Correspondence: Susan M.Chambers, University of Western Sydney, Parramatta North Campus, Locked Bag 1797, Penrith Sth DC, NSW 1797, Australia. Tel.: +61 2 9685 9919; fax: +61 2 9685 9915; e-mail: firstname.lastname@example.org
Fungi were isolated from the roots of 17 plant species from the families Apiaceae, Cunoniaceae, Cyperaceae, Droseraceae, Fabaceae-Mimosoideae, Lomandraceae, Myrtaceae, Pittosporaceae, Proteaceae and Stylidiaceae at a sclerophyll forest site in New South Wales, Australia. Internal transcribed spacer (ITS) restriction fragment length polymorphism (RFLP) and sequence comparisons indicated that the isolated fungi had affinities to a range of ascomycetes, basidiomycetes and zygomycetes. Four RFLP types had closest affinities to previously identified Helotiales ericoid mycorrhizal (ERM) or Oidiodendron spp. Isolates representing six RFLP types, which were variously isolated from all 17 plant species, formed ERM coils in hair root epidermal cells of Woollsia pungens (Ericaceae) under gnotobiotic conditions. Three of these isolates formed intercellular hyphae, intracellular hyphae and/or microsclerotia, which are typical of dark septate endophyte infection, in roots of Stylidium productum (Stylidiaceae), indicating an ability to form different types of association with roots of different hosts. Overall the data indicate that a broad range of plant taxa may act as repositories for ERM fungi in sclerophyll forest soil.
Despite previously being regarded as associating only with Ericaceae roots, there is increasing evidence that ERM ascomycetes may associate with roots of other plant taxa. Helotiales ascomycetes isolated from roots of the grass Deschampsia flexuosa have, for example, been shown to colonize roots and enhance nitrogen uptake by Calluna vulgaris (Ericaceae) seedlings and vice versa (Zijlstra et al., 2005). Although the nature of the infection was not reported in this case, a Helotiales ascomycete isolated from Quercus ilex roots was shown to form typical ERM coils in hair roots of Erica arborea (Ericaceae) (Bergero et al., 2000). Such observations suggest that roots of non-Ericaceae plant taxa may serve as repositories for Helotiales ERM fungi and that the fungi may form a beneficial association with non-Ericaceae hosts. Moreover, the fact that identical genotypes of a Helotiales ascomycete were isolated from roots of Q. ilex and E. arborea at a woodland site raises the possibility that these fungi may form hyphal links between the two plant taxa. This is supported by the observation that an R. ericae complex isolate simultaneously formed ERM with Vaccinium myrtillus (Ericaceae) and an ectomycorrhizal association with Pinus sylvestris under gnotobiotic conditions (Villarreal-Ruiz et al., 2004).
Several Ericaceae taxa, including Woollsia pungens and Epacris pulchella, occur as common understorey shrubs in some south-eastern Australian mixed sclerophyll forests (Robinson, 1991). Previous studies have revealed that these taxa form ERM associations with a range of typical ERM fungi which includes Oidiodendron sp. and unnamed Helotiales ascomycetes (Chambers et al., 2000; Midgley et al., 2004; Bougoure & Cairney, 2005a). In order to test the hypothesis that roots of non-Ericaceae plants act as repositories for ERM fungi, we have investigated fungal endophyte communities associated with roots of a broad taxonomic range of plant taxa at a sclerophyll forest site where W. pungens and E. pulchella form an important part of the plant community.
Materials and methods
Plant collection and fungal root endophyte isolation
Juvenile specimens of 17 plant species (including sedges, herbs, small shrubs, large shrubs, small trees and large trees) were collected from a sclerophyll forest at Lovers Jump Creek Reserve, Turramurra, NSW, Australia (33°44′S 151°06′E), between May 2005 and February 2006. Plants comprised species from the families Apiaceae (Xanthosia pilosa, Xanthosia tridentata), Cunoniaceae (Bauera rubioides, Callicoma serratifolia, Ceratopetalum gummiferum), Cyperaceae (Schoenus melanostachys), Droseraceae (Drosera spatulata), Fabaceae-Mimosoideae (Acacia terminalis), Lomandraceae (Lomandra longifolia), Myrtaceae (Leptospermum polygalifolium, Eucalyptus sp., Angophora costata), Pittosporaceae (Pittosporum undulatum), Proteaceae (Banksia spinulosa, Hakea sericea, Grevillea linearifolia) and Stylidiaceae (Stylidium productum). Plants, with adhering soil, were transported to the laboratory where soil was carefully removed by first soaking (if strongly adhered to roots) and then washing under gently running tap water for c. 15 min. Roots were excised from stems and then surface sterilized in 20% bleach (4.5% available chlorine) with 100 μL L−1 of the surfactant polyoxyethylene (20) sorbitan monolaurate (Tween 20, Sigma) for 30 s. Roots were then soaked in a solution containing 70% ethanol with 100 μL L−1 Tween 20 for 30 s. This was followed by 3 × 1 min rinses in sterile distilled water, before sterile roots were cut into 0.5 cm pieces and plated on 2% potato dextrose agar (PDA, Oxoid) containing streptomycin (15 mg L−1, Sigma), gentamycin (15 mg L−1, Sigma) and tetracycline (12 mg L−1, Sigma) (Pearson & Read, 1973) in 90-mm Petri dishes. Petri dishes were incubated at 20 °C in the dark and slow-growing fungi were subcultured on PDA. Cultures were maintained in the dark at 21 °C and were subcultured every 16 weeks. Cultures were grouped on the basis of morphology and a representative from each culture morphotype from each plant was selected for DNA extraction.
DNA extraction, ITS-restriction fragment length polymorphism (RFLP) and sequence analysis
DNA was extracted from cultures using the FastDNA®Kit (Bio 101) following the manufacturer's instructions. The ITS region of DNA obtained from fungal endophyte cultures was amplified using 25 pmol of each of the primers ITS1 and ITS4 (White et al., 1990) in 50-μL reaction volumes following the protocol of Anderson et al. (1998). Cycling parameters were 30 cycles of 94 °C for one min, 50 °C for 1 min and 72 °C for 1 min followed by a final 10 min extension. Amplification products were electrophoresed on 2% agarose gels, stained with ethidium bromide and visualized under UV light.
Approximately 1.0 μg of each PCR product was digested with the restriction enzymes HinfI, HaeIII or RsaI (Promega) at 37 °C for 3 h. Digestion products were electrophoresed on 3% agarose gels and visualized as above. Isolates were grouped based on their restriction digest patterns for all three enzymes. Representative isolates from each RFLP group were then chosen for sequencing. Direct sequencing of PCR products was conducted using the primer ITS1 or ITS4. PCR products were purified using the Wizard SV Gel and PCR Clean-up System (Promega) and sequenced using an ABI3730XL DNA analyzer using single primer extension. Sequencing was conducted under BigDye™ terminator cycling conditions. Sequences were analysed using the blastn program (Altschul et al., 1997) on the NCBI site (http://www.ncbi.nlm.nih.gov/BLAST/). Putative taxonomic affinities were assigned conservatively to RFLP types based on blastn expected values and identities of the closest several sequence matches obtained from the blastn searches as described by Bougoure & Cairney (2005a), as well as the blast distance tree of results.
Investigation of fungus : plant interactions under gnotobiotic conditions
Interactions between selected fungi and roots of W. pungens (Ericaceae) and S. productum (Stylidiaceae) were investigated under gnotobiotic conditions. Woollsia pungens was selected as a model Ericaceae host as it occurs commonly at the Lovers Jump Creek Reserve site, can be propagated under sterile conditions (D.S. Bougoure and J.W.G. Cairney, unpublished observations) and the fungi associated with its roots have been studied previously (Chambers et al., 2000; Midgley et al., 2002, 2004). Stylidium productum was selected as a non-Ericaceae plant with which to investigate infection by putative ERM fungi as it is a common herb at the Lovers Jump Creek Reserve site and can readily be propagated under sterile conditions.
Seeds of W. pungens and S. productum were collected from the Lovers Jump Creek Reserve site, surface sterilized for 1 min in a 20% bleach solution (4% available chlorine) containing Tween 20® and rinsed 3 × in sterile distilled water, before being soaked overnight in sterile distilled water. Seeds were then plated out in 9.0-cm-diameter Petri dishes on sterile, moist filter paper. Seeds were incubated at 22 °C for 16 h day : 8 h night period, and W. pungens seed started to germinate after c. 6 weeks, while S. productum seed germinated after 2 weeks. Seedlings were transferred to 70-mL polycarbonate vials containing c. 20 g peat/sand mixture (80% dry/wet sand, 20% dry/wet peat) and kept moist with sterile distilled water. Seedlings were incubated as above.
In preparation for seedling inoculation, plugs of fungus taken from the actively growing edges of cultures on 2% PDA were placed in liquid modified Melin Norkrans medium (MMN) (Marx & Bryan, 1975) in 9.0-cm-diameter Petri dishes and incubated in the dark at 22 °C for c. 4 weeks. Fungal mycelium was then rinsed in sterile MilliQ water and placed in screw-topped tubes containing Lysing matrix C (Bio 101), before being placed in a bead beater and shaken for 5 s at 4 m s−1. Macerated fungal mycelium (0.5 mL) was then inoculated into the root systems of 2–5-cm-tall W. pungens or S. productum plants in sterile peat/sand. Woollsia pungens seedlings were inoculated with a representative isolate of each RFLP type and S. productum seedlings with a representative isolate RFLP type that had closest sequence matches to fungi previously shown to form ERM associations with Ericaceae hosts (RFLP types 1, 3, 7, 14 and 22), along with RFLP types 11 and 23 that formed ERM coils with W. pungens roots in the gnotobiotic infection experiment.
Inoculated plants were incubated as above for 3 months before harvesting. Soil was washed from roots that were then cleared and stained following a method modified from McLean & Lawrie (1996). Roots were cleared in 10% KOH for 30 min at 90 °C, stained in 1.0% Trypan Blue (Sigma) at 90 °C for 90 min and then destained in a 1 : 1 : 1 mixture of lactic acid, glycerol and deionized water at room temperature for 25 min (Hutton et al., 1994). Roots were mounted in glycerol on slides and infection was assessed using a Leica DMRBE light microscope with spot advanced 3.4 software (Diagnostic Instruments Inc.).
Following grouping of morphologically identical isolates, a total of 197 culture morphotypes were obtained from roots of the 17 plant species, with 34 distinct RFLP types identified after ITS-RFLP and sequence analysis (Table 1). Most RFLP types (15) were identified as present from roots of only 1 plant species, whereas 12 were identified from roots of two or three plant species. The remaining eight RFLP types were obtained from the roots of six or more plant taxa, with RFLP type 1 being isolated from 15 of the 17 plant species studied (Table 1). Three to 12 RFLP types were obtained from the roots of each plant species, with the largest number being identified from roots of H. sericea (11 RFLP types) and Bauera rubioides (nine RFLP types). Fewest RFLP types (three) were obtained from roots of D. spatulata and G. linearifolia (Table 1).
Table 1. Details of the plant species from which the 34 fungal RFLP types were isolated
B. rubioides, C. serratifolia, C. gummiferum, L. longifolia, X. pilosa, X. tridentata, L. polygalifolium, Eucalyptus sp., A. costata. B. spinulosa, H. sericea, G. linearifolia, S. melanostachys, P. undulatum, S. productum
B. rubioides, D. spatulata, H. sericea
C. serratifolia, L. longifolia, X. pilosa, X. tridentata, S. productum, A. terminalis
L. longifolia, X. pilosa, X. tridentata, L.polygalifolium, H. sericea, G. linearifolia, S. productum, A. terminalis
C. serratifolia, C. gummiferum, L. longifolia, X. tridentata, Eucalyptus sp., B. spinulosa, H. sericea, G. linearifolia, S. melanostachys, P. undulatum, S. productum, A. terminalis
L. longifolia, S. melanostachys, A. terminalis
C. serratifolia, C. gummiferum, X. pilosa, X. tridentata, L. polygalifolium, Eucalyptus sp., H. sericea, S. melanostachys, S. productum
C. gummiferum, X. pilosa
B. rubioides, Eucalyptus sp., A. costata
B. rubioides, S. melanostachys
B. rubioides, C. serratifolia, Eucalyptus sp., A. costata. B. spinulosa, S. melanostachys, P. undulatum
B. rubioides, C. serratifolia, D. spatulata, L. longifolia, X. tridentata, L. polygalifolium, H. sericea, S. melanostachys, P. undulatum, S. productum, A. terminalis
D. spatulata, X. tridentata
X. tridentata, H. sericea
A. costata, B. spinulosa
C. serratifolia, B. spinulosa
Eucalyptus sp., H. sericea
A. costata, S. productum
A. costata, H. sericea
Sequence comparisons indicated that the 34 RFLP types had closest ITS sequence affinity to a broad range of ascomycetes, basidiomycetes and zygomycetes (Table 2). Four RFLP types (1, 7, 11 and 22) had closest affinities to previously identified Helotiales ERM endophytes, while RFLP type 3 was most similar to Oidiodendron spp. and RFLP type 12 to Cryptosporiopsis spp. Several RFLP types had closest ITS sequence identities to a variety of ascomycetes that have been isolated from roots of Australian Ericaceae, but for which the nature of the association with the plant host remains unclear (Table 2). Five RFLP types had closest affinity, albeit with low affinity, to basidiomycetes including ectomycorrhizal (Cortinarius sp.), saprotrophic (Panellus sp.) or pathogenic (Inonotus sp.) taxa (Table 2).
Table 2. Putative taxonomic affinities of sequence types from root systems of 17 plant species inferred from blastn searches of ITS sequences in the GenBank/EMBL/DDBJ databases
Isolate closest match
Putative taxonomic affinity
Indicates isolates that formed typical ERM structures in Woollsia pungens hair roots under gnotobiotic conditions in the present study.
Closest matches in bold indicate isolates that, in previous studies, produced ERM under gnotobiotic conditions.
While isolates of most RFLP types did not form ERM structures in roots of W. pungens, isolates of six RFLP types (3, 4, 7, 12, 14 and 23) formed typical ERM coils in hair root epidermal cells during three months in gnotobiotic culture with W. pungens seedlings (Table 3). In the case of RFLP type 14, the majority of hair root epidermal cells were infected by ERM coils (Fig. 1) whereas fewer epidermal cells were infected by RFLP types 3, 4, 7 and 23. RFLP type 12 formed coils only in occasional epidermal cells (Fig. 1). No infection of W. pungens hair roots was observed for RFLP types 11 and 22, whereas infection by RFLP type 1 was limited to inter- and intracellular hyphae, with no evidence of ERM coil formation (Table 3).
Table 3. Interactions between isolates of seven RFLP types and Woollsia pungens or Stylidium productum seedlings assessed after 3 months in gnotobiotic culture
Extensive inter- and intracellular hyphae
Extensive inter- and intracellular hyphae
Hyphae on root surface. Inter-and intracellular hyphae. Extensive ERM coils
Hyphae on root surface. Intercellular hyphae
Hyphae on root surface. Inter- and intracellular hyphae. ERM coils
Extensive intra- and intercellular hyphae. Extensive ERM coils
Hyphae on root surface. Extensive inter- and intracellular hyphae. Microsclerotia
Hyphae on root surface. Inter- and intracellular hyphae. ERM coils
Hyphae on root surface. Extensive intracellular and intercellular infection. Extensive ERM coils
Hyphae on root surface. Inter- and intracellular hyphae. Microsclerotia
Hyphae on root surface. Inter- and intracellular hyphae. ERM coils
No infection observed
No ERM-like structures were formed in S. productum roots by isolates of the seven RFLP type tested. Hyphae of RFLP types 3, 7 and 14 were observed on root surfaces, whereas production of intercellular hyphae, intracellular hyphae and/or microsclerotia was observed for RFLP types 1, 3, 7 and 14 (Table 3, Fig. 1). No hyphae of RFLP type 23 were observed within S. productum roots or on root surfaces (Table 3).
It has previously been demonstrated that ERM fungi occur in roots of ectomycorrhizal trees, with some isolates being capable of forming both ERM and ectomycorrhizal associations with different hosts (Bergero et al., 2000; Vrålstad et al., 2002b; Villarreal-Ruiz et al., 2004). The data presented here indicate that ERM fungi are common root associates of a diverse array of plant taxa across a range of life forms within a plant community. Thus, fungi capable of forming ERM-like structures in epidermal cells of W. pungens were isolated from roots of all 17 plant species examined. These comprised an understorey sedge (Schoenus melanostachys), a rush (L. longifolia), herbs (Stylidium productum and D. spatulata), small shrubs (B. rubioides, X. pilosa, X. tridentata), large shrubs (H. sericea, G. linearifolia, L. polygalifolium, Banksia spinulosa), small trees (A. terminalis, P. undulatum, Ceratopetalum gummiferum and Callicoma serratifolia) and canopy trees (A. costata and Eucalyptus sp.).
Fungi from a broad taxonomic range were isolated from the 17 plant species, with six RFLP types shown to form typical ERM coils in epidermal cells of W. pungens hair roots, suggesting that they probably represent ERM fungi. Of these, three RFLP types (1, 7 and 14) were identified as Helotiales ascomycetes, with strong ITS sequence similarity to fungi that have previously been isolated from the roots of Australian Ericaceae (Midgley et al., 2004; Bougoure & Cairney, 2005a). This observation is consistent with previous reports that a range of unnamed Helotiales ascomycetes are common ERM endophytes of Ericaceae worldwide (e.g. Berch et al., 2002; Cairney & Ashford, 2002; Bougoure & Cairney, 2005a, b; Bougoure et al., 2007). On the basis of ITS sequence similarity, RFLP type 3 was identified as an Oidiodendron species. Although Oidiodendron spp. appear to exist as saprotrophs, there are numerous reports that they form ERM associations with Ericaceae hosts (see Rice & Currah, 2006). Oidiodendron sp. has also previously been isolated from hair roots of Australian Ericaceae, including W. pungens at the Lovers Jump Creek field site (Chambers et al., 2000; Bougoure & Cairney, 2005a, b).
ITS sequence similarity indicated that RFLP type 12 is probably a Cryptosporiopsis species, and an isolate of this RFLP type formed dense ERM-like coils in occasional cells in W. pungens hair roots under gnotobiotic conditions. Cryptosporiopsis spp. are known root-inhabiting fungi, and Cryptosporiopsis rhizophila has been isolated from Ericaceae roots in The Netherlands and shown to colonize Ericaceae roots as an endophyte (Verkley et al., 2003). The nature of the infection was not, however, specified by the authors. Sigler et al. (2005) cultured Cryptosporiopsis ericae and Cryptosporiopsis brunnea from surface-sterilized roots of Canadian Ericaceae, but no ERM structures were formed between these fungi and Ericaceae hosts under gnotobiotic conditions. While more detailed structural investigation of the interaction between RFLP type 12 and W. pungens roots is clearly required, the fact that the fungus formed well-developed ERM-like hyphal coils in epidermal cells strongly suggests that this putative Cryptosporiopsis species has the potential to form an ERM association with this host.
Three of the fungal RFLP types (3, 7 and 14) that formed typical ERM coils in epidermal cells of W. pungens hair roots also formed inter- and/or intracellular infections, along with microsclerotia (RFLP types 7 and 14), in roots of S. productum. These structures are consistent with descriptions of dark septate endophyte (DSE) infection (Jumpponen & Trappe, 1998; Sieber & Grünig, 2006). DSE occur worldwide in a variety of habitats and often coexist with mycorrhizal fungi (Mandyam & Jumpponen, 2005), and have, for example, been identified in roots of a wide range of plants in a single woodland community (Menoyo et al., 2007). They are regarded as forming a continuum of associations with plant hosts that ranges from parasitism to mutualism (Jumpponen, 2001). Understanding the likely nature of their association with S. productum and the other non-Ericaceae plant taxa at the Lovers Jump Creek Reserve will therefore require further investigation.
The data presented here indicate that fungi represented by three RFLP types are capable of forming different types of association with roots of different hosts, viz ERM with W. pungens and DSE with S. productum. Importantly, because at least one of these RFLP types was isolated from each of the 17 plant taxa, fungi with this ability appear to be relatively common at Lovers Jump Creek Reserve. It has previously been reported that some ascomycete isolates from ectomycorrhizal tree roots can form ERM coils in Ericaceae roots (Bergero et al., 2000, 2003; Villarreal-Ruiz et al., 2004), whereas an isolate of the DSE Heteroconium chaetospira formed ERM structures in roots of Rhododendron obtusum (Usuki & Narisawa, 2005). Although the nature of the infection was not described, root endophytes from the grass Deschampsia flexuosa in Dutch heathland/forest were further shown to infect and enhance shoot nitrogen content of Calluna vulgaris and vice versa (Zijlstra et al., 2005).
While some ERM endophytes may persist in soil in habitats that lack ericaceous hosts, possibly as free-living saprotrophs (Bergero et al., 2003; Rice & Currah, 2006), it has been suggested that ectomycorrhizal roots of Q. ilex may act as a source of inoculum for ERM infection of establishing Ericaceae seedlings (Bergero et al., 2000). The observation from the present investigation, that fungi with the ability to form ERM associations with an Ericaceae host were present in roots of all plant taxa examined from a single plant community, indicates that a broad range of plant taxa may act as repositories for ERM fungi in, at least, the sclerophyll forest habitat that we investigated. As suggested by Bergero et al. (2000), this may be of particular importance in establishment of Ericaceae seedlings, especially following a disturbance such as fire in the sclerophyll forest habitat, where fire is a common natural phenomenon and can negatively impact Ericaceae abundance (Morrison, 2002).
Many questions remain regarding relationships between the fungi that form ERM and DSE associations (Girlanda et al., 2006; Schulz, 2006), particularly with regard to the benefits conferred upon plants by DSE infection (Jumpponen, 2001). The ability to infect multiple hosts raises the possibility that ERM structures in Ericaceae roots and DSE structures in non-Ericaceae roots might represent parts of a common mycelial network as has been proposed for ERM and ectomycorrhizal associations in certain forest habitats (Bergero et al., 2000; Bougoure et al., 2007). While direct evidence for such mycelial linkages is currently lacking, if demonstrated, this could influence our understanding of carbon and nutrient cycling in habitats such as the sclerophyll forest at Lovers Jump Creek Reserve, where Ericaceae represent common understorey plants.
We thank Ku-ring-gai Municipal Council for granting permission to collect plant specimens and seeds from Lovers Jump Creek Reserve and staff from the Plant Identification Counter at the National Herbarium of NSW for confirmation of plant species identification. This research was funded by an Australian Research Council Discovery Projects Grant to J.W.G.C.