A novel plant–fungus symbiosis benefits the host without forming mycorrhizal structures


  • Khalil Kariman,

    1. School of Earth and Environment M087, The University of Western Australia, Crawley, WA, Australia
    2. School of Plant Biology M084, The University of Western Australia, Crawley, WA, Australia
    Search for more papers by this author
  • Susan J. Barker,

    1. School of Plant Biology M084, The University of Western Australia, Crawley, WA, Australia
    2. Institute of Agriculture M082, The University of Western Australia, Crawley, WA, Australia
    Search for more papers by this author
  • Ricarda Jost,

    1. School of Plant Biology M084, The University of Western Australia, Crawley, WA, Australia
    Search for more papers by this author
  • Patrick M. Finnegan,

    1. School of Plant Biology M084, The University of Western Australia, Crawley, WA, Australia
    2. Institute of Agriculture M082, The University of Western Australia, Crawley, WA, Australia
    Search for more papers by this author
  • Mark Tibbett

    Corresponding author
    1. School of Earth and Environment M087, The University of Western Australia, Crawley, WA, Australia
    2. Department of Environmental Science and Technology (B37), School of Applied Sciences, Cranfield University, Cranfield, Bedfordshire, UK
    Search for more papers by this author


  • Most terrestrial plants form mutually beneficial symbioses with specific soil-borne fungi known as mycorrhiza. In a typical mycorrhizal association, fungal hyphae colonize plant roots, explore the soil beyond the rhizosphere and provide host plants with nutrients that might be chemically or physically inaccessible to root systems.
  • Here, we combined nutritional, radioisotopic (33P) and genetic approaches to describe a plant growth promoting symbiosis between the basidiomycete fungus Austroboletus occidentalis and jarrah (Eucalyptus marginata), which has quite different characteristics.
  • We show that the fungal partner does not colonize plant roots; hyphae are localized to the rhizosphere soil and vicinity and consequently do not transfer nutrients located beyond the rhizosphere. Transcript profiling of two high-affinity phosphate (Pi) transporter genes (EmPHT1;1 and EmPHT1;2) and hyphal-mediated 33Pi uptake suggest that the Pi uptake shifts from an epidermal to a hyphal pathway in ectomycorrhizal plants (Scleroderma sp.), similar to arbuscular mycorrhizal symbioses, whereas A. occidentalis benefits its host indirectly. The enhanced rhizosphere carboxylates are linked to growth and nutritional benefits in the novel symbiosis.
  • This work is a starting point for detailed mechanistic studies on other basidiomycete–woody plant relationships, where a continuum between heterotrophic rhizosphere fungi and plant beneficial symbioses is likely to exist.


More than 80% of land plant species are involved in beneficial associations with diverse fungal taxa known as mycorrhizal symbioses (Smith & Read, 2008; Brundrett, 2009), which are considered to be a critical factor in terrestrial ecosystems for sustained productivity (van der Heijden et al., 2008), nutrient cycling (Read & Perez-Moreno, 2003; Finlay, 2008) and plant–soil feedback (Bever, 2002). At the root–fungal interface, the heterotrophic symbiont colonizes plant roots, where it receives organic carbon (mainly glucose and fructose) supplied by the phototrophic partner (Smith & Read, 2008). In return, extraradical hyphae (ERH) explore the soil beyond the roots for > 10 cm in arbuscular mycorrhizal (AM) symbioses (Li et al., 1991) and up to at least 40 cm in ectomycorrhizal (ECM) symbioses (Finlay & Read, 1986), scavenge nutrients from the soil, assimilate them and subsequently transport them to the host plant.

Seven categories of mycorrhizas have been recognized based on the morphological structures developed and the plant–fungal species involved. These include AM, ECM, ectendomycorrhizas, orchid mycorrhizas, ericoid mycorrhizas, arbutoid mycorrhizas and monotropoid mycorrhizas (Smith & Read, 2008). Dark septate endophytes (Jumpponen, 2001) and Sebacinales (Weiß et al., 2011) can colonize plant tissues and establish beneficial associations, which are debatably referred to as mycorrhizas or endophytes. However, plant–fungus associations with unconventional or undeveloped structures have also been documented that cannot be classified into any of the above mycorrhizal categories and are of ambiguous functional status (Warcup & McGee, 1983; Kope & Warcup, 1986; Brundrett, 2009). Our previous work showed that Austroboletus occidentalis substantially promoted growth and nutrient acquisition of jarrah (Eucalyptus marginata) seedlings, without forming mycorrhizal structures (Kariman et al., 2012). A positive growth response has also been reported for seedlings of Eucalyptus camaldulensis inoculated with the ECM fungus Pisolithus tinctorious without the formation of ECM structures (Neumann, 1959).

The existing literature has typically viewed the function of mycorrhizas with respect to specialized mycorrhizal structures and hyphal-mediated nutrient uptake, which might not always reflect the functionally relevant factors of an association. Recent studies, however, highlighted the role of fungi in biogenic weathering of soils in most forest ecosystems including mineral dissolution and phosphate (Pi) solubilization, leading to enhanced nutrient availability for ERH and ECM roots (Landeweert et al., 2001; Finlay et al., 2009).

In the current study, we aimed to determine the source of, and mechanism by which the plant host accessed, enhanced mineral nutrient in the absence of root colonization in the symbiosis observed between jarrah and A. occidentalis. We used radiolabeled Pi (33Pi) and multiple compartments to unravel the underlying functional mechanisms for the nutritional benefits received by the host. We also considered the effects of conventional ECM formation by Scleroderma sp. and the noncolonizing fungal partner A. occidentalis on the transcript profiles of three jarrah high-affinity Pi transporter (PHT1) genes in roots. There were two major questions to be addressed: are there any similarities in nutritional mechanisms and benefits between the novel symbiosis and the classic ECM symbiosis, and does ECM symbiosis shift Pi transport from a root epidermal to a hyphal pathway?

Materials and Methods

Plant species and seed germination

Jarrah (Eucalyptus marginata Donn ex Smith) seeds were obtained from a single tree in Dwellingup, Western Australia. Seeds were surface-sterilized with 70% (v/v) ethanol for 1 min, rinsed with sterile water and then treated with 4% (w/v) sodium hypochlorite for 30 s. Subsequently, seeds were rinsed thoroughly with sterile water and germinated on wet, sterile filter papers in Petri dishes by incubating at 15°C in the dark for 2 wk.

Fungal isolates and inoculum production

Austroboletus occidentalis Watling & N.M. Greg. and Scleroderma sp. isolates were collected from a jarrah forest rehabilitation site at Langford Park, Western Australia and Banksia woodland at Piney Lakes, Western Australia, respectively. The fungal inocula were produced using a vermiculite-based carrier method (Marx & Bryan, 1975). Polyethylene jars were half-filled with 200 ml of a mixture of medium-grade vermiculite and Lithuanian peat moss (5 : 1 v/v) and autoclaved. The contents of the jars were moistened with 125 ml of liquid growth medium (Lambilliotte et al., 2004) and re-autoclaved. Jars were subsequently inoculated under sterile conditions with 10 mycelial plugs (5 mm in diameter) from the edge of actively growing cultures on PDA plates and incubated at 23°C for 2 months. Colonization studies were performed as previously described (Kariman et al., 2012).

Experimental design

A double polyester mesh bag (with 40-μm pores) was used to separate the root–hyphal compartment (RHC) from the hyphal compartment (HC) in which we placed a radiation compartment (RDC) containing 33Pi-labeled sand. In this system, plant roots were confined to the RHC and only fungal hyphae could penetrate the mesh bag and reach the 33Pi source located 7–12 cm away from the root systems (Fig. 1a). The radiolabeled Pi (orthophosphoric acid) was purchased from PerkinElmer (Boston, MA, USA) and used to investigate the ability of fungi to transport 33Pi from the RDC to jarrah plants. For this purpose, 250 g of sand containing 928 kBq 33Pi was placed in plastic vials (7 × 7 cm) at the bottom of cylindrical PVC pots (15 × 35 cm). Vials were topped up with 100 g of clean sand (2 cm) before being buried in 3.250 kg of double-pasteurized washed river sand. A double-thickness polyester mesh bag was filled with 4.150 kg of a mixture of fungal inoculum (peat-vermiculite substrate cultures) and double-pasteurized washed river sand (1 : 10 v/v) to form the RHC and placed above the HC containing the RDC. The nonmycorrhizal (NM) control received sterilized inoculum to equalize the amount of nutrients and organic matter among the treatments. Three pre-germinated jarrah seeds were planted in each pot and the pot surface was covered with 3 cm of sterile plastic beads to minimize cross or environmental contaminations. Plants were grown for 16 wk in a controlled environment cabinet with 16 h light (750 μM m−2 s−1) : 8 h dark, 24 : 18°C temperature and 60 : 70% relative humidity. All plants received the 1× modified (Cavagnaro et al., 2001) Long Ashton solution minus P once a fortnight started 2 wk after planting (10 ml kg−1 soil): K2SO4 2 mM, MgSO4.7H2O 1.5 mM, CaCl2.2H2O 3 mM, FeEDTA 0.1 mM, (NH4)2SO4 4 mM, NaNO3 8 mM, H3BO3 46 μM, MnCl2.4H2O 9 μM, ZnSO4.7H2O 8 μM, CuSO4.5H2O 0.3 μM and Na2MoO4.2H2O 0.01 μM.

Figure 1.

Jarrah (Eucalyptus marginata) shoot biomass greatly enhanced by both novel (Austroboletus occidentalis) and ectomycorrhizal (ECM; Scleroderma sp.) symbioses. (a) Schematic diagram showing the three-compartment system used to study the potential functional pathways in mycorrhizal symbioses. Jarrah roots were confined to the root hyphal compartment (RHC), while extraradical hyphae (ERH) could pass through the polyester mesh bag (PMB) and colonize both hyphal (HC) and radiation (RDC) compartments. (b) Shoot (black bars) and root (gray bars) dry biomass of jarrah plants. Bars are significantly different from nonmycorrhizal (NM) controls at: **, < 0.05; *, < 0.10. Error bars are SE (= 3).

Measuring the 33P activity

A Geiger counter (Mini 900 Ratemeter; Thermo Scientific, Waltham, MA, USA) was used to nondestructively track the 33Pi uptake in shoots of growing plants. Shoot radioactivity at harvest was determined in tissues oven-dried at 70°C for 72 h. A measured quantity (c. 200 mg) of dried ground shoot tissue was digested in a mixture of nitric : perchloric acid solution (4 : 1 v/v) and diluted to 10 ml with water. A 1.0-ml portion of the diluted digest was mixed with 3.0 ml of scintillation fluid (Irga-Safe Plus; Perkin Elmer) for assessment (TR 1500 liquid scintillation counter; Packard Instrument Co., Downers Grove, IL, USA). The radioactivity measurement was corrected for decay and background radiation. To measure 33Pi in soil, the plant-available Pi was extracted using a NaHCO3-based method (Rayment & Higginson, 1992) before measuring the radioactivity as already described.

Plant nutrient analysis

The acid-digested samples from the 33Pi measurements were used to measure the shoot concentrations of phosphorus (P), sulphur (S), magnesium (Mg), iron (Fe), zinc (Zn) and copper (Cu) using inductively coupled plasma optical emission spectrometry (ICP-OES; Optima 5300 DV; Perkin Elmer). A measured amount (c. 60 mg) of dried ground shoot material was used to measure the nitrogen (N) percentage in a combustion analyzer (Elementar Vario Macro, Hanau, Germany). The plant-available Pi was extracted using a NaHCO3-based method (Rayment & Higginson, 1992) before the P concentration was determined using a Malachite green-based colorimetric assay (Motomizu et al., 1983).

Quantification of fungal hyphae in different compartments

Fungal hyphae were extracted from soil by an aqueous solution and membrane filter technique (Jakobsen et al., 1992) with a slight modification (5% ink (Black Sheafer) in white vinegar was used to stain hyphae instead of lactoglycerol-trypan blue). Tennant's formula (Tennant, 1975) was used to calculate the hyphal length on each nitrocellulose gridded filter (0.45-μm Millipore, Billerica, MA, USA).

Cloning of jarrah PHT1 genes and real-time PCR

Total RNA was isolated from jarrah roots using a cetyltrimethylammonium bromide (CTAB)-based protocol (Korimbocus et al., 2002) with a slight modification. Sodium D-isoascorbate was added to the extraction buffer just before use to a final concentration of 100 mM. Total RNA (1.0 μg) was treated with DNase I (RQ1 RNase-free DNase; Promega) to remove genomic DNA contamination. Oligo(dT)18 was used to prime cDNA synthesis (GoScript reverse transcriptase; Promega) according to the supplier's instructions. PHT1 cDNA fragments were amplified by PCR using combinations of two forward primers (EcPT1-F and EcPT2-F) and two reverse primers (EcPT4-R and EcPT5-R) (Table 1) that were designed based on conserved sequences in the PHT1 genes of E. camaldulensis (Koyama et al., 2006). PCR amplifications were carried out in 50 μl containing 3 μl of cDNA equivalent to 60 ng of total RNA, 75 μM dNTP mix, 0.5 μM of each forward and reverse primer, 2 mM MgCl2, 2.5 U of Taq DNA polymerase (Biotaq; Bioline, Boston, MA, USA) in 1× PCR buffer (Bioline). The PCR program included 1.30 min of initial denaturation at 94°C followed by 30 cycles of denaturation at 94°C for 40 s, annealing of the primers at 56°C for 40 s and extension at 72°C for 2 min. The amplicons were terminated by a final extension at 72°C for 10 min. PCR products were purified (Sure Clean Plus; Bioline) and ligated into pJET1.2 (CloneJET PCR Cloning Kit; Fermentas, Glen Burnie, MD, USA). Plasmids with inserts of the expected size were submitted to sequencing (Australian Genome Research Facility, Perth, Australia). The identities of PHT1 sequences were confirmed by comparison to publicly available sequence databases and used to design primers to amplify the 5′ and 3′ ends of PHT1 cDNAs by rapid amplification of cDNA ends (RACE; SMART RACE cDNA Amplification Kit and Advantage 2 Kit; Clontech, Palo Alto, CA, USA). RACE products were purified, ligated into pJET1.2 and sequenced using the same procedure as outlined for PCR products.

Table 1. Oligonucleotide primer pairs used
PrimerDirectionSequence (5′-3′)Product size (bp)PCR efficiency
  1. na, not applicable.

Primers used to clone the transcripts
Primers used for real-time PCR

Fragments of five EmPHT1 cDNAs were isolated from a whole-root cDNA library. The identified fragments encoded amino acid sequences that are highly conserved among dicot PHT1 sequences. All the jarrah PHT1 sequences obtained were deposited in GenBank: EmPHT1;1 (KC172372; 1052 bp), EmPHT1;2 (KC172373; 980 bp), EmPHT1;3 (KC172374; 932 bp), EmPHT1;4 (KC172375; 964 bp) and EmPHT1;5 (KC172376; 752 bp). Transcript quantification was carried out for three jarrah PHT1 sequences including EmPHT1;1, EmPHT1;2 and EmPHT1;5, which sequences were available at the time of performing this assay. Gene-specific primers were designed for RT-PCR (Table 1). In addition, a fragment of a jarrah actin gene (ACT) (KC172376; 470 bp) was amplified using Actin-F and Actin-R primers (Webster, 2008) and sequenced to allow the design of gene-specific primers to be used as an internal control for the relative quantification assay (Table 1). The SYBR green-based real-time PCR (q-PCR) was carried out to quantify the transcripts from PHT1 genes relative to ACT transcripts. When preparing cDNAs, a subsample of the cDNA synthesis reaction was removed before adding reverse transcriptase. These samples were included in q-PCR assays to check for the presence of genomic DNA contamination in the RNA samples after DNase I treatment. The q-PCR reactions were performed in 96-well plates in 10-μl reactions containing 0.3 μM of each gene-specific primer and 2.5 μl of cDNA equivalent to 50 ng of total RNA in 1× PCR cocktail (SYBR Green PCR master mix; Applied Biosystems, Foster City, CA, USA). All q-PCR experiments were performed on a 7500 FAST Real-time PCR System (Applied Biosystems). The q-PCR results were expressed on a log2 scale, represented by 40 – ∆CT (Bari et al., 2006), where CT is the value of the threshold cycle for the individual PCR reactions and ∆CT is the difference in CT values between the target gene and the reference gene, ACT. This difference is subtracted from 40, the maximum number of cycles used, so that all values are positive.

Quantification of root carbohydrates

Carbohydrates were extracted from freeze-dried root samples using 80% ethanol (Cononoco, 2002). High-performance liquid chromatography (HPLC) was used to determine the concentration of fructose and mannitol in root extracts (Smith et al., 1986; AOAC, 1999). Root extracts were injected into an HPLC separation module (W2695; Waters) using acetonitrile/water as the mobile phase and separation was achieved using a high-performance carbohydrate column (5 μm × 3.9 mm × 300 mm; Waters). Sugars were detected using a refractometer (2414 Refractive Index (RI) detector; Waters, Milford, MA, USA) with individual peak areas quantified. Sugar speciation and concentration were determined after comparison to primary compound standard sets covering each analyte.

Measuring the carboxylate concentration in rhizosphere soil

The root system was shaken carefully to remove the bulk soil. The soil adhering to the root was defined as the rhizosphere soil (Veneklaas et al., 2003). The root system was gently shaken in a measured volume of 0.2 mM CaCl2 to wash off the rhizosphere soil. Syringe filters (0.2-μm Supor membrane; Pall Acrodisc, East Hills, NY, USA) were used to filter a subsample of extract into 1-ml HPLC vials. The filtered extracts were acidified with concentrated phosphoric acid (0.1% v/v) and frozen at −20°C until analysis. The carboxylate concentration of rhizosphere extracts was determined by HPLC (Cawthray, 2003) using 600E pump, 717 plus autosampler, a 996 photodiode array detector (Waters) and a C-18 reverse-phase column (Alltima; Alltech Associates, Deerfield, IL, USA).

Statistical analysis

Plants were grown in a completely randomized design with three replicates. The two fungal treatments were Scleroderma sp. and A. occidentalis. The NM control received sterilized inoculum to equalize the amount of nutrients and organic matter among treatments. Another set of controls was also assessed. These had no 33Pi and were used to calculate the background radiation in soil and jarrah shoot tissues. ANOVA was performed (sas, version 9.2; SAS Institute, Cary, NC, USA). Means were separated using the least significant difference at a 5% significance level unless otherwise stated.


Jarrah growth and nutrition

We examined plant growth responses to each fungal partner using a three-compartment system (Fig. 1a). The presence of a fungal partner greatly improved jarrah growth (Fig. 1b). Positive responses were achieved regardless of the root-colonizing ability of the fungus. There was no sign of root–hyphal association or a short-root morphology change for jarrah seedlings grown with A. occidentalis, while the Scleroderma sp. treatment achieved 77% ± 11% (mean ± SE, n = 3) ECM colonization. However, A. occidentalis was as effective as Scleroderma sp. in enhancing the shoot biomass of plants. Furthermore, jarrah plants inoculated with either fungus had larger root systems than NM controls, with the largest root system belonging to the Scleroderma sp. treatment (Fig. 1b).

Both types of symbiosis effectively increased the shoot nutrient content of plants compared with NM controls (Table 2). Shoots of plants inoculated with A. occidentalis had almost equal contents of P, S, Mg and Cu and even more N than the Scleroderma sp. treatment. The root-colonizing fungus Scleroderma sp. seems, however, to be more effective than A. occidentalis at facilitating uptake of the micronutrients Fe and Zn. Compared with NM controls, inoculation with A. occidentalis enhanced the shoot P and Mg concentrations and inoculation with Scleroderma sp. enhanced the shoot concentrations of P, S, Mg, Fe and Zn (Table 2).

Table 2. Shoot nutrient status of jarrah (Eucalyptus marginata) seedlings
Treatmentsmg per potμg per pot
  1. Data represent mean values (= 3) ± SE. Values of each nutrient are significantly different from nonmycorrhizal (NM) controls at: **, < 0.05; *, < 0.10.

NM control0.54 ± 0.0943.14 ± 2.864.84 ± 1.088.95 ± 2.2383.68 ± 10.5558.93 ± 14.1419.26 ± 1.65
Austroboletus occidentalis 1.86** ± 0.5659.95** ± 1.7311.49** ± 1.4321.17 ** ± 3.54165.97** ± 23.32146.27** ± 13.5243.52** ± 3.02
Scleroderma sp.1.91** ± 0.2244.46 ± 0.8812.61** ± 1.3920.31** ± 3.41341.48** ± 26.47202.10** ± 19.7648.37** ± 1.71
 mg g−1 DWμg g−1 DW
NM control0.19 ± 0.0116.26 ± 3.491.65 ± 0.043.00 ± 0.1030.38 ± 4.1819.88 ± 0.757.21 ± 1.48
A. occidentalis 0.29*± 0.0310.37 ± 1.451.93 ± 0.153.49**± 0.0727.63 ± 1.2624.80 ± 2.297.42 ± 0.78
Scleroderma sp.0.31** ± 0.037.52** ± 1.212.05** ± 0.073.26* ± 0.0856.30** ± 3.9233.19** ± 2.218.10 ± 1.04

33Pi uptake from a radiation compartment

To examine the source of enhanced P accumulation in inoculated jarrah seedlings, the content of 33Pi from the radiation compartment (RDC) (Fig. 1a) was assessed in the shoots of the 16-wk-old plants (Fig. 2a). Scleroderma sp. transported 33Pi to jarrah plants, while there was only background radiation in shoot tissues from the A. occidentalis and NM treatments. Nondestructive monitoring of shoot radioactivity during the growth period indicated that the fungal-mediated 33Pi uptake in the Scleroderma sp. treatment was detectable 9 wk after planting. Furthermore, the specific activity of radiolabeled sand remaining after harvest in the Scleroderma sp. treatment was significantly lower than for the NM and A. occidentalis treatments (Fig. 2a), supporting the hypothesis that 33Pi was taken up by ERH. The plant-available Pi extracted from the root hyphal compartment (RHC) and hyphal compartment (HC) did not contain radioactivity in any of the treatments, showing that 33Pi did not move or leak from the RDC. Some 33Pi was, however, in transit from the RDC to jarrah shoots within ERH and roots, but this was unextractable by the method used and therefore has been excluded here.

Figure 2.

Functional evidence for root-mediated Pi uptake in the novel (Austroboletus occidentalis) symbiosis and hyphal-mediated Pi uptake in the ectomycorrhizal (ECM; Scleroderma sp.) symbiosis. (a) Specific activity of 33Pi in shoot and 33Pi-labeled sand in the radiation compartment (RDC). (b) Hyphal length density in the three different compartments. (c) Plant-available Pi in different compartments. (d) Expression of three high-affinity phosphate transporter genes (PHT1) genes in jarrah (Eucalyptus marginata) roots relative to expression of an internal reference gene (ACT) determined by real-time PCR. The scale on the vertical axis is a log2 scale based on ΔCT (the threshold cycle (CT) of the target PHT1 gene minus the CT of the reference gene). Thus, a difference of one CT value corresponds to a two-fold difference in transcript abundance. The ∆CT value is subtracted from 40, the maximum number of PCR cycles, to make all values positive for ease of comparison (Bari et al., 2006). Bars with different letters are significantly different at < 0.05. Bars within each compartment are significantly different from nonmycorrhizal (NM) controls at: **, < 0.05; *, < 0.10. Error bars are SE (= 3).

Hyphal length densities in different compartments

Our results are consistent with two possibilities. Austroboletus occidentalis hyphae might grow throughout the compartments, but not transfer 33Pi to the RHC, or A. occidentalis might not colonize the RDC. To investigate this further, we quantified the hyphal length in the different compartments (Fig. 2b). Hyphae from the A. occidentalis treatment were thinner (1.5–4 μm) than those found in the Scleroderma sp. treatment (3.5–14 μm). As shown in Fig. 2(b), both inoculated treatments had significantly higher hyphal length densities in the RHC compared with the NM controls. However, the Scleroderma sp. treatment had significantly greater hyphal length in both the HC and the RDC than found for either A. occidentalis or NM treatments that did not differ significantly from each other for either compartment.

Phosphate (Pi) uptake from different compartments

To ascertain the source of P exploited by plants and fungi and transferred to the shoots of inoculated plants, we determined the concentration of plant-available Pi in soil from the various compartments (Fig. 2c). The A. occidentalis treatment had the lowest available Pi in the RHC of all the treatments. The available Pi in the HC and RDC did not differ significantly between A. occidentalis and NM control treatments, whereas the Scleroderma sp. treatment had lower available Pi in both the HC and the RDC compared with NM controls. Thus, the A. occidentalis-treated plants did not access Pi from the HC or RDC to any greater extent than the NM control plants.

Transcript abundance of jarrah high-affinity Pi transporter (PHT1) genes

A real-time PCR transcript assay was used to quantify the abundance of transcripts from three jarrah EmPHT1 genes currently available for assay relative to an internal ACT reference gene (EmACT1) in response to the fungal symbionts. The results revealed that there were significantly lower abundances of EmPHT1;1 and EmPHT1;2 transcripts in roots from the Scleroderma sp. treatment compared with the NM control and A. occidentalis treatments (Fig. 2d). By contrast, there was no difference in the transcript profiles for the three PHT1 genes in roots of plants from NM and A. occidentalis treatments. The transcript abundance for EmPHT1;5 was similar across treatments.

Root carbohydrate concentration

Mannitol and fructose were measured in roots by HPLC (Fig. 3). Here, mannitol was only found in roots from the Scleroderma sp. treatment, where hyphae formed typical ECM structures. No mannitol was found in roots of either NM control or A. occidentalis treatments, both of which had no sign of fungal colonization of roots. There was a constant amount of fructose in roots across all three treatments.

Figure 3.

Jarrah (Eucalyptus marginata) roots contain mannitol only in the Scleroderma sp. treatment where ectomycorrhizal (ECM) colonization occurred. Nonmycorrhizal control, black bars; Austroboletus occidentalis, gray bars; Scleroderma sp., hatched bars. Determinations for each sugar indicated by a different letter are significantly different at < 0.05. ND, not detected. Error bars are SE (= 3).

Carboxylate concentration in the rhizosphere soil

The HPLC analysis showed that citrate was the main carboxylate present in the rhizosphere soil, which, along with lower concentrations of fumarate and shikimate, was present in all three treatments (Table 3). The rhizosphere from the A. occidentalis treatment had significantly higher concentrations of citrate and fumarate compared with both Scleroderma sp. and NM treatments. The Scleroderma sp. treatment had the lowest citrate among the treatments, while fumarate and shikimate concentrations in that treatment did not differ from those of NM controls. Furthermore, traces of maleate were detected in NM and Scleroderma sp. treatments, and all three treatments contained traces of oxalate. However, the values for these two carboxylates were very near or below the detection limit and therefore were not considered for the quantitative analysis.

Table 3. Carboxylate concentration in the rhizosphere soil (μmol g−1 root DW)
  1. Values within a column followed by different letters are significantly different at  0.05. Data are mean ± SE (= 3).

Nonmycorrhizal (NM) control1234b ± 11323b ± 587a ± 13
Austroboletus occidentalis 2525a ± 12059a ± 1082a ± 17
Scleroderma sp.660c ± 7918b ± 378a ± 7


Our research has uncovered distinct mechanisms for plant host nutritional benefit operating in the two fungal symbioses examined. The results indicate that the range of beneficial fungal interactions with jarrah is wider than the classic definition of mycorrhizal symbiosis. A novel symbiosis clearly exists between jarrah and A. occidentalis, where the fungus does not penetrate the plant root but has defined beneficial consequences for plant growth and nutrition that differ in their mechanistic basis from the similar benefits derived by the ECM symbiosis that we examined. Based on the results for 33Pi uptake (Fig. 2a), shoot P content (Table 2), hyphal extent (Fig. 2b), the plant-available Pi in soil (Fig. 2c) and the expression of jarrah PHT1 genes (Fig. 2d), we conclude that jarrah plants involved in the novel symbiosis mainly accessed Pi from the rhizosphere soil and vicinity (RHC), whereas the plants inoculated with Scleroderma sp. (ECM) obtained Pi from all root-accessible (RHC) and hypha-accessible (HC and RDC) compartments.

The abundance of transcripts from numerous plant PHT1 genes typically increases under low plant P status and decreases under high plant P status. AM colonization has been shown to induce a remodeling of PHT1 transcript profiles, such that there is preferential accumulation of transcripts from specific PHT1 genes in cells in close proximity to the fungal symbiont at the expense of transcripts from other PHT1 genes in root epidermal cells not in direct contact with fungal hyphae (Karandashov & Bucher, 2005). The outcome of the remodeling is thought to be the preferential delivery of Pi from the mycorrhizal symbiont to root cells instead of the uptake of Pi directly from the bulk soil solution. Here, the reduced abundance of EmPHT1;1 and EmPHT1;2 transcripts in roots subsamples of plants exposed to Scleroderma sp. (Fig. 2d), combined with the apparent hyphal delivery of 33Pi, indicates that these two PHT1 genes are likely to be responsible for the direct pathway of Pi uptake by plant roots. Some plant PHT1 gene products have been shown to have reduced in AM (Karandashov & Bucher, 2005) and ECM (Loth-Pereda et al., 2011) symbioses, which presumably results in a reduction in Pi uptake via the direct (root) pathway. However, it is not still clear whether the reduced expression of PHT1 genes in mycorrhizal roots is triggered by improved P nutrition of plants or happens in response to the symbiosis (Javot et al., 2007). Our research has advanced understanding of the ECM symbiotic Pi uptake mechanism from that inferred by Loth-Pereda et al. (2011). Based on the active involvement of Scleroderma sp. (ECM) hyphae in 33Pi uptake, Pi uptake from both hypha-accessible compartments (HC and RDC), and the reduced expression of two plant PHT1 genes, we suggest that Pi uptake shifts from a root epidermal to a hyphal pathway in ECM plants, as documented to occur in AM symbiotic plants (Smith et al., 2011). This finding may have profound effects on our understanding of P nutrition in woody plants, which are highly dependent on ECM associations for their growth, nutrition and survival (Read & Perez-Moreno, 2003; van der Heijden et al., 2008; Smith & Read, 2008).

In contrast to the Scleroderma sp. treatment, there was no difference in the transcript profiles for the three PHT1 genes between roots of plants from NM and A. occidentalis treatments (Fig. 2d). This lack of remodeling of the PHT1 transcript pool in response to A. occidentalis, accompanied by the absence of hyphal-mediated 33Pi uptake and a physical contact (no colonization) between the two partners, indicates that active Pi uptake occurs directly from the soil solution via plant roots in the proposed novel symbiosis.

Mannitol is a sugar alcohol typically found in fungi, serving various physiological functions such as carbon storage (Koide et al., 2000) and cryoprotection (Tibbett et al., 2002). Plants generally do not synthesize mannitol but there are documented exceptions (Stoop et al., 1996; Merchant et al., 2006). The exclusive presence of mannitol in mycorrhizal roots has been attributed to the fungal structures associated with roots (Lewis & Harley, 1965; Stribley & Read, 1974; Smith & Read, 2008). We observed that mannitol was also present in mycelia from axenic cultures of both Scleroderma sp. and A. occidentalis growing on solid nutrient medium, which was 529 ± 337 (mean ± SE, n = 3) and 1414 ± 215 (mean ± SE, n = 3) μg g−1 FW, respectively. Combined with the observation that NM jarrah roots lacked mannitol, but it was easily detected in roots from the Scleroderma sp. treatment, the lack of mannitol in roots from the A. occidentalis treatment (Fig. 3) is physiological support for the absence of microscopic fungal structures in and on the roots here and in our previous work (Kariman et al., 2012), supporting our conclusion that A. occidentalis does not colonize jarrah roots.

Carbon flow from roots into the rhizosphere is a key phenomenon for carbon cycling in soil and occurs via a variety of processes, including carbon flow from roots to root-associated symbionts such as mycorrhizal fungi (Nehls et al., 2007; Jones et al., 2009; Lambers et al., 2009). ECM plants support the carbon nutrition of their fungal partners through direct supply of sucrose (and hexoses) at the root/fungal interface or via repression of their root monocsaccharide importers to allow carbon flow from soil towards hyphae, or a combination of these pathways (Nehls et al., 2007). The washed river sand used in this experiment was very poor in organic matter, and therefore the direct supply of fixed carbon from the plant to the ECM partner (Scleroderma sp.) is most likely under these experimental conditions. Here, ECM roots had the same concentration of fructose as NM roots, which could be attributable to the rapid conversion of fructose to mannitol by intraradical hyphae (Stribley & Read, 1974). With regard to A. occidentalis, as hyphae do not penetrate roots, we hypothesize that the products of sucrose breakdown (glucose and fructose) are exuded into the rhizosphere soil, which is the primary fungal habitat (Fig. 2b). Clarified understanding of carbon nutrition of the fungal partner in this novel symbiosis is required to assist in determination of whether this is a commensal or a mutualistic association.

Biological weathering of soil to dissolve minerals and release plant nutrients (except N) is accomplished mainly by plant roots and soil microbes through the exudation of metabolites such as carboxylates, protons, respiration-derived CO2, phenolic compounds and siderophores (Landeweert et al., 2001; Lambers et al., 2009). It has been established that ECM fungi exude carboxylates (Landeweert et al., 2001) and protons (Hoffland et al., 2004) to mobilize nutrients from primary silicate minerals, such as the sand used here. Low levels of carboxylates in rhizosphere soil of ECM plants (Leyval & Berthelin, 1993; Wallander, 2000a,b) have been attributed mainly to high microspatial variability in concentration and rapid microbial consumption (Landeweert et al., 2001). Here, we propose another factor that might be associated with the low carboxylate concentration in the rhizosphere of ECM plants. The active mineral nutrient uptake in mycorrhizal symbioses occurs via fungal hyphae rather than plant roots (Marschner & Dell, 1994; Cameron et al., 2006; Finlay, 2008; Smith & Read, 2008). Carboxylates in ECM symbioses have been shown to originate primarily from fungal hyphae dissolving nutrients surrounding hyphal tips (Wallander, 2000b). Therefore, in ECM plants, carboxylates are presumably released by extensive hyphal networks into a much broader soil volume (hyphosphere) rather than into the immediate rhizosphere soil, and consequently the carboxylates are diluted in the bulk soil surrounding the roots. Our data also show that, in a classic ECM symbiosis (Scleroderma sp.), the plant exudation of citrate is suppressed, whereas, in the presence of the novel symbiosis, exudation of citrate is enhanced (Table 3), although we cannot exclude the possibility that some part of the additional citrate is of fungal origin. Likewise, the significant enhancement of fumarate in the A. occidentalis treatment compared with the NM control and Scleroderma sp. treatments could suggest that the additional fumarate may be of fungal origin.

Higher concentrations of carboxylates in the rhizosphere of NM controls and A. occidentalis treatments compared with the Scleroderma sp. treatment could be attributable to the fact that active uptake of mineral nutrients takes place directly via plant roots in the former two treatments. The available Pi was very low (Fig. 2c) in the washed river sand used for plant culture. Although the soil total P was not measured in this study, it is considerably higher in different soil types, typically between 100 and 3000 mg P kg−1 soil (Frossard et al., 2000). Phosphorus release from sand particles by aid of carboxylates could be behind the improved P nutrition of jarrah plants associated with A. occidentalis. Our results support previous studies suggesting that citrate is the main carboxylate associated with nutrient (P, in particular) release and mobilization from primary minerals such as apatite and biotite (Wallander et al., 1997; Olsson & Wallander, 1998; Lambers et al., 2009).

Austroboletus occidentalis enhanced the shoot N content of jarrah plants, while no positive effects were observed with Scleroderma sp. Improved N nutrition is a well-known benefit in ECM symbioses, which is mainly associated with the capability of ECM fungi to exude nutrient-mobilizing enzymes such as acid phosphatases, proteinases, and polyphenol oxidase to release nutrients, including N-containing nutrients, from organic sources in soil (Bending & Read, 1995; Tibbett & Sanders, 2002; van der Heijden et al., 2008). The Scleroderma sp. treatment had the largest root system among the treatments, but the lowest shoot N concentration. This is apparently a result of the dilution of N in the higher shoot biomass, as the content of N per pot was the same as in NM controls (Table 2). The failure of the ECM hyphae to improve the N nutrition of jarrah plants was probably a result of the lack of organic matter in the washed river sand used for this study.

Free-living N fixers are ubiquitous in various environments, including soil, sediment and river habitats (Stewart, 1975; Affourtit et al., 2001; Franche et al., 2009). N-fixing bacilli and clostridia produce dormant endospores (Stewart, 1975; Leggett et al., 2012), which are highly resistant to heat (Henriques & Moran, 2000; Leggett et al., 2012) and therefore might have survived the heat-based pasteurization process applied to the river sand in this experiment. On the other hand, citrate has been shown to significantly promote bacterial growth and activity in soil (Olsson & Wallander, 1998). Accordingly, we hypothesize that the higher soil carboxylates (citrate in particular) produced by the novel symbiosis underlie the enhanced N nutrition of jarrah, perhaps by stimulating the growth of diazotrophic bacteria. Another possible explanation for the lack of N response in ECM compared with the novel symbiosis is that Scleroderma sp. produced much more hyphae (Fig. 2b) than A. occidentalis. Therefore, a greater proportion of the available N could be taken up and used by the extensive hyphal network in the Scleroderma sp. treatment, leading to lower N availability for the shoot tissues.

The present study shows clear functional differences between the jarrah–A. occidentalis symbiosis and previously described mycorrhizal associations. In this novel symbiosis, fungal hyphae are found and function in the vicinity of root systems, but do not colonize plant roots. Moreover, the 33Pi baiting experiment did not support the presence of ‘hyphal-mediated nutrient uptake’ which is a well-established strategy in mycorrhizal symbioses. The enhanced carboxylate release into the rhizosphere soil correlates with higher shoot nutrient content in plants involved in this novel symbiosis and may be the underlying mechanism. It will be of interest to determine how much of the enhanced exudation can be attributed to either plant or fungus. More investigations are warranted to unearth other potential strategies that might be involved in this novel symbiosis such as secretion of protons, nutrient-mobilizing enzymes or other unknown pathways.


The authors are grateful to Tim Morald who collected an A. occidentalis mushroom from a jarrah forest rehabilitation site (2007, Langford Park, Jarrahdale, Western Australia) and Neale Bougher who kindly identified the Austroboletus species. The authors are also thankful for technical advice from Lindsey Loweth, Gregory Cawthray, Michael Smirk, Hazel Gaza, Stuart Pearse, Megan Ryan, Hossein Khabaz-Saberi, Basu Regmi and Robert Creasy. We appreciate the University of Western Australia postgraduate scholarships (SIRF/UIS) awarded to K.K. and also financial support and grants from the Centre for Land Rehabilitation at the University of Western Australia (M.T.) and the Australian Research Council (M.T., P.M.F.).