Carbohydrate and Amino Acid Metabolism in the Eucalyptus globulus-Pisolithus tinctorius Ectomycorrhiza during Glucose Utilization

Abstract

The metabolism of [1-¹³C]glucose in Pisolithus tinctorius cv Coker & Couch, in uninoculated seedlings of Eucalyptus globulus bicostata ex Maiden cv Kirkp., and in the E. globulus-P. tinctorius ectomycorrhiza was studied using nuclear magnetic resonance spectroscopy. In roots of uninoculated seedlings, the ¹³C label was mainly incorporated into sucrose and glutamine. The ratio (¹³C3 + ¹³C2)/¹³C4 of glutamine was approximately 1.0 during the time-course experiment, indicating equivalent contributions of phosphoenolpyruvate carboxylase and pyruvate dehydrogenase to the production of α-ketoglutarate used for synthesis of this amino acid. In free-living P. tinctorius, most of the ¹³C label was incorporated into mannitol, trehalose, glutamine, and alanine, whereas arabitol, erythritol, and glutamate were weakly labeled. Amino acid biosynthesis was an important sink of assimilated ¹³C (43%), and anaplerotic CO₂ fixation contributed 42% of the C flux entering the Krebs cycle. In ectomycorrhizae, sucrose accumulation was decreased in the colonized roots compared with uninoculated control plants, whereas ¹³C incorporation into arabitol and erythritol was nearly 4-fold higher in the symbiotic mycelium than in the free-living fungus. It appears that fungal utilization of glucose in the symbiotic state is altered and oriented toward the synthesis of short-chain polyols.

Carbohydrate metabolism in ectomycorrhizae has received considerable attention (for review, see Hampp and Schaeffer, 1995; Smith and Read, 1997). Using carbohydrates for storage, for increasing biomass, and for conversion into metabolic energy, ectomycorrhizal fungi create strong assimilate sinks. Photoassimilates move into the phloem of trees primarily as Suc and reach the ectomycorrhizal tissues in this form (Jakobsen, 1991; Hampp and Schaeffer, 1995). Suc is the main labeled carbohydrate in root cells but is not detected in symbiotic fungal tissues, where mannitol, trehalose, and glycogen are the main labeled carbohydrates (Söderström et al., 1988; Hampp and Schaeffer, 1995; Smith and Read, 1997). Glc resulting from Suc catabolism is thought to be the primary source of carbon for the generation of ATP, reducing power, and carbon skeletons for biosynthetic pathways in ectomycorrhizae (Hampp and Schaeffer, 1995). The metabolic pathways leading to the synthesis of major fungal carbohydrates such as mannitol and trehalose have been characterized in several free-living ectomycorrhizal fungi (Martin et al., 1985, 1988; Ramstedt et al., 1989). These carbohydrates have also been found in ectomycorrhizae (Ineichen and Wiemken, 1992), but metabolic routes converting Suc to fungal carbohydrates and other metabolites in symbiotic tissues have not been characterized.

There is evidence that ectomycorrhizal symbiosis brings about considerable modification of carbon metabolism in the host roots and in the mycobiont forming the association (Martin et al., 1987; Hampp and Schaeffer, 1995). An important question in relation to the physiology of ectomycorrhizal associations concerns the extent to which each partner contributes to the metabolism of carbohydrates. A full understanding of the metabolic fate of Glc in ectomycorrhizae requires the characterization of (a) the metabolic pathways converting Glc to other carbohydrates and metabolites, (b) the carbon compounds accumulated in symbiotic tissues, and (c) the changes induced by the symbiosis on the partner metabolism. We used NMR spectroscopy in conjunction with [1-¹³C]Glc labeling to study carbohydrate and amino acid metabolism in a eucalypt (Eucalyptus globulus subsp. bicostata) and in Pisolithus tinctorius, growing separately and in mycorrhizal association. The results demonstrated significant mutual effects on fungal and host-plant metabolism.

MATERIALS AND METHODS

Biological Material and in Vitro Synthesis of Ectomycorrhizae

Eucalyptus globulus subsp. bicostata ex Maiden Kirkp. seeds (Kylisa Seeds Co., Weston Creek, Australia) were sterilized with 20% calcium hypochloride (v/v) for 20 min, rinsed with four changes of sterile water, and plated on low-sugar Pachlewski medium (2.7 mm di-ammonium tartrate, 7.3 mm KH₂PO₄, 2.0 mm MgSO₄·7H₂O, 5 mm Glc; 2.9 μm thiamine-HCl, and 1 mL of a trace-element stock solution [Kanieltra 6Fe, Hydro Azote Co., Ambès, France]; Hilbert et al., 1991) in 2.0% (w/v) agar. After 7 d, aseptically germinated seedlings were placed on the edge of 14-d-old fungal mats of the ectomycorrhizal gasteromycete Pisolithus tinctorius Coker & Couch isolate 441, grown on low-sugar Pachlewski medium in 2.0% agar, and left for 7 d in a controlled-environment growth chamber with 16 h of light (25°C, 150 μmol m⁻² s⁻¹) and 8 h of dark (Hilbert et al., 1991) for ectomycorrhiza formation. After 7 d, the ectomycorrhizal sheath and Hartig net were differentiated on the main root and lateral roots of seedlings (Dexheimer et al., 1994). Petri dishes of free-living mycelium and uninoculated control seedlings were grown under the same conditions. Fungal colonization of root tissues was measured by the ergosterol assay (Martin et al., 1990). Uninoculated and 7-d-old ectomycorrhizal seedlings, together with the edges of 21-d-old fungal mats, were then sampled and preincubated in 5 mL of Pachlewski medium containing 5 mm Glc for 1.5 h prior to [¹³C]Glc labeling.

Labeling Studies

[1-¹³C]Glc labeling of uninoculated and ectomycorrhizal seedlings was carried out with the roots fully submerged in 5 mL of Pachlewski medium containing 5 mm [1-¹³C]Glc (99 atom % ¹³C, Sigma-Aldrich) for 6 to 30 h. Fungal mats were floated on the surface of the labeled solution. Samples (approximately 0.1 g dry weight) were taken for natural abundance NMR analysis immediately before the addition of the labeled [1-¹³C]Glc (0-time sample). After this, the labeled samples (roots and mycelium) were rinsed thoroughly with Glc-free Pachlewski medium to remove any remaining labeled Glc, blot-dried, frozen in liquid nitrogen, and lyophilized. Lyophilized samples were extracted using cold (−20°C) methanol:water (70:30, v/v; Martin and Canet, 1986) and processed for NMR analysis of water-soluble carbon compounds (Martin, 1991). Two labeling experiments were performed for each time course, with very similar results (±5% to ±10%); in each case data from one of the replicates are shown.

NMR Spectroscopy

¹H-decoupled ¹³C-NMR spectra were recorded at 100.55 MHz using a spectrometer (Unity Plus 400, Varian Instruments, Sugarland, TX) with a superconducting magnet (Oxford Instruments, Oxford, UK). Spectra were recorded at 25°C with the following spectrometer conditions: proton decoupling by WALTZ-16 composite pulse sequence, 25,000-Hz spectral width, quadrature phase detection, 16 K data storage array, 11.6–μs observation pulse (corresponding to a 45° flip angle), 2.38-s recycle time, and 25,000 free induction decays. The lock signal was obtained from the 99% (v/v) D₂O in which the extract was dissolved. Spectra were processed with 3.0-Hz exponential line broadening.

Chemical shifts are quoted relative to the Suc C5′ resonance (82.48 δ ppm; plant extracts) or the trehalose C1 resonance (94.1 δ ppm; fungal and ectomycorrhizal extracts) and expressed in 100 δ ppm downfield from tetramethylsilane. The resonances were assigned by comparing observed chemical shifts with previously published values for carbohydrates (Dijkema et al., 1985; Martin et al., 1985, 1988; Shachar-Hill et al., 1995; Fan, 1996) and amino acids (Martin and Canet, 1986; Fan, 1996). The identification of each carbohydrate and amino acid component was also made by peak matching with authentic samples, spiking with authentic samples, or analyses of heteronuclear single-quantum coherence spectroscopy spectra.

Acquisition conditions that do not permit complete relaxation of all carbon signals between pulses were used. However, this had a constant effect on intensities throughout the series of experiments, and the relative intensities could therefore be measured accurately. To compare the amounts of ¹³C incorporated into metabolites in ectomycorrhizal and fungal extracts sampled during the time-course experiments, the peak intensities of the various ¹³C resonances was standardized to the peak intensity of the natural abundance of mannitol C2,5 within each spectrum and then to mannitol C2,5 resonance in the natural abundance (T₀) spectrum. The ¹³C enrichment (atom % ¹³C) of mannitol C1,6 and C3,4 was calculated by comparison with natural abundance ¹³C enrichment (1.1%) of mannitol C2,5 within the same spectrum. Suc ¹³C enrichment was evaluated by comparing the intensity of the C1 and C1′ resonances (62.4 and 93.3 δ ppm) with the intensity of the unlabeled natural abundance resonance of the C5′ (82.4 δ ppm). Absolute ¹³C content in trehalose C1 was determined from the ¹H-¹³C satellite spectra, and ¹³C labeling in C6,6′ was calculated from the C6,6′-to-C1,1′ ratio in ¹³C spectra.

RESULTS

Glc Assimilation in Free-Living P. tinctorius

Methanolic extracts of P. tinctorius harvested after growth in a medium containing 5 mm [1-¹³C]Glc for 29 h gave rise to the ¹³C-NMR spectra shown in Figure 1. The largest resonances observed in the carbohydrate region of Figure 1A arose from C1,6 of mannitol and C1,1′ of trehalose. The C1,6 of mannitol was the most highly labeled component in the time-course experiment (Fig. 4A), incorporating 31% of total NMR observable ¹³C at 29 h. It is possible to deduce the percentage of mannitol C1,6 labeling by comparing natural abundance mannitol C2,5 with the C1,6 intensity. Mycelium showed 8.5%, 20%, and 29% ¹³C enrichment (i.e. atom % ¹³C) after 6, 21, and 29 h, respectively. This distribution of label is consistent with the hypothesis that mannitol, the most prominent soluble carbohydrate of free-living P. tinctorius (natural abundance spectrum not shown), was synthesized by a direct route from labeled [1-¹³C]Glc.

Figure 1.

¹³C-NMR spectra of intracellular ¹³C-carbohydrates (A) and free ¹³C-amino acids (B) in P. tinctorius mycelium obtained after feeding [1-¹³C]Glc (99 atom %) for 29 h. G, Glc; T, trehalose; M, mannitol; E, erythritol; A, arabitol. Subscripts refer to the carbon positions.

Figure 4.

Time dependence of the ¹³C content of metabolites identified in P. tinctorius mycelium (A), uninoculated E. globulus roots (B), and P. tinctorius-E. globulus ectomycorrhizae (C) incubated in 5 mm [1-¹³C]Glc for the indicated times. The ¹³C contents are from the peak heights of the carbohydrate and amino acid resonances in the NMR spectra. ○, Trehalose; •, mannitol; ▪, arabitol; □, erythritol; ▵, Suc; ▴, Gln; X, Glu; ⋄, Ala. ¹³C content in fungal and ectomycorrhizal spectra were standardized to the natural abundance mannitol C2,5. Average values are from duplicate experiments (±5% to ±10%).

Mannitol C3,4 showed a very low ¹³C enrichment (approximately 2%), indicating that the flux of ¹³C from Glc through the pentose phosphate pathway was limited (Martin et al., 1988). Apart from the rapid incorporation of ¹³C into mannitol, a number of changes occurred with time, notably the marked increase in intensity of trehalose C1,1′ (Figs. 1A and 4A). From the ¹H-¹³C satellite spectrum of trehalose (data not shown), it appears that the ¹³C enrichment of this C1,1′ position was 61%, whereas the C6,6′ position exhibited a 14% ¹³C enrichment. The occurrence of ¹³C labeling at the C6,6′ position of trehalose indicates an isotopic scrambling between C1,1′ and C6,6′ positions, suggesting that Glc carbon used to form trehalose was cycled through the metabolically active mannitol pool or the pentose phosphate cycle enzyme transaldolase (Martin et al., 1988; Pfeffer and Shachar-Hill, 1996). However, the ¹³C6,6′-to-¹³C1,1′ ratio of trehalose was approximately 0.18, suggesting low randomization of the ¹³C label. After 29 h of labeling, mannitol, trehalose, erythritol, arabitol, and Glc accounted for 54%, 26%, 9%, 6%, and 5% of the total carbohydrate ¹³C, respectively.

Figure 1B shows an expanded ¹³C-NMR spectrum of the amino acid region after 29 h of labeling. The most intense peaks of this spectrum had chemical shifts that correspond to C2, C3, and C4 of Gln and C3 of Ala. Glu positions were weakly labeled. In Gln and Glu, C4 exhibited a greater ¹³C content than the C2 or C3 positions of these amino acids (Fig. 1B). The ratio (¹³C3 + ¹³C2)/¹³C4 of Gln was approximately 1.0 during the time-course experiment. The proportion of the label entering the free amino acid pools represented 22% (6-h time sample) to 43% (29-h time sample) of the ¹³C observed by NMR. After 29 h, Gln (59%) accounted for the largest incorporation in the amino acid pool, followed by Ala (30%) and Glu (11%). Multiple ¹³C resonances of Gln C3 at 27 ppm showed that the synthesis of multilabeled [¹³C3-¹³C4]Gln occurred. This indicates cycling of amino acid precursors through the Krebs cycle (Malloy et al., 1988).

Glc Assimilation in Roots of Uninoculated E. globulus

Suc is the main soluble carbon metabolite in uninoculated eucalypt roots, as shown by the presence of natural abundance ¹³C resonances of this disaccharide (Fig. 2A). After addition of [1-¹³C]Glc, the majority of labeling was incorporated into the glucosyl C1 and fructosyl C1′ moieties of Suc. These positions showed about 15% ¹³C enrichment after 29 h. The labeling of the C1 of Fru (the precursor of the C1′ of Suc) and C1 of the Glc used to synthesize Suc were identical when the contribution of free Glc α C1 was subtracted. This suggests a rapid labeling of the Fru pool from absorbed Glc. There was a weak scrambling between the C1,1′ of Glc into the C6,6′ positions of Suc.

Figure 2.

¹³C-NMR spectra of intracellular ¹³C-carbohydrates (A) and free ¹³C-amino acids (B) in noncolonized E. globulus roots obtained after feeding [1-¹³C]Glc (99 atom %) for 28 h. G, Glc; S, Suc; F, Fru; T, trehalose; M, mannitol; E, erythritol; A, arabitol.

Gln was the only free amino acid detected in uninoculated eucalypt roots (Figs. 2B and 4B). The ratio (¹³C3 + ¹³C2)/¹³C4 of Gln was approximately 1.0 during the time-course experiment, indicating equivalent contributions of PEP carboxylase and pyruvate kinase/pyruvate dehydrogenase to the production of Krebs cycle intermediates.

Glc Assimilation in Ectomycorrhizae

Based on the ergosterol assay (Martin et al., 1990), the analyzed ectomycorrhizae contained approximately 25% to 30% fresh weight of mycelium (data not shown). As shown in Figure 3, ectomycorrhiza formation had a dramatic effect on Glc metabolism in the mycobiont and the host roots. After a 27-h incubation in [1-¹³C]Glc, resonances detected in the carbohydrate region (Fig. 3A) predominantly arose from C1,1′ of trehalose, C1,6 of mannitol, C1,5 of arabitol, and C1,4 of erythritol. The ¹³C natural abundance resonances of Suc were detected, but ¹³C incorporation into the glucosyl C1 and fructosyl C1′ moieties of the disaccharide was very low (about 1% above the natural abundance). The C1,1′ of trehalose was labeled to about 75% from the ¹³C spectrum (comparison of the C5,5′ position at 73.1 δ ppm and the C1,1′ position at 94.1 δ ppm; Fig. 3A) and approximately 81% from the ¹H-¹³C satellite spectrum (data not shown). Trehalose C6,6′ was more intense than the natural abundance resonances of this carbohydrate at 73.4, 73.1, 71.9, and 70.5, indicating the occurrence of an isotopic scrambling (i.e. cycling) between the C1,1′ and C6,6′ positions. The ¹³C6,6′-to-¹³C1,1′ ratio of trehalose was approximately 0.11, implying low randomization of the ¹³C label, as observed in the free-living mycelium (Fig. 1A). The amount of newly incorporated ¹³C into trehalose was nearly 2-fold higher in symbiotic tissues than in free-living mycelium. The C1,6 of mannitol was also highly labeled over the time-course experiment (Fig. 4C), reaching 21% ¹³C enrichment after 27 h.

Figure 3.

¹³C-NMR spectra of intracellular ¹³C-carbohydrates (A) and free ¹³C-amino acids (B) in P. tinctorius-E. globulus ectomycorrhizae obtained after feeding [1-¹³C]Glc (99 atom %) for 27 h. G, Glc; S, Suc; F, Fru; T, trehalose; M, mannitol; E, erythritol; A, arabitol.

Apart from the rapid synthesis of [¹³C]trehalose and [¹³C]mannitol, a striking feature of [¹³C]Glc assimilation in ectomycorrhizae was the rapid and intense accumulation of [¹³C]erythritol and [¹³C]arabitol (Figs. 3A and 4C). The distribution of ¹³C label is consistent with the hypothesis that these polyols were synthesized by a direct route from [1-¹³C]Glc. These sugar alcohols, which barely accumulated in free-living P. tinctorius (Fig. 1A), represented a large part of the labeling in the soluble ectomycorrhizal carbon compounds (Fig. 4C). After 27 h, trehalose, mannitol, erythritol, and arabitol accounted for 31%, 25%, 26%, and 18% of NMR-observable ¹³C incorporated in fungal carbohydrates, respectively, whereas root Suc was barely detectable. The amount of newly incorporated ¹³C in total polyols (i.e. mannitol, erythritol, and arabitol) of symbiotic tissues was about 1.6-fold higher than in free-living mycelium. The increased synthesis of erythritol plus arabitol in the symbiotic mycelium was even higher: 4.4-fold higher than in free-living mycelium.

As found in free-living P. tinctorius, the most intense peaks observed in the amino acid region (Fig. 3B) had chemical shifts corresponding to C2, C3, and C4 of Gln and C3 of Ala. Glu positions were weakly labeled. In Gln, C4 exhibited a greater ¹³C content than the C2 and C3 positions of these amino acids (Fig. 3B). The ratio (¹³C3 + ¹³C2)/¹³C4 of Gln was approximately 1.1, indicating that the ¹³C flux through anaplerotic carboxylases was still high in symbiotic tissues and was not affected by ectomycorrhiza development. As observed in the free-living mycelium, multiplet ¹³C resonances of Gln C3 revealed the presence of Gln isotopomers. The proportion of the label entering the free amino acid pools represented 19% (6-h time sample) to 25% (27-h time sample) of the ¹³C observed by NMR. After 27 h, Gln (72%) accounted for the largest incorporation, followed by Ala (15%) and Glu (13%). The proportion of ¹³C incorporated into Ala was thus 50% lower in symbiotic mycelium compared with the free-living mycelium.

DISCUSSION

¹³C Metabolism in Free-Living P. tinctorius

In ectomycorrhizal ascomycetes (Martin et al., 1985, 1988), basidiomycetes (Martin et al., 1984; Söderström et al., 1988; Ramstedt et al., 1989; Ineichen and Wiemken, 1992; Hampp and Schaeffer, 1995), and other fungi (Lewis and Smith, 1967; Dijkema et al., 1985), trehalose and various polyols (e.g. mannitol and arabitol) have been reported to be present during active growth. These carbohydrates form endogenous storage pools that are continuously metabolized and contribute to the osmotic stabilization of the hyphae. Mannitol contains the highest proportion of carbon from assimilated Glc in Cenococcum geophilum and Sphaerosporella brunnea (Martin et al., 1985, 1988), whereas trehalose is prominent in Piloderma croceum (Ramstedt et al., 1989) and Laccaria bicolor (Martin, 1991). This indicates that mannitol and trehalose are important components of carbohydrate conversion and biosynthesis.

Insoluble glycogen was not detectable in the extracts, but likely contributed to the carbohydrate pools (Martin et al., 1985, 1988). The extensive labeling of mannitol (31% of total NMR-observable ¹³C at 29 h) and trehalose (17%) in P. tinctorius is consistent with this scheme. Free-living mycelium of P. tinctorius strain Lelly/Marx 298 showed high contents of trehalose and arabitol (Ineichen and Wiemken, 1992), whereas [¹³C]arabitol was barely detectable (6% at 29 h) in the strain used in the present study. The fact that trehalose showed only a weak isotopic scrambling between the C1 and C6 positions in hexose pools in free-living P. tinctorius and symbiotic fungal cells (Figs. 1A and 3A) is in marked contrast to observations in free-living C. geophilum and S. brunnea. In these ectomycorrhizal ascomycetes, cycling through the mannitol cycle is high and leads to intense isotopic scrambling (Martin et al., 1985, 1988).

Free amino acids also represent an important sink of absorbed and assimilated carbon in P. tinctorius (43% of total ¹³C at 29 h). This value was similar to the proportion of ¹³C entering the free amino acids of other ectomycorrhizal fungi (France and Reid, 1983; Martin and Canet, 1986; Martin et al., 1988). Under the conditions of nitrogen (5.4 mm) levels in the growth medium, a large proportion of the carbon is therefore shifted toward production of amino acids. Labeled Gln was already abundant after 6 h of feeding (Fig. 4A) and its ¹³C content increased rapidly to 25% of the soluble ¹³C. In [1-¹³C]Glc-fed C. geophilum (Martin and Canet, 1986) and P. croceum (Ramstedt et al., 1989), Gln was also rapidly synthesized. Because Gln has a strong signal and its labeling pattern reflects the isotopic distribution of α-ketoglutarate, it could be used to track the label through Krebs cycle intermediates (Martin, 1991; Pfeffer and Shachar-Hill, 1996).

The intramolecular ¹³C-labeling pattern of Glu/Gln in P. tinctorius is in agreement with the operation of the Krebs cycle. α-Ketoglutarate, used to synthesize Glu and Gln, would therefore arise from sequential action of citrate synthase, aconitase, and isocitrate dehydrogenase. There is evidence of about 32% multiple labeling of Gln, as indicated by ¹³C-¹³C spin-spin coupling (e.g. resonance at 27.1 ppm of Gln C3; Fig. 1B) showing that α-ketoglutarate used to form amino acids does cycle to a minor extent through the Krebs cycle (Malloy et al., 1988). The ¹³C isotopic distribution in Gln was used to estimate the contribution of the anaplerotic C flux to malate via pyruvate carboxylase, as previously demonstrated by Martin and Canet (1986). During the time-course experiment, the ratio (¹³C3 + ¹³C2)/¹³C4 of Gln indicated that the contributions of pyruvate carboxylase and pyruvate dehydrogenase to the production of Krebs cycle intermediates (Martin and Canet, 1986; Martin, 1991) were similar. Anaplerotic CO₂ fixation is therefore an important component of Glc metabolism in free-living mycelium. It is likely that this anaplerotic role is particularly significant under conditions of amino acid accumulation to replenish intermediates of the Krebs cycle that are drawn off for biosynthesis during active growth. ¹³C intramolecular enrichment of Glu and Gln in other ectomycorrhizal fungi (Martin and Canet, 1986; Martin et al., 1988; Ramstedt et al., 1989) also suggested high activity of anaplerotic carboxylases during rapid Glc utilization.

Ala synthesis was a significant fate for the Glc carbon in P. tinctorius. The high ¹³C labeling of the C3 of Ala is in agreement with the synthesis of this amino acid via pyruvate kinase and Ala aminotransferase.

¹³C Metabolism in Uninoculated Eucalypt Roots

The majority of labeling from [1-¹³C]Glc was incorporated into C1 of the glucosyl and fructosyl moieties of Suc in the uncolonized roots. Gln was the only amino acid detected and it represented an important sink of absorbed and assimilated carbon (17% at 20 h). As shown by the (¹³C3 + ¹³C2)/¹³C4 ratio of Gln, PEP carboxylase and pyruvate dehydrogenase contributed equally to the production of Krebs cycle intermediates. It is now well documented in plant cells that both malate and pyruvate act as the point of entry for glycolytic carbon into the Krebs cycle and that malate is favored during rapid respiration, such that a significant fraction of glycolytic products enters the Krebs cycle via the combined action of PEP carboxylase and malate dehydrogenase (Wiskich and Dry, 1985). Edwards et al. (1998) showed that PEP carboxylase contributed 62% of the malate synthesized in respiring maize root tips. The flux through PEP carboxylase is comparable in magnitude in eucalypt roots. This high PEP carboxylase anaplerotic activity likely sustains the synthesis of Gln.

¹³C Metabolism in Ectomycorrhizae

The utilization patterns of the Glc source by seedlings and mycelium was dramatically influenced by mycorrhizal colonization, with a greater allocation of carbon to short-chain polyols, arabitol, and erythritol and to trehalose in the mycelium and a suppression of Suc synthesis in the roots. The labeling of Suc by the host cells was suppressed in the mycorrhizal roots despite the significant level of [1-¹³C]Glc supplied (5 mm). This finding does not seem to be a result of the preferential interception of labeled Glc by the hyphal network ensheathing the roots, because previous studies using ³⁵S-labeled Met and Cys have shown that synthesis of proteins is taking place at a high rate from the exogenous precursors in the root cells of ectomycorrhizal eucalypt seedlings (Hilbert et al., 1991; Burgess et al., 1995). Compared with uninoculated roots, the level of Suc was reduced by 50% in spruce roots colonized by either Amanita muscaria or C. geophilum (Schaeffer et al., 1995). The mycobiont, which lacks sucrolytic enzymes (Schaeffer et al., 1995), could possibly induce Suc breakdown to meet its carbohydrate supply. This could be achieved by inducing a higher acid-dependent Suc breakdown (Schaeffer et al., 1997). It is clear that the carbohydrate metabolism of host roots is regulated by the presence of a fungal partner that is able to induce strong additional carbon sinks (Shachar-Hill et al., 1995; Schaeffer et al., 1997).

Conversely, E. globulus-P. tinctorius ectomycorrhizae contain a high amount of fungus-specific carbohydrates. The disaccharides trehalose and mannitol are the prominently labeled carbon compounds, and it appears that fungal metabolism dominates the assimilation of exogenous carbohydrates into symbiotic tissues. ¹³C incorporation into arabitol and erythritol was nearly 4-fold higher than in the free-living mycelium. Arabitol and erythritol accumulated 6% of the total ¹³C in free-living hyphae, whereas these polyols incorporated 25% of ¹³C detected in symbiotic tissues. Arabitol accumulation in P. tinctorius during spruce ectomycorrhiza formation has also been reported (Ineichen and Wiemken, 1992).

The initial steps in the ectomycorrhizal interaction include the swelling of hyphal tips and the formation of fan-like structures on the root surface (Jacobs et al., 1989; Kottke et al., 1997). The hyphal tip then produces the force to break the root surface and penetrates between epidermal cells to initiate the Hartig net (Gea et al., 1994). The internal turgor pressure is believed to be generated by an influx of water caused by the osmotic gradient produced in the fungal cell (Smith and Read, 1997). It is tempting to speculate that the accumulation of the osmolytes arabitol and erythritol, together with the up-regulation of the synthesis of cell wall hydrophobins that takes place in P. tinctorius during eucalypt mycorrhiza development (Tagu et al., 1996), provides a simple mechanism for plant infection. Arabitol and erythritol may be the compatible solutes responsible for generating the hydrostatic pressure. The rice blast fungus (Magnaporthe grisea) simultaneously accumulates a high amount of glycerol and hydrophobins during the formation of its appressorium (De Jong et al., 1997).

Dark CO₂ fixation by fungal and root carboxylases contributes substantially to fulfilling the demands for carbon compounds in ectomycorrhizae (France and Reid, 1983; Martin et al., 1988; Hampp and Schaeffer, 1995). In free-living mycelium of C. geophilum (Martin and Canet, 1986), S. brunnea (Martin et al., 1988), and P. tinctorius (present study), a large part of the α-ketoglutarate for Glu and Gln biosynthesis is also provided by anaplerotic CO₂ fixation. There was a large accumulation of Gln, which displayed a (¹³C3 + ¹³C2)/¹³C4 ratio in agreement with a high anaplerotic carboxylase activity (Fig. 3B), in symbiotic tissues. The spectrum (Fig. 3B) does not directly show whether the amino acids labeled in mycorrhizal roots are of plant or fungal origin. However, the amino acid labeling (e.g. intramolecular labeling of Gln) in symbiotic tissues was similar to those observed in the free-living fungus, suggesting that most of the labeling took place in the fungus. The high labeling in Gln is in agreement with the known high activity of the Gln synthetase/Glu synthase cycle in eucalypt ectomycorrhizae (Turnbull et al., 1995), and is likely related to the high NH₄⁺ concentration used in the growth medium. However, the proportion of ¹³C allocated to Gln and Ala was lower by 30% and 50%, respectively, in symbiotic tissues than in the free-living mycelium.

In conclusion, the assimilation of [¹³C]Glc in free-living P. tinctorius and E. globulus-P. tinctorius ectomycorrhizae resulted in the production of a large amount of labeled polyols, trehalose, Gln, and Ala, whereas E. globulus roots mainly accumulated Suc and Gln. Ectomycorrhiza development induces striking alterations in the carbohydrate pools, including enhanced synthesis of arabitol and erythritol. Whether this accumulation of polyols is linked to the aggregation of hyphae to form the ectomycorrhizal sheath and/or penetration of root surface by the hyphal tips will await further biochemical and molecular analyses.