifuge. The DNA pellet was washed twice with 70% ethanol and dried using SpeedVad (AES 1010; Savant, Holbrook, NY, USA) for 10 min or until dry. The DNA pellet was then resuspended in 100 μl TE (10 mM Tris–HCl, 1 mM EDTA).

The ITS (ITS1, 5.8 S, ITS2) regions were amplified using primer pairs ITS4 and ITS5 (White et al. 1990). Amplification was performed in a 50 μl reaction volume which contained PCR buffer (20 mM KCl, 10 mM (NH4)2SO4, 2 mM MgCl2, 20 mM Tris–HCl, pH 8.4), 200 μM of each deoxyribonucleotide triphosphate, 15 pmols of each primer, c. 100 ng template DNA, and 2.5 units of Taq DNA polymerase. The thermal cycling program was as follows: 3 min initial denaturation at 95ºC, followed by 35 cycles of 40 sec denaturation at 94ºC, 50 s primer annealing at 52ºC, 1 min extension at 72ºC, and a final 10 min extension at 72ºC. A negative control using water instead of template DNA was included in the amplification process. Four microliters of PCR products from each PCR reaction were examined by electrophoresis at 75 V for 1 h in a 1% (w/v) agarose gel in 1× TAE buffer (0.4 M Tris, 50 mM NaOAc, 10 mM EDTA, pH 7.8) and visualized under UV light after staining with ethidium bromide (0.5 μg/ml).

Isolates whose sequences had a similarity greater than 97% were considered to belong to the same species or species complex (O’Brien et al. 2005). Sequence-based identifications were made by searching with Blastn in the EMBL/GenBank database of fungal nucleotide sequences.

Data analysis

Colonization rate was calculated as the total number of plant tissue segments infected by one or more fungi, divided by the total number of incubated segments (Petrini et al. 1982). Relative frequency was calculated as the number of particular taxon divided by the total number of all taxa in each treatment. Shannon-Weiner index (H’) was calculated according to the formula: \( H\prime = - \sum\limits_{i = 1}^k {{P_i} \times \ln {P_i}} \), where k is the total species number of one plot, and P i is the relative abundance of endophytic fungus species of one plot (Pielou 1975). To evaluate the degree of community similarity of endophytic fungi between the two treatments, the Sorenson’s coefficient similarity index (C S ) was employed and calculated according to the following formula: C S  = 2j/(a+b), where j is the number of endophytic fungus species co-existing in two treatments, a is the total number of endophytic fungus species in one treatment, and b is the total number of endophytic fungus species in another treatment.

All statistical analyses used one-way analysis of variance (ANOVA) to determine any significant difference (SPSS for windows, version 11.5, SPSS Inc, Chicago, USA). ANOVA were used to test whether there were significant differences in colonization rates and Shannon-Weiner diversity index of endophytes between different treatments and tissues. Statistical significance was determined at the p < 0.05 level.

Results

Colonization rate

A total of 1536 tissue segments collected from 96 individuals of S. grandis in the four treatments were processed, and 571 fungal isolates were recovered. Of these, 148 (35 from leaves and 113 from roots) were isolated in treatment I, 143 (28 from leaves and 115 from roots) in treatment II, 125 (26 from leaves and 99 from roots) in treatment III, and 155 (39 from leaves and 116 from roots) in treatment IV (Table 1). The colonization rates of endophytic fungi were significantly higher in roots than in leaves, but there was no significant difference of endophytic fungi within roots or in leaves among the four treatments.

Table 1 Colonization rate and Shannon-Weiner diversity index (H’) of endophytic fungi isolated from leaves and roots of Stipa grandis in the four treatments
Full size table

Composition of endophytic fungi

Of the 571 isolates, 251 (44%) sporulated and were identified into 13 taxa based on the morphological characteristics. The remaining 320 isolates (56%) did not sporulate and were divided into 60 morphotypes (Lacap et al. 2003) based on the cultural characteristics, and were further identified into 21 taxa by means of ITS sequence analyses (Table 2). Thus, 34 taxa were determined based on morphology and molecular analyses.

Table 2 Tentative identification of sterile mycelia based on ITS sequences
Full size table

Alternaria sp. 1 and Pyrenophora sp. were the dominant species isolated only from leaves in the four treatments (Table 3). In the roots, Fusarium redolens was dominant in treatments I and II, and Phialophora sp. was dominant in treatments III and IV.

Table 3 Relative frequencies of endophytic fungi isolated from leaves and roots of Stipa grandis in the four treatments
Full size table

Nine taxa were isolated from leaves and 25 from roots and no taxa occurred in both leaves and roots (Table 3). Of the nine fungi isolated from leaves, Lewia sp. and Sclerostagonospora sp. were only present in treatment I, Preussia sp. 1 and Stagonospora sp. only occurred in treatment IV, and the remaining five taxa appeared in at least two treatments. Of the 25 fungi isolated from roots, Fusarium sp. 2 was only present in treatment I, Monosporascus sp., Pyrenochaeta sp., Thielavia appendiculata and Tricholomataceae sp. only in treatment II, Alternaria sp. 2 in treatment III, and Microdochium sp. and Pleosporales sp. 1 only in treatment IV.

The endophyte species numbers in leaves were respectively 5, 4, 4, and 6 in treatments I–IV and 14, 18, 17, and 13 in roots (Table 3). The Sorenson’s coefficient similarity index for endophytic fungi in the leaves was highest between treatments II and IV and was lowest between treatments I and IV, whereas, for endophytic fungi in roots it was highest between treatments I and IV and lowest between treatments II and IV (Table 4). The Shannon-Weiner diversity index of endophytic fungi was significantly higher in roots than in leaves in all treatments, but there was no significant difference among the treatments (Table 1).

Table 4 Sorenson’s coefficient similarity index of the endophytic fungi isolated from leaves and roots between the four treatments
Full size table

Discussion

In this study we used traditional techniques to isolate endophytes from S. grandis. The resulting isolates were identified by morphology and if they were mycelia sterilia we also used molecular sequence data to aid the identification of taxa. This method has become commonplace in endophyte studies and is responsible for placing names on many more endophyte strains than was previously possible (Guo et al. 2003; Wang et al. 2005; Huang et al. 2008, 2009; Tao et al. 2008). We identified 13 taxa using morphology and a further 21 taxa following analysis of sequence data. Although the additional use of sequence data is preferable to only using morphological identification of endophyte isolates, the method still relies on the isolation and growth of endophytes using traditional techniques (Hyde and Soytong 2008). The endophytes isolated from grasses in this and other studies differ from those recorded as saprobes (Poon and Hyde 1998; Wong and Hyde 2001), indicating that most endophytes of grasses may have functionally different roles.

Effect of tissue types and treatments on colonization rate

The colonization rates of endophytic fungi were significantly higher in roots than in leaves in the four treatments, but there was no significant difference in roots and in leaves between the treatments. This is perhaps because the roots of S. grandis are perennial, while leaves are annual. This is consistent with a hypothesis of predominantly horizontal transmission of endophytes, as old plant tissues would have had more time to accumulate endophytes from the environment (Bertoni and Cabral 1988; Taylor et al. 1999; Wang and Guo 2007; Guo et al. 2008 ).

Effect of tissue types on endophytic fungi

Thirty four fungal taxa were isolated from S. grandis, which is lower than that reported for leaves and roots of Ammophila arenaria (75 species) and Elymus farctus (54 species) from the Atlantic coasts of Europe (Sánchez Márquez et al. 2008). This may be due to the arid climate in the Inner Mongolia steppe or lower diversity of surrounding plant species.

No endophytic species occurred in both leaves and roots. Similar results were reported in previous studies where roots and leaves of one plant have been studied. Suryanarayanan and Vijaykrishna (2001) found that the composition of endophytic species differed between leaves and roots of Ficus benghalensis. There was little overlap in endophyte communities between the roots and leaves of Tripterygium wilfordii, with the exception of Pestalotiopsis disseminata (Kumar and Hyde 2004). It might be a reflection of the recurrence of endophytic fungi in different tissues, and might also reflect their capacity for surviving within a specific substrate (Carroll and Petrini 1983). The distribution of endophytic fungi has also been found to be affected by different tissues in other studies (Petrini and Carroll 1981; Arnold et al. 2001; Wang and Guo 2007).

Several dominant genera were ubiquitous endophytes present in other grasses and plant families, e.g., Alternaria and Fusarium (Sánchez Márquez et al. 2007, 2008). Some species identified as endophytes in this survey are previously described as pathogens in other plants, e.g., Fusarium oxysporum is a pathogen on many plants (Nemat Alla et al. 2008), Pyrenophora sp. is a pathogen of wheat (Colson et al. 2003). Previous studies have shown that latent pathogens can live within plants for some time as symptomless endophytes (Brown et al. 1998; Photita et al. 2004; Promputtha et al. 2007). Since we grouped species based on 97% ITS sequence similarity it is more likely we have identified species complexes. Fusarium oxysporum has been show to be a species complex comprising at least 50 species (Kvas et al. 2009).

There have been fewer studies on endophytes colonizing roots when compared to those in leaves and even fewer have examined both roots and leaves (e.g. Kumar and Hyde 2004; Sánchez Márquez et al. 2008). Our results indicate that roots harbour a higher diversity of endophytic fungi than leaves and this was the case in all treatments. Other studies however, have shown that endophyte diversity in roots was similar to that of leaves. For example, there were 15 and 18 endophytic taxa in the roots and leaves of Tripterygium wilfordii, respectively (Kumar and Hyde 2004). Fifty one taxa were isolated from leaves and 38 from rhizomes of Ammophila arenaria, while 36 taxa were isolated from leaves and 34 from rhizomes of Elymus farctus (Sánchez Márquez et al. 2008).

One possible reason for the higher biodiversity in roots of S. grandis is that the leaves are slender, thus reducing colonization opportunities of endophytic fungi which are often transmitted via wind dispersal (Fröhlich et al. 2000). Although roots are also narrow, they are often in close contact with each other, just like a net, which is favourable for horizontal endophyte transmission. The leaves are also short-lived, dying during the long cold winter, whereas, roots persist from one year to the next. Leaves are more biochemically dynamic, environmentally variable, more critical for photosynthesis, and subject to damage by sucking and chewing herbivores (Arnold 2007). Accordingly, endophytes inhabiting leaf foliage are under a suite of selective pressures distinct from those facing xylem endophytes, or endophytes associated with tissues such as inner bark and roots (Arnold 2007; Oses et al. 2008).

Effect of treatments on endophytic fungi

In this study, the relative frequencies of endophyte taxa in the different treatments were varied. The dominant species in the roots was also not consistent in the four treatments, for example, Fusarium redolens was dominant in treatments I and II, while Phialophora sp. was dominant in treatments III and IV. Some species only occurred following certain treatments, e.g., Fusarium sp. 2, Lewia sp. and Sclerostagonospora sp. were only present in treatment I, Monosporascus sp., Pyrenochaeta sp., Thielavia appendiculata, and Tricholomataceae sp. were only present in treatment II, Alternaria sp. 2 in treatment III, and Microdochium sp., Pleosporales sp. 1, Preussia sp. 1 and Stagonospora sp. only in treatment IV. These results show that different treatments affect the composition of the endophytic fungus community and this has not been reported previously. The removal of different functional plant groups decreased plant diversity which in turn caused a change in composition of the endophyte community. Arnold (2008) also indicated that plant taxa harbored different endophyte communities and that fungal diversity increases with host plant diversity. Thus neighbouring plants can affect the endophyte diversity present in a host.

Endophytes can be horizontally transmitted between different hosts. Diverse ascomycetes or basidiomycetes can be transmitted between roots (Rodriguez et al. 2009). There has not been sufficient ecological studies to permit a full understanding of the distribution and abundance of fungal endophytes in the rhizosphere (Rodriguez et al. 2009), but they should have important ecological benefits to plants. Horizontal transmission can broaden their distribution.

This study has shown that colonization rates of endophytic fungi in S. grandis were significantly higher in roots than in leaves in the four treatments. This finding differs from most previous studies and may be due to the small size of the leaves which grow annually, as compared to the roots which persist from year to year under the ground. The study also provides some proof that horizontal transmission of endophytes may occur between the same and different grass species. This would normally occur through the roots, again accounting for the higher diversity. The study has also shown that some endophytes have a certain degree of tissue specificity, which has also been observed in many other endophyte studies. The surrounding plant diversity or plant make-up was also observed to affect endophyte communities, e.g., the number of fungal species, relative frequencies of some fungi, and species composition. If endophyte communities alter with change of functional plant groups, then this is likely to affect the dynamics of ecosystem functioning. Global warming and human activities can increase species extinction, therefore, if some functional groups disappear, then the fungi communities will also change. We isolated endophytes using traditional methodology. Direct environmental sampling of endophytes from the leaves and roots would enhance the present study (Duong et al. 2006; Seena et al. 2008).