The importance of individuals: intraspecific diversity of mycorrhizal plants and fungi in ecosystems


Author for correspondence:
David Johnson
Tel: + 44 1224 273857



II.The potential importance of intraspecific diversity in ecosystems615
III.How diverse are mycorrhizal plant and fungal populations in ecosystems?615
IV.What is the magnitude of physiological variation within species?617
V.The need to understand links between mycorrhizal plant and fungal traits618
VI.How do assemblages of individuals affect ecosystems?619
VII.How do interactions between mycorrhizal plants and fungi affect intraspecific diversity?620
VIII.Can we predict the conditions when intraspecific diversity matters most for regulating ecosystem processes?621
IX.Population genomics: the future to understanding the importance of fungal intraspecific diversity?621
X.Conclusion: the need for a ‘community genetics’ approach in future mycorrhizal symbiosis research624


A key component of biodiversity is the number and abundance of individuals (i.e. genotypes), and yet such intraspecific diversity is rarely considered when investigating the effects of biodiversity of mycorrhizal plants and fungi on ecosystem processes. Within a species, individuals vary considerably in important reproductive and functional attributes, including carbon fixation, mycelial growth and nutrient utilization, but this is driven by both genetic and environmental (including climatic) factors. The interactions between individual plants and mycorrhizal fungi can have important consequences for the maintenance of biodiversity and regulation of resource transfers in ecosystems. There is also emerging evidence that assemblages of genotypes may affect ecosystem processes to a similar extent as assemblages of species. The application of whole-genome sequencing and population genomics to mycorrhizal plants and fungi will be crucial to determine the extent to which individual variation in key functional attributes is genetically based. We argue the need to unravel the importance of the diversity (especially assemblages of different evenness and richness) of individuals of both mycorrhizal plants and fungi, and the need to take a ‘community genetics’ approach to better understand the functional significance of the biodiversity of mycorrhizal symbioses.

I. Introduction

Plants and their interactions with associated mycorrhizal fungi are crucial components of the biodiversity of ecosystems that regulate biogeochemical cycles, biological diversity and the behaviours and activities of above- and below-ground organisms (Smith & Read, 2008). Yet, we still know remarkably little about how biodiversity of mycorrhizal plants and fungi affects ecosystem functioning. Testing the importance of mycorrhizal plant and fungal diversity in ecosystems requires identification of the most appropriate component of diversity on which to focus. For example, species richness has important consequences for a variety of attributes (Reiss et al., 2009), although the exact pattern of response and generality of findings are often enthusiastically debated (Grime, 2002). Several studies have found that other components of biodiversity, such as evenness (Hillebrand et al., 2008) or particular species combinations (Johnson et al., 2008), are important drivers of ecosystem functioning. ‘Functional type’ diversity, whereby phylogenetic classification gives way to dissimilarity among key traits, is considered by many to be more useful (e.g. De Deyn et al., 2008). By contrast, fewer studies have considered the importance of the diversity of individuals within species (i.e. intraspecific or genotypic diversity) for regulating ecosystem processes, or how variation at the level of the individual contributes to functional type diversity. This is an important gap to fill because only by understanding how the fundamental building blocks of biodiversity, that is, genes and genotypes, interact with their physical, chemical and biological environment will we really begin to fully understand how biodiversity in its broadest sense drives ecosystem functioning (Hughes et al., 2008). In the case of mycorrhizal fungi, all three levels of diversity (functional type, species and genotype) lead to difficulties in their measurement. For example, which traits are most relevant and under what conditions should they be defined? How is intraspecific diversity measured, especially for those groups that may contain cryptic species or where the reproductive strategies are poorly resolved (e.g. arbuscular mycorrhizal (AM) fungi)? How precise are current neutral markers in separating species from genotypes?

The requirement to consider biodiversity at the level of the individual is conceptually difficult because most plant species never exist in the absence of mycorrhizal fungi. Thus, past studies have ignored the fact that the majority of plants are tightly coevolving ‘units’ (Hoeksema, 2010) comprising plant and fungi (and their cortege of bacteria). The situation is even more complex because ‘multiple coevolution’ is probably operating, whereby an individual plant associates with several mycorrhizal fungi (Thrall et al., 2007), and an individual mycorrhizal fungus can associate with several host plants (Beiler et al., 2010). We must consider the functional basis of interactions between individual mycorrhizal plants and fungi, and here we argue the need to consider physiological and genetic factors simultaneously. Allied to this need is careful consideration of the experimental approaches used. The inability to culture many mycorrhizal fungi makes some manipulative experimentation difficult, but this approach is often essential to determine cause and effects, which correlative approaches will never resolve.

We therefore tackle these issues in this review with an emphasis on the interactions between mycorrhizal plants and fungi. Although there is clearly a need to better combine evolutionary and ecological processes in mycorrhizal systems (Hoeksema, 2010), we focus on the mechanisms through which intraspecific diversity of mycorrhizal plants and fungi can affect ecosystems. We also draw on the latest developments in whole-genome sequencing and population genomics to highlight how these data can be used to better understand how genotypic diversity drives ecological processes.

II. The potential importance of intraspecific diversity in ecosystems

Hughes et al. (2008) highlighted four cases when intraspecific diversity could have substantial effects on ecosystem functioning: (1) when communities are dominated by a small number of keystone species; (2) when genetic diversity of one species directly affects the abundance, distribution or function of a keystone species; (3) when genetic diversity is reflected by trait diversity; and (4) in changing and variable environments. Given that climate change is predicted to continue to have an impact throughout the globe, this point becomes relevant for all ecosystems. In mycorrhizal systems, there are probably no cases where more than one of these situations does not apply; for example, case 2 could be applied to all mycorrhizal systems because of the very tight coupling between plant and fungal partners largely driven by resource exchange (Fig. 1). Here, intraspecific diversity can have not only indirect effects on ecosystems (the second case made by Hughes et al.), but also direct effects. This latter process is likely to be exaggerated under case 3, where individuals exhibit large variation in key traits. Whether the pathway is direct or indirect, the underlying mechanisms are the same, namely niche differentiation, facilitation and selection effect. The first two mechanisms can lead to complementarity, that is, the response of mixtures compared with the performance of monocultures; thus the sign can be either positive or negative (Loreau & Hector, 2001). Whilst these mechanisms are quite well studied for plants, the same cannot be said for mycorrhizal fungi. Even less information is available concerning the strength of these mechanisms in response to combinations of mycorrhizal plant and fungal genotypes when in symbiosis; it is essential that we redouble efforts to address these gaps. A further crucial aspect of intraspecific diversity of mycorrhizal plants and fungi concerns the interactions with other levels of ‘vertical’ diversity that occur both above and below ground. Finally, this entire suite of interactions is driven by larger-scale processes, such as atmospheric deposition of pollutants, land use and climate change. The following sections discuss our current understanding of intraspecific diversity in mycorrhizal systems and the extent to which this knowledge supports the pathways and mechanisms illustrated in Fig. 1.

Figure 1.

Direct and indirect effects of intraspecific diversity of plants and mycorrhizal fungi on ecosystem functioning. Identifying whether there are predictable relationships (perhaps based on guilds of functional traits) among intra- and interspecific diversity of plants and mycorrhizal fungi (dashed box), and the factors that regulate them, is a key challenge. The ability of particular genotypes of plants and fungi to exchange resources (‘resource stoichiometry’; Johnson, 2010) and to maintain this under changing environmental and climatic conditions is predicted to play a central role. Intraspecific diversity should be considered a key component of ‘horizontal diversity’, while above- and below-ground biological diversity both represent ‘vertical diversity’ (Duffy et al., 2007).

III. How diverse are mycorrhizal plant and fungal populations in ecosystems?

To test whether the cases 1 and 2 outlined in the previous section hold true in mycorrhizal systems, we need to ask two very basic questions: how many individuals of a particular species are found in communities; and how likely are they to interact? Whilst these questions have been answered for many plant communities (e.g. Widén et al., 1994; Bruederle et al., 1996; Suzuki et al., 1999), the situation for fungi is less well known, which is what we focus on here. These questions also raise the ongoing issue of how we define individuals, especially in clonal organisms where it is difficult to distinguish ramets and genets (genet = several ramets). This is particularly difficult in fungi, especially AM fungi (see Section IV), where we have limited knowledge of fungal ramet size and limits. Here we use the term individual to mean genet.

The number of mycorrhizal fungal individuals found in a given habitat is likely to depend on a range of factors that includes plant community composition and age, soil chemical, physical and biological properties, and climate, meaning that considerable variability can be expected. It is crucial that we gain a better understanding of the relative importance of these factors, and this requires more effort to quantify intraspecific diversity of mycorrhizal fungi (Douhan et al., 2011). Much recent work has tended to focus on the distribution of genotypes of one or two species over small areas (e.g. Kretzer et al., 2003) or on root systems where it is clear that individual seedlings can support several genotypes of a given species (Gryta et al., 1997; Guidot et al., 1999). The abundance of ectomycorrhizal (ECM) fungal genotypes per unit area seems more variable, and some genotypes have been found over large distances (Bonello et al., 1998; Anderson et al., 2001; Lian et al., 2006). In some cases, the contrast in abundance of genets can be substantial: Fiore & Martin (2001) identified > 100 genets of Laccaria amethystina and only a single genet of Xerocomus chrysenteron in the same plot of a mature beech forest. These contrasting findings highlight the need to adopt rigorous sampling strategies that can account for small- and large-scale spatial variation that is prevalent in mycorrhizal systems (Wolfe et al., 2009).

Of particular importance is integrating this knowledge with spatial analyses, and the presence of other genotypes and species of mycorrhizal plants and fungi, because complementarity effects are strongest when individuals come into close proximity. In forests, mapping genets of Rhizopogon vinicolor and Rhizopogon vesiculosus in relation to host plant identity in situ within a 30 × 30 m plot provided clear demonstration of the chaotic nature of interactions between plants × fungal genotypes (Beiler et al., 2010). Here, a total of 401 Rhizopogon ectomycorrhizas were recovered and were matched to 67 Douglas fir (Pseudotsuga menziesii) trees. This, and other studies investigating fungal species that form rhizomorphs, demonstrates the potential for individual genets to connect trees and explore habitats at large spatial scales. The ability of fungal genotypes to colonize several individual plants is likely to be essential for the evolution of mycoheterotrophy (Leake, 2004), whereby nonphotosynthetic plants are wholly or partially dependent on nutrient exchange from neighbouring green plants via common mycelial networks (Taylor & Bruns, 1997; Selosse et al., 2002). The observation that some ECM fungal species can exhibit spatial structure both vertically and horizontally (i.e. their distributions are nonrandom; Koide et al., 2005; Pickles et al., 2010) clearly suggests that space is one factor that could generate patterns and affect biodiversity in natural systems (see section VII on selection mosaics). At small (cm) vertical and horizontal scales, the distribution of mycelium and root tips of some ECM species correlates, whilst other species do not (Genney et al., 2006). Such fine-scale analyses of mycorrhizal fungi at the genotype level may provide insight into the biotic (e.g. root tip abundance) and abiotic (e.g. nutrient content) factors that regulate distribution of individuals (Zhou et al., 2001).

Intraspecific diversity of AM fungi in temperate systems is even less well understood than for ECM fungi because of the difficulties in unequivocally identifying the fungi to species level and their obligately biotrophic habit. However, a similar picture is emerging, whereby there is considerable variation across several scales (Stukenbrock & Rosendahl, 2005). In AM agroecosystems, intraspecific diversity can be sensitive to disturbance (Rosendahl & Matzen, 2008), whereas for ECM systems the effects of disturbance seem difficult to predict. Rosendahl & Matzen (2008) demonstrated that there were significant differences in the genotype frequencies, with one or two genotypes showing clear dominance that indicated replacement of particular genotypes, indicating that the different land use types may be one of the main factors leading to changes in the relative abundance of the genotypes.

Spore-based cultures of AM fungi enable measurement of functional differences among genotypes. In a single agricultural field, 17 distinct genotypes were identified from a total of 41 isolates of Glomus intraradices and there was significant clustering of genotypes according to the three species of trap plant used, that is, the genotypes appeared to exhibit strong host preferences (Croll et al., 2008). At the species level, host preferences have also been observed under field (Vandenkoornhuyse et al., 2003), mesocosm (Johnson et al., 2004), pot (Bever et al., 2009) and axenic (Kiers et al., 2011) conditions, but whether genotype and species level preferences occur with similar strength and propensity remains to be established. Regardless, there is clearly considerable potential for feedback among fungal genotypes and host plant species with consequences for community composition and ecosystem function (see Section VII).

It is now known that ericaceous plant roots have the potential to be colonized by a more diverse range of fungi than was hitherto thought (Vrålstad et al., 2000, 2002; Hambleton & Sigler, 2005). In addition, a global analysis of ericaceous vegetation has revealed that the Sebacinales are common associates of their roots (Selosse et al., 2007) and that fungi in this group also colonize both ECM plants and orchids (Selosse et al., 2009) as well as being found as endophytes in nonmycorrhizal plant roots. Individual roots may also support several species of ericoid mycorrhizal (ERM) fungi (Curlevski et al., 2009) and Archeorhizomycetes (Rosling et al., 2011), while the distribution of ERM fungi at the whole habitat scale is largely unknown. The most recent information comes from analysis of genotypes associated with both Vaccinium vitis-idaea hair roots and adjacent Piceirhiza bicolorata ectomycorrhizas of Pinus sylvestris (Grelet et al., 2010). Across two locations, a total of 70 intersimple sequence repeat (ISSR) haplotypes were isolated from a total of 124 isolates, with most belonging to the Helotiales. The data showed that the spatial structure of identical genets isolated from both overstorey trees and adjacent understorey shrubs was small (< 13 cm), but that the intraspecific diversity of these fungi was large. Curlevski et al. (2009) have similarly reported that a single genotype of a Helotiales ERM fungus was present in root systems of understorey Epacris pulchella in an Australian dry sclerophyll forest.

Inevitably, the development of molecular markers is technically challenging and species-specific, and this is one reason why our knowledge of intraspecific diversity in nature tends to be focused on just a handful of the world’s mycorrhizal fungal species. Nevertheless, the emerging picture is that there is considerable scope for individuals to interact at scales from cm to several metres, and that the number of genotypes can be similar to the number of species. Yet, we still lack clear evidence of whether species diversity is related to genotypic diversity (see Section VI). In order to understand how intraspecific diversity affects ecosystem functioning, we must explore the magnitude of physiological variation found within species.

IV. What is the magnitude of physiological variation within species?

According to the third statement (see Section II) made by Hughes et al. (2008), for intraspecific diversity to play an important direct role in regulating ecosystem processes (Fig. 1), individuals must exhibit variation in physiological traits. Whilst variation in plant traits between individuals of a species is relatively easy to determine (e.g. Garnier et al., 2001), the situation in fungi is far more challenging. The ability of many ECM and ERM fungi to grow in pure culture permits large-scale quantification of traits under simple but standard conditions. There is no doubt that many important traits can vary considerably amongst genotypes (Cairney, 1999), supporting the existence of both indirect and direct effects of intraspecific diversity on ecosystem processes and on the potential for genotypes to exploit particular niches (Fig. 1). Yet, direct comparisons of the performance of genotypes and species of mycorrhizal fungi remain rare. Comparison of the coefficient of variation (CV) in the biomass of eight genotypes of Paxillus obscurosporus and eight species of ECM grown in pure culture at three N availabilities (C : N ratios, 10 : 1, 20 : 1 and 40 : 1) provides some insight into similarities between intra- and interspecific diversity (Fig. 2). Here, the CV associated with only the different C : N ratios was always much smaller than the CV resulting from genotype or species richness. Whilst undertaken under very simple conditions, these data indicate that the range of responses to intraspecific diversity is similar to those of species diversity. Such analyses, especially when combined with genomic and transcriptomic information (see Section IX) would be a key first step in testing the importance of trait variation at the level of the individual. Whilst the applicability of pure culture experiments to natural systems is often debated, we emphasize the need to embrace these approaches to devise innovative experiments to generate hypotheses that can then be tested under more ecologically meaningful conditions.

Figure 2.

Coefficient of variation (CV) of biomass produced by genotypes of the same species (open bars) and different species (closed bars) of ectomycorrhizal (ECM) fungi in pure culture with substrates of differing C : N ratio (C held constant). CV is greater between eight genotypes of Paxillus obscurosporus and eight ECM species (Cenococcum geophilum, Hebeloma crustuliniforme, Cortinarius glaucopus, Suillus bovinus, Amanita muscaria, Lactarius rufus, Laccaria bicolor and Paxillus involutus) than between a fourfold change in substrate C : N ratio. See Wilkinson et al. (2010) for full experimental details.

The challenge to determine the extent of intraspecific physiological variation is particularly apparent in AM fungi because of our poor and sometimes conflicting knowledge of their unique reproductive mechanisms (Rosendahl, 2008; Croll & Sanders, 2009). Because they are multigenomic, segregation of nuclei may occur (Angelard et al., 2010) and lead to recombination in newly developed spores. Alternatively, fusion of genetically distinct spores into a common mycelial network may also lead to recombination by a simple mixing process (Giovannetti et al., 1999, 2001). Recent evidence suggests that in nature, some individuals from the same habitat have undergone recombination whilst others are developed from clonal lineages (Croll & Sanders, 2009). Thus, traditional concepts of species and individuals may be difficult to apply to AM fungi (Rosendahl, 2008), but there can be no doubt that there is often considerable genetic variation within a putative AM fungal species (Koch et al., 2004). This may mean that ‘intraspecific’ diversity as it is currently defined has greater importance in AM fungi than other mycorrhizal types, and that genetic contents may fluctuate with time in vegetative ramets of AM fungi. Efforts to understand the importance of AM fungal diversity for regulating communities and ecosystem processes must ultimately consider ways to quantify this, and so it is pertinent to review some of the recent advances made in measuring intraspecific diversity in the Glomeromycota.

It is increasingly apparent that host plants respond very differently to individual isolates of AM fungi (Munkvold et al., 2004; Avio et al., 2006). In many cases, plant responses varied among isolates, and this may reflect both phenotypic variation arising from host plant genotype and growth conditions and genetic variation between individuals. The distinction between these two sources of variation is important, but particularly so in AM fungi because theory predicts that ecological generalists have considerable variation in functional traits (such as competitive abilities to colonize roots) within individuals (Bolnick et al., 2007). For example, AM fungi may well exhibit intraindividual variation, whereby an individual may produce spores of different nucleotypes (i.e. genetically different nuclei) with subsequent specificities or traits (Kuhn et al., 2001). Testing the consequences for plant performance and ecosystem processes of intraindividual variation requires sophisticated experimental design in which single spore cultures are maintained so that environmental and culturing effects (phenotypic effects) can be minimized from genetic effects. Using such a design, Koch et al. (2004) showed that hyphal lengths and spore formation by 16 individuals (i.e. derived from a single spore) varied considerably over and above tillage treatment. Subsequent experiments using the same isolates led to both positive and negative growth responses of Prunella vulgaris and Brachypodium pinnatum host plants (Koch et al., 2006). Later experiments found strong genotype × environment (phosphorus availability) interactions and that host plant identity also affected fitness-related growth traits of AM fungi (Ehinger et al., 2009). Host plant species can therefore affect the fate of different nucleotypes following segregation in AM fungi. This finding suggests that both nucleotype segregation and host plant species diversity can maintain genetic diversity among and within species of AM fungi (and probably vice versa – see section VII). The ecological consequences of AM fungal diversity are therefore likely to be operating at three levels: species, individual and within-individual. This idea raises many fascinating questions as to the control and spatial allocation of resources, primarily carbon (C), nitrogen (N) and phosphorus (P), within individual networks of AM fungi and the relative contribution of the different levels of AM fungal diversity in regulating ecosystem processes. This is especially interesting if the fungi have formed large common mycelial networks linking individual plants and if there is a genetic basis to the control (from either plant or fungi) of resource movement through highly genetically diverse AM fungal networks.

V. The need to understand links between mycorrhizal plant and fungal traits

From a mycorrhizal functioning standpoint, plant intraspecific diversity will matter most when there is large variation in traits that have a role in regulating fungal colonization or function, including those related to root signalling, for example, production of fungal branching factors (Akiyama et al., 2005). Remarkably, it remains unclear the extent to which below-ground C allocation to mycorrhizal fungi is related to subsequent development and functioning of the symbioses. Variation in many traits is known to occur at the species level (e.g. Wright et al., 2004) and variation in trait diversity has recently been proposed to be the main mechanism by which plant community composition drives ecosystem functioning (De Deyn et al., 2008). In this model, soil C storage is best predicted by identifying key traits, including the propensity to form associations with mycorrhizal fungi. Whether these predictions are supported in nature remains to be tested. A recent analysis of nine plant species from three families found no relationship between mycorrhizal colonization and a range of other root traits, including diameter and length, although the extent of mycorrhizal colonization contributed to the separation of plant families using principal components analysis (Roumet et al., 2008). Variation in traits may also affect other processes, including interactions among individuals and species within a community, which may have consequences for shaping mycorrhizal community composition and functioning.

We highlight the emerging picture that intraspecific diversity can lead to variation in traits of a similar magnitude to that seen between species (at least within broad functional or taxonomic groups). A study of F. ovina clones in a 10 × 10 m area in a lightly grazed limestone grassland showed maximum differences between clones to be 3.5-fold in biomass and 2.4-fold in canopy height (Bilton et al., 2010). Trait variation among genotypes of at least one plant species in this system significantly predicted the abundance of that species (Whitlock et al., 2010), which in turn appeared to drive changes in genotypic diversity that occurred in other coexisting plant species. Other work has confirmed that c. 50% of variation in plant traits occurs at the level of the individual (e.g. Garnier et al., 2001; Jung et al., 2010). Differences in colonization among genotypes may be one of the reasons for the fine-scale variation in allocation of recent plant assimilates seen in mixed species swards in upland grassland (Grieve et al., 2006). A more comprehensive view of plant intraspecific diversity in natural ecosystems coupled with more rigorous analyses (e.g. Jung et al., 2010) of traits directly relevant to formation and functioning of mycorrhizas are clearly required. We advocate these approaches to directly address recent calls to combine trait analyses of ECM fungi with their phylogeny and function (Parrent et al., 2010). In AM fungi, phenotypic variation in hyphal length, growth rate, and extent of anastomoses have been shown to be strongly related to host plant performance (Avio et al., 2006). Such traits are likely to regulate the ability of individual genotypes of mycorrhizal fungi to colonize plant roots. Yet, we know remarkably little about the interactions (positive or negative) that occur among individuals that affect the extent of colonization, although whole-genome sequencing is beginning to inform us about its genetic basis (Martin et al., 2008; see Section IX).

VI. How do assemblages of individuals affect ecosystems?

The potential for considerable intraspecific variation in certain traits may also contribute to the wide range in responses to species manipulations seen in past experiments, illustrated with a hypothetical example in Fig. 3. It is remarkable that there have been limited attempts to test hypotheses centred on the effects of manipulation of the diversity (i.e. relative abundance or richness) of genotypes of organisms, especially fungi (but see, e.g., Wilkinson et al., 2010). Nevertheless, we consider here the observations made thus far, primarily from experimental manipulations of plants, which highlight the importance of assemblages (leading to different composition, richness and evenness) of individuals, and the mechanisms through which they affect ecosystem processes.

Figure 3.

Hypothetical positive response of productivity to diversity, highlighting the potential importance of intraspecific diversity (genotypes per species) as an important contributor to variation seen in manipulations of species diversity. Regardless of the pattern of response to diversity (positive, negative or neutral), considering overall genetic diversity is likely to result in more precise estimates of biodiversity-ecosystem function relationships.

There is compelling evidence that gradients of intraspecific diversity may have indirect effects on ecosystem functioning mediated via changes in species diversity (Fig. 1). For example, in experimentally manipulated limestone grassland mesocosms, it was found that genetically impoverished plant communities became more divergent in their species composition and that they lost species diversity faster than did communities comprising more genotypes of each species (Booth & Grime, 2003; Whitlock et al., 2007). The genetically impoverished communities supported a less diverse AM fungal community, although it was suggested that this occurred indirectly via changes in plant species richness (Johnson et al., 2010a).

Biodiversity theory predicts that the diversity at one level (i.e. genotypes) will positively covary with diversity at another (i.e. species), giving rise to the idea that ‘diversity begets diversity’ (Vellend, 2008). This is because individuals and species may respond to the same variable or because of direct effects at one level on another (Fig. 1). Empirical evidence from both natural (He & Lamont, 2010) and model (Fridley et al., 2007) communities suggests that such positive relations can occur, although there may be other factors that influence this. In situations where some individuals produce allelopathic compounds, facilitative interactions lead to plant genotypic diversity playing a key role in maintaining species richness of the community (Lankau & Strauss, 2007). Some simulation models also suggest that genetic diversity increases species coexistence only when the fitness of the individual is dependent on a neighbouring or competing species (Vellend & Geber, 2005; Vellend, 2006), and this is most likely to occur with plants of similar functional type. Other models suggest that the relationship between genotypes and species is more consistent (Bown et al., 2007), but these models have thus far ignored mutualisms.

But can intraspecific diversity directly affect the way ecosystems function, or do the effects of intraspecific diversity have to be moderated via changes in species diversity? A limited amount of research undertaken to date does indeed show that, for some groups of organisms, intraspecific diversity per se is an important regulator of ecosystem functioning. Phenotypic variation within species can be greater than that among species (Bangert et al., 2005), and where this variation is manifested through ecologically important traits (Section V), intraspecific diversity becomes as important as, or more than, interspecific diversity (Hughes et al., 2008). For example, when genotypic richness of the tall goldenrod (Solidago altissima) was increased, a parallel effect was seen on annual net primary productivity (Crutsinger et al., 2006). Furthermore, multitrophic impacts were also observed, such as an increased richness of predators and herbivores. In marine seagrass (Zostera marina), the communities comprising a greater number of genotypes were more productive and supported an increased abundance of invertebrate herbivores, and therefore had consequences for second-order effects (Reusch et al., 2005) and contributed to ‘vertical diversity’ (Fig. 1). These studies provide strong evidence that ecosystem functioning can be affected by intraspecific diversity through both indirect pathways that occur via changes in species diversity and direct pathways that occur because of variation in key traits at the individual level (Fig. 1).

The lack of information on the effects of assemblages of individual fungi arises partly from the experimental complexity necessary to unequivocally test effects of diversity. At the species level, there is evidence that richness of fungi matters for plant performance (Van der Heijden et al., 1998), although effects can be dependent on a variety of factors, including soil fertility and host plant species (Jonsson et al., 2001). Using gradients of one to five saprotrophic fungi, Tiunov & Scheu (2005) found that the positive effects of species richness were driven by both ‘selection effect’ (i.e. the greater probability of using a functionally dominant organism in diverse combinations) and complementarity, but that the strength of these effects depended on substrate complexity. At the level of the genotype, Wilkinson et al. (2010) conducted a pure culture experiment using the ECM fungus Paxillus obscurosporus in 15 replicated treatments with genotypic richness of one, two, four and eight genotypes in factorial combination with N availability. A striking outcome was that the biomass and efflux of CO2 from the populations was driven mainly by genotypic richness rather than N availability or ‘genotypic composition’; the DW of the fungi was, on average, 50% greater in mixtures of eight different genotypes than in mixtures of eight identical genotypes. It is known that fungal genotypes can exhibit striking morphological and growth responses when they come in to contact with each other (Malik & Vilgalys, 1999) and it follows that such responses could have significant impacts on the key functions they undertake. The experiment of Wilkinson et al. (2010) was undertaken under very simple conditions and so raises numerous questions about the pattern and magnitude of responses that would be observed in more natural conditions, for example, with host plants of contrasting species and genotype, heterogeneous nutrient supply, greater options for spatial exploration, and perturbation event. Based on past evidence from manipulation of both species and genotypic diversity of plants, most of these factors would be expected to exaggerate the intraspecific richness effect. In addition, it may be that experimental manipulation of evenness is similarly important, and possibly more ecologically relevant. For example, perturbations like climate change and N deposition are more likely to alter evenness rather than richness of populations and communities, and from an experimental perspective, manipulations of evenness eliminate the need to consider sampling effect (Huston, 1997).

VII. How do interactions between mycorrhizal plants and fungi affect intraspecific diversity?

Here we discuss how interactions among symbionts themselves affect the diversity and function of mycorrhizal plants and fungal genotypes in the environment. We also highlight recent evidence supporting the importance of plant–fungal genotypic interactions, along with other important biological factors, before discussing the larger-scale effects of climate on these interactions.

There is a paucity of information focusing on the relationship between intraspecific diversity of plants and mycorrhizal fungi. However, because of the wide variation in traits amongst plant and fungal genotypes (see Sections IV and V), evidence from manipulations at the species level may be relevant to formulate hypotheses concerning how variation of genotypes may affect ecosystem processes. Bever et al. (1997) predicted that feedback between plants and mycorrhizal fungi is central to maintaining the species diversity of each group. Negative feedback is predicted to maintain species richness, whilst positive feedback is predicted to have the opposite effect. These results have been borne out by experimentation (Bever, 2002, 2003), but there is clear scope for extending such approaches. In particular, the mechanism underpinning such interactions is less well known, but recent work suggests that the C demand of particular AM fungi may be one factor (Bever et al., 2009). We now need to discover the extent to which these interactions occur at the level of the genotype (of both plant and fungi) and how these interactions go on to influence the responses at the species level (Fig. 1). There is some theoretical underpinning to this question based on resource stoichiometry that Tilman, (1982)developed, albeit for species, into the resource use (R*) concept. Here the differing abilities of organisms to acquire limiting nutrients relative to competitors may help to maintain species diversity, and this may be one of the mechanisms by which diversity is maintained at the level of the individual. Such differing ability to obtain resources has recently been put in a mycorrhizal context in a ‘trade balance model’ (Johnson, 2010). Here the direction and magnitude of resource (C, N and P) exchange between plant and mycorrhizal fungus drives the nature and strength of interactions. This idea has recently been borne out experimentally under axenic conditions, where allocation of C from transformed carrot roots to mycorrhizas was dependent on the ability of AM fungal partners to provide the best nutritional benefit (Kiers et al., 2011). These models and experiments point to the need to undertake more sophisticated functional studies of mycorrhizal systems, in which experiments are designed to identify the performance of individuals rather than species.

A key first step to discovering whether genotypic diversity of symbiotic partners matters is to look for patterns when the intraspecific diversity of either plants or fungi is manipulated and held constant. In the field, Korkama et al. (2006) found that ECM species composition differed among the roots of clones of Norway spruce (Picea abies). They suggested that plant growth traits were one factor explaining the observed differences in ECM species composition, indicating a key role of resource exchange. Such plant genotype × fungal genotype (GG) interactions have been tested experimentally with Pinus muricata and Rhizopogon occidentalis under a range of conditions (Piculell et al., 2008). This experiment demonstrated that plant traits, such as the extent of colonization by ECM, were driven by interactions among plant and fungal genotype, soil type (either a commercial inocula or field collected soil) and microbial community composition. This work suggests that environmental heterogeneity can affect the outcome and strength of GG interactions leading to ‘selection mosaics’ (Thompson, 2005). This prediction was demonstrated recently using ecotypes of the grass Andropogon gerardii Vitm. Plants inoculated with native (i.e. ‘home’) AM fungal communities and grown in native soil had greater numbers of arbuscules in their roots than plants grown in nonnative (i.e. ‘away’) communities and soils, suggesting a greater capacity of the fungi to acquire C (Johnson et al., 2010b). These sorts of elegant cross-inoculation experiments offer considerable promise; but one pressing issue concerns the effects of the addition of multiple genotypes on an individual host plant root. As the complexity of mutualists increases (i.e. more genotypes colonising the same plant), there may be a greater chance of ‘cheaters’ that exploit the host plant (Thrall et al., 2007). We might also expect that colonization by fungal genotypes exhibiting a range of traits would reduce the likelihood of selection mosaics because plants may divert resources into those fungi that offer the best ‘resource exchange rate’ (Johnson, 2010) for the prevailing conditions. We emphasize the urgency to quantify resource exchanges in experimental systems that systematically increase in complexity using first, plants colonized with a single fungal genotype; second, plants colonized by several genotypes of fungi; and third, combinations of several ‘units’ of plant and fungi. Ultimately, these measures of resource exchange, and other indices of performance need to be reconciled with fitness.

There are many potential candidate variables that could influence GG interactions, and some of these have already been identified. For example, the susceptibility of P. edulis to herbivory by the stem-boring moth (Dioryctria albovittella) was shown to lead to changes in the composition, but not species richness, of ECM communities (Sthultz et al., 2009). This is one of the few clear examples of the importance of ‘vertical diversity’ (Fig. 1) in ecosystems that cascade through trophic levels both above and below ground, and which are largely missing from current theoretical frameworks (Duffy et al., 2007). Sthultz et al. (2009) were able to discount direct effects of the moth on the changes in ECM communities and instead suggested they were driven by the genetic basis underpinning susceptibility. Other important factors that may affect GG interactions include soil nutritional characteristics, and interactions with other soil organisms, including root and fungal herbivores, that can forage selectively. Disentangling the genetic vs environmental drivers of phenotypic variation, or the ability of particular genotypes of plants and mycorrhizal fungi to interact, is crucial if we are to fully understand the importance of intraspecific diversity in ecosystems.

VIII. Can we predict the conditions when intraspecific diversity matters most for regulating ecosystem processes?

Hughes et al. (2008) concluded that individuals that exhibit variation in traits are most likely beneficial when conditions are unstable or unpredictable, for example, in response to drought and temperature fluctuations. The urgency to understand how global climate change impacts ecosystems means that we must redouble our efforts in this area. We need to understand not only how climate change might impact on allelic frequency and gene expression, but also the extent to which intraspecific diversity provides resistance and resilience against perturbations. From a plant perspective, it has been suggested that a combination of frequent sexual reproduction and high intraspecific genetic diversity enables Salix arctica and Spartina patens to survive rapidly changing climatic conditions (Steltzer et al., 2008) or to adapt to strongly divergent environments (Silander & Antonovics, 1979). Intraspecific diversity may therefore be expected to impart resilience to communities and ecosystems (Jump et al., 2009). This is precisely what has been found in experimental marine eelgrass (Zostera marina) communities. Here, the effects of intraspecific diversity were principally seen after extreme climatic events; that is, intraspecific diversity increased biomass, density and faunal abundance even in extreme water temperatures (Reusch et al., 2005; Ehlers et al., 2008). Such findings suggest a key role of intraspecific diversity in providing resistance and resilience to anthropogenic stressors like climate change. However, recent work has also highlighted the possibility that climate change may feed back to affect plant genetic diversity. Small-scale genetic variation in Fagus sylvatica has been linked to variation in temperature (Jump et al., 2006), while genetic differentiation was significantly greater in seedlings subjected to drought and warming than in seedlings from an identical population grown in control conditions (Jump et al., 2008). Whether there is a ‘mycorrhizal mechanism’ behind such observations is not known. By contrast, individual-based modelling has predicted that mutualistic strategies (i.e. mycorrhizal plants and fungi) might suffer to a greater extent than competitive strategies (i.e. nonmycorrhizal plants and fungi) under rapidly fluctuating climates, primarily because range expansion of mutualists is limited (Brooker et al., 2007). Thus, plasticity in the ability of mycorrhizal plants and fungi to expand their ranges is important; inevitably this may select against host specificity.

Despite the advances already discussed, we are still a long way from determining whether similar patterns occur in fungi. In the case of drought, ECM fungal diversity has a potentially important role, because there is evidence that moisture availability can affect several functional attributes of the mycorrhizal symbiosis (Wu et al., 1999; Kennedy & Peay, 2007); some ECM species are more affected by drought than others (Jany et al., 2003); a survey of 55 isolates from 18 species revealed considerable variation in growth rate both within and between species in response to water stress (Coleman et al., 1989); and some species have key roles in mediating water transport to host plants whereas others do not (Duddridge et al., 1980; Dietz et al., 2011). For many species of mycorrhizal fungi, the possibility of undertaking the necessary experiments (e.g. reciprocal transplants along climate gradients; manipulation of mycorrhizal plant genotype × mycorrhizal fungus genotype × climate) can be easily realized.

IX. Population genomics: the future to understanding the importance of fungal intraspecific diversity?

Our limited understanding of the importance of intraspecific diversity of fungi has in part been driven by technical issues in defining species and individuals. The internal transcribed spacer (ITS) region of the ribosomal RNA (rRNA) gene cluster has been the most widely targeted genomic region in molecular investigations of fungi (Horton & Bruns, 2001; Anderson & Cairney, 2004). However, the ITS region represents a single genomic target and is thus an extremely small fraction of the genetic, and hence biological, potential of any fungal species. For the recently sequenced ECM basidiomycete Laccaria bicolor (which has a 65 MB genome and contains 50 copies of the rRNA genes; Martin et al., 2008; Martin & Selosse, 2008), the ITS region represents < 0.1% of the entire genome and therefore can only ever provide a very limited view of genetic diversity. Whilst a range of molecular techniques have led to significant advances in understanding the extent of intraspecific diversity in ecosystems, these still have limitations including the separation of genets and ramets. We have now progressed into another era in which population genomics is being used to gain unparalleled information on the diversity of individuals, and the genes involved in key processes. In this section, we highlight how these techniques could bring about a step change in our understanding of the importance of intraspecific diversity of mycorrhizal plants and fungi in ecosystems.

1. Current state-of-the-art of genome sequencing of mycorrhizal plants and fungi

Moving away from single molecular markers towards the direct comparison of whole genomes will aid in our understanding of both sequence and biological variations within and between species of mycorrhizal fungi. While only a limited number of fungal genomes (c. 100) have been published so far (Martin et al., 2010; Grigoriev et al., 2011), already the genomes of a number of species and strains of mycorrhizal plants and fungi are in the process of being sequenced (Supporting Information, Table S1).

Even though the majority of fungi targeted thus far have largely been selected based on their contrasting nutritional lifestyles and phylogenies, important differences between individuals are being discovered (Martin et al., 2008, 2010; Plett & Martin, 2011). In particular, the comparison of genomes among fungi with saprotrophic, mycorrhizal and pathogenic modes of nutrition (e.g. Spanu et al., 2010; Duplessis et al., 2011; Eastwood et al., 2011) may help us to understand the genetic basis of mycorrhiza formation. For example, in symbiosis L. bicolor and G. intraradices expressed genes that encode for a suite of small secreted proteins (mycorrhiza-induced small secreted proteins; MiSSPs) that are similar to those produced by pathogens to subvert host plant cell defences (Plett et al., 2011). By contrast, the phylogenetically distant ECM fungus Tuber melanosporum does not induce MiSSPs, suggesting this species has an alternative strategy for mediating symbioses. Another important discovery has been the apparent loss of degradative enzymes in ECM fungi compared with saprotrophs, which reflects the need of the former to colonize host roots without causing damage to their cell walls (Eastwood et al., 2011; Plett & Martin, 2011).

2. The move to population genetics

Will whole-genome sequencing reveal important functional insights among strains of mycorrhizal plants and fungi? To answer this question, it is crucial to have fully annotated reference genomes for diverse species, which has already been achieved for L. bicolor (Martin et al., 2008), T. melanosporum (Martin et al., 2010), Paxillus involutus, Hebeloma cylindrosporum, Piloderma croceum and Oidiodendron maius (Mycorrhizal Genomics Consortium, unpublished; Table S1), which will enable comparison among strains without the need to undertake complete annotations. Furthermore, population genomics also improves the potential for genet delineation among related individuals and provides more precision in measuring diversity at the chromosomal level. Interrogation of the genome enables unequivocal identification of SSRs that can be used to rapidly identify genotypes, and this has already been achieved for L. bicolor (Labbéet al., 2011b) and T. melanosporum (Murat et al., 2011). Genome resequencing of L. bicolor strains from different geographic accessions revealed a significant sequence nucleotide polymorphism (SNP) in protein-coding sequences, with > 180 000 SNPs within the L. bicolor S238N haploid progeny and > three million SNPs amongst geographic accessions. Single nucleotide variants were found in symbiosis-regulated genes, including MiSSPs (Francis Martin, unpublished).

Comparison among strains of other eukaryotic microorganisms is already beginning to reveal important differences in responses to environmental variables and interaction with host plants. These may be human-induced, as shown in domestication and subsequent artificial selection of the yeast Saccharomyces cerevisiae in the wine and brewing industry. Here, whole-genome sequencing of six strains revealed the genetic basis that characterized strains used either in brewing or in wine-making (Borneman et al., 2011). Whole-genome sequencing is also beginning to reveal important differences among strains under less extreme selection pressures. In the ascomycete Neurospora crassa, whole-transcriptome (the subset of the genome transcribed and processed as mRNA) sequencing of 48 individuals revealed two cryptic populations. Analysis of chromosomal divergence suggested that the populations developed in response to contrasting prevailing temperatures at the two locations (Ellison et al., 2011). In plants, genomic diversity of wild populations of soybean was found to be greater than that of cultivated varieties, which may have been driven by human selection (Lam et al., 2010), while analysis of Arabidopsis lyrata populations has revealed candidate mutations related to tolerance to serpentine-rich soils (Turner et al., 2010). We predict that similarly elegant analyses will provide links between fine-scale phylogenetic information and key environmental, climatic and biological drivers in mycorrhizal systems, whether caused by natural variation or human intervention.

The focus of initial investigations of genetic diversity of mycorrhizal fungi at the intraspecific level aims to unravel differences among mating systems and geographically isolated strains. In the former, it has been shown that T. melanosporum strains belong to two distinct mating types (i.e. they are heterothallic) that are nonselfing (Rubini et al., 2010b). Whilst the mating types can initially form mycorrhizas on the same root system, individuals ultimately develop to form spatially segregated populations (Rubini et al., 2010a). The spread and ultimate dominance of a particular mating type in a local area is likely to be driven by C from host plants because genomic analyses have revealed that T. melanosporum has a limited repertoire of enzymes required for a saprotrophic lifestyle (Martin et al., 2010), although direct competition among individuals may also be involved. This study demonstrated the significance of intraspecific diversity in T. melanosporum populations to maintaining reproductive fitness, which has clear implications for maximizing fruit body production of this economically and gastronomically important fungus. Current experiments are focusing on comparison of gene coding sequences and transcriptomes of different strains of T. melanosporum from geographically isolated regions. Analyses of the type illustrated in Fig. 4 highlight potential areas of the genome exhibiting intraspecific variation.

Figure 4.

(a) Summary of the three main levels of control between mycorrhizal plant (shaded) and fungal (unshaded) cells (modified from Plett & Martin, 2011). Novel signalling pathways between plant and ectomycorrhizal (ECM) fungi involve auxins, tyrosine kinases and ethylene. Mycorrhiza-induced small secreted proteins (MiSSPs) are thought to control cell defences and, once established, the symbiosis is driven by reciprocal exchange of resources (C, N and P). In different strains of some ECM fungi (e.g. Tuber melanosporum), the genes of opposite mating types are located on the same chromosomal locus but are dissimilar in sequence: one encodes a protein with an α-box domain (MAT1-1-1), whereas the other encodes a high mobility group (HMG) protein (MAT1-2-1; Debuchy et al., 2010). Because these loci contain different genes and therefore are not allelic, the two versions of a MAT locus (MAT1-1 and MAT1-2) are called idiomorphs. Mating types may be found on the same root system, in the same habitat or across wide geographical ranges. (b) DNA polymorphism in the mating type locus of T. melanosporum as revealed by Illumina genome resequencing of different geographic accessions (France, Italy and Spain). The idiomorphic region (i.e. where two mating-type genes typically occupy the same chromosomal location in different genomes but lack sequence similarity) is boxed. Scaffold 247 with gene models in the MAT locus are indicated as a violet line, with gene models in red and repeated elements in dark blue. Plots depicting numbers of sequence reads (log-scale) across the genomic region carrying the mating-type locus (MAT1-2-1) coding for a HMG domain containing transcriptional factor are shown for the six strains. The colour reflects the number of Illumina reads aligning to the genomic sequence (scaffold 247) with lowest to highest sequence similarities (blue to red to green to yellow). Highly polymorphic regions corresponding to the divergent idiomorphic regions (the MAT1-2-1 gene, the gene GSTUMT00001092001 coding for an hypothetical protein and the surrounding intergenic regions; see Rubini et al., 2010b for details) showed little read alignment (C. Murat & F. Martin, unpublished).

3. Linking genomic data from mycorrhizal plants and fungi

Ultimately we need to know more about how individuals (of both plants and fungi) drive ecosystem functioning. In the immediate future, we therefore advocate the need to focus genomic analyses on host plant × fungal interactions, particularly in resource transfers between partners and loci involved in interaction with the host, such as effector proteins (e.g. MiSSP7 and SP7 in L. bicolor). These studies may require marriage of advanced ecophysiological techniques to quantify nutrient fluxes (inevitably using stable isotopes and radioisotopes) with sequencing and transcriptome analyses. It is essential that such investigations consider in equal measure the role of host plants and mycorrhizal fungi. At the genomic level, this is possible because an increasing number of host plants are being sequenced. However, there is clearly an imbalance between the number of compatible plants and fungi: most plants that have been, or are about to be, sequenced associate with AM fungi, while the majority of fungi targeted form ectomycorrhiza (Table S1). In addition, large-scale resequencing of fungi is only being undertaken on L. bicolor and T. melanosporum, Tricholoma matsutake, and G. intraradices (Table S1) and this may further restrict our ability to test hypotheses concerning the role of intraspecific mycorrhizal diversity in the future.

The genome of the ecologically and economically important ECM species Populus trichocarpa (Tuskan et al., 2006) has enabled identification of quantitative trait loci in an F1 cross between P. deltoides and P. trichocarpa and their parents (Labbéet al., 2011a). In total 1543 expressed genes were found to have significantly different transcript abundances in the cross compared with the parents, which included genes thought to have roles in host plant defence mechanisms and pathogen resistance. This supports the consensus that mycorrhiza formation is, at least in part, under the control of the host plant (Tagu et al., 2001, 2005). Given that we are now in a position to analyse genomic data from a number of host plants, there is a pressing need to combine ecophysiological and genomic analyses of both partners. One promising avenue of central importance in ECM-dominated boreal and north temperate forests is the processes contributing to soil organic matter formation and C storage (Podila et al., 2009). This includes identification of genes that regulate production of key enzymes involved in organic matter turnover, such as laccases, phenoloxidases and cellulases, and processes regulating the transport of labile C into the rhizosphere (Courty et al., 2010). Other important areas to focus on include the genes that control symbiosis development, including in planta accommodation and down-regulation of plant defence reactions, mineral nutrient acquisition, resource trading between partners (sensu Kiers et al., 2011), and induction and formation of reproductive organs (sporophores). L. bicolor genets have been found to fruit asynchronously, with the suggestion that fruiting is under genetic rather than environmental control (Selosse et al., 2001).

X. Conclusion: the need for a ‘community genetics’ approach in future mycorrhizal symbiosis research

A central question in ecology and evolution concerns understanding the mechanisms and controlling factors that lead to speciation; indeed, it is paradoxical that theory generally predicts low diversity and yet, in many natural ecosystem species, diversity is very high. Although the idea of ‘community genetics’ (i.e. considering genetic variation within both populations and species) has been established for some time (Antonovics, 1976), intraspecific diversity is beginning to be recognized as being a key factor that reconciles this paradox (e.g. Vellend & Geber, 2005). For example, a recent meta-analysis of tree species data from southeastern USA demonstrated that genotypic variation can maintain species diversity and allow for coexistence (Clark, 2010). One could envisage a similar mechanism occurring in mycorrhizal fungi, given the wide variation in traits seen within many species (see Sections IV and V). We therefore emphasize the need to obtain tightly linked data on the population, community and function (or traits) of organisms within a habitat; this has rarely, if ever, been considered in the case of mycorrhizal fungi but we must redouble efforts to make progress in this area. Some advances have been made (e.g. Parrent et al., 2010), but it is important to ensure consistency in the identification and measurement of what are considered key traits. We suggest focusing on traits directly related to the exchange of resources, primarily C, N and P, which underpin most mycorrhizal symbioses and many important ecosystem functions.

There is clear emerging evidence that intraspecific variation of both mycorrhizal plants and fungi is a crucial component of biodiversity. In some cases, intraspecific variation is of equal importance to interspecific variation, especially in its response to perturbations. We therefore conclude by emphasizing the need to consider the large variation in genotypes of both mycorrhizal plants and fungi, regardless of mycorrhizal type, when investigating the importance of biodiversity for ecosystem processes. We identify a sequence of broad research areas, each of which requires a suite of approaches and sustained effort to fully understand the role of individuals in ecosystems:

  • 1Identify natural genotypic diversity of both plant and fungal partners in ecosystems, and understand how this relates to species diversity;
  • 2Undertake controlled experimentation that can unravel how genetic diversity identified in (1) affects ecosystem processes (e.g. productivity, biogeochemical cycles);
  • 3Quantify the relative contributions of genetic diversity and phenotypic plasticity to variation measured in (2).
  • 4Determine how genotypic diversity of mycorrhizal plants and fungi regulates resource stoichiometry (i.e. the flux of mineral nutrients and C among genotypes of mycorrhizal plants and fungi);
  • 5Understand the context-dependent nature of the effects of genetic diversity (i.e. genotype × genotype × environment interactions) because individuals may interact very differently depending on the environmental conditions. This is probably best addressed using both controlled experimentation and in situ ecosystem approaches.

These core areas of research are of equal importance across orchid, AM, ERM and ECM symbioses and in the diverse ecosystems in which they occur. However, some specific questions are more easily addressed in particular ecosystems or plant/fungus combinations. This is especially likely to be the case in the short term in relation to genomic questions, given the limited number of genomes of mycorrhizal fungi that have been sequenced to date (Table S1), although the rapidity of technological advances is such that whole-genomic information of mycorrhizal fungi from environmental samples is a real prospect in the near future. Experimental manipulation of gradients of both genotypes and species of fungi and plants may enable us to determine whether there are optimum degrees of diversity that are required to maintain key ecosystem functions. For a biodiversity-ecosystem function perspective, the total genetic diversity may ultimately be the most tractable measure (Fig. 3). Moreover, the tractability and ecological importance of mycorrhizal systems makes them ideal models to test and develop biodiversity theory.


We thank Prof. Ian Sanders and Dr Lucy Gilbert for useful discussion, the editor (M-A. Selosse) and referees for their constructive comments, the Natural Environment Research Council, the Genomic Science Program of the US Department of Energy, Office of Science, Biological and Environmental Research (Plant-Microbe Interface project), the INRA and the University of Western Sydney for funding.