• A fungal isolate was obtained from Piceirhiza bicolorata-like ectomycorrhizas on Pinus sylvestris in a 160-yr-old natural woodland.
• The fungus was identified by sequencing the PCR-amplified rDNA ITS regions. The sequence was compared with similar known taxa and grouped with Cadophora finlandia in the Hymenoscyphus ericae aggregate.
• The fungus formed P. bicolorata-like ectomycorrhizas in aseptic synthesis with P. sylvestris seedlings. When seedlings of Vaccinium myrtillus were exposed to mycelium arising from these ectomycorrhizas, or to mycelium in pure culture, the hyphae entered the cells of the hair roots and formed coils characteristic of ericoid mycorrhizas. The presence of the fungus stimulated Vaccinium root growth and altered root architecture.
• This is the first full report of the ability of a fungus from the H. ericae aggregate simultaneously to form both ectomycorrhizas and what appear to be ericoid mycorrhizas.
In the forests of the north temperate and boreal zone, the ectomycorrhizal trees in the overstorey often occur with an understorey of ericaceous dwarf shrubs including Calluna vulgaris (L.) Hull and Vaccinium spp., which form ericoid mycorrhizas. A distinctive black ectomycorrhiza, Piceirhiza bicolorata (Brand et al., 1992), is a common component of the ectomycorrhizal fungal communities of these forests, and has been recorded on a number of ectomycorrhizal host species (Vrålstad et al., 2000). It is, for example, frequent in the native Pinus sylvestris L. forests of north-east Scotland (L. Villarreal-Ruiz, unpublished data).
Vrålstad et al. (2000) were able to show that the fungi forming P. bicolorata ectomycorrhizas were genetically similar to the ascomycete Hymenoscyphus ericae (Read) Korf and Kernan. This was a significant discovery because H. ericae is known to form structurally different ericoid mycorrhizas on ericaceous host plants. On this basis, Vrålstad et al. (2000) hypothesized that ericoid and ectomycorrhizal plants share some common mycobionts in the H. ericae aggregate. One of the fungi in this aggregate is the so-called dark septate endophyte (Jumpponen, 2001) of tree roots, formerly known as Phialophora finlandia Wang and Wilcox and transferred by Harrington & McNew (2003) to Cadophora as a new combination following Gams (2000). A strain of Cadophora finlandia (Wang & Wilcox) Harrington & McNew (UAMH7454) has been shown by Wilcox & Wang (1987a, 1987b) to form ectomycorrhizas and ectendomycorrhizas on woody hosts. Monreal et al. (1999) subsequently reported that the same strain ‘in limited trials’ formed typical ericoid mycorrhizas with the ericaceous shrub Gaultheria shallon Pursh. If this were to be confirmed, it would add weight to Vrålstad et al.'s (2000) hypothesis. However, in commenting on these findings Read (2000) sounded a note of caution, pointing out that a test of Vrålstad's hypothesis would require ‘cross-inoculation experiments, to demonstrate first that under ecologically realistic conditions mutually compatible associations are formed by identified genotypes, and second that similar nutritional or other fitness-related responses are seen in both plant partners’.
Vrålstad et al. (2002a) subsequently performed inoculation trials between selected isolates of the H. ericae aggregate and potential ectomycorrhizal and ericoid hosts. Ectomycorrhizal formation was largely restricted to C. finlandia-like isolates of ectomycorrhizal origin, and none of the ectomycorrhiza-forming isolates produced any structures that resembled ericoid mycorrhizas when inoculated onto Vaccinium vitis-idaea L. Overall, their data supported the idea that both the ericoid and the ectomycorrhizal habit have evolved within the H. ericae aggregate, but there was no support for the hypothesis that single isolates could form both types of mycorrhiza. However, genetically identical fungi have been found in both ectomycorrhizal and ericoid mycorrhizal roots in two instances: Quercus robur L. (ectomycorrhizal) and Calluna vulgaris (ericoid); or Salix herbaceae L. (ectomycorrhizal) and Vaccinium vitis-idaea (ericoid) (Vrålstad et al., 2002b). This finding is in line with that of Bergero et al. (2000), who found that ectomycorrhizal roots of Quercus ilex L. harboured fungi capable of forming typical ericoid mycorrhizas with Erica arborea L.
In summary, there is good evidence that fungi in the H. ericae aggregate which form ericoid or ectomycorrhizas are closely related, and that fungi capable of forming ericoid mycorrhizas are found in ectomycorrhizal root systems. However, the only evidence for a single isolate forming both types of mycorrhiza is that of Monreal et al. (1999), where the synthesis was widely separated in time and the evidence for ericoid mycorrhizal formation was said to be limited and still to be confirmed.
In an attempt to clarify further the nature of ericoid and ectomycorrhizal formation by fungi in the H. ericae aggregate, we report here the outcome of dual and tripartite in vitro culture of a fungal isolate from a P. bicolorata-like ectomycorrhiza with seedlings of P. sylvestris and Vaccinium myrtillus L.
Materials and Methods
Site description and sampling procedure
Piceirhiza bicolorata-like ectomycorrhizas were extracted in November 2000 from soil cores randomly sampled from a permanent plot in a 160-yr-old stand (plot 11: 57°01′45.5″Ν, 002°53′27.9″W, 333 m a.s.l.) in Glen Tanar National Nature Reserve, Aboyne, Aberdeenshire, north-east Scotland. Soil cores were wrapped with aluminium foil and placed in sealed plastic bags, stored in a cold room at 4°C, and processed within 72 h. Samples were soaked with tap water and cleaned by wet sieving. Ectomycorrhizal tips were cleaned of soil particles and organic debris under a dissecting microscope and placed in petri dishes with deionized water.
Mycobiont isolation and culture
Vital P. bicolorata-like root tips (sensuBrand et al., 1992: tissue light in colour when sectioned) were shaken for 1 min in 0.2% aqueous solution of Tween 20 (v/v) and rinsed in sterile distilled water, then transferred to 30% aqueous H2O2 (v/v) for 30 s and immediately rinsed three times in sterile deionized water. Surface-sterilized ectomycorrhizal tips were placed on sterilized microscope slides and dissected into small pieces, then placed in petri dishes with modified Melin–Norkrans agar media (MMN; Marx, 1969; Marx & Bryan, 1975) plus chlorotetracycline (30 mg l−1) and incubated in the dark at 20°C. Emergent fungal colonies were transferred individually to fresh MMN agar petri dishes without antibiotics. Adjacent P. bicolorata-like ectomycorrhizas from the same core were fixed in FEA solution (formaldehyde : ethanol 70% : acetic acid (5 : 90 : 5 v/v/v)) as voucher material (Agerer et al., 2000).
DNA was extracted from a 30-d-old axenic culture of isolate LVR4069 and from a synthesized ectomycorrhiza. Briefly, 0.5 g mycelium or one synthesized ectomycorrhizal tip was placed in a Bio-101 red lysing matrix tube (Qbiogene, Cambridge, UK) along with 0.5 ml hexadecyltrimethylammonium bromide (CTAB) extraction buffer (100 mm Tris, 1.4 m NaCl, 20 mm EDTA, 2% CTAB, 1% PVP, 0.2%β-mercaptoethanol) and 0.5 ml phenol/chloroform/isoamyl alcohol (25 : 24 : 1, pH 8.0). The sample was then lysed for 2 × 15 s at a speed of 5.5 m s−1 in a FastPrep bead beating system (Qbiogene, Cambridge, UK) and the aqueous phase was separated by centrifugation at 13 500 r.p.m. for 7 min at 4°C. The sample was extracted further with 600 µl chloroform before precipitating the nucleic acids with sodium acetate (3 m) and isopropanol by centrifugation at 13 500 r.p.m. for 30 min. Pelleted nucleic acids were then washed with cold 70% (v/v) ethanol and air-dried overnight before resuspension in 100 µl TE buffer (pH 7.4).
ITS PCR, sequencing and phylogenetic analysis
Internal transcribed spacer (ITS) regions were amplified using the primers ITS1 and ITS4 (White et al., 1990). Polymerase chain reactions were carried out on a PTC-200 Thermal Cycler (MJ Research, Reno, NV, USA) using 50 µl reaction volumes each containing c. 50 ng template DNA, 20 pmol of each primer, 2 mm MgCl2, 250 µm of each dATP, dCTP, dGTP and dTTP, 10× buffer [20 mm Tris–HCl pH 7.5, 100 mm KCl, 1 mm dithiothreitol, 0.1 mm EDTA, 0.5% Tween 20 (v/v), 0.5% Nonidet P40 (v/v), 50% glycerol (v/v)] and 2.5 U Taq DNA polymerase (Bioline, London, UK). Cycling parameters were 30 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 30 s with a final extension at 72°C for 10 min. Reactions were performed in duplicate, and negative controls (containing no DNA) were included in each reaction. Products were electrophoresed in 1.5% agarose gels, stained with ethidium bromide and visualized under UV light. PCR products were purified using the Qiaquick purification kit (Qiagen, Crawley, UK) and sequenced using the BigDye Terminator Cycle Sequencing Kit with an automated DNA sequencer (ABI model 377, Applied Biosystems, Warrington, UK) using ITS1 and ITS4 primers. DNA sequences generated from isolate LVR4069 and synthesized ectomycorrhiza were edited, and a consensus sequence obtained for both, using the sequencher software package (ver. 3.0; Gene Codes Corporation, Ann Arbor, MI, USA). Closest matches were identified in the GenBank/EMBL/DDBJ nucleotide databases using the fasta 3.0 program (Pearson & Lipman, 1998) and were included in the phylogenetic analysis along with other ITS sequences of taxa within the H. ericae aggregate (Vrålstad et al., 2002b) (Table 1). All sequences were aligned using clustalw (ver. 1.8.2; Thompson et al., 1994) and manual adjustments were made to the alignment where necessary. The transition/transversion ratio and the gamma distribution parameter were estimated using tree-puzzle (ver. 5.0) before conducting a neighbour-joining analysis using the F84 model in paup (ver. 4.0b10; Swofford, 2002) with 1000 bootstrap replicates. Mollisia cinerea (Batsch:Fr.) P. Karst. (AJ430222) and Phialocephala fortinii Wang and Wilcox (AY394921) were used as the outgroup taxa in the analysis.
Table 1. GenBank accession number, fasta closest matches of isolate LVR4069, mycobiont origin and mycotrophic status of ascomycetes included in the phylogenetic analysis
Phialocephala fortinii, axenic culture from hemlock root tip, Canada
Ectomycorrhizas were synthesized aseptically using a low-carbon and -nutrient agar culture system. Plastic petri plates (14 cm diameter) were prepared containing autoclaved modified Ingestad's solution for P. sylvestris (Ingestad, 1979; Ingestad & Kähr, 1985) solidified with 0.8% agar, supplemented with 0.01% glucose (MISAG), and overlaid with sterilized cellophane. From the edge of a 30-d-old mycobiont colony, six mycelium plugs (5 mm) were spaced over two-thirds of a MISAG–cellophane petri plate surface. The dish was sealed and incubated for 2 wk at constant temperature (20°C) in the dark to promote hyphal growth. Surface-sterilized Scots pine seeds were placed in petri plates with autoclaved MISAG, and germinated in the dark for 2 wk. Five axenic germinated seedlings were transferred to the petri plates previously inoculated with the fungi, sealed and placed in a growth chamber (photoperiod 18 h; light 500 µmol m−2 s−1 PAR; temperature 20°C/15°C day/night; RH 75%) for 8 months.
Axenically cultured V. myrtillus seedlings were placed in petri plates with autoclaved MISAG in the growth chamber until used. The cross-infection experiments were performed in culture systems similar to those described above for ectomycorrhizal synthesis. In the dual synthesis, eight V. myrtillus seedlings were transferred aseptically to each of the petri plates previously inoculated with nine mycelial plugs on fresh autoclaved MISAG–cellophane. In the tripartite synthesis, six V. myrtillus seedlings were transferred with one eight month-old Scots pine seedling bearing P. bicolorata-like ectomycorrhizas, synthesized as described above, to three replicated fresh MISAG–cellophane plates. Two plates with six to eight noninoculated V. myrtillus seedlings were left as a control. Plates were placed for 3 months in a growth chamber (photoperiod 18 h; light 200 µmol m−2 s−1 PAR; temperature 18°C/8°C day/night; RH 75%).
Growth response of Vaccinium myrtillus
After 3 months the V. myrtillus seedlings were harvested and the shoot and root carefully separated with a scalpel. Shoots were oven-dried at 60°C for 48 h and their dry weight recorded. The excised hair root system of each seedling was digitized using a Win-Rhizo LA 1600 Scanner (Regent Instruments, Quebec, Canada). The analysis of digitized hair root images was performed using win-rhizo ver. 3.9 software at a scanning resolution of 400 dpi (Polomski & Kuhn, 2002). In order to describe hair root architecture, the following morphometric variables were measured: (a) total hair root length; (b) total hair root surface area; (c) total hair root volume; (d) total number of hair root tips; (e) hair root branch density (root tip number/total root length) (Gamalero et al., 2002). Data were obtained for each seedling, and the mean value for all surviving seedlings in a dish (n = 5–8) was used for statistical comparisons.
Staining and microscopy
A sample of synthesized ectomycorrhizal root tips were excised and split into three subsamples: (1) preserved at 4°C as voucher material in 3% glutaraldehyde in 0.1 m phosphate buffer pH 6.8; (2) placed in CTAB extraction buffer and frozen at −81°C for molecular analyses; (3) morpho-anatomically characterized using freehand longitudinal and cross sections and mantle scrapes, mounted in 85% lactic acid and observed and photographed using a Carl Zeiss Axiophot D-7082 photomicroscope.
Each Vaccinium hair root system from the synthesis plates was cleared for 5 min with 10% (w/v) KOH in a water bath (Grant Instruments, Barrington, Cambridge, UK) at 90°C, acidified with 0.1 m hydrochloric acid for 1 h, stained in 0.05% (w/v) trypan blue for 15 min at 90°C, and destained overnight with lactic acid : glycerol : deionized water (14 : 1 : 1 v/v/v). Whole roots were mounted in the destaining solution, observed, and photographed as above.
Percentage of infection
The magnified intersections method (McGonigle et al., 1990) was adapted to quantify the percentage of infection of Vaccinium hair roots, using hyphal coils in epidermal cells as the sampling target. From each petri plate, three Vaccinium seedlings were chosen at random and the whole root system was examined under a microscope equipped with a cross-hair eyepiece graticule using differential interference contrast at ×1000 magnification. Where possible, at least 100 intersections per seedling root system were scored; on the smaller root systems a minimum of 44 intersections were scored. Counts were recorded as percentage of root length colonized (RLC) by the mycobiont using the formula:
%RLC = 100 × number of intersections with coils/total number of intersections counted.
Synthesized ectomycorrhizal root tips and 1 cm fragments of Vaccinium hair roots (control and uninfected) from each cross-infection experimental plate were aseptically excised and transferred in triplicate to PDA agar, grown in the dark at 20°C for 4 wk, and screened every week for comparisons with the original inoculated fungal strain. The fungal ITS region from synthesized ectomycorrhiza was PCR-amplified and sequenced for comparison with the inoculated fungus.
Molecular identification and phylogeny
FASTA analysis of the ITS sequence of the P. bicolorata-like isolate LVR4069 (GenBank accession number AY579413) revealed that it was most closely related to ITS sequences of ascomycetes within the H. ericae aggregate (Table 1). Phylogenetic analysis using ITS sequences of the closest database matches, along with sequences representing the four main clades of the H. ericae aggregate (as described by Vrålstad et al., 2002b), clustered the P. bicolorata-like mycobiont isolate LVR4069 in clade IV along with fungi referable to C. finlandia (Fig. 1). The clade was strongly supported in a bootstrap analysis (96% bootstrap support) and the sequences within this group were 94.7–99.7% similar to each other. The P. bicolorata-like (LVR4069) sequence was most closely related (98.9% sequence similarity) to AJ430136 (P. bicolorata root tip on Betula pubescens Ehrh., Norway), and the two sequences clustered together within clade IV. The remaining sequences clustered into three further clades representing clades I, II and III described by Vrålstad et al. (2002b). The ITS sequence for the P. bicolorata-like mycobiont had 91.6–92.8%, 93.3–94.1% and 89.9–90.0% similarity with sequences in clades I, II and III, respectively.
Synthesis of mycorrhizas
After 8 months a complete series of ectomycorrhizal developmental stages were present on Scots pine seedlings, ranging from unbranched, through dichotomously branched, to sparsely branched (Figs 2a, 3b–d) with or without a hyaline tip. The colour and appearance changed with age from orange-brown, grey, dark-grey with emanating hyaline hyphae to fully melanized charcoal black and dark-reddish-brown ectomycorrhizas. The emanating hyphae from the sheath (up to 1025 µm long, 1.5–2.0 µm diameter) had rounded tips and were narrow and septate, hyaline, amber to dark brown, thick walled, smooth or asperulose to sparsely verrucose (individual warts up to 0.5 µm) connected to a dense dark mantle (12.5–45 µm) over tannin cells. A well developed, palmetti-type Hartig net was observed. The morpho-anatomical characteristics of synthesized mycorrhizas were similar to those of mycorrhizas collected in the field (Fig. 3a).
In both dual and tripartite inoculations (Fig. 2b,c), the fungus produced structures that appeared to be typical ericoid mycorrhizas in the epidermal cells of hair roots of healthy Vaccinium plants (Fig. 3e,f). The control plants remained uninfected during the experiment. In the dual inoculation dishes (V. myrtillus + fungus) the mycobiont formed vigorous mycelial fans covering parts of the hair roots (Fig. 2d). In the tripartite inoculation (V. myrtillus + P. sylvestris + fungus) these vigorous mycelial fans were not observed. In this case the aerial hyphae arising from P. sylvestris ectomycorrhizas ran along hair root surfaces and entered the epidermal cells. In both cases the mycobiont produced melanized and verrucose runner hyphae (1.5–2.0 µm diameter) growing along the hair root system, which gave rise to fine, hyaline, smooth hyphae (1 µm). These runner hyphae gave rise in some places to dense melanized mycelial patches with some occasional loops and gangliform hyphae. The smooth, hyaline hyphae entered the epidermis and formed intracellular hyphal coils, which ranged from simple circinate hyphae with obvious organelles (Fig. 3f) to coils of convoluted hyphae (Fig. 3e), which could be either hyaline or melanized.
The infection of the Vaccinium hair roots was between 20 and 25% RLC, and was not significantly different whether the fungus colonized from colonies on MISAG–cellophane or from pine ectomycorrhizas. There was no relationship between the size of individual seedling root systems and RLC (data not shown).
Fungal effect on plant growth and hair root architecture
All the V. myrtillus seedlings were green and healthy at the end of the 3 month growth period, and there were no differences in shoot dry weight between the treatments (Table 2). However, there were significant effects of fungal presence on Vaccinium root development, although these differed in intensity between dishes where the fungus alone was present, and those where the fungus was also colonizing Scots pine (Table 2). The most profound effects were on total hair root length (eightfold increase in the presence of the fungus alone) and the number of hair root tips (sixfold increase in the presence of the fungus alone) (Fig. 4a–c). The effect of the fungus was in the same direction, but was of lesser magnitude when infection was by mycelia that were also connected to the ectomycorrhizas on Scots pine.
Table 2. Growth of Vaccinium myrtillus under axenic conditions (V.myr) or in the presence of a fungal (Fun) isolate from the Hymenoscyphus ericae aggregate inoculated as a pure culture (V.myr + Fun), or as synthesized mycorrhizas on Scots pine (Sp) seedlings (V.myr + Sp + Fun)
V.myr (n = 2)
V.myr + Fun (n = 3)
V.myr + Sp + Fun (n = 3)
Means with the same letter are not significantly different (P ≤ 0.05, Student–Newman–Keuls). Replicates (n) are on petri plates containing five to eight Vaccinium seedlings.
Shoot dry weight (mg)
Hair root length (cm)
Number of hair root tips
Hair root surface area (cm2)
Hair root volume (mm2)
25 ± 9.84
20 ± 5.69
Post-harvest fungal reisolation
Synthesized ectomycorrhiza and fragments of V. myrtillus hair roots gave rise to dark, slow-growing colonies on PDA, which appeared microscopically identical to the original isolate. The ITS sequence from synthesized ectomycorrhiza was 100% identical to that of the inoculated isolate.
Systematic position of the isolate from Piceirhiza bicolorata-like mycorrhizas
Phylogenetic analysis of the P. bicolorata-like mycobiont (LVR4069) based on its ITS sequence grouped the isolate in clade IV of the H. ericae aggregate described by Vrålstad et al. (2002b). The majority of sequences in this clade come from fungi detected or isolated from the roots of ectomycorrhizal (ECM) hosts reported from Norway (Vrålstad et al., 2002b), Sweden (Rosling et al., 2003) and Finland (Wang & Wilcox, 1985), and include C. finlandia (AFO 11327).
In contrast, clade III of the phylogenetic tree is predominantly, but not exclusively, represented by fungi detected in the roots of ericaceous plants, and includes the isolates considered to be the putative teleomorph (H. ericae/UAMH6735) and anamorph (Scytalidium vaccini/UAMH5828) (Pearson & Read, 1973; Read, 1974; Egger & Sigler, 1993; Hambleton et al., 1999). This distinction between the clades based on host, and hence ericoid mycorrhizal (ERM) or ECM, origin has been discussed previously (Vrålstad et al., 2002b). This conclusion has been supported to some extent by data from resynthesis experiments using some isolates from the two clades with ERM and/or ECM hosts (Wilcox & Wang, 1987a, 1987b; Dalpéet al., 1989; Monreal et al., 1999; Vrålstad et al., 2000, 2002a, 2002b). However, many more isolates need to be tested in dual and tripartite cross-infection experiments before any conclusion can be drawn on the separation of these clades based on mycorrhizal status. Jumpponen & Trappe (1998) pointed out that any outcome from pure culture synthesis needs to be interpreted with caution because the experimental conditions can influence the nature of the symbiotic response.
Dual infection of Vaccinium and Pinus roots
This study has demonstrated for the first time that an isolate from the H. ericae aggregate obtained from wild P. bicolorata-like (sensuBrand et al., 1992) mycorrhizas on Scots pine from native ecosystems can be used to synthesize ectomycorrhizas in vitro on P. sylvestris. We also show that the same isolate in dual and tripartite culture enters the hair root epidermal cells of V. myrtillus and forms hyphal coils, similar to the typical ericoid mycorrhizal structures described by Bonfante & Gianinazzi-Pearson (1979) and Read (1983). Under our experimental conditions, colonization of V. myrtillus induced a substantial growth response in the root system. It must be emphasized that our designation of these structures as putative ericoid mycorrhizas is based on the presence of fungal coils and the positive growth effects seen in Vaccinium growing in our experimental system. Much more detailed experimentation will be required to satisfy fully the criteria of Read (2000) and to demonstrate that mutually beneficial metabolite transfer takes place between the partners.
Previous resynthesis attempts have produced contradictory evidence about the ability of individual isolates from the H. ericae aggregate to form both ecto- and ericoid mycorrhizas (Monreal et al., 1999; Bergero et al., 2000; Vrålstad et al., 2002a). Monreal et al. (1999) suggested (but presented no evidence) that an ectendomycorrhizal isolate of C. finlandia could form ERM with Gaultheria. In the same context, Bergero et al. (2000) demonstrated that identical fungal genets were present in ECM and ERM hosts in Mediterranean ecosystems, but did not demonstrate ECM or ERM formation by the same mycobiont. Finally Vrålstad et al. (2000, 2002b) were able to demonstrate that, on the one hand, there is strong phylogenetic relationship between H. ericae (ERM) and C. finlandia (ECM) as well as the presence of identical or nearly identical ITS genotypes in ECM and ERM hosts; but on the other hand, none of the ECM-forming strains produced ERM on Vaccinium.
Positive response of Vaccinium to fungal presence
Under the experimental conditions used, infection of epidermal hair roots cells of V. myrtillus in dual culture led to a significant increase in root length and number of root tips, suggesting a benefit to the host. This could be a response to improved nutrition, as has been demonstrated in ericoid mycorrhizal associations (Mitchell & Read, 1981). Alternatively, Berta & Gianinazzi-Pearson (1986) suggested stimulation of plant hormone production or the production of indole-3-acetic acid (IAA) by the fungus as a possible cause of the substantial change in root length and number of hair roots in C. vulgaris seedlings infected with H. ericae. IAA has been identified in culture filtrates from H. ericae strains (Gay & Debaud, 1986) and recently in a ‘Phialophora sp.’ (Rommert et al., 2002) subsequently identified as Phialocephala fortinii based on molecular evidence (B. Schulz, Technical University of Braunschweig, Germany, personal communication). Rommert et al. (2002) also reported a positive effect on the growth of Larix decidua plants and an increase in root system branching, both when their P. fortinii isolate was applied as a fungal extract and when it colonized the roots. It should be noted that the production of IAA was demonstrated only in the presence of tryptophane (Gay & Debaud, 1986), which was not added to our culture media.
The effect of the fungus on Vaccinium hair root growth was greater when the infection arose from colonies on MISAG–cellophane rather than from pine mycorrhizas. The reasons for this are unclear, but were not related to the extent of infection of the hair roots, which was the same in both cases. The fungus appeared more vigorous when growing from colonies on MISAG–cellophane, and this might affect the production of IAA.
The success of ericaceous plants in heathland ecosystems is the result of the ability of the plant/fungal symbiosis to succeed in conditions of extreme mineral N and P limitation and accumulation of recalcitrant organic matter (Cairney & Meharg, 2003; Read & Pérez-Moreno, 2003). In addition, boreal and mediterranean biomes support ericaceous plants as understoreys of ectomycorrhizal conifer or broadleaf forests that have been interpreted as ‘relict populations’ trapped by recurrent glaciation events (Read, 1991; Rendell & Ennos, 2002). In these two situations, environmental and edaphic conditions may operate in a different fashion, and can be expected to produce different selection pressures on both photo- and mycobionts along a gradient of increasing organic matter decomposability from open moorlands to woodlands (Cornelissen et al., 2001; Wilson & Puri, 2001). Soil fungal community structure derived from ITS- denaturing gradient gel electrophoresis (DGGE) profiles shows a distinct shift along a moorland–pine forest environmental gradient (Anderson et al., 2003), as do soil physical, chemical and biological properties (Chapman et al., 2001, 2003). Our observations support the notion that members of the well supported clade III of the H. ericae agreggate may have evolved as strict ericoid endophytes in moorlands, but that members of clade IV (including C. finlandia) are primarily ECM fungi, but may have the potential to be ERM symbionts in woodlands containing both ericaceous and ectomycorrhizal plants. This raises the intriguing possibility that understorey and overstorey plants are linked by a common mycorrhizal fungal mycelium. Identifying the functional significance, if any, of such linkages will require careful experimentation in the field.
Brundrett (2002) suggested that the evolution of mycorrhizas in Ericales is represented by a shift from arbuscular mycorrhizal ancestors towards ericoid systems (and subsequently to ‘exploitative’ ECM in Monotropoidae). He claims that the important event in the evolution of ericoid mycorrhiza was a ‘switch to a new fungal lineage’ in which the first ericoid endophyte evolved from an ECM fungus, probably from the H. ericae aggregate. Our results support this notion in as much as we have demonstrated that a strain from the H. ericae aggregate has the potential to produce both ectomycorrhizas and what appear to be ericoid mycorrhizas.
We are grateful to Glen Tanar Estate for permission to sample on the field sites, to CONACYT for a scholarship to the senior author, and to Colegio de Postgraduados–IREGEP, Mexico. I.C.A. is funded by the Scottish Executive, Environment and Rural Affairs Department. We thank Trude Vrålstad and two anonymous referees for valuable comments on an earlier version of the manuscript.