Diversity of root-associated fungi of Vaccinium mandarinorum along a human disturbance gradient in subtropical forests, China

Abstract

Aims

Ericaceous plant species can host diverse fungi in their roots, including ericoid mycorrhizal fungi (ERMF), endophytes, pathogens and some species with unknown functions. However, how this diversity of fungi responds to different human disturbances is not well understood.

Methods

In this study, we examined the effects of different human disturbance on fungal diversity in hair roots of Vaccinium mandarinorum, an ericaceous plant. Fungal DNA was extracted from hair roots of V. mandarinorum and high-throughput sequencing was applied to detect the diversity of root-associated fungi along a human disturbance gradient in subtropical forests in Gutianshan National Nature Reserve (GNNR) in East China. The four forest types with different disturbance regime were: old growth forest (OGF), secondary forest with once cut (SEC I), secondary forest with twice cut (SEC II) and Cunninghamia lanceolata plantation (PLF).

Important Findings

The results showed that: (i) diverse fungal operational units (OTUs) were detected in hair roots of V. mandarinorum in the four types of forests, covering fungal phyla of Ascomycota, Basidiomycota, Chytridiomycota, Glomeromycota and Zygomycota; (ii) Community composition of root-associated fungi of V. mandarinorum in PLF was distinct from those in the other three forest types, and two types of secondary forests had similar fungal community composition; (iii) Different fungal families respond differently to human disturbances: fungal families with significant preference to OGF were ectomycorrhizal or saprophytic fungi while fungal families with higher relative abundance in PLF were plant pathogenic or saprophytic fungi; (iv) The first principal component (PC1) of plant community had a significant effect on composition of root-associated fungal community, while edaphic parameters showed no significant effect on fungal community composition in roots of V. mandarinorum. Our results help to better understand the responses of both ericaceous plants and their fungal partners to human disturbances and forest managements.

INTRODUCTION

Ericaceae is a large plant family, which has ca. 4500 species within 125 genera (Hazard et al. 2014), including trees and shrubs growing all over the world (Kron et al. 2002), especially under harsh environments such as high latitude, high elevation, poor nutrients and acidic soils (Cairney and Meharg 2003; Perotto et al. 2002; Read 1996; Stehn et al. 2011; Verbruggen et al. 2014; Walker et al. 2011). Most species in Ericaceae can form ericoid mycorrhiza (ERM) with soil fungi, helping them to survive in the low nutrient and stressful environments (Smith and Read 1997), except for a few groups that have other kinds of mycorrhiza, such as Arbuscular mycorrhiza (AM) (Obase et al. 2013), Monotropoid mycorrhiza (MM) or Arbutoideae with arbutoid mycorrhizas (ARM) (Dickie et al. 2013).

Ericaceous plants have been considered to form ericoid mycorrhizal associations with only a relatively narrow range of ascomycetous fungi (McLean et al. 1999; Midgley et al. 2004; Smith and Read 1997). However, recent studies reported that a broad range of ascomycetous fungi, such as Rhizoscyphus, Meliniomyces, Oidiodendron, Capronia and Cryptosporiopsis and some basidiomycetous fungi, especially Sebacinales, can also form mycorrhizal associations with ericaceous plants (Allen et al. 2003; Berch et al. 2002; Bougoure and Cairney 2005; Bougoure et al. 2007; Grelet et al. 2009; Kottke et al. 2008; Selosse et al. 2007; Sun et al. 2012; Weiß et al. 2011). Roots of ericaceous plants also host some other functional fungal species, such as dark septate endophyte (DSE) (Currah et al. 1993; Cázares et al. 2005; Tian et al. 2011), fungal pathogens (Grunewaldt-Stöcker et al. 2013) and saprophytes (Piercey et al. 2002; Read 1996; Wurzburger et al. 2012). The co-existence and shifts of diverse fungal species in ericaceous roots are expected to be related to performance of host plants in different habitats and to their resistance to disturbance (Cairney and Meharg 2003; Dickie et al. 2013; Kernaghan and Patriquin 2011; Perotto et al. 2002; van der Heijden and Horton 2009; Vohník and Albrechtová 2011).

Plant communities may directly affect diversity and composition of root-associated fungal communities through their composition diversity, age and host specificity (Carney and Matson 2006; Monika et al. 2014), or indirectly through changing micro-climatic parameters (e.g. soil moisture, temperature and pH) (Carney and Matson 2006). Bougoure et al. (2007) investigated variation of ERMF communities associated with Calluna vulgaris along natural vegetation gradients from heathlands to mature pine forests, and their results suggested that habitats and host specificity determine the diversity and structure of the ERMF community. Ishida and Nordin (2010) also found that dominant tree species and forest types could affect assemblages of fungi associated with two Vaccinium species. Colonization and diversity of mycorrhizal fungi could also be influenced by physical and chemical properties of soils, such as soil texture, soil moisture, pH and nutrients (Hofland–Zijlstra and Berendse 2010; Jantineke 2006; Nielsena et al. 2009; Sadowsky et al. 2012; Toberman et al. 2008).

Human disturbances and forest management activities, such as clear cut and select cut, fertilization or introduction of commercial trees could affect both plant compositions and soil parameters in forests, and therefore affect diversity and structure of soil and root-associated fungal communities (Fichtner et al. 2014; Lentendu et al. 2011; Verbruggen et al. 2014). In a recent study, McGuire et al. (2014) compared soil fungal communities in original forests, secondary forests and plantations in Malaysia and their results showed that fungal compositions in all sites were significantly different, and fungal community structure in regenerative forests were more similar to that in the ancient forests, rather than plantations. As for fungi associated with ericaceous plants, results of some studies also showed that they varied with different land use (Hazard et al. 2014), different habitats (Bougoure et al. 2007; McLean and Lawrie 1996), changes of dominant plants and soil conditions (Urcelay et al. 2003).

There are many ericaceous plants in China, including 826 species of 22 genera (Fang et al. 2005). Vaccinium mandarinorum is a native evergreen shrub or small tree, widely distributing in subtropical forests from 100 to 1600 m a.s.l. (or even 1800–2900 m a.s.l. on Yunnan plateau) in China (Fang et al. 2005). V. mandarinorum is sensitive to forest managements, and the total area at breast height (TABH) and mean diameter at breast height (DBH) per individual of V. mandarinorum are significantly reduced by human disturbances (online supplementary Table 1).

In order to understand the responses of fungal communities associated with ericaceous plants to human disturbances, diversity and community compositions of root-associated fungi of V. mandarinorum in old growth forests, secondary forests and plantation forests were examined in this study, and our hypotheses are: (i) human disturbances may affect root-associated fungal community of V. mandarinorum; (ii) different fungal taxa may response differently to human disturbances; (iii) fungal community may be related to both plant community and edaphic factors along the human disturbance gradient.

MATERIALS AND METHODS

Study sites

The study site is located at Gutianshan National Nature Reserve (GNNR), Zhejiang Province in East China (29°10′19.4″N–29°17′41.4″N, 118°03′49.7″E– 118°11′12.2″E), and the total area of the reserve is about 81 km². Annual mean temperature is 15.3ºC, and annual precipitation ranges from 1793 to 1960mm. Subtropical red soil with granite or deeply weathered granite as parent rock is the dominant soil type (Zhang et al. 2011). Subtropical evergreen broad-leaved forest is the typical vegetation in GNNR (Yu et al. 2001), with Castanopsis eyrei and Schima superba being the dominant tree species.

Four types of forests with different disturbance history in the GNNR were studied in the present work: old growth forests (OGF), secondary forests with once cutting (SEC I), secondary forests with twice cutting (SEC II) and Cunninghamia plantations (PLF). Within each type of forests, three 1-ha (100×100 m) plots were randomly selected. SEC I was clear-cut about 50 years ago, while SEC II was clear-cut about 50 years ago and then selectively cut about 20 years ago. Cunninghamia plantations (PLF) were established about 20 year ago after clear cutting of secondary forests. Stands in both types of secondary forests and plantations have been undergoing natural recovery since their last anthropogenic disturbances. OGF is undisturbed forests that did not experience tree-felling during the last 100 years and is generally located at the core zone of GNNR (Song et al. 2011).

Sampling procedure

Hair roots of four individuals of V. mandarinorum were sampled from each plot. In total, root samples from 48 individuals were collected (4 forest types × 3 plots per forest type × 4 individuals per plot). In detail, the terminal portion of the fine roots (typical ericaceous ‘hair roots’) was retrieved from soils at four directions around the trunk of each V. mandarinorum individual. The roots were washed carefully after 1h soaking in sterile water. Hair roots were then cut into 1cm segments and 20 hair root segments were selected randomly from each V. mandarinorum individual sample. Each root segment was put into a centrifuge tube and kept in 70% alcohol at −70ºC before DNA extraction. A 200g soil sample was taken from the top 10cm of soil adjacent to each plant sampled for elemental analyses.

Soil samples used for elemental analyses were ground to pass through an 80-mesh sieve (opening size: 0.18mm) using a mechanical mill (Retsch MM 400, Retsch GmbH & Co KG, Haan, Germany). Soil organic carbon (SOC) was determined using a H₂SO₄–K₂Cr₂O₇ oxidation method (Zhou et al. 2009). Subsamples were digested in H₂SO₄–H₂O₂ (Bennett et al. 2002) and subjected to total N and total P analysis. Total N content was analyzed with an Alpkem autoanalyzer (Kjektec System 1026 Distilling Unit, Sweden). Total P of the digest was measured colorimetrically at 880nm after reaction with molybdenum blue (Cui et al. 2010).

DNA extraction

DNA was extracted from hair roots of V. mandarinorum following the protocol of DNA secure Plant Kit (TIANGEN Biotech Co. Ltd), with slight modifications. In detail, hair root segments were put into a sterile centrifuge tube containing 20 μl 2× CTAB extraction buffer solution (2% CTAB, 100mM Tris–HCl pH8.0, 20mM EDTA pH 8.0, 1.4M NaCl), and ground with a plastic pestle on ice. A 630μl aliquot of 2× CTAB extraction buffer solution was added into the centrifuge tube. Samples were warmed at 65ºC for 1h, and then shaken for 10min. A 630 μl aliquot of chloroform/isoamyl alcohol (24:1) was added into the tubes. After being fully mixed, the samples were centrifuged at 13 201×g for 8min at room temperature. The supernatant was transferred into a new centrifuge tube and an equal volume of chloroform/isoamylol (24:1) was added. After centrifuging at 13 201×g for 8min again, the supernatant was transferred into a new centrifuge tube and DNA was precipitated with a double volume of 100% alcohol at 4ºC for 1h. After centrifuging at 17 968×g for 8min, the supernatant was discarded. DNA precipitate was washed twice using 70% ethanol, dried in a vacuum desiccator and dissolved in 30ml sterile ddH₂O at 4ºC. DNA samples were stored at −22ºC prior to downstream analyses.

PCR amplifications

The ribosomal internal transcribed spacer 1 (ITS1) region of fungi were amplified by PCR (95ºC for 2min, followed by 25 cycles at 95ºC for 30s, 55ºC for 40s and 72ºC for 50s and a final extension at 72ºC for 5min) using primers 1723F 5′-barcode-CTTGGT CATTTAGAGGAAGTAA-3′ and 2043R 5′-GCTGCGTTCTTCATCGATGC-3′, where barcode is an eight-base sequence unique to each sample. PCR reactions were performed in triplicate 20 μl mixture containing 4 μl of 5× FastPfu Buffer, 2 μl of 2.5mM dNTPs, 0.8 μl of each primer (5 μM), 0.4 μl of FastPfu Polymerase and 10ng of template DNA.

Illumina MiSeq sequencing and processing of sequencing data

Amplicons were extracted from 2% agarose gels and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) following the manufacturer’s instructions and quantified using QuantiFluor™-ST (Promega, USA). Purified amplicons were pooled in equimolar and paired-end sequenced (2×250) on an Illumina MiSeq platform adopting the standard protocols.

Raw fastq files were demultiplexed and quality-filtered using QIIME (ver 1.7) with the following criteria: (i) The reads were truncated at any site receiving an average quality score <20 over a 10-bp sliding window, discarding the truncated reads that were shorter than 50bp. (ii) Exact barcode matching was used, allowing two nucleotide mismatch in primer matching, and reads containing ambiguous characters were removed. (iii) Only sequences that overlapped longer than 10bp were assembled according to their overlap sequence. Reads that could not be assembled were discarded.

Open reference OTU picking was done with pick_open_reference_otus.py using the default uclust method and ITS 12-11 dataset (97% similarity cutoff was used, alpha release, download from web site of QIIME http://qiime.org/home_static/dataFiles.html), and singletons and doubletons were removed during OTU picking. The phylogenetic affiliation of each sequence was analyzed by RDP Classifier (http://rdp.cme.msu.edu/) against ITS 12-11 dataset using confidence threshold of 80%. Diversity analyses were performed by running a workflow on QIIME, with the script core_diversity_analyses.py and specifying a sampling depth of 11 500.

Statistical analysis

One-way ANOVA was carried out to test the difference in soil parameters, individual density and TABH and mean DBH per individual of V. mandarinorum between four forest types using SPSS (ver. 16.0, SPSS Inc.). Indicator species analysis was carried out using otu_category_significance.py in Qiime to determine the indicator fungal species for four forest types. Non-metric multidimensional scaling (NMDS) was performed using the metaMDS function in the R package ‘vegan’ to determine the community structure of root-associated fungi. ‘Envfit’ function was used to identify the main factors influencing root-associated fungal community. Principal components analysis (PCA) was performed using ‘rda’ function in the R package ‘vegan’ and the first three components (PC1, PC2 and PC3) were used as plant parameters in ‘envfit’. Fungal OTUs were divided into four groups according to their occurrence pattern in all four forest types, i.e. group 0 for fungal OTUs present in all four forest types, group 1 for those absent in one forest type, group 2 for those absent in two forest types and group 3 for those present in only one forest type. Abundances of fungal OTUs were compared between four groups using ‘TukeyHSD’ function in R package ‘stats’ and 0.999 was used as confidence level.

RESULTS

Differences of edaphic and vegetation parameters

Edaphic parameters of four forest types in GNNR are shown in Table 1. Both soil organic carbon (SOC) and soil total nitrogen (STN) were significantly higher in OGF than those in disturbed forests. There were no significant differences in soil total phosphorus (STP), ammonia nitrogen (NH₄⁺-N) and soil pH between different forest types. In addition, significant differences in nitrate nitrogen (NO₃⁻-N) and available phosphorus (AP) were found between OGF and SEC II.

Individual density, TABH and mean DBH per individual of V. mandarinorum in four forest types were shown in online supplementary Table S1. Individual density was significantly higher in SEC I than those in SEC II and PLF, and no significant differences were found between individual density in OGF and those in other forest types. TABH in SEC II and PLF were significantly lower than that in OGF, and V. mandarinorum individuals in OGF had a higher DBH than those in disturbed forests.

Diversity of root-associated fungi in different forest types

Rarefaction curves of observed fungal OTUs in roots of V. mandarinorum were shown in Fig. 1. The observed number of fungal OTUs was the highest in OGF and lowest in PLF, but the differences among the four forest types were not significant.

Total OTUs in four forest types were 1216, of which only 97 OTUs were shared by all forest types (Fig. 2). OTUs that occurred specifically in only OGF, SECI, SECII, and PLF were 208, 120, 183 and 122, respectively. The total OTUs number in OGF, SECI, SECII and PLF were 642, 420, 522 and 557, respectively, suggesting that the diversity of OTUs in OGF was richer. The number of shared OTUs by OGF with SECI, SECII and PLF were 284, 276 and 197, indicating that root-associated fungal community in OGF was more similar to those in secondary forests, rather than that of plantations.

Community composition of root-associated fungi

Community compositions of root-associated fungi are shown in Fig. 3. Fungal community compositions were very similar between SECI and SECII. Fungal community composition of PLF was different from the other three forest types. Two dominant fungal phyla in all forest types were Ascomycota and Basidiomycota. Chytridiomycota, Glomeromycota and Zygomycota together accounted for less than 4% of the total number of OTUs in each type of forests. The common classes in all forest types included Agaricomycetes, Leotiomycetes, Eurotiomycetes, Dothideomycetes and Sordariomycetes. The results of NMDS supported the above results that SECI and SECII were more similar in fungal community composition, and PLF was distinct from the other three forest types (Fig. 4).

Preferences of identified fungal families are shown in Fig. 5. Totally 68 fungal families were observed in all 12 plots, and relative abundance of 58 fungal families were compared between different forest types. Relative abundances of identified fungal families were similar between OGF and secondary forests (SEC), and only RA of Mycenaceae and Atheliaceae were significantly higher in OGF than those in SEC (Fig. 5A). Comparing OGF and PLF, we found that relative abundances of Atheliaceae, Herpotrichiellace and Russulaceae were significantly higher in OGF, while relative abundances of Vibrisseaceae, Trichocomaceae, Nectriaceae and Amphisphaeriaceae were significantly higher in PLF (Fig. 5B). Relative abundances of Vibrisseaceae, Trichocomaceae, Nectriaceae, Amphisphaeriaceae and Pleosporaceae were higher in PLF than those in SEC (Fig. 5C). Of the fungal families with preferences to OGF, Russulaceae are typical ectomycorrhizal taxa, Mycenaceae and Herpotrichiellace are saprobic, and Atheliaceae contains both saprobic and ectomycorrhizal genera (Tedersoo et al. 2014). Of the fungal families with preferences to PLF, Pleosporaceae is plant pathogenic; Vibrisseaceae and Trichocomaceae are saprobic; Nectriaceae and Amphisphaeriaceae contain both plant pathogen and saprobic fungi (Tedersoo et al. 2014).

There were a total of 19 indicator fungal species found in all forest types: six, two, six and five for OGF, SECI, SECII and PLF, respectively (Table 2). Of the indicator species, Helotiales 1, Helotiales 2, Helotiales 3, Sebacinales 1, Sebacinales 2, Herpotrichiellaceae 1, Herpotrichiellaceae 2, Hypocrea sp1, Sordariales 1, Chaetothyriales 1 and Chaetothyriales 2 are likely putative ERM fungi. It is interesting that all indicator species for OGF were putative ERM fungi and only two or three indicator species were putative ERM fungi for SEC II or PLF.

Determinant factors of root-associated fungal community

Of all plant community parameters, only PC1 of plant community had significant effects on community composition of root-associated fungi (Table 3). Plant species that had significant contributions to PC1 of plant community were Cunninghamia lanceolata, Castanopsis eyrei, C. carlesii, Rhododendron ovatum, R. latoucheae, V. carlesii, Cyclobalanopsis glauca, Lithocarpus glaber, Chimonanthus salicifolius. Of these plant species, Cunninghamia lanceolata is dominant in PLF, and Castanopsis eyrei, C. carlesii, Cyclobalanopsis glauca and Lithocarpus glaber are dominant plants in OGF. Rhododendron ovatum, R. latoucheae and V. carlesii are dominant ericaceous plants in the study site.

All edaphic parameters had no significant effects on root-associated fungal community of V. mandarinorum (Table 3).

DISCUSSION

Diverse fungal species occurred in hair roots of V. mandarinorum

Diverse fungal taxa were observed in hair roots of V. mandarinorum, which may be owing to the progress of high-throughput sequencing. We observed 1216 fungal OTUs in our study sites, while in researches using isolation and culture or local database based ITS-RFLP (Internal Transcribed Spacer—Restriction Fragment Length Polymorphism), Zhang et al. (2009), Tian et al. (2011), Sun et al. (2012) found only 17, 12 and 35 fungal taxa in similar forests of China, respectively. Compared with traditional approaches, application of high-throughput sequencing makes it possible to detect organisms with very low-abundance (Begerow et al. 2010; Ekblom and Galindo 2011; Lindahl et al. 2013). Diverse fungal taxa were also observed when studying other mycorrhizal fungal community using high-throughput sequencing. Buscardo et al. (2015) found that the number of ectomycorrhizal fungal taxa was 8-fold higher by using 454 pyrosequencing than by Denaturing Gradient Gel Electrophoresis (DGGE), and Oja et al. (2015) obtained 5805 OTUs in orchid mycorrhizal symbionts by 454 pyrosequencing.

The dominant fungal phyla were Ascomycota and Basidiomycota. Of Ascomycota, members belonging to Leotiomycetes, Dothideomycetes, Eurotiomycetes and Sordariomycetes were found frequently. Helotiales is considered as a typical fungal order of ERM fungi, which has been found in hair roots of not only species from Vaccinium (Berch et al. 2002; Gorman and Starrett 2003; Gorzelak et al. 2012; Grelet et al. 2009; Ishida and Nordin 2010; Vralstad et al. 2002; Walker et al. 2011), but also from other genera of ericaceous plants, such as Rhododendron (Piercey et al. 2002; Selosse et al. 2007; Sun et al. 2012; Wurzburger et al. 2012; Zhang et al. 2009), Calluna (Berch et al. 2002; Hofland-Zijlstra and Berendse 2009), Gaultheria (Berch et al. 2002; Bougoure et al. 2007; Xiao 1994), Erica (Bergero and Girlanda 2002; Hofland-Zijlstra and Berendse 2009), Woollsia (Chambers et al. 2008; Midgley et al. 2004), Empetrum (Hofland-Zijlstra and Berendse 2009; Vralstad et al. 2002) and Epacris (Bougoure and Cairney 2005; McLean et al. 1998; Perotto et al. 2002). Rhizoscyphus ericae (syn. Hymenoscyphus ericae) complex, a putative ERMF of Helotiales, can form ERM with a variety of species in many continents (Ishida and Nordin 2010). Rhizoscyphus ericae was found in only several samples, showing that this species was not dominant in ERM fungal communities. Fungal species from Eurotiomycetes have been observed in hair roots of ericaceous plants all over the world (Allen et al. 2003; Bergero and Girlanda 2002; Bougoure et al. 2007; Walker et al. 2011; Zhang et al. 2009). Oidiodendron spp. of Eurotiomycetes is found in our plots which frequently found to colonize on Ericaceae plants and is considered as common ERM endophytes in subtropical forests of China (Sun et al. 2012; Zhang et al. 2009). Some ascomycetous genera that have been identified as ERMF, such as Bionectria, Neonectria, Chalara and Plectosphaerella (Tian et al. 2011), as well as Acremonium (Midgley et al. 2004) of Sordariomycetes, Cryptosporiopsis (Walker et al. 2011; Zhang et al. 2009), Lachnum (Walker et al. 2011), Meliniomyces (Grelet et al. 2009; Walker et al. 2011) and Scytalidium (Xiao 1994) of Leotiomycetes was also observed in this investigation.

For the fungal phylum of Basidiomycota, the typical order, Sebacinales, has been found in roots of many ericaceous plants (Allen et al. 2003; Ishida and Nordin 2010; Wurzburger et al. 2012). Sebacinales was proven to be ERM endophytes much later than Ascomycota fungi because of the difficulty in culturing them (Tian et al. 2011; Verbruggen et al. 2014). In subtropical forests of China, fungal species of Sebacinales have also been observed (Sun et al. 2012; Zhang et al. 2009). Species for Agaricales were found in our study, which were also reported as ERM endophytes in a number of ericaceous shrubs, e.g., V. angustifolium (Straker 1996), Gaultheria shallon (Allen et al. 2003), Rhododendron macrophyllum (Selosse et al. 2007) and R. maximum (Wurzburger et al. 2012), though many species in this order are considered as ECMF as well (Wurzburger et al. 2012).

Phialocephala fortinii was also found in our study, and it could be identified as a typical DSE when it was inoculated with Ericaceae plants (Jantineke 2006). It has been noticed that ERM and DSE often coexist in hair roots of Ericaceae plants (Cázares et al. 2005; Sadowsky et al. 2012; Vohník and Albrechtová 2011). DSE could help host plants in absorbing P and N (Vohník et al. 2003), but the functions of DSE is still in debate.

Except ERMF and DSE, other mycorrhizal fungi were detected in our study. Fungal OTUs from Glomeraceae are often considered as AM fungi (Vohník and Albrechtová 2011), and fungal species of Boletales, Russulales, Thelephorales are usually considered as ECM fungi (Kohout et al. 2011; Wurzburger et al. 2012). ECM fungal taxa were also observed in roots of some other ericaceous plants (Allen et al. 2003; Smith et al. 1995; Wurzburger et al. 2012), indicating that they may form symbiotic structures with both ERM and ECM host plants. Researches of the same study site have shown that root-trait variations and mycorrhizal fungal diversity may drive consistence and productivity of plant community (See Bu et al. 2017; Shi et al. 2017). For the four forest types in our study, the dominant tree species of canopy layer are usually ECMF hosts (e.g. Fagaceae) in secondary and old growth forests, and Cunninghamia lanceolata in the plantations is host of AM fungi. As ericaceous plants are dominant in the shrub layer of subtropical forests, sharing mycorrhizal species and possible mycorrhizal connections with dominant species of canopy layer may be essential for species coexistence and ecosystem functioning in subtropical forests.

Root-associated fungal community changes in response to different human disturbances

The comparisons of identified fungal families between OGF, SEC and PLF showed that different fungal taxa may have specific responses to human disturbances (Fig. 5). Relative abundances of most identified fungal families of SEC were similar to those in OGF, showing that these fungal families in roots of V. mandarinorum were relatively resistant to logging or could be restored during succession after logging. There were more fungal families showing forest-type preferences when comparing PLF and OGF (Fig. 5C). Fungal families preferring to OGF are usually ectomycorrhizal or saprophytic fungi, while fungal families preferring to PLF are usually saprophytic fungi or plant pathogens. The preferences of ectomycorrhizal fungal families to OGF are reasonable since the dominant tree species are ectomycorrhizal in OGF and it has been documented that ericaceous plants could host ectomycorrhizal fungi in forests (Allen et al. 2003; Wurzburger et al. 2012). It is interesting that three fungal families with many plant pathogens have higher relative abundances in PLF than those in OGF, which might be a reason for the inhibited growth of V. mandarinorum in plantations (supplementary Table 1).

Our results showed that root-associated fungal communities of V. mandarinorum in two secondary forests were more similar to those in old growth forest rather than plantation (Figs 3–5), indicating that introduction and growth of Cunninghamia lanceolata significantly changed the fungal community in roots of V. mandarinorum. In a study on fungal communities in Malaysia, McGuire et al. (2014) also found fungal community in secondary forests were more similar to that in the original forests rather than plantations. Some researches on ectomycorrhizal or endophytic fungal communities have also shown that fungal communities in ecosystems at different succession stages may be distinct (Osono and Trofymow 2012; Tejesvi et al. 2010). Moreover, with the increase of successional time since disturbances, microbial communities tend to become more similar to those in undisturbed ecosystems (Fichtner et al. 2014; Jangid et al. 2010, 2011).

As for ERM fungi, Hutton et al. (1997) observed that ericoid mycorrhizal colonization in disturbed sites returned to levels comparable to undisturbed sites after 12 years regeneration from disturbance. Wurzburger et al. (2012) also reported that ericoid mycorrhizal fungal communities might be more diverse in late successional stage when a mature organic horizon was developed. Agreeing with their results, we found the total number of OTUs in OGF is higher than that in the disturbed forests, with more specific species, and putative ERM fungus as indicator species, suggesting that the fungi associated with V. mandarinorum in undisturbed plots are abundant and their communities develop better than those in disturbed plots. In this study, the similarity we found in root-associated fungi community between the two secondary forests and the old growth forest is possible though clear cut and selected cut did actually change the organic horizon based on the soil parameters compared with undisturbed forests, because some fungal species can persist in soil even the hosts are absent (Bergero and Girlanda 2002).

Plant community and edaphic parameters in determining root fungi

Human disturbances and forest managements usually cause alteration in both plant community compositions and edaphic parameters, which may in turn affect fungal community in plant roots. Changes of plant diversity may have direct or indirect effects on microbe related ecosystem functions such as fine-root productivity and wood decomposition (see Eichenberg et al. 2017; Sun et al. 2017). Our results indicated that plant community, rather than edaphic factors, is the main determinant for root-associated fungal community of V. mandarinorum (Table 3; Fig. 4). The plant species that had significant effects on root-associated fungal community of V. mandarinorum included a commercial species in PLF, i.e. Cunninghamia lanceolata, some dominant ECM plant species in subtropical forest, i.e. Castanopsis eyrei, C. carlesii, Cyclobalanopsis glauca and Lithocarpus glaber, as well as some common ERM plant species, i.e. R. ovatum, R. latoucheae and V. carlesii. These results indicated that both dominant trees and common ERM neighbors may affect ERM fungal community of V. mandarinorum.

Our results suggested that edaphic parameters are not key factors in determining ERM fungal community of V. mandarinorum, which does not agree with some earlier research on nitrogen addition and pH gradient. Those studies found that N added into soil would affect Ericaceae plants and its mycorrhizal aggregation (Hofland–Zijlstra and Berendse 2010; Jantineke 2006). And soil pH has been considered an important factor in determining fungal community structure (Fujimura and Egger 2012). The insignificant effects of edaphic factors may be due to that the disturbances in our forests did not result in significant changes in edaphic factors.

CONCLUSION

A total of 1216 fungal OTUs associated with hair roots were found from V. mandarinorum in four different human disturbance forests (old growth forest, secondary forests with once or twice disturbed and plantations) in subtropical forest of China. The dominant phyla were Ascomycota and Basidiomycota, and the common classes were Leotiomycetes, Dothideomycetes, Eurotiomycetes, Sordariomycetes and Agaricomycetes. Community composition of root-associated fungi of V. mandarinorum in PLF was distinct from those in the other three forest types. Two types of secondary forests had more similar fungal community composition, displaying more similarity to that in old growth forests rather than plantations. Different fungal families respond differently to human disturbances: fungal families with significant preference to OGF are ectomycorrhizal or saprophytic, while fungal families with higher relative abundance in PLF are plant pathogen or saprophytic fungi. PC1 of plant community had significant effects on composition of root-associated fungal community, and dominant plants in PLF and OGF as well as dominant ericaceous plants had significant contributions to PC1 of plant community.

SUPPLEMENTARY MATERIAL

Supplementary material is available at Journal of Plant Ecology online.

FUNDING

This study was supported by the National Natural Science Foundation of China (31170469, 31170495 and 31470565) and Technology division of Shaoxing (2013B70040). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

ACKNOWLEDGEMENTS

We are deeply thankful to Yefei Jin and Li Han who contributed to the experiments and samples processing during this project.

Conflict of interest statement. None declared.

REFERENCES

Author notes