Progress towards understanding the extent to which mycorrhizal fungi are involved in the mobilization of nitrogen (N) and phosphorus (P) from natural substrates is reviewed here. While mycorrhiza research has emphasized the role of the symbiosis in facilitation of capture of these nutrients in ionic form, attention has shifted since the mid-1980s to analysing the mycorrhizal fungal abilities to release N and P from the detrital materials of microbial faunal and plant origins, which are the primary sources of these elements in terrestrial ecosystems. Ericoid, and some ectomycorrhizal fungi have the potential to be directly involved in attack both on structural polymers, which may render nutrients inaccessible, and in mobilization of N and P from the organic polymers in which they are sequestered. The advantages to the plant of achieving intervention in the microbial mobilization–immobilization cycles are stressed. While the new approaches may initially lack the precision achieved in studies of readily characterized ionic forms of N and P, they do provide insights of greater ecological relevance. The results support the hypothesis that selection has favoured ericoid and ectomycorrhizal systems with well developed saprotrophic capabilities in those ecosystems characterized by retention of N and P as organic complexes in the soil. The need for further investigation of the abilities of arbuscular mycorrhizal fungi to intervene in nutrient mobilization processes is stressed.
More than 10 years have elapsed since it was proposed (Read, 1991) that because plants with ericoid (ERM), ecto-(ECM) and arbuscular (AM) mycorrhizas typically predominate under distinctive soil conditions, their respective fungal associates might, in each case, be making contrasting types of contribution to plant nutrition. At the heart of the original proposition was the notion that at the global scale, shifts from lower to higher latitudes or altitudes, through their effects directly upon climate and indirectly upon decomposition, led to fundamental changes in the nature of the soil as a nutrient source for plants. On the basis that these gradients involved progressive increases in the extent to which the major nutrients nitrogen (N) and phosphorus (P) were sequestered in organic forms which were unavailable to autotrophs, it was hypothesised that at any point along its length, natural selection would have favoured associations between plants and those fungal mutualists which were capable of unlocking the two key growth limiting resources.
Soil scientists acknowledge that, even in agricultural soils, mineralisation of N and P involves a sequence of processes for which the living microbial biomass provides the enzymes and dead microbial material much of the substrate (Mengel, 1996; Appel & Mengel, 1998). In general, however, they have not appreciated that a large proportion both of the microbial biomass and necromass of soils in most terrestrial ecosystems consists of the mycelium of mycorrhizal fungi (Finlay & Söderström, 1992; Read, 1992; Högberg & Högberg, 2002) or that access to photosynthate from their autotrophic partners releases this component of the microbial biomass from carbon limitation so providing it with the potential to play a major role in nutrient mobilisation. The consequences of an ability of mycorrhizal fungi to intercede in the processes otherwise leading to release of N and P in ionic form would be considerable, not only in terms of provision for their host plants of preferential access to the nutrients but also because the early intervention in events normally leading to mineralisation would reduce the downstream supply of these elements in the ionic forms required by nonmycorrhizal competitors.
Unfortunately, and probably under the influence of the extensive agricultural literature, research carried out through much of the last century emphasised the roles of mycorrhizas in the capture of mineral nutrient ions (see Harley & Smith, 1983). The possibility that fungal symbionts might be themselves involved in direct attack upon nutrient containing organic polymers, with the exception of the visionary speculation of Frank who proposed the ‘Organic Nitrogen Theory’ in 1894 (Frank, 1894), was largely ignored until the mid-1980s. As a result a view has emerged which sees two entirely separate functional groups of fungi in soil, one, the decomposers, being alone involved in destruction of organic substrates and release of any nutrients contained in them and the other, the fungal mutualists, which absorb mineral ions released by the decomposition processes.
Research over the last two decades has begun to throw considerable doubt both on the validity of any such functional separation between decomposers and mutualists, and upon the view that mycorrhizal fungi have access only to mineral nutrient ions. In so doing it has lent further support to the hypothesis that the different mycorrhizal types are selected according to their abilities to mobilise essential nutrients from polymeric sources. The main distinguishing feature of this period of research is the move away from reductionist approaches which consider only simple mineral ions as possible sources of the major elements, to an emphasis upon the abilities of ERM, ECM and AM fungi to exploit N and P contained in substrates representative of those occurring in their distinctive natural environments. Where model compounds have the potential to provide insights these have been used in the new generation of experiments, but increasingly materials from the rooting zones of the plants themselves have been employed. To the extent that some these of have not received full physico-chemical characterisation, the novel approaches may initially lack some of the precision obtainable when employing simple mineral ions, but it is the contention here (Fig. 1), that the gains in terms of ecological relevance will outweigh the disadvantages.
In this paper the nutritional status of the three major types of mycorrhizal symbiosis is reassessed and factors determining their separate occurrence or coexistence are considered in the light of the new observational and experimental evidence.
Nutrient cycling by ericoid mycorrhizal fungi in heathland ecosystems
Heathlands are characterised on the basis of the predominance of dwarf shrubs of the order Ericales and of the prevalence of acidic raw humus soils in which mineralisation processes are inhibited to the extent that the essential nutrients N and P are present almost exclusively in combination with organic residues. While a number of structural attributes of above-ground tissues, in particular sclerophylly and leaf retention, contribute to nutrient conservation in ericaceous plants, there is a accumulating evidence that it is the ericoid mycorrhizal symbiosis which provides the key to releasing such resources from detrital materials. The epidermal cells of the ‘hair roots’ of the dominant members of these communities are extensively penetrated by hyphae of a small number of distinctive ascomycetous fungi. Amongst these the most widely occurring and most extensively investigated are strains of the Hymenoscyphus ericae-Scytalidium vaccinii complex (Helotiales) (Egger & Sigler, 1993; Read, 1996) and of Oidiodendron maius (Onygenales) (Hambleton et al., 1998). The mycorrhizal cells, and their emanating hyphae, occupy an interface between the potential nutrient resources of the raw humus within which the roots proliferate and the body of the plant. The fungi forming ericoid mycorrhiza can be readily cultured enabling experimental studies of their abilities to degrade organic polymers. These have confirmed that ericoid fungi have considerable saprotrophic capabilities, which, if expressed in soil, would be expected to facilitate nutrient mobilisation from the residues exploited by the mycorrhizal roots and mycelium. For the most part, such studies have used model compounds and so describe only the potential of the mycelium for in situ mobilisation. They are none the less important because they emphasise the biochemical versatility which is accessible to the ericaceous plant as a result of its selective association with this group of fungal symbionts.
Mobilisation of N and P from ‘model’ compounds
Two classes of polymer degrading enzymes can be recognised, one which provides the potential to cleave those organic molecules such as lignins, polyphenols and tannins which may protect or precipitate essential nutrients (Table 1) and another which facilitates direct attack upon the nutrient-containing molecules themselves. The former contains lignases and polyphenol oxidases which are of particular interest because they would be expected to contribute in a major way to decomposition as well as to plant nutrition. Despite the demonstration (Haselwandter et al., 1990) using respirometric methods, that H. ericae was capable of ring-cleavage of aromatic constituents of lignin, subsequent searches (Bending & Read, 1996b, 1997; Burke & Cairney, 1998) have failed to find evidence for the production of the enzymes lignin (LiP) and manganese (MnP) peroxidase which would be required to enable direct attack, of the kind achieved by ‘white rot’ fungi, upon the aromatic structures which make up the polymer. However, recognition that H. ericae releases hydrogen peroxide (Bending & Read, 1997) and hydroxyl radicals (Burke & Cairney, 1998) into culture media confirms that the fungus may contribute to a form of lignin degradation, similar to that seen in ‘brown rot’ fungi, in which the radicals facilitate fragmentation of the lignin polymer by demethoxylation and side chain oxidation (Burke & Cairney, 1998; Cairney & Burke, 1998). It is of interest that in addition to their impacts upon lignin, hydroxyl radicals have been implicated in the degradation of cellulose by ‘brown rot’ fungi (Joseleau et al., 1994).
Table 1. Extra cellular enzymes, known to be produced by the ericoid mycorrhizal fungus Hymenoscyphus ericae, which would be expected to provide the ability to degrade structural components of plant litter, thereby contributing to decomposition processes and to ‘unmasking’ of nutrients
Italics indicate results based upon indirect methods of observation. For older literature see Leake & Read (1997).
Polyphenolic compounds also characteristically reach high concentrations in the heathland soils (Jalal & Read, 1983). Enrichment of their surface horizons by residues rich in phenolic compounds which can reach over 30% of tissue dry weight in plants such as Calluna vulgaris (Jalal et al., 1982), is an inevitable consequence of the accumulation of these tertiary metabolites in ericaceous plants. It has implications both for pedogenic processes and for the plants which forage for nutrients within such soils. A key to understanding the nutritional contribution of ericoid mycorrhizas in these phenol rich environments may lie in the now proven ability of their fungal partners not only to use many monomeric phenolic compounds as carbon sources (Leake & Read, 1989) but also to release the enzymes laccase and catechol oxidase (Bending & Read, 1996a,b, 1997) that are involved in the degradation of hydrolysable polyphenols. These attributes, which are developed far more extensively in ericoid than in ectomycorrhizal fungi so far examined (Bending & Read, 1996a,b), are now thought to be of significance from the perspectives both of raw humus formation and plant nutrition. When the fungus is grown on culture media containing, as sole carbon source, the model polyphenol tannic acid (Bending & Read, 1996a), it utilises the carbon contained in the substrate, and releases as a residue, dark-coloured reactive quinone-like compounds which are thought to be the precursors of the recalcitrant humic and fulvic acid polymers that form the bulk of raw humus (Stevenson, 1982).
From the nutritional standpoint it is the propensity of phenolic compounds to form complexes with substrates containing the major elements nitrogen and phosphorus which is likely to be of greatest significance. Here, the abilities of H. ericae to assimilate individual phenolic monomers thus reducing the extent of complexation and, by cleaving polymers, to secure release of nutrients coprecipitated with polyphenols must both be considered. Bending & Read (1996b) demonstrated that, in addition to degrading tannic acid-protein precipitates by releasing polyphenol oxidase, H. ericae was able to acquire nitrogen from these complexes through the expression of extracellular acid protease activity. This property, which the fungus shares with Oidiodendron, was not shown by any of five ectomycorrhizal fungi examined. Since it became recognised (Bajwa et al., 1985) that ERM fungi could mobilise the N contained in model proteins and assimilate the released amino-acids, it has been revealed that representatives of other functional types of plants which are often associated with ericoid shrubs in tundra and boreal forest environments, notably in the Cyperaceae (Chapin et al., 1993; Kielland, 1994; Raab et al., 1996) and Gramineae (Näsholm et al., 1998; Näsholm & Persson, 2001) could readily assimilate amino-acid N. Such observations indicate that the ability to absorb amino acids and hence to bypass the N mineralisation process may be widespread in organic soils. However, they should not be allowed to obscure the essential ecophysiological distinction between the capacity for amino-acid scavenging which is an attribute of some nonmycorrhizal as well as mycorrhizal plants, and the ability conferred upon ERM (and some ECM – see below) plants by their mycorrhizal symbionts, to carry out direct attack upon the sources of soil amino-N compounds, which are the proteins themselves. It has been proposed (Read, 1996) that the overwhelming predominance of the ericoid mycorrhizal functional group in heathland ecosystems may provide some indication of the relative effectiveness of the two strategies for N acquisition.
Mobilisation of N and P from ‘natural’ substrates
While experiments using model compounds as surrogate substrates can be enlightening there is clearly a need to consider naturally occurring materials. In species-poor heathland systems the residues reaching the soil are derived largely from the ericaceous plants themselves and their fungal symbionts. The shoot residues form a distinct litter (L) layer at the surface whereas the fine root and fungal residues are concentrated immediately below this in the fermentation (F) layer in which they selectively proliferate while alive. Ideally, these materials would be harvested and used as substrates in comparative analyses of the growth of mycorrhizal and nonmycorrhizal plants but because they carry inoculum of ERM fungi, sterilisation is required in order to maintain control plants in the nonmycorrhizal condition. Since all forms of sterilisation so far employed on these organic materials have produced changes in their quality such approaches have little ecological appeal.
Faced with these difficulties advances towards relevance have been made by a multifaceted approach using the microbial and plant tissues themselves, produced aseptically, and killed naturally by air drying, to provide ‘necromass’, this then being supplied as the sole course of these elements to axenically grown M and NM plants (Fig. 1).
Analysis of the nutritional role of the fungal cell wall polymer chitin, which is known to be an important potential N source in raw humus soils (Kerley & Read, 1998) has been achieved using a combination of these approaches. It was first demonstrated (Leake & Read, 1990a; Mitchell et al., 1992), using crustacean chitin as a model compound, that H. ericae could cleave the polymer to its constitutent subunits N-acetyl glucosamine and N-acetyl galactosamine, both of which were readily assimilated by the fungus. A significant proportion of the N so acquired was subsequently shown (Kerley & Read, 1995) to be transferred to aseptically grown mycorrhizal plants of Vaccinium macrocarpon. Natural sources of the polymer, in the form of mycelial necromass of H. ericae itself, were later used as substrates (Kerley & Read, 1997). Hyphae of the fungus were supplied to M and NM plants either as purified cell wall fractions or in the unfractionated form as sole sources of N. In the M condition plants gained access to the N contained in both pure and unfractionated wall fractions, their yields and N contents on the former being almost as high as those on the latter. The ability to effectively recycle N originally sequestered in the fungus will contribute to a lessening of competition between the partners in the symbiosis. Similar mycorrhiza-dependent pathways for nitrogen transfer were demonstrated (Kerley & Read, 1998) using aseptically produced shoot and root necromass of the ericaceous plants themselves.
Clearly, caution must be exercised when interpreting results obtained from use of aseptically produced necromass as nutritional substrates. As they have not been subject to the biological and physico-chemical processes associated with normal senescence and decomposition, these materials are not truly representative of those found in the litter layer. However, as surrogates for such materials they can contribute new insights when used as part of a broader screening of natural substrates.
Relatively little attention has been paid to the acquisition of P from polymeric sources by ericoid plants. However, both DNA in purified form (Leake & Miles, 1996) and entire nuclei (Myers & Leake, 1996) have been shown to constitute useable sources of the element for V. macrocarpon providing the plants are in the mycorrhizal condition. These polymers were selected as surrogates for the organic P sources which must be continuously released into soil as the tissues of microbial and plant populations senesce. In these cases mobilisation of P is dependent upon production by H. ericae of phosphodiesterases (Leake & Read, 1997).
Collectively, these studies confirm, using model compounds, naturally occurring components of heathland ecosystems, or their intermediates, that ERM colonisation will enable intervention in the immobilisation processes of soil N and P cycles and facilitate a direct pathway for transfer of these key growth limiting elements to the autotrophs.
Nutrient cycling by ectomycorrhizal fungi in boreal and temperate forest ecosystems
It is a prerequisite of any treatment of the nutritional capabilities of ECM's to recognise that over 5000 species of fungi are probably capable of forming symbioses of this type (Molina et al., 1992), and that only a fraction of this number are readily culturable and physiologically characterised. Hence if we generalise we do so on the basis of knowledge of a small number of unfastidious species that can be easily manipulated under laboratory conditions. Since even these, which are likely to be ‘ruderal’ strategists, have been shown to express a range of ‘decomposer’ capabilities (Table 2) there is the likelihood that the full potential of this functional group of organisms is greatly underestimated. Awareness of this limitation has increased in recent times in particular because DNA fingerprinting methods (White et al., 1990; Gardes & Bruns, 1993, 1996; Gardes et al., 1991) have revealed the importance of groups of fungi which have been hitherto overlooked because they produce inconspicuous hypogeous or resupinate fruit bodies. These can occupy a large proportion of the ECM root population in boreo-temerate forests (Baar et al., 1999; Erland & Taylor, 1999; Kõljalg et al., 2000; Dahlberg, 2001). Amongst the basidiomycetes, members of the genera Tomentella(Kõljalg, 1996), Tomentellopsis (Thelephoraceae) (Kõljalg et al., 2000) and Tylospora (Corticiaceae) (Taylor et al., 2000) are prominent. The new findings are challenging, nor least because many of the fungi involved have hitherto been regarded as decomposers of woody debris. If these organisms expressed the ability to degrade structural and nutritional polymers when in the mycorrhizal condition, the conventional distinction between decomposers and symbionts as identifiable functional groups would be as blurred in ECM as they are in the ERM. Some molecular phylogenetic studies of ECM fungi are already pointing in this direction (Hibbett et al., 2000). Clearly there is an urgent need to identify the roles played by these recalcitrant fungi in the nutrient economies of forest systems. It is to be expected that in addition to discovery of distinctive hydrolytic capabilities in some of these organisms, there may be others which are found to lack these attributes, so conforming more closely to the conventional view of ECM fungi as being largely nonsaprotrophic.
Table 2. Extra cellular enzymes, known to be produced by selected ectomycorrhizal fungi, which would be expected to provide some abilities to degrade structural components of plant litter thereby contributing to decomposition processes and to ‘unmasking’ of nutrients. Italics indicate observations based upon gene presence rather than enzyme expression. For older literature see Leake & Read (1997)
Italics indicate results based upon indirect methods of observation.
An emerging possible solution to the problem of unculturability is to extract and sequence DNA from dried basidiomes in the search for the genes which encode key enzymes. Using this approach, Chen et al. (2001) screened for the presence of LiP and MnP genes in more than 40 ECM fungi. Around 68% of the species examined were found to possess at least 1 LiP gene and several species including Piloderma croceum, Cortinarius rotundisporus and Tylospora fibrillosa yielded the MnP gene, which had previously been identified in T. fibrillosa (Chambers et al., 1999). On the basis of such findings Chen et al. are justified in drawing attention to the potential of these fungi for decomposition of lignin in temperate forests, but it is only a potential! The need to establish and quantify the extent of gene expression is paramount. Timonen & Sen (1998) examined gene expression in identified functional components of entire mycorrhizal systems associated with Pinus sylvestris. They observed differential expression of isozymes, the activities of some of which, notably polyphenol oxidase and acid phosphatase, were markedly increased in the hyphal fronts of P. involutus and S. bovinus mycelial systems as they advanced into forest humus. Observations made under these more ecologically realistic circumstances have much in their favour. There is the potential to extend them to recalcitrant fungi some of which can successfully be grown with plants in microcosms even though they are not readily cultured in vitro (Kõljalg et al., 2002). Some caution will still be required when interpreting the results of such studies since up-regulation of genes is not necessarily an indication that their products will be expressed in the external environment.
Of even greater importance than the essentially qualitative questions of gene or enzyme presence, are the quantitative issues relating to the extent of activity of the enzymes in the environment. When comparisons are made between levels of expression of those exoenzymes likely to be deployed in decomposition of complex organic polymers by ECM and by saprotrophic fungi, the latter group invariably show greater activity when the fungi are grown under identical conditions (Maijala et al., 1991; Bending & Read, 1996a, 1997; Colpaert & Van Laere, 1996; Colpaert & Van Tichelen, 1996). This suggests that in competitive circumstances ECM fungi will be relatively poorly equipped to exploit such substrates. Their prevalence in specific parts of the soil profile, notably the fermentation horizon (FH) horizon, is likely to reflect preferential exploitation of substrates of a particular quality as reflected by their C : N ratio (Read, 1991). Substrate specialisation would lead to avoidance of competitive relationships with specialist decomposers. The nutrient foraging interactions between one such group, the white-root fungi, and ECMF are for the first time now being investigated under seminatural circumstances (Lindahl et al., 1999; Leake et al., 2001, 2002) and more work is required before generalisations about their outcomes can be made.
Mobilisation of N and P from ‘model’ compounds
As with studies of ERM, research on the roles of ECM in plant nutrition, have increasingly focussed in recent times on the ability to mobilise resources from model polymers. Recognition of the proteolytic capabilities of ERMF in the mid 1980s led to a search for these properties in ECM systems (Smith & Read, 1997; Chalot & Brun, 1998). Model proteins have since been extensively used (Leake & Read, 1990b, Finlay et al., 1992; Ryan & Alexander, 1992; Turnbull et al., 1995) as substrates to determine the potential of ECM fungi and their plant associates to mobilise N from polymeric macromolecules. The main extracellular acid carboxy proteinase involved in the mobilisation was purified and characterised (El-Badaoui & Botton, 1989; Zhu et al., 1990). While the ECM fungi so far examined have generally demonstrated lower levels of proteolytic activity than those seen in ERM fungi, and some, the so-called ‘nonprotein’ fungi express no such activity, many have been shown readily to cleave the protein and to facilitate transfer of the N contained in it, to mycorrhizal plants. Some are also able to gain access to N contained in protein which had been precipitated with phenolic acid (Bending & Read, 1996a) but again not to the extent seen in ERM.
There is evidence from studies along a north-south transect through Europe (Schulze et al., 2000; Taylor et al., 2000) that both the proteolytic capabilities (Fig. 2a,b) and the biodiversity (Fig. 2c,d) of ECM fungal communities are greater in raw humus soils of northern boreal forests, where nitrification is undetectable, than in more southerly locations where mineral N enrichment occurs, either as a result of natural or anthropogenic inputs. Taylor et al. (2000) used the plant protein gliadin as a model to screen for proteolysis in 31 different isolates obtained from roots or carpophores along the gradient. In the more southerly environments accumulation of nitrate appeared to select in favour of a relatively small number of fast growing species with unspecialised nutrient requirements (R strategists) at the expense of those with the ability to mobilise N from the recalcitrant source (S strategists). Studies over more localised gradients of N deposition in North America (Lilleskov et al., 2002a,b) appear to show similar selective effects. According to Tibbett et al. (1998) strains of the ECM fungus Hebeloma originating from cold environments have inherently greater proteolytic potential than those from temperate regions. Such observations, coupled with others (Tibbett et al., 1999) indicating that in the same genus a thermal optimum for activity of protease may be as low as 0–6°C, suggest that selection favours ‘protein fungi’ in arctic and boreal environments where low temperatures, acidity and poor resource quality combine to inhibit N mineralisation.
In theory, the natural abundance of the heavier isotope of N, δ15N, in plant and fungal tissues should reflect that of the nitrogenous substrates on which the organisms have been grown. Some reports (Michelsen et al., 1996) have indeed suggested that the N signatures of field grown plants with ERM, ECM and AM colonisation demonstrate the hypothesised preferences for organic vs mineral N sources in natural soil. However analyses using amino acids and protein of predetermined δ15N signatures as model substrates for fungi and plants representative of these mycorrhizal groups (Emmerton et al., 2001a,b) have dampened earlier enthusiasm for this indirect approach to identification of substrate exploitation. It was shown that extensive fractionations of substrate N occur in the course of assimilation by the organisms and that these are of sufficient magnitude to distort the signatures seen in sink tissues.
In view of the fact that, as in the case of N, most of the soil P in boreo-temperate systems is in organic combination, there is a lamentable ignorance of the extent to which, and the mechanisms whereby, these sources are accessible to plants. Inositol hexaphosphate (IHP) used as a model polymeric source of P, has been shown to be accessible to a number of ECM fungi (Antibus et al., 1992). Since IHP is believed to be one of the major repositories of P in soil organic matter, albeit not in pure form, use of this polymer provides information of direct ecological relevance.
Mobilisation of N and P from ‘natural’ substrates
In recent times, attention has been turned to assessment of the abilities of ECM systems to access N and P sequestered in naturally occurring rather than ‘model’ substrates. Recognition of the propensity of these fungi to form dense mycelial mats when encountering particular types of soil organic matter both in the field (Hintikka & Naykki, 1967; Cromack et al., 1979; Griffiths et al., 1991) and in microcosms (Read, 1991; Bending & Read, 1995a,b) has led to analyses of the microbial dynamics of these areas of enhanced fungal growth. Ingham et al. (1991) showed that mat soils had higher levels of microbial biomass while Cromack et al. (1988) observed that populations of protozoans, nematodes and microarthropods in mat and nonmat soil were qualitatively and quantitatively different. Using microcosms it has also been shown that the patterns of distribution and the population structures of bacterial communities associated with ECM mycelium are distinctive (Nurmiaho-Lassila et al., 1997). Recognition that diverse communities of organisms are apparently selectively present in association with ECM mycelia represents an important conceptual advance but much remains to be learned about the contribution of the different trophic groups to nutritional dynamics of these systems. Assays of enzyme activity have confirmed enhanced levels of cellulase, peroxidase, proteinase and phosphatase in mat relative to nonmat soils (Griffiths et al., 1991) while in microcosms, Bending & Read (1995b) observed higher levels of protease and polyphenol oxidase together with short-term elevation of phosphatase activity in patches of soil occupied by dense growth of Paxillus mycelium, relative to that seen in adjacent unoccupied soil.
Occupation by dense mats of mycorrhizal mycelium, whether of soil in nature (Entry et al., 1991) or material collected from the FH and placed as discrete blocks in microcosms (Bending & Read, 1995a Perez-Moreno & Read, 2000) is associated with reductions in N and P contents of these substrates. Entry et al. (1991) reported reductions of both these elements of around 33% while in adjacent nonmat soils over the same period losses were around 17%. Using the microcosm approach Bending & Read (1995a) observed that the extent of depletion of N and P in FH material differed between fungal symbionts. When growing in association with Pinus sylvestris, dense patches of Suillus bovinus mycelium produced reductions of c23% in both N and P concentrations of pine litter, while Thelephora terrestris reduced N contents by only 13% and failed to provide a significant reduction of P. In these cases analyses of control litter samples incubated free of ECM fungal colonisation showed that mineralisation was insufficient to account for gains of N or P seen in the plants.
The importance of differences between combinations of host plants and fungi is further emphasised by the study of Perez-Moreno & Read (2000). In this case a similar experimental design was employed but with a wider range of litter types and Betula pendula-Paxillus involutus as the photo and mycobionts (Fig. 3a,c). As in the earlier study (Bending & Read, 1995a) exploitation of the FH materials by the mycorrhizal fungi facilitated significant increases of yield (Fig. 4a), photosynthetic area (Fig. 4b), N and P acquisition (Fig. 5) in colonised relative to nonmycorrhizal plants. However, in this case amounts of P extracted from the litters were considerably greater than those of N compared with respective controls. Values of P reduction ranged from 35% in pine litter to c40% in that of birch, these being not dissimilar to those reported for mat soils by Entry et al. (1991). The lesson to be learned from such observations is that even within the very small proportion of ECM plant-fungus combinations that have been examined to date, there is a wide range of nutritional attributes. Further, in view of the likelihood that so many species of fungi are capable of forming ECM it is necessary to be cautious when generalising on the basis of particular plant–fungus combinations.
While organic residues collected from soil horizons like the FH in which ECM mycelia and roots selectively proliferate have the advantage of being realistic candidates as nutritional substrates, they suffer the disadvantage that their chemical constituents are poorly characterised. Hence, again, precision is lost in the service of relevance. However, some of the components that contribute significantly to the nutrient fund of FH can be identified, isolated and chemically characterised and so have the potential, when used as ECM substrates, to provide both greater precision and relevance. Amongst these, mycelial biomass of mycorrhizal fungi, pollen, and the cadavers of organisms making up the microfaunal community have now been examined as potential substrates.
Andersson et al. (1997) grew the ECM fungus Suillus variegatus on a diet of 15N enriched ammonium and used the lyophilised and powdered mycelium as a source of organic N for pine plants which were either colonised by P. involutus or nonmycorrhizal. The mycorrhizal plants acquired significantly more 15N from the labelled source than did their nonmycorrhizal counterparts. Though this study did not employ the fractionation of the mycelium carried out by Kerley & Read (1997) in their study of ERM’s, it's results suggest that ECM colonisation will enable, at the very least, recycling of the N contained in the more labile organic components of the fungal mantle and external mycelium.
Pollen is deposited annually in huge quantities onto soils of all those ectomycorrhizal forests, such as those of the boreal and temperate regions, which are dominated by anemophilous tree species. Only a minute fraction of this pollen is involved in the reproductive process the remainder, the production of which constitutes an enormous nutritional cost in such impoverished environments, being returned to the soil. Estimates of the amount of this annual pollen ‘rain’ range from 10 to 80 kg ha (Koski, 1970; Lee et al., 1996). Since its nutrient content is well characterised in terms both of quantity (N 2–3%, P 0.4%) and quality (90% of the N and most of the P being in organic form) (Greenfield, 1999; Oleksyn et al., 1999) it is possible to estimate the annual depositional inputs involved for each macro element. At 80 kg ha these are equivalent to 1.6 kg N and 0.32 kg P ha year. Stark (1972) hypothesised that the pollen rain would constitute a significant seasonal source of N and P for litter decay fungi and Hutchison & Barron (1997) observed that 41 of 147 fungal saprotrophs could use pollen as a nutrient resource, but until recently the possible contribution of ECM fungi to the processes of nutrient recovery by the trees had not been considered.
Perez-Moreno & Read (2001a) supplied pollen as the only major potential nutrient source and at levels calculated to be representative of those seen in nature, to microcosms supporting plants of Betula pendula which were either mycorrhizal with Paxillus involutus or nonmycorrhizal. P. involutus exploited the pollen intensively producing typical mat-like mycelial surfaces over the substrate (Fig. 3b) the N and P contents of which were reduced by 75 and 97%, respectively, after 115 days (Fig. 6). Of this 29% of the N and 25% of the P were transferred to the plants. In nonmycorrhizal microcosms, only 42 and 35% of N and P, respectively, were lost from the pollen, presumably as a result of export by fungal saprotrophs, and only 12 and 7%, respectively, were transferred to the plants (Fig. 7). It was concluded that the contribution of ECM fungi to recovery of the nutrients invested by plants in their annual cycle of pollen production might have a major impact on sustainability both of their growth and reproductive capabilities.
While it has been recognised by those carrying out research on mycorrhizas that the soil fauna represents a very large biomass it has been the practice to emphasise the potential of these organisms, particularly through their mycovorous activities, to have adverse affects upon the functioning of the symbiosis (Fitter & Sanders et al., 1992; Gange & Brown, 2002; Gehring & Whitham, 2002). However, recent studies involving two of the largest faunal groups of forest soils, nematodes (Perez-Moreno & Read, 2001) and collembolans (Klironomos & Hart, 2001) have emphasised the likely importance of ECM fungi in facilitation of recycling of the nutrients contained in faunal necromass. Perez-Moreno & Read (2001b) employed pure cultures of Heterorhabditis megadis which had been killed by air drying. This necromass was supplied as the major potential source of N & P for mycorrhizal and nonmycorrhizal plants of Betula in microcosms (Fig. 3d). Based upon the estimates of Petersen (1982) that the biomass of nematodes in temperate forest soils ranged between 15 and 3000 mg dry weight m2, aliquots of nematode necromass equivalent to 2220 mg m2 were added to mycorrhizal and-nonmycorrhizal microcosms as sole major sources of nutrient. In mycorrhizal systems Paxillus involutus reduced N and P contents of the necromass by 68 and 65%, respectively, in 195 days (Fig. 8). Equivalent values for nonmycorrhizal plants were 37 and 24%, respectively. N and P gains by mycorrhizal plants were approximately double those seen in nonmycorrhizal microcosms (Fig. 9). Gains of nitrogen were particularly high and could not be accounted for simply in terms of N added in the necromass. In this case, the possibilities that exploitation of nutrients contained in the nematodes enabled more effective scavenging of surrounding peat by an invigorated mycelium, or that N-fixing activities had been enhanced both require further investigation.
Because the N and P requirements of boreal forest stands are known (Cole, 1986) it is possible, with a number of assumptions, to calculate the approximate contributions that mobilisation of these elements from pollen and nematodes by mycorrhizal fungi could make to the annual budgets of trees in these ecosystems. Taken together, values of 12.3% of N and 15.3% of P requirement are obtained (Table 3). Since these are only two amongst the wide range of natural substrates, which are potentially accessible to ECM, it is possible that a considerable proportion of the plants annual requirement of both N and P could be satisfied through breakdown of organic polymers by their fungal symbionts.
Table 3. Calculated contribution of mycorrhizal mobilisation to annual N and P requirement of boreal forest systems
Annual production dry matter kg/ha−1/yr−1
Equivalent annual input of nutrients kg/ha−1/ yr−1
Estimated nutrient mobilisation by mycorrhiza(4) kg/ha−1/yr−1
Contribution of mycorrhizal to mobilisation to annual elemental requirement (%)(5)
Klironomos & Hart (2001) suggested that the fungus Laccaria bicolor not only facilitated transfer of N from collembolan cadavers to its mycorrhizal partner but that the fungus might be directly predating the animals. The mechanisms whereby the presence of L. bicolor in microcosms induced increased death rates in the collembolan population were not determined but the observation is fascinating, reversing as it does previous conceptions of the nature of the relationship between this group of animals and mycorrhizal fungi.
Nutrient cycling by arbuscular mycorrhizal fungi in grassland ecosystems
The gradient based explanation for distribution of mycorrhizal types (Read, 1991) envisaged a syndrome of effects in warmer climates, leading to the replacement of nitrogen by phosphorus as the main growth limiting plant nutrient and hence to the selection of systems dominated by the AM symbiosis with its known propensity for P acquisition. Amongst the most important drivers of the shift from N to P limitation were considered to be, on the one hand, increased rates of mineralisation and nitrification with release of N in the form of mobile nitrate ions, and, on the other, the sequestration of phosphate ions with salts, associated with greater rates of evapotranspiration and/or raised pH. This view was in accordance with the perception that AM fungi, in contrast to their ECM and ERM counterparts, lacked the saprotrophic capability to enable N mineralisation but possessed effective P-scavenging attributes. While little has emerged over the past 10 years to radically change this view, a refreshing new awareness has emerged of the possibility that, under some circumstances, AMF may be involved both in decomposition processes and in the capture of the less mobile amino-acids or ammonium ions. Hodge (2001) added 15N/13C labelled glycine as a localised organic N source to mineral soil and grew plants with 3 different AMF species or in the nonmycorrhizal condition. None of the fungi responded to the presence of the glycine by hyphal proliferation and neither N nor C capture were increased in the M relative to NM plants. By contrast, experimental analysis of ammonium uptake indicates that both capture and transport of this ion to the plant can be enhanced by AMF (Mäder et al., 2000). Provision of access to ammonium ions may well be of ecological importance for plants in biomes with acidic organic soils, such as those prevailing in some tropical environments, which can support abundant AM colonisation (Moyersoen et al., 2001).
There is a single report (Hodge et al., 2001) of enhanced decomposition and increased N capture by AMF from organic necromass. Dual 15N/13C labelled grass leaves were supplied as an organic residue which could be colonised by fungal mycelium extending through a mesh barrier from colonised roots. AM hyphae were shown to facilitate enhancement of N capture from the litter, the N gain in the plants being linearly related to hyphal density in the organic matter. Enhanced release of 13C in the AMF compartment was also indicative of increased decomposition rates. Since analysis of the microbial community structure in the organic patches showed no qualitative differences between AMF colonised and uncolonised material it was concluded that the fungal symbionts may be directly involved in the decomposition processes. The possibility remains, however, that allocation of photosynthate by these fungi might facilitate enhancement of the activity of microbial generalists in the AMF compartments. Since AM mycelia provide extensive conduits for allocation of carbon to the soil system (Johnson et al., 2001, 2002) the extent to which they act directly, rather than as facilitators, of decomposition and nutrient release urgently needs to be determined. Here, as in the case of ERM systems (Kerley & Read, 1998) studies under monoxenic conditions may offer a fruitful way forward.
In view of the preoccupation with phosphorus capture shown in studies of the AM symbiosis over the years there have been remarkably few analyses of the ability of AM fungi to mobilise P from the organic residues which, even in the mineral soils that frequently support AM communities can be the main potential source of the element for plants. The deficiency is compounded by the fact that so many studies of AM systems do not provide adequate discrimination between nutrient mobilising activities of the roots and their fungal symbionts. Tarafdar & Marschner (1994) separated the two activities by allowing an AM fungus to grow from a host plant (Triticum aestivum) through a root-excluding mesh. Throughout the hyphal compartment so obtained, acid phosphatase activities were much higher in the mycelial compartment and a strong correlation between this activity and AM hyphal length provided circumstantial evidence that the enzyme was produced by the fungal symbiont rather than by the background microflora. In this study, of the total P uptake by the plant, the mycorrhizal contribution accounted for 22–33% when P was supplied in inorganic form and 48–59% with P in organic form. It should be said, however, that when Joner et al. (1995) repeated this type of experiment using a similar design they could find no influence of the presence of AM hyphae on soil phosphatase activity in spite of high hyphal densities in the mycelial compartments of their microcosms. A failure to detect increases in phosphatase activity in AM mycorrhizal systems is in some ways surprising since it has been shown (Koide & Kabir, 2000) under axenic conditions that the extra radical hyphae of one such fungus, Glomus intraradices, was readily capable of hydrolysing exogenously supplied organic P sources. Similarly, Joner et al. (2000), again under circumstances where other microbes were excluded from the assay system, showed that an AM fungus was capable of mobilisation and transfer of 32P from the organic compound AMP to a plant.
Many more studies of this kind are required before we will be in a position adequately to discriminate between direct and facilitated attack on soil organic resources by AM fungi. Of particular value would be analyses of P starved tropical rain forest ecosystems in which AM roots, probably colonised by so-far undescribed fungal symbionts, are known to be intimately associated with decaying organic matter (Newbery et al., 1997; Moyersoen et al., 1998).
The conventional view that all plants are dependent for their supplies of N and P upon the mineralising activities of microbial generalists is increasingly challenged by the emerging evidence. Rather, it appears that in environments where rates of mineralisation are sufficiently slow to threaten the fitness of autotrophs, selection has favoured associations with fungal symbionts that are physiologically equipped to facilitate capture of these elements from their locally predominant organic sources. These nutrient mobilising attributes are selectively deployed largely in superficial soil environments where attacks upon substrates of particular quality can pre-empt both the immobilisation of N and P by saprotrophs, and their incorporation into the heterocyclic humic compounds which are the long-term repositories of these elements if they reach the lower parts of the profile. The mechanism is one in which plants invest carbon, a currency not in short supply for most autotrophs, to enable mobilisation, by mycorrhizal fungi, of the fitness threatening resource.
The broader implications of this relatively recent appreciation of nutrient cycling processes in ecosystems require emphasis. Though the proportion of plant species associated with the ERM or ECM fungi known to be capable of these activities is relatively small, perhaps around 10% of those examined to date (Trappe, 1987), as dominants of the tundra, taiga, and boreo-temperate forest zones, they cover up to 70% of the terrestrial surface of the northern hemisphere and exploit soils containing the largest global capital of organic carbon (Post et al., 1982). By occupying the interface between the plants which are the sites of carbon fixation, and the soil in which the element is stored, these fungi have the potential profoundly to influence the carbon source-sink relationships upon which global climate systems ultimately depend. In particular through their involvement in the removal of N and P from organic polymers they inevitably increase the C : N and C : P ratios of the residual materials and thus will contribute to C retention in soil.
The major potential sources of N and P are the residues of the locally dominant symbiotic and free-living microflora, of the micro and meso-fauna, and of the plants that characterise the particular ecosytem or biome. It is clear that ERM and many ECM fungi have the abilities to degrade some of the structural units within such residues which might otherwise restrict access to these elements. They then can attack many of the nutrient containing polymers themselves. There is little, in contrast, to indicate that mycorrhizal fungi, even if some of them possess modest ligninolytic capabilities, have the potential to compete with ‘white rot’ fungi in the essentially N and P-poor environments provided by large woody debris. Since their energy requirements are satisfied by autotrophs there would, indeed, seem to be little advantage for them to compete in such carbon-rich habitats.
The results obtained over the last decade from experimental analysis of the physiological capabilities of mycorrhizal fungi have provided much support for the hypothesis (Read, 1991) that on a global scale these heterotrophs may be making significant contributions which are distinctive at latitudinal or altitudinal scales to ecosystem nutrient cycling (Fig. 10). Some recent studies at a local scale have added refinement to the perspective. Giesler et al. (1998) examined plant–microbe–nutrient interactions across an extreme gradient within the northern boreal zone in which soil conditions changed, over a distance of only 90 m, from extreme acidity (pH 3.5) and an absence of nitrification at one end, to near neutrality (pH 6.4) with high levels of NO3 in a localised zone of ground water discharge at the other. Because levels of available P were consistently low across this pH-N gradient the transition, in effect, was one of decreasing P : N ratio. Along the gradient the ground flora was seen to change from an exclusive dominance of ERM shrubs in the absence of nitrification in the acidic region to a dominance of potentially AM colonised tall herbs at the other. Although, as would be expected at this northerly latitude an overstorey of spruce was maintained across the gradient, quantitative analyses of microbial fatty acid signatures indicated that the extent of ECM fungal mycelial development was greatly reduced as productivity increased in the nitrogen enriched part of the system. Giesler et al. (1998) propose that the success of distinctive types of understory plant at particular points along this short gradient are likely to be determined by the abilities of their mycorrhizal symbionts to access the locally limiting macro-element. It remains to be demonstrated experimentally that AM fungi contribute to enabling the gradual increase in representation of herbs along such gradients of decreasing P : N ratio. In particular, analyses of the responsiveness of putatively AM herbs to colonisation by their naturally selected fungal endophytes under realistic nutrient regimes are long overdue.
At a scale intermediate between the global (Read, 1991) and local (Giesler et al., 1998), the occurrence of distinctive understory associations dominated by ericacaeous, mixed shrub-herb or herb systems has long been recognised at the regional level in boreal biomes, where it has underpinned plant community (Cajander, 1926) or nutrient-based (Dahl et al., 1967; Lahti & Vaisanen, 1987) descriptions of forest types. More recently, Cornellissen et al. (2001) used 83 British plants species of known functional and mycorrhizal categories to test the hypothesis that classification of the plant functional types according to mycorrhizal association could help to explain nutritional feedbacks between plant productivity and litter turnover. They observed within this particular subset of the temperate flora, that ericoid and ectomycorrhizal strategies were linked to low and arbuscular mycorrhizal species to high ecosystem turnover of carbon and nutrients. This study provides a complementary link between the global and local scales and largely confirms their predictions. The emerging appreciation of the full physiological potential of each of the major mycorrhizal types should help to improve our mechanistic understanding of the factors enabling dominance of distinctive plant communities at local, regional and global scales.
While Giesler et al. (1998) described a pristine boreal environment which was likely to be in long-term nutritional equilibrium, decreases in P : N ratio are occurring across much of the northern hemisphere as a result of anthropogenic N enrichment. These changes should also favour the dominance of AM-based communities (Aerts, 2002). Clearly, there is a need to examine the extent of involvement of mycorrhizal fungi in the changing patterns of plant distribution which are arising globally in response to the nutritional disturbances induced by human activities.