Nuclear magnetic resonance (NMR) studies of mycorrhizal symbioses have illuminated a number of functional aspects of these complex associations. Here we review studies of the two main types of mycorrhiza (ectomycorrhizas and arbuscular mycorrhizas) to which NMR has been applied. Although the physiological questions addressed in each case are frequently the same, these two mutualistic symbioses are sufficiently different to justify separate discussion. In conjunction with isotopic labelling NMR is able to examine the transfer of substrates between the symbionts both in vivo and in vitro, as well as the production of secondary metabolites in response to colonization. In addition, this methodology is capable of determining the locations of the biosynthesis and translocations of storage compounds, such as polyphosphates, lipids and carbohydrates, in mycorrhizal fungi both in the free-living and in the symbiotic stages of their life cycle. NMR has been useful in analysing metabolism, transport and energetics, and the results of such studies have practical and ecological significance. Models of transport and physiology to which NMR has contributed form the necessary foundation for functional genomic exploration.
Advances in NMR spectroscopy in recent years have been astounding. With the advent of higher magnetic fields have come much greater sensitivity and spectral resolution. Modern high-speed computers have made the application and analysis of complex multidimensional spectroscopy routine. These innovations add significantly to the usefulness of NMR spectroscopy for compound identification and for studying metabolism, transport and physiology. The impact of NMR in these fields is evident in the considerable current literature that deals with in vivo and in vitro studies of biological systems including mammals, plants, and microorganisms. For reviews of NMR spectroscopic applications to studies of ectomycorrhizal fungi, to plant-microbe symbioses and to plant metabolism in general see Martin (1991), Pfeffer & Shachar-Hill (1996) and Ratcliffe & Shachar-Hill (2001), respectively, where practical aspects of NMR as well as its advantages and limitations for such investigations are discussed.
Some of the most interesting and practically important questions about mycorrhizas concern the nutritional benefits that these symbioses confer on each participant. These benefits hinge on the structural and physiological intimacy between the symbionts, therefore detailed information on metabolism, transport, and functional anatomy of intact systems is required. Although techniques like mass spectrometry and radioisotope labelling can give valuable information and are often more sensitive than NMR, for example in detecting dilute compounds or labelling in metabolically active compounds, they yield much less information on the position of isotopic enrichments within each molecule. Such information frequently allows one to trace in detail the biochemical pathways by which a given molecule was made (Bago et al., 1999; Pfeffer et al., 1999). NMR spectroscopy yields labelling information directly and quantitatively and this is a significant advantage of NMR spectroscopy for metabolic studies.
The potential to spectroscopically differentiate host from fungal metabolites in vivo or in crude extracts without the need for separation or chemical derivatization is another useful aspect of NMR measurements in studying the biochemistry of mycorrhizas. This is because unexpected or labile compounds are more likely to be observed under such circumstances. Finally, the sensitivity of in vivo NMR signals to the intracellular environment has allowed useful insights to be gained about compartmentation, mobility and binding of important molecules. NMR has therefore contributed significantly to such questions as the biochemistry of polyphosphates and the pathways and regulation of carbon and nitrogen metabolism in mycorrhiza, as well as the identification of secondary metabolites made in response to the formation of mycorrhizas and/or to xenobiotics.
Here we review studies on the two types of mycorrhiza to which NMR has been applied: ectomycorrhizas and arbuscular mycorrhizas. Although the physiological questions addressed in both cases are frequently the same, these two mutualistic symbioses are sufficiently different to justify separate discussion. For a thorough review of these symbioses see Smith & Read (1997). Ectomycorrhizal (EC) fungi can grow in a free-living state so we can follow and compare their metabolism both under symbiotic and asymbiotic conditions. In contrast, arbuscular mycorrhizal (AM) fungi are obligate biotrophs that stop growing unless they establish a functional symbiosis with a host root (Azcón-Aguilar et al., 1998). Moreover, culture of arbuscular mycorrhizas in aseptic synthetic media (AM monoxenic cultures) has been achieved only fairly recently (Bécard & Fortin, 1988; St Arnaud et al., 1996). This, together with the fact that the AM fungus only represents a small fraction of the tissue volume in AM roots, has been a serious hindrance to studying this particular system. As a result, much less is known about the metabolism of arbuscular mycorrhizas than about ectomycorrhizas.
Phosphate is at the heart of nutrient exchange in mycorrhizas. Because 31P NMR gives information on the amounts, forms and locations of phosphorus metabolites it has been useful in studying P nutrition in both EC and AM systems. Martin and co-workers carried out the first 31P study of axenically cultured ectomycorrhizal fungi (Cenococcum geophilum and Hebeloma crustuliniforme) in 1983. This report demonstrated the direct observation of polyphosphate (polyP), whose central phosphates gives a distinctive peak at −22 ppm in the 31P NMR spectrum. Such spectra also show that polyP (the form in which phosphate is usually stored by fungi; Jennings, 1995) represents much of the mobile phosphorous present in the EC fungal tissue. This study also showed that polyP contained only a fraction (approx. 8%) of the total phosphate in the mycelium, most of the rest being in phospholipids, DNA and other immobile forms of P that do not contribute to such in vivo31P NMR spectra. In later studies the NMR-observable (mobile, low molecular weight) polyP was found to correspond to a minimum of 80% of the acid-extractable polyP in the mycelium of H. crustuliniforme, and in vivo31P spectroscopy revealed that under phosphorous starvation this fungus degrades polyP to Pi (Martin et al., 1985b). The average chain length of the mobile polyP can be deduced from the ratio of the sizes of the peaks of the terminal and internal phosphate peaks in 31P spectra; thus, in H. crustuliniforme this comparison yielded an average chain length of 11 phosphate residues (Martin et al., 1985b). These values were in agreement with those obtained previously from chemical analysis of polyP extracted from this fungus (Rolin et al., 1984). More recently, polyP chain lengths of a similar size (15 units/chain) have been characterized in Pisolithus tinctorius also by 31P NMR (Ashford et al., 1993).
Further information on the physical state of polyP in axenically cultured EC fungi was derived from measurements of the ‘spin-lattice’ relaxation rate for the 31P NMR signals from polyP (Martin et al., 1985b). This spectroscopic parameter can give information on the mobility of molecules in vivo and the spin lattice relaxation behaviour of polyP was consistent with the formation of ‘supramolecular aggregates’ of these short polyP chains in free-living H. crustuliniforme.
Concerning the symbiotic state of EC fungi, Loughman & Ratcliffe (1984) first demonstrated the presence of polyP, and active incorporation of Pi into mobile polyP in living mycorrhizal beech (Fagus sylvatica) root tips. Subsequently in vivo31P NMR spectra of synthetic mycorrhiza of birch (Paxillus involutus) showed that levels of polyP were somewhat higher (10–40% of the total soluble phosphate) than those of free-living mycelium of the fungus (Grellier et al., 1989). In a study of intact mycorrhizal red pine roots (Pinus resinosa ×Hebeloma arenosa) MacFall et al. (1992) showed that the Pi : polyP ratio remained the same (approximately 1.8) when plants were grown for 19 wk in either P-amended or unamended soil. These findings were interpreted as supporting the idea that the fungus, by controlling the synthesis of polyP, can modulate Pi flow within the mycorrhizal root, and thereby regulate the amounts of Pi available for transport to the shoots. Gerlitz & Werk (1994) also followed the levels of Pi and polyP by 31P NMR in mycorrhizal pine and beech roots. Their findings showed that mobile polyP (chain lengths < 100) increased by 5–10% over a period of 20 min after exposure to Pi and then returned to their initial levels. The authors suggest that during this short period of time incoming Pi is rapidly incorporated into mobile polyP, and that it then becomes NMR-invisible as the chain length increases and/or it becomes immobilized. Recently, in an in vivo31P study of chestnut (Castanea sativa) mycorrhizas Martins et al. (1999) have demonstrated the accumulation of phosphate, mostly in the form of polyP, over a period of 3 months, in contrast to nonmycorrhizal roots, which accumulated only Pi (Fig. 1). The level of orthophosphate in mycorrhizal roots was significantly lower than in nonmycorrhizal ones, and the authors interpreted this in terms of dependence of the host on fungal polyphosphates as a major source of phosphate. The ratio of Pi : polyP in this study was not constant over time, contrasting with the findings of MacFall et al. (1992) where this ratio did not vary.
Taken together 31P NMR studies of ectomycorrhizas and EC fungi provide measurements of short and long-term uptake and incorporation of phosphate into inorganic phosphate and polyP pools. The dynamics and ratios of P in different pools differ between systems and such differences as well as the fact that not all polyP is NMR-visible necessitate care when making physiological and ecological interpretations of some findings. Nevertheless by discriminating different chain lengths and by allowing real-time measurements of different P pools, NMR has provides useful data on the dynamics of P in ectomycorrhizas.
The information available from NMR spectra on the physico-chemical state of small molecules has also proven fruitful in studies of aluminium (Al) uptake by ectomycorrhizas and EC fungi. Under acid soil conditions Al can cause severe disruption in root physiology and function, including P nutrition (Haug, 1984; Pfeffer et al., 1987; Kochian & Shaff, 1991), and mycorrhizas can moderate this toxicity. In a unique study using 27Al NMR Martin et al. (1994b) demonstrated that Al is chelated by polyP within the mycelium of the EC fungus Laccaria bicolor. The presence of four different peaks in the 27Al spectrum showed that the Al3+ ion exists in at least 4 different states and that these are probably confined to the fungal vacuole. These polyP-Al complexes were found to be very resistant to remobilization, even when the mycelium was transferred to Al-free medium. Thus the sequestration of Al by polyP could serve to protect mycorrhizal plants. Subsequently, Gerlitz (1996) demonstrated using 31P NMR that an Al-adapted EC fungus (Suillus bovinus) had a greater ability for phosphate uptake, and contained mobile polyP of shorter chain length than nonadapted fungus. Mobile shorter polyP chains were suggested to complex Al, rendering the fungus and host plant resistant to this ion’s toxic effects.
By comparison with ectomycorrhizas, few 31P NMR studies have been carried out in arbuscular mycorrhizas, mostly because of the much smaller quantities of fungal tissue available from these systems. Axenically germinating spores of the AM fungus Glomus etunicatum contain polyP of quite small chain length (Shachar-Hill et al., 1995) with the 31P signals from central, penultimate and terminal phosphate residues having intensities in the ratio 1 : 2 : 2. This points to an average chain length of 5 phosphate units. Also, and in contrast to what was found for EC fungi – in which polyP dominates (see above) – asymbiotic G. etunicatum was found to contain relatively high levels of Pi, as indicated by the large peak at 2.0 ppm corresponding to an internal compartment at pH 6.9 (Fig. 2). The presence of small, mobile polyP units and of high Pi levels may reflect the active synthesis of phosphorylated metabolites in the germination stage of the AM fungal life cycle. A recent 31P study on G. etunicatum hyphae by Rasmussen et al. (2000) found a significantly lower level of Pi relative to polyP, and this may reflect changes during the development of G. etunicatum since the earlier observations were on spores with very little hyphal material present.
In vivo31P spectra of excised roots colonized by G. etunicatum revealed the presence of a polyP peak (at −22 ppm, Fig. 3), whereas spectra of uncolonized roots did not show this peak (Shachar-Hill et al., 1995; Rasmussen et al., 2000). The absence (Shachar-Hill et al., 1995) or presence of very weak signals (Rasmussen et al., 2000) of polyP from terminal and penultimate residues suggests that polyP chains are much longer (> 50 residues) in symbiotic G. etunicatum than in the asymbiotic fungus (see above). Vacuolar and cytoplasmic pH’s (measured from the position of the 31P peaks of Pi in these compartments) as well as nucleotide levels in colonized and control roots were indistinguishable. The mobile polyP content of the colonized root was estimated as equivalent to approx. 10 mM phosphate in the AM fungus, based on a comparison with nucleotide peaks and the assumption that the mycobiont occupies one-tenth of the volume of the host cytoplasm. Despite the small quantities of AM fungal tissue within the host root, the relatively high concentration and unique peaks of fungal polyP mean that NMR spectroscopy can be used to nondestructively detect the presence of AM fungi and can yield information on the nature of the stored fungal polyP (Shachar-Hill et al., 1995; Rasmussen et al., 2000).
Since the NMR signals of nitrate, ammonium, amine and amide nitrogens are separate and readily identified in the spectrum, 15N-NMR has proven useful for studies of N metabolism. Thus, NMR time course experiments allow the kinetics of labelling in amine and amide positions of Gln and other amino acids (AA) to be straightforwardly monitored. The low sensitivity of 15N-NMR relative to 31P or 13C has however, to date restricted studies to ectomycorrhizal systems where sufficient tissue quantities are available.
It is known that EC symbioses enhance the acquisition by plants of soil nitrogen (Hogberg, 1989), especially that of ammonium (Finlay et al., 1988). However, it is not known if EC hyphae absorb nitrate and ammonium more efficiently than those of colonized roots. Martin (1985) and Martin et al. (1994a) used in vivo15N-NMR to follow the metabolism of ammonium in free-living C. graniforme and L. bicolor, and demonstrated that ammonium is predominantly incorporated by these fungi via glutamine synthetase/glutamate synthase (GS/GOGAT), with a smaller proportion of the N metabolized via glutamate dehydrogenase (GDH). However, under N-limiting conditions or when the inhibitor MSX was used in these EC fungal species, 15N labelling patterns showed that the flux through GDH increased several-fold (Martin, 1985; Martin et al., 1994a), so that the relative contributions of the GDH pathway to N assimilation by EC fungi increases with decreasing extracellular NH4+ concentrations. This is not the case for higher plants, which exclusively use the GS/GOGAT pathway for NH4+ assimilation (Robinson et al., 1991). From the NMR data, it has been suggested (Martin & Botton, 1993) that the activity of both GDH and GS/GOGAT in the assimilation of ammonium by EC fungi accounts for glutamate biosynthesis with little if any contribution from other glutamine transaminase or transamidase activities.
After addition of 15N label to the culture medium, Arg was substantially labelled in C. graniforme, which is consistent with the operation of the ornithine cycle (Martin, 1985). Since signals from labelled Arg were not detectable in vivo but were clear in spectra of extracts of the same cultures of C. graniforme, it was concluded that this AA may form a large, relatively immobile, NMR-invisible complex with polyP located in fungal vacuoles. The author points out that this Arg/polyP association would mean that the movement of polyP within hyphae could be responsible for a joint translocation of nitrogen and phosphorus to the root. This case serves as another example of the usefulness of comparing spectra acquired in vivo with spectra of extracts.
Beech roots forming ectomycorrhizas with Lactarius sp. and Russula sp. gave somewhat different metabolic responses to NH4+ feeding compared to the free-living EC fungi (Martin et al., 1986). Using 15N NMR to monitor the progress of specific labelling of AAs the authors demonstrated that, as in free-living hyphae, the amide-N of glutamine was most enriched, consistent with the operation of the GS/GOGAT pathway. In this case treating the tissue with MSX and albizine (both inhibitors of GS/GOGAT) blocked incorporation of 15N into amino acids, and caused an accumulation of NH4+. Thus it appears that, in contrast to what happens during their free-living state, EC fungi do not utilize the GDH pathway in NH4+ assimilation, that is, symbiosis establishment suppresses fungal GDH. However, the authors caution that it was not possible to obtain a good sampling of the 15N-labelled AA’s from the fungal mantle, therefore fungal AA’s may have been underrepresented in the overall analysis.
The relative insensitivity of 15N NMR has prevented the kind of observations described above for ectomycorrhizas being made on nitrogen assimilation of AM. However, evidence on N metabolism in AM fungi has been obtained using 13C NMR. Arg peaks appeared in 13C-spectra of extracts of G. intraradices spores that had germinated in the presence of 13C labelled substrates (Bago et al., 1999). The 13C labelling patterns indicated that Arg is made by the usual metabolic pathway. In the mycorrhizal state, 13C peaks from Glu were observed in extracts of extraradical hyphae of fed with different 13C labelled compounds (Pfeffer et al., 1999). This and mass spectroscopic studies are consistent with active AA metabolism during the symbiotic state of AM fungi. It is interesting to note that NMR studies of N and C metabolism have inspired the proposal of models for the uptake translocation and transfer of N in both ectomycorrhizas (Martin et al., 1986) and in the AM symbiosis (Bago et al., 2001). These models certainly incorporate information from other methods and remain to be tested by further NMR and other analyses but they serve to show the usefulness of NMR in generating data relevant to the integration of metabolism and transport in mycorrhizal function.
Carbon metabolism and movement
Much information on C metabolism is available from studies of axenically cultured EC fungi (Martin et al., 1985a, 1988). In EC fungi as with some other filamentous fungi (Jennings & Burke, 1990), mannitol becomes highly labelled when labelled glucose (Glc) is supplied in the culture media. The substantial accumulation and turnover of this typical fungal polyol suggests that it probably plays important, if still disputed metabolic/physiological role(s) (Martin et al., 1988; Jennings & Burke, 1990; Koide et al., 2000). These studies also indicated that there is rapid cycling of carbon between mannitol and hexose. This was deduced from the substantial fraction of label supplied as 13C1 Glc that was scrambled between the C1 and C6 positions by the conversion of hexose into mannitol (whose molecular symmetry makes the C1 and C6 equivalent) and then returning to hexose. This is an example of the usefulness of the positional labelling information obtained from NMR in allowing deductions to be made about the likely metabolic pathway by which a metabolite or metabolic pool is formed or turned over.
There is evidence that the establishment of the EC symbiosis brings about considerable modification of carbon metabolism, both in the host root and the mycobiont (Martin et al., 1987; Hampp et al., 1995). Important questions include: the ways in which each partner contributes to the metabolism of carbohydrates; and the extent to which the metabolism of each partner is affected in the symbiosis. A recent study of Eucalyptus globulus × Pisolithus tinctorius mycorrhiza exemplifies how 13C NMR can help address these questions (Martin et al., 1998). Among the findings: sucrose accumulation by the host was decreased in the colonized roots compared with nonmycorrhizal control plants (as previously observed in AM see below); and in the fungus the incorporation of 13C into arabitol and erythritol was more than fourfold higher and the proportion of label appearing in trehalose was also increased compared to the free-living state (Koide et al., 2000). Fungal synthesis of short chain polyols may be involved in generating turgor pressure and thereby mechanical force for entry into, and growth within, the host root. Recent work in pathogenic fungi (Thines et al., 2000) on the regulation of carbon metabolism for turgor production during appresorium formation shows the importance of analysing carbon fluxes as a function of the life cycle.
An important unanswered question about AM fungi is why they do not complete their life cycle in the absence of symbiosis with a host root. Under these conditions AM fungal germ-tubes stop growing without producing new resting spores, even when any one of a large number of C forms is added to the culture medium (reviewed by Azcón-Aguilar et al., 1998). One explanation proposed is that there is a failure or deficiency in C metabolism in the asymbiotic state, so that understanding C metabolism in this stage of AM fungal development is of potential practical importance. However relatively little is known about carbon metabolism in the AM symbiosis, and NMR has been used to identify the forms of carbon taken up and to delineate metabolic pathways active at different stages of the AM fungal life cycle.
Using the natural-abundance 13C NMR spectra from AM spores of several species, Bécard et al. (1991) found that trehalose is the major carbohydrate storage compound. Metabolism in the asymbiotic phase (spore germination and germ tube growth) of the AM fungus G. intraradices has also been investigated using 13C-labelled substrates (Bago et al., 1999). Such substrates can be added to the germination medium, or alternatively the spores’ internal carbon stores can be labelled with 13C by supplying the labelled substrates to the mycorrhiza during sporulation, and the utilization of this internal label can be analysed after germination of the spores. Analysis of the resultant 13C and 1H NMR spectra (Fig. 4) shows that during the asymbiotic phase of this fungus: sugars are made from stored lipid; trehalose (but not storage lipid) is synthesized as well as degraded; glucose and fructose, but not mannitol, can be taken up and utilized; dark fixation of CO2 is substantial; and Arg and other amino acids are synthesized. The labelling patterns are consistent with significant carbon fluxes through gluconeogenesis, the glyoxylate cycle, the tricarboxylic acid cycle, glycolysis, the pentose phosphate pathway, nonphotosynthetic one-carbon metabolism, and most or all of the urea cycle. Therefore most of the basic C metabolic pathways appear to be active in the germinating spore, except for synthesis of storage lipids (triglycerides). This failure or blockage in fatty acid synthesis clearly distinguishes the germinating spore from other phases of AM fungal metabolism (see below) and might be important in AM fungal growth arrest under asymbiotic conditions (Bago et al., 1999, 2000).
13C-NMR spectroscopy has also been used to follow the uptake, metabolism and translocation of C during AM symbiosis. Results of in vivo experiments in AM leek (Allium porrum ×G. etunicatum;Shachar-Hill et al., 1995) indicated that glucose is taken up directly by the intraradical fungal mycelium, that glycogen and trehalose are the main short-term sinks for hexose taken up by the fungus, and that the presence and status of fungus and host strongly influence one another’s metabolism of hexose. In subsequent experiments the labelling of fungal trehalose from exogenous glucose was found to be substantially less when the number of arbuscules was low (Pfeffer & Shachar-Hill, 1996). Thus the presence of arbuscules may be necessary for C uptake and/or is correlated with times of high fungal activity that require more hexose uptake from the host.
Recently the use of in vitro AM monoxenic root cultures (St Arnaud et al., 1996) has allowed longer term labelling experiments without the interference of other organisms and has also facilitated the selective application of labelled substrates to either the extraradical mycelium or to the mycorrhizal roots. The use of this system for NMR studies involves spectroscopic analysis of extracts and is illustrated in Fig. 5. The main analytical advantages of this approach are the selective extraction of mycorrhizal root vs extraradical hyphal tissues and of lipids vs carbohydrates and other water-soluble compounds. Equally important is the more accurate information on labelling levels and on compounds present at lower concentrations that is available from the NMR analysis of extracts compared with spectra taken in vivo. Such experiments (Pfeffer et al., 1999) have been useful in revealing pathways and locations of C metabolism and transfer in AM. These studies have demonstrated that hexose is taken up by the fungus within the host root, converted to triacylglycerol, and then moved to the external mycelium where it is stored in spores or converted to carbohydrate (Pfeffer et al., 1999). Recent data indicate that some of the C acquired by the AM fungus from the host root is also exported to the extraradical mycelium in the form of glycogen (Bago et al., unpublished). Although working models of C metabolism in AM have been developed from the results of NMR experiments and other approaches (Bago et al., 2000), we remain ignorant about the activities of certain important pathways and about metabolic regulation in AM. The analysis by NMR of other carbon sinks such as chitin, nucleic acids and proteins that require hydrolysis for thorough analysis should yield further valuable information.
NMR contributions to models of mycorrhizal function
NMR studies of metabolism and transport have inspired working models of how mycorrhizas function in metabolizing and transporting nitrogen phosphorus and carbon (Martin et al., 1987; Martin & Botton, 1993; Bago et al., 2000, 2001). Such models are of course not based solely on observations made with NMR and indeed some of their features are largely based on data from other methods. However certain features were clearly elucidated by NMR as exemplified in Fig. 6 which shows a combined model of the metabolism, interactions and transport of N, P and C. The model is based on those we have recently proposed (Bago et al., 2000; 2001) and the figure aims to emphasize those features based on findings that were made by NMR.
Xenobiotics and secondary metabolism
31P NMR has been used by Sukarno et al. (1998) to assess the effects of fungicides on AM symbiosis. The study compared the effects of Aliette (fosetyl-Al) on mycorrhizal and nonmycorrhizal onion (Allium cepa) with the effects of the degradation products, Al and phosphonate. One significant conclusion from this study is that the recovery of mycorrhizal plants from the toxic effects of the fungicide is due to the improved ratios of phosphate to phosphonate compared with those in equivalent nonmycorrhizal plants. NMR has proven to be a powerful tool for the study of xenobiotics (Ratcliffe & Roscher, 1998) and further applications to mycorrhizal systems would undoubtedly increase our understanding in this area.
NMR spectroscopy has also played a key role in identifying a number of potentially important secondary metabolites in both ecto- and arbuscular mycorrhizas. For example, using 1H NMR, Zarb & Walters (1994) confirmed the identity of aminopropylcadaverine and N,N bis(3-aminopropyl) cadaverine in the ectomycorrhizal fungi L. proxima, Paxillus involutus, Thelephora terrestris and H. mesophaeum. These studies demonstrated that mycorrhizal fungi can convert lysine into higher homologues of cadaverine. The pathway for the synthesis of these compounds from L-aspartic-semialdehyde and cadaverine was shown to proceed through a Schiff base.
In arbuscular mycorrhizas, Maier et al. (1995) used one- and two dimension heteronuclear 13C and homonuclear 1H NMR to identify a new terpenoid glycoside, blumenin. This compound was continuously accumulated in barley (Hordeum vulgare) colonized by G. intraradices. The accumulation was correlated with the degree of root infection and its concentration is thought to be associated with internal signalling during the development of the symbiosis. By contrast, the levels of four phenolic amides that were later identified by 1H NMR and MS in the same mycorrhizal barley roots, were shown to increase only transiently only during the early stages of colonization. Thus these compounds may have roles in the initial defense response of the host (Peipp et al., 1997) that is initiated but then suppressed during AM colonization. Another secondary metabolite, known as the ‘yellow pigment’, which has been often used to quantify the level of mycorrhizal infection (Becker & Gerdemann, 1977), was identified in AM gramineous plants using 13C and 1H NMR. It turned out to be E-4,9-dimethyldodeca-2,4,6,8,10 pentaenedioic acid, an oxidative degradation product of a carotenoid (Klingner et al., 1995). Maier et al. (2000) have extended their NMR identification of secondary metabolites to six glycosylated C13 cyclohexanone derivatives in various tobacco and tomato species and cultivars. The authors speculate that such compounds may have a role similar to blumenin.
From the identification of metabolites to following metabolic pathways and transport processes in both plant and fungal partners, NMR spectroscopy has contributed significantly to our understanding of the physiology of mycorrhizal symbioses. The studies discussed above illustrate some of the advantages of NMR spectroscopy: the nondestructive chemical analysis of living tissues or crude extracts, so unique fungal and plant metabolites can be detected; the sensitivity of NMR signals to the physico-chemical environment of the molecules involved, so the immobilization or location of small molecules is frequently observable; and the potential to distinguish and quantify labelling levels in different atomic positions of metabolites that are formed upon exposure to substrates labelled with stable isotopes, so the activities of particular metabolic pathways can be deduced. The main limitation of NMR spectroscopy is its insensitivity, and peaks are generally detected only from metabolites or other chemical species present in the sample at concentrations above 10−4M. A significant constraint on the use of NMR spectroscopy in vivo concerns the geometry of samples, so that in situ studies of intact mycorrhizas are not generally feasible and only excised tissues or axenically growing fungus have been studied. As with any method, some of the most exciting future contributions may be expected from the use of NMR spectroscopy in conjunction with other methods such as molecular genetic approaches, enzymology and cytohistochemical microscopy. The isolation of genes responsible for transport and metabolism and determining the location and extent of their expression is particularly important for testing hypotheses and models to which NMR has contributed significantly (Lammers et al., unpublished).