Soil fungal diversity in three nature reserves of Jiuzhaigou County, Sichuan Province, China

Abstract

Jiuzhaigou County is located at the southern transition zone of Sichuan Basin and the Qinghai–Tibet Plateau and is the site of three famous nature reserves, namely, the Jiuzhaigou Nature Reserve (JZNR), Baihe Nature Reserve (BHNR) and Wujiao Nature Reserve (WJNR). The soil fungal diversity in this region has not yet been investigated. In this study, we collected 25 soil samples from these three nature reserves. Soil fungi were isolated using the soil dilution plate technique and Rose Bengal agar medium. The culturable soil fungal density based on analysis of the 25 samples ranged from 2.18 log to 4.38 log CFU g⁻¹ dry weight soil, with the fungal density being highest in samples from JZNR and lowest in those from BHNR. Based on morphological characters and the results of phylogenetic analysis of the internal transcribed spacer (ITS) of the rDNA operon, we identified 38 genera (two genera could not be identified) belonging to Ascomycota, Zygomycota and Basidiomycota. The dominant genera were Penicillium, Humicola, Aspergillus and Trichoderma. The species richness index S, biodiversity index H′ and evenness index E of the 25 sampling sites were in the range 10–29, 1.96–3.05 and 0.74–0.95, respectively. The highest mean values of the S, H′ and E indices were in soil samples from BHNR, where the values of these indices were 20.00, 2.66 and 0.90, respectively. These results indicate that the diversity of culturable fungi in these three nature reserves was high. Furthermore, a total 14 Trichoderma isolates were tested for their antagonism activity against mycelium growth of three pathogens: Bipolaris maydis, Curvularia lunata, Rhizoctonia solani. The results showed that six Trichoderma isolates had good antagonistic effects on the three pathogenic fungi.

Introduction

It has been estimated that there are 712,000 extant fungal species worldwide (Schmit and Mueller 2007). Soil is an important fungal habitat, and the great majority of fungal species spend at least some part of their life cycles in the soil environment (Bridge and Spooner 2001). Fungi play a crucial role in the terrestrial ecosystem, and they are responsible for many key steps in the maintenance of ecosystem stability, particularly by recycling the soil organic matter and mineral elements (e.g. cellulose, lignin, carbon, nitrogen, phosphorus) (Gadd 2007; Barrico et al. 2010; Hollister et al. 2010), and their diversity and activity reflect soil health (Mueller et al. 2004; Wardle et al. 2006; Singh et al. 2007). Fungal functional diversity is intimately related to the taxonomic community structures (Zak and Visser 1996; Deacon et al. 2006). On the one hand, the structure of soil fungal communities vary depending on different ecological factors (Buee et al. 2011), including altitude (Laganà et al. 1999; Pan et al. 2009), climate (Persiani et al. 1998; Satish et al. 2007; McGuire et al. 2012), species and age of the vegetation (West and Jones 2000; Dong et al. 2004; Curlevski et al. 2010; Nie et al. 2012), soil nutrients (Lejon et al. 2005; Kara and Asan 2007; Thoms et al. 2010) and human disturbances (Cabello and Arambarri 2002; Bastias et al. 2006). On the other hand, they are significantly affected by fertilizer (Schneider et al. 2010; Jirout et al. 2011) and tillage management (Wu et al. 2007, 2008; Entry et al. 2008; Wang et al. 2010; Zhang et al. 2011) of agricultural land. These findings suggest that soil fungal community structures are good tools for monitoring the changes in environmental conditions (Grishkan et al. 2008; Nie et al. 2012).

However, it is still difficult to isolate all soil fungi due to the limitations in available culture methods. The soil dilution plate method is a relatively reliable and widely employed approach by which to characterize the communities of culturable soil fungi (West and Jones 2000; Cabello and Arambarri 2002; Grishkan et al. 2006, 2009; Nesci et al. 2006; Pan et al. 2009; Wang et al. 2010; Arenz et al. 2011; Arenz and Blanchette 2011). Traditional fungal identification is mainly based on morphological criteria. However, some fungi do not form spores or fruiting body when cultured on/in an artificial substrate (Lim et al. 2005), causing great difficulties for identification purposes. Developments in biotechnology have resulted in new and improved methods for identifying fungi, including; these include the use of the internal transcribed spacer (ITS) of the rDNA operon, beta-tubulin gene and translation elongation factor 1 alpha gene (Alves et al. 2008; Phillips et al. 2008; Tanaka et al. 2009). The ITS region is widely used for identifying fungi and is regarded as the most powerful and reliable tool for the accurate identification of fungi (Henry et al. 2000; Anderson et al. 2003; Anderson and Parkin 2007; Ortega et al. 2008; Wang et al. 2008). However, nucleotide databases do not cover all fungal taxa, especially when the fungal sequences show only low similarity with sequences in nucleotide databases, thereby hampering the application of molecular identification techniques (Lim et al. 2005). It would therefore appear that the use of morphological characteristics coupled with molecular analysis can provide a more accurate identification of fungi.

The Jiuzhaigou Nature Reserve, Baihe Nature Reserve and Wujiao Nature Reserve are located in Jiuzhaigou County, at the southern edge of Sichuan Basin and the Qinghai-Tibet Plateau transition zone, China. Newly explored and unexplored habitats are important potential sites for discovering new fungal species (Hawksworth and Rossman 1997), and the soil fungal diversity in this region is noteworthy. Given that this area is characterized by its geographical location and low human disturbance, we have investigated the diversity of culturable soil fungi in the three nature reserves.

Materials and methods

Study site and soil sample collection

Jiuzhaigou County (area 5,290 km²) is located at the southern edge of Sichuan Basin and the Qinghai–Tibet Plateau transition zone, China. Average temperature changes range between −3.7 °C and 16.8 °C (annual average 7.3 °C). Total annual rainfall is approximately 700–800 mm, with most precipitation occurring between May and September. Many of the sites are covered with forest, with an altitudinal variation ranging from mixed forest through coniferous forest to alpine meadows. The Jiuzhaigou Nature Reserve (hereinafter referred to as JZNR), Baihe Nature Reserve (hereinafter referred to as BHNR) and Wujiao Nature Reserve (hereinafter referred to as WJNR) are located in the southwest, northwest, and southeast parts of Jiuzhaigou County, respectively. JZNR is national nature reserve and famous for its State 4A-level scenery and its designations as a World Natural Heritage site and World Bio-sphere Reserve site. BHNR and WJNR are provincial nature reserves. All three nature reserves located far from industrial centers and human impact is low.

Soil samples were collected in July 2007 across the range of soil types, vegetation and altitudes of the three nature reserves. Global Positioning System technology (GPSMap76; Garmin Ltd., USA) was used to determine the sampling locations. We collected 25 soil samples in total (8 from JZNR, 8 from BHNR and 9 from WJNR). For each sample, five soil sub-samples were collected from the topsoil (depth 0–15 cm) from random positions of approximately 1.0 m² in size; these were then mixed to make up one sample for analysis.

After the removal of vegetation debris, approximately 300 g of the soil sampled from each sample site was immediately collected in sterile plastic bags, kept in the icebox, transported to the laboratory within 48 h and then stored at 4 °C. Details on the 25 soil sample sites are given in Table 1.

Table 1 The characteristics, biodiversity indices and density of 25 sampling sites

Fungal isolation and morphological identification

Soil fungi were isolated using the suspension plating method (Mueller et al. 2004). In brief, 10 g soil of each sample was added to 90 ml sterilized water, producing a soil slurry of 10⁻¹ (w/v); this soil suspension was shaken for 15 min and diluted to final concentrations of 10⁻² and 10⁻³. Suspensions (1 ml) of different concentrations (10⁻¹, 10⁻² and 10⁻³) were placed in 90-mm diameter petri plate, and then the Rose Bengal agar medium (approximately 40 °C) was added and mixed evenly with the suspension. Sterilized water was used as the controls. The plates were kept in the dark at 25 °C for 5–7 days. Only plates containing 10–100 colony forming units (CFU) were used for counting CFU g⁻¹ dry weight (DW) soil (Nesci et al. 2006; Wang et al. 2010). Three replicates were made for each concentration.

Single fungal colonies were transferred onto potato dextrose agar medium (PDA) for purification and then kept in tube slants of PDA for further taxonomic identification. All isolates were sub-cultured and initially grouped into morphotypes based on morphological characters. Those fungi which could produce fruiting bodies on PDA or sporulation induction media (such as Czapek’s medium, Sabouraud’s medium, Oat medium, water agar, etc.) were first identified to genus or species level based on the morphological characteristics reported in original taxonomic papers and relevant taxonomic keys (Wei 1979; Barnett and Hunter 1987; Bissett 1984, 1991a, b, c; Qi 1997; Zhang 2003; Domsch et al. 2007; Kong 2007; Visagie 2008; Jaklitsch 2009, 2011; Samson et al. 2011; Li et al. 2012; Jurjevic et al. 2012). The remaining non-sporulating isolates were identified based only on ITS sequence comparison.

DNA extraction, PCR amplification and sequencing analysis

Genomic DNA of the representative isolates of 111 morphotypes was extracted from pure cultures as described by Barnes et al. (2001). The fungal ITS region was amplified using the primers ITS1 and ITS4 (Tanaka et al. 2009). The PCR reaction mixtures (50 μl) contained 1 μl of genomic DNA (about 100 ng), 1 μl of each primer (10 mM), 22 μl of sterile deionized water, 25 μl of 2× Taq PCR Mastermix (0.05 U/μl Taq DNA polymerase, recombinant); 4 mM MgCl₂; 0.4 mM dNTPs) (Sangon Biotech, China). The PCR amplification program consisted of 94 °C for 3 min, followed by 35 cycles of 94 °C for 45 s, 58 °C for 45 s and 72 °C for 1 min, with a final extension of 72 °C for 10 min. The PCR products were sequenced by Sangon Biotech on an ABI-PRISM3730 automated sequencer (Applied Biosystems, USA).

The obtained ITS sequences were compared by BLAST on GenBank of NCBI. The sequences of all samples and their closest matches were aligned by ClustalX (ver. 1.7) with some other reference ITS sequences (Wang et al. 2007), but ambiguous regions on both sides were excluded from the analysis. The phylogenetic tree was inferred from the neighbor–joining algorithm by MEGA5 with 1,000 bootstrap replicates (Tanaka et al. 2009; Tamura et al. 2011).

Soil fungal diversity analysis

The density of soil fungi was denoted in units of CFU g⁻¹ DW soil (here expressed in log CFU) (Wang et al. 2010). The following diversity indices were calculated at the species level: (1) species richness (S, the number of different species in a soil sample); (2) the biodiversity index (Shannon–Wiener) ($H\prime =-{\displaystyle \sum_{I=1}^s pi \ln pi}$, where pi is the proportion of total species I in a soil sample); (3) the Pielou’s evenness index (E=H'/Hmax, H'max = lnS); (4) relative abundance, which is the number of species (or genus) × 100/number of total species (or genus).

Measure of the Trichoderma antagonism

To measure the inhibitory effect of dual culture, we tested 14 Trichoderma isolates (JZ-8, JZ-25, JZ-66, JZ-67, JZ-69, JZ-77, JZ-82, JZ-90, JZ-129, JZ-132, JZ-149, JZ-161, JZ-165, JZ-179) for antagonism activity against three pathogens, namely, Bipolaris maydis, Curvularia lunata and Rhizoctonia solani, which were provided by the Department of Plant Pathology, Sichuan Agricultural University, China. The experiment was carried out according to the method described by Patil et al. (2012) with some minor modifications. A 5-day-old mycelial disc (diameter 5 mm) was cut out from the margin of actively growing cultures, including pathogens and Trichoderma spp. The mycelial discs of Trichoderma strain and pathogen were placed opposite to each other at a distance of approximately 4.5 cm on a PDA petri plate. The plate with only pathogen was used as the control. Experiments were repeated at least three times. All plates were incubated at 26 °C, and the mycelial growth of pathogens was measured after 6 days or the growth of control treatment covered the petri plate, whichever came first. The percentage inhibition of mycelial growth was calculated according to the formula:

$$ \mathrm{D}\left(\%\right)=\left[\left(\mathrm{L}1-\mathrm{L}2\right)/\mathrm{L}1\right]\times 100\% $$

where D = the percentage of growth inhibition; L1 = the radial growth measurement of pathogen in the control plate; L2 = the radial growth of the pathogen in the presence of Trichoderma spp.

The method of Galletti et al. (2008) was used to measure the inhibitory effect of the Trichoderma culture filtrates. Briefly, Seven 5-day-old mycelial discs (diameter 5 mm) of Trichoderma were inoculated into 100-ml sterilized potato dextrose broth and cultured for 9 days in the dark at 26 °C with shaking (180 rpm. The liquid cultures were then centrifuged at 3,500 rpm for 10 min and passed through a filter membrane (pore size 0.45 μm); 4 ml sterilized culture filtrate was then added to 16 ml PDA (final concentration of culture filtrate 20 %, v/v) and poured into a 9-cm-diameter petri plate. After the agar had solidified, an approximately 5-day-old mycelial disc (diameter 5 mm) of pathogen was placed on the center of the dish and the dish incubated at 26 °C. PDA plates with only pathogen were used as the control, and there were three replicates for each treatment and control. The mycelial growth of pathogens was measured after 6 days or the growth of control treatment covered the petri plate, whichever came first. The percentage inhibition was calculated using the following formula:

$$ \mathrm{F}\left(\%\right)=\left[\left(\mathrm{L}1-\mathrm{L}2\right)/\mathrm{L}1\right]\times 100\% $$

where F = percentage of growth inhibition; L1 = colony growth on the control plate; L2 = colony growth on the treatment plate.

Results

Abundance of soil fungi

There were obvious differences in fungal density in 25 sampling sites of Jiuzaigou County, ranging from 2.18 log to 4.38 log CFU g⁻¹ DW soil (Table 1; Fig. 1); the mean fungal density of the 25 sampling sites was 3.52 ± 0.57 log CFU g⁻¹ DW soil (Table 2). The highest and lowest values were found in site JZNR8 and site WJNR7, respectively (Table 1). The mean densities of soil fungi in the three nature reserves was ordered as follows: JZNR (3.84 ± 0.45 log CFU g⁻¹ DW soil) > WJNR (3.40 ± 0.70 log CFU g⁻¹ DW soil) > BHNR (3.35 ± 0.43 log CFU g⁻¹ DW soil) (Table 2).

Fig. 1

Density of soil fungi in 25 soil samples taken from sampling sites at different altitudes (m a.sl.). CFU Colony forming units

Table 2 The biodiversity indices and density of soil fungi in the three nature reserves

Identification and phylogenetic analysis

A total of 2,143 fungal isolates were obtained from top soils in three nature reserves. We grouped all isolates into 111 different morphotypes and then selected 111 representative isolates from the 111 morphotypes for morphological and molecular characterization. Based on their morphological characters and ITS sequences, 109 morphotypes were identified as 38 known genera. The other two morphotypes (representative isolates JZ-13, JZ-72) were non-spore forming types, and their closest BLASTN sequences were from unidentified fungi. Phylogenetic analysis of the ITS sequences demonstrated that JZ-13 and JZ-72 were classified into different genera of Ascomycota (Fig. 2). Therefore, a total of 40 genera (73 species) were obtained from the topsoil in Jiuzhaigou County (Tables 3, 4).

Fig. 2

Phylogenetic tree of 111 strains (bold font) with their closest BLASTN matches and other related taxa based on the internal transcribed spacer (ITS) sequence. Numbers on branching points ≥50 % bootstrap values of a bootstrap test of 1,000 runs, number after species names GenBank (ITS) accession number

Table 3 The identification, closest BLASTN matches and relative abundance of isolated fungi

Table 4 Relative abundance of isolated genera

The phylogenetic relationships of the 111 isolates with their closest BLASTN matches and other reference nucleotide sequences were inferred from the neighbor-joining analysis. In Fig. 2, these ITS sequences are classified into three groups (Ascomycota, Basidiomycota, Zygomycota). Except for JZ-70, all experimental sequences had strong affinities to their closest BLASTN sequences, which indicated that the phylogenetic tree coincided with the results based on morphological identification. Although the branch length of JZ-70 with Absidia repens (FJ849793) was >0.05 (Fig. 2), the morphological characteristics revealed that JZ-70 belonged to genus Absidia and that Absidia glauca (AY944880) also had a long-distance (>0.05) relationship with Absidia repens (FJ849793).

A high diversity of ITS sequences occurred in Penicillium and Trichoderma (Hypocrea), and the close phylogenetic relationship between the experimental and reference sequences indicated that these Penicillium and Trichoderma (Hypocrea) isolates were correctly distinguished or identified. However, the ITS variation was low between Penicillium griseofulvum, P. turbatum and P. chrysogenum (Fig. 2), which is in agreement with previously reported results (Skouboe et al. 1999). Isolates JZ-10, JZ-11, JZ-43 and JZ-53 had a strong affinity to described species. However, the four Penicillium isolates were not clearly identified based on phenotype. As we know, the asexual state of all Hypocrea is Trichoderma (Chaverri et al. 2001), and this point was also supported by our molecular data. Surprisingly, JZ-27 and JZ-28 did not form sexual and asexual spores, and they showed no Trichoderma morphological characters; therefore the two isolates were identified using ITS sequences and termed as Hypocrea sp. (Fig. 2).

Based on observable morphological characters (e.g. the color of colonies or other cultural characteristics, conidia shape and their formation), we classified the ten Aspergillus isolates into three different species and the 11 Humicola isolates into three different species. The ITS sequences of the Aspergillus and Humicola isolates identified in our study formed six and four groups in the phylogenetic tree, respectively. The bootstrap values among these groups were relatively low, indicating that the genetic diversity in these Aspergillus or Humicola species is quite low. According to microscopic analysis, the different species of Aspergillus or Humicola were clustered together into a clade by the neighbor-joining phylogenetic analysis (Fig. 2), which is in accordance with the clustering obtained in earlier studies (Zhao et al. 2008; Varga et al. 2010).

Fungal community composition

Forty genera with 73 species were collected from Jiuzaigou County and classified as Ascomycota, Zygomycota and Basidiomycota at a ratio of relative abundance of 219.65:17.44 :1, respectively. Thirty–five genera were classified into Ascomycota with 66 species, three genera into Zygomycota with five species and two genera into Basidiomycota with two species.

The relative abundance of Ascomycota, Zygomycota and Basidiomycota in JZNR was 90.28, 9.19 and 0.57 %, respectively. In BHNR and WJNR, the relative abundance of Ascomycota was 91.98 and 93.88 %, respectively. The relative abundance of Ascomycota colonies in JZNR was 9.82- and 158.39-fold higher than that of Zygomycota and Basidiomycota colonies, respectively. This pattern was broadly similar in BHNR and WJNR, in which the ratios of Ascomycota:Zygomycota:Basidiomycota were 766.50:65.75:1 and 149.02:8.75:1, respectively.

Genera (or species) with a relative abundance of >5.0 % were considered to be dominant (Wang et al. 2010). In JZNR, among the 35 genera (56 species) isolated, the range in relative abundance of species was 0.19 to 6.57 %, and only three species showed values of >5.0 %, whereas ten species showed the lowest values (Table 3). The dominant genera detected were Trichoderma, Penicillium, Humicola, Aspergillus and Mortierella (Table 4). In BHNR, among the 33 genera (54 species) isolated, the relative abundance of species was 0.12–6.17 %, with two Penicillium species having a relative abundance of >5.0 % and six species having the lowest percentages (Table 3). The dominant genera were Penicillium, Humicola, Aspergillus, Trichoderma, and Fusarium (Table 4). In WJNR, 31 genera (50 species) were isolated, and the relative abundance of all species ranged from 0.13 to 4.88 %, with no species having a relative abundance of >5.0 %, and six species having a relative abundance of >4.0 % (Table 3). The main genera detected were Penicillium, Aspergillus, Humicola, Trichoderma, and Myrothecium (Table 4).

For Jiuzhaigou County as a whole (all three nature reserves, 25 soil samples), the most dominant genus was Penicillium, with 11 species accounting for 24.83 % of isolates, followed by Trichoderma, with 11 species accounting for 12.18 %, Humicola with three species accounting for 11.76 % and Aspergillus with three species accounting for 11.48 % (Table 4). The other three main genera were Fusarium (4.62 %), Mortierella (4.48 %) and Myrothecium (3.92 %) (Table 4). The seven genera represented together 73.27 % of total abundance. All seven of the main genera were found in all three nature reserves where they represented the main populations.

Diversity characteristics

For the 25 sampling sites, the range in values for species richness (S), the biodiversity index (H') and evenness (E) were 10–29, 1.96–3.05 and 0.74–0.95, respectively (Tables 1, 2). Among the three sites, site BHNR3 showed a larger S value (29), and sites WJNR5 and BHNR8 had the highest E value (0.95), whereas site BHNR4 showed the higher H'′ value (Table 1).

Among the three nature reserves, the highest S value was found in the BHNR (20.00), followed by JZNR (19.38) and WJNR (18.11). Similarly, the highest H′ value was observed in BHNR (2.66), followed by JZNR (2.54) and WJNR (2.49). The highest evenness (E) value was also found in BHNR (0.90). This means that the culturable fungal species of BHNR were better distributed than those of other regions, with the higher contribution to biodiversity (H′).

Trichoderma antagonism

A total 14 Trichoderma isolates were tested for their antagonism activity against three pathogens (Bipolaris maydis, Curvularia lunata, Rhizoctonia solani). Table 5 shows the inhibition effect of three pathogens for six Trichoderma isolates which gave inhibition values of >20 %; the other eight Trichoderma isolates showed little effect to the three pathogens according to the dual culture or culture filtrate tests.

Table 5 Effect of six Trichoderma species on mycelial growth inhibition of three pathogens according to dual culture and Trichoderma culture filtrates on potato dextrose agar medium

For the dual culture tests, the inhibition values of mycelium growth of the three pathogens by these six Trichoderma isolates were all >22 % (Table 5). These results indicate that the six Trichoderma isolates can clearly inhibit the growth of B. maydis, C. lunata and R. solani. In particular, JZ-77 (T. harzianum) showed the highest inhibition value to R. solani (62.44 %) and C. lunata (56.97 %), and the highest inhibition value of B. maydis was induced by T. koningii (40.13 %).

In terms of the inhibitory effect of the Trichoderma culture filtrates, growth inhibition ranged from 20.28 to 52.14 %, from 27.13 to 70.44 % and from 27.15 to 67.79 % for B. maydis, C. lunata and R. solani, respectively (Table 5). These values demonstrate that the highest inhibition values of C. lunata and R. solani were induced by T. koningii (70.44 and 67.79 %, respectively) and that the highest inhibition value of B. maydis was induced by T. hamatum (52.14 %).

Discussion

In this study, a conventional method (soil dilution plate method) and Rose Bengal agar medium were used to assess the diversity of culturable soil fungi in three nature reserves of Jiuzhaigou County. Among the 25 sampling sites, a sharp change in fungal density appeared in two pairs at two nearby sites (BHNR1 and WJNR7; BHNR5 and WJNR8) (Fig. 1) which showed the same vegetation and little change in altitude (≤20 m) (Table 1; Fig. 1). On the contrary, some sites showed significant differences in vegetation and altitude (e.g. JZNR7 and BHNR7; WJNR5 and BHNR4; BHNR2 and WJNR3), but little change in fungal density (<0.05) (Table 1; Fig. 1). It is therefore difficult to draw clear inferences of the relationship among fungal density in terms of vegetation or altitude. Although we could not rule out the existence of differences of micro-environmental characteristics within each sampling site, previous research revealed that quantitative changes in plant rhizodeposition, geographical environment, pH and carbon availability may affect fungal abundance (Grishkan et al. 2006, 2009; Kara and Asan 2007; Hannula et al. 2010; Jirout et al. 2011).

For culturable fungi, the ranges in the values of species richness (S), biodiversity index (H′) and evenness (E) of the 25 sampling sites were 10–29, 1.96–3.05 and 0.74–0.95, respectively (Table 1). Sites BHNR3 and BHNR4 showed the highest value of S (29) and H′ (3.05), respectively. The highest E value (0.95) was found in BHNR8 and WJNR5. However, because of the isolation technique used, the S, H′ and E of the 25 sampling sites only depend upon their culturable fungal fraction. Thus, to ensure a reliable assessment of soil fungal diversity characteristics, a wider range of research techniques should be employed in further studies.

It has been estimated that there are approximately 3,150 species of soil fungi, many of which have a worldwide distribution (Gams 2007). The culturable fungal community of the 25 sites analyzed in our study was dominated by Ascomycota, accounting for 92.25 % of isolates, with 66 species (35 genera), followed by Zygomycota, accounting for 7.33 %, with five species (3 genera), and Basidiomycota, which were only rarely isolated (0.42 %). These results confirm those reported previously on soil fungal communities (Grishkan et al. 2009; Arenz and Blanchette 2011). Some isolates belong to ubiquitous genera which have been reported earlier to be soil fungi (e.g., Absidia, Aspergillus, Fusarium, Mortierella, Penicillium, Trichoderma and Geomyces). In our study, Penicillium, Trichoderma, Humicola and Aspergillus species comprised a higher proportion of the fungal isolates and were particularly common (representing 24.83, 12.18, 11.76 and 11.48 % of total isolates, respectively); this result is similar to that previously reported and supports the notion that these species can utilize soil organic matter more widely than other species and inhabit various soil environments (Grishkan et al. 2009).

Compared to past research, our data differ with respect to the occurrence of some genera; for example, culturable isolates belonging to Alternaria spp., Botrytis spp., Rhizopus spp., Mucor spp., Phialocephala spp. among others, were not identified in our study. In addition, although the environments among the three nature reserves are similar, we found significant differences among fungal species and relative abundance: for example, Umbelopsis sp., Coprinellus sp., Trematosphaeria sp. and Fungal sp.2 (strain JZ-72) were isolated only from JZNR sites, Podospora sp., Seimatosporium sp. and Fungal sp.1 (strain JZ-13) were not observed at JZNR and BHNR sites and a sharp increase of Myrothecium genus relative abundance (mainly due to M. roridum) was observed at WJNR sites, possibly indirectly providing a significant signal of plant disease caused by this pathogen. Overall, different soil characteristics (e.g. pH, aeration, water content, plant litter quality input, among others) would appear to influence the existence and culturability of fungal species in the different regions.

In our previous publication (Zhou et al. 2013), the isolate SC-12 (JN662930) was identified as a new Trematosphaeria species and named as Trematosphaeria terricola, which is the first Trematosphaeria species from alpine soil in China. As we know, this study represents the first record of the presence of Podospora sp. and Seimatosporium sp. in soil of China. It is interesting that some of the species isolated in our study were obtained from different hosts or habitat and showed saprophytic or parasitic character in comparison to results of earlier studies. This suggests that fungi can shift their ecological role with changes in ecological features. Some isolated fungi are important plant pathogens, such as, Stachybotrys sp., Fusarium sp., Myrothecium sp., Cladosporium sp., Phoma sp., Aspergillus sp., Penicillium sp., Scopulariopsis sp., Cylindrocarpon sp., among others. To the contrary, Aspergillus spp., Penicillium spp., Acremonium spp. and Trichoderma spp. are among the best known fungal agents of bio-control and perhaps play an important role by controlling or preventing soil-borne fungal diseases. In fact, some Trichoderma isolates did show bio-control potential against pathogens in our test. The antagonistic mechanisms used by Trichoderma species to control plant pathogens include competition, antibiosis, mycoparasitism and so on (Sánchez et al. 2007; Patil et al. 2012). In addition, there are reports of other fungi playing an important role in ecological functioning. For example, Mortierella sp. can solubilize soil immobile phosphorus (Osorio and Habte 2001), Humicola sp. is a strongly cellulolytic microfungi (Deacon et al. 2006) and Bjerkandera sp. is regarded as the white rot fungi because of its lignin degradation function (Dorado et al. 2001). Therefore, these species are possible important factors for ecological balance in the three nature reserves.

The ITS sequences of JZ-103, JZ-104, JZ-123 and JZ-169 are most closely related to the BLASTN sequence of Lecythophora sp. (AY219880), Lecythophora luteoviridis (DQ404354) and L. mutabilis (HM036599) (Table 3), while these sequences appeared to be clustered into two clades with a more distant relationship (Fig. 2). These results suggest that Lecythophora species are polyphyletic even though they showed a similar morphology. This phenomenon has been observed in other genera, such as Lulworthia (Campbell et al. 2005), Lignincola (Pang et al. 2003) and Lophiostoma (Zhang et al. 2009). Although the ITS data play an important role in current mycological taxonomy, the phylogeny inferred from any gene may not really reflect the evolution history of organisms. According to Uilenberg et al. (2004), the polyphasic taxonomy should include genotypical and phenotypical characteristics. Therefore, the JZ-103, JZ-104, JZ-123 and JZ-169 were identified as Lecythophora spp. Penicillium has been characterized using molecular techniques in many studies, but the data from these are insufficient too provide a statistical solution for classifying infrageneric Penicillium (Samson et al. 2004). Therefore, morphological features were the main appraisal method for Penicillium species.

Many studies have used molecular techniques to detect fungal diversity by direct DNA extraction from environmental samples and separation of amplicons obtained by different methods, including denaturing gradient gel electrophoresis, terminal restriction fragment length polymorphism and single–strand conformation polymorphism (Viaud et al. 2000; Jumpponen 2003; Anderson and Cairney 2004; Jeewon and Hyde 2007). Although these methods are able to analyze the diversity independently of culturability, their results may still show biases due to the processes of sample DNA extraction or PCR amplification (Kirk et al. 2004).

The number of isolated fungal species is clearly related with the isolation method and culture medium (Cabello and Arambarri 2002). According to Bridge and Spooner (2001), only about 17 % of known fungal species can be cultured using conventional microbiological culture techniques, and most remaining fungi are missed. The conventional cultural techniques only detect a small fraction of soil fungi—those which can grow and sporulate on the isolation medium used (Cabello and Arambarri 2002; Jeewon and Hyde 2007), and the selective culture medium is a major determining factor in the isolation of soil fungal species (Zhang et al. 2013). Thus, it is impossible to accurately detect fungal diversity in soil samples using the single isolation method and medium, and soil fungal diversity might be greatly underestimated when based only on the traditional cultural techniques used in this study. Consequently, there are limitations to the techniques used in this study for the detection of the true diversity in any chosen environment. However, we were able to provide a basic diversity measurement of soil fungi which can readily grow on the culture medium used, as well as obtain the pure cultures for further study. It is clear that a combination of traditional and molecular approaches will provide a comprehensive picture of fungal diversity in sampling sites because many species of “unculturable” fungi have already been detected using molecular techniques (Jeewon and Hyde 2007).