The majority of land plants form symbiotic associations with communities of arbuscular mycorrhizal fungi (AMF). These mutualistic soil fungi usually promote plant growth (Smith & Read, 1997; Klironomos, 2003) and AMF communities influence a number of important ecosystem processes, including plant productivity, plant diversity and soil structure (Grime et al., 1987; van der Heijden et al., 1998, 2006; Vogelsang et al., 2006). Recent studies have provided some fascinating insights into the structure and diversity of AMF communities in the field. Differences in AMF communities have been found between plant species, ecosystems, locations and seasons (Bever et al., 2001; Husband et al., 2002; Öpik et al., 2006), but also between different parts of the same root system, such as the roots and root nodules of legumes (Scheublin et al., 2004). However, most studies remain at the level of observation, and it is difficult to link the identification of AMF communities in the field with the functional significance of these AMF communities. Here we evaluate the problems that can be encountered in linking identification and functional significance of AMF and discuss possible approaches to deal with these problems. We present a list of 13 different functional traits that could be used to determine the existing functional diversity present within AMF communities. This mycorrhizal functional trait diversity could subsequently be used to predict the impact of specific AMF communities on plant growth and ecosystem functioning. Although knowledge of the composition of AMF communities in plant roots in the field is expanding, unfortunately this knowledge is rarely connected to studies on the function of the AMF symbiosis (Helgason et al. 2002; Read, 2002). It appears difficult to make this link between AMF community composition and function for several reasons: The species concept of AMF is still poorly developed and it has not been possible to show that particular AMF taxa have specific functions, as with animal and plant species, where a species is a genetically, morphologically and functionally distinct entity. Several recent studies have shown that AMF species or taxa, identified either by spore morphology or by ribosomal DNA, are highly variable in several functional traits (Hart & Reader, 2002a; Munkvold et al., 2004; Koch et al., 2006; T. R. Scheublin et al. unpublished). For example, Munkvold et al. (2004) compared the effects of 24 AMF isolates on plant growth and phosphorus (P) nutrition. They observed that growth and P uptake varied as much between isolates of the same AMF species as between different AMF species. Hence, these results indicate that the effects of AMF communities on plant growth cannot be predicted based on the species composition of AMF communities. Several molecular studies have shown that unknown and uncultured AMF types are abundant in the field, especially in undisturbed ecosystems (Fitter, 2005; Stukenbrock & Rosendahl, 2005). For example, Helgason et al. (2002) reported that sequences of c. 60% of AMF types detected in the roots of woodland plants had no cultured representative. The functional significance of these uncultured AMF types is unknown and it is not possible to determine their impact on plant growth or ecosystem functioning. Hence, further studies should attempt to cultivate these uncultured AMF in order to identify their ecological function. Apart from these problems in linking AMF community composition with functional significance, there are also some difficulties in the determination of AMF communities itself, which could disturb the subsequent link with function: Some of the AMF types that are identified in the plant roots are never found as spores, and some of the spore types are never found in roots (Clapp et al., 2002). Hence a characterization of AMF communities in the field based on either spore communities or molecular profiling of AMF communities in plant roots alone is insufficient to cover the whole spectrum of AMF present within a community. Molecular methods are, in our view, more suitable to characterize AMF communities because these methods can target active AMF communities that are present within plant roots. However, molecular characterisation of AMF communities in species-rich natural communities can be time consuming because roots of many plant species need to be analysed to cover potential host specific AMF types (Bever et al., 2001) and because AMF types might be present in specific soil horizons (Oehl et al., 2005) or during specific times in the year (Merryweather & Fitter, 1998). Hence, in order to assess the total AMF community present at a specific site, it will be useful to use both methods because they can complement each other. In addition to this, the abundance of the mycelial networks of the different AMF types is generally not measured, and is likely to be very important for AMF traits such as nutrient uptake, and therefore for the functioning of the AMF symbiosis. Most studies have used qualitative approaches to investigate AMF communities and only a few have actually attempted to determine the relative abundance of various AMF types using quantitative molecular techniques (Alkan et al., 2006). It is necessary to know which AMF types are rare or abundant because ecological theory predicts that dominant species often have the largest impact on ecosystem functioning. Most studies have assessed the composition of AMF communities at one time-point and at one place; few have searched for repeated patterns measuring throughout seasons, at several locations and sampling enough plant individuals to determine natural existing variation. It will be difficult to link AMF community composition to a particular ecological function if AMF communities in plant roots are highly dynamic and rapidly changing. Hence, for a better understanding of AMF function in nature it is essential to search for repeated patterns. The use of DNA microarrays for the rapid and simultaneous detection of thousands of genes in environmental samples such as plant roots (van Straalen & Roelofs, 2006) provides a powerful technology to search for such repeated patterns. Similarly, it is important to know whether AMF isolates have the same impact on plant growth when experiments are repeated. Surprisingly, few ecological experiments with specified AMF isolates have been performed to test for such repeated patterns and this is something that needs attention. Thus, the search for repeated patterns in mycorrhizal ecology is necessary for finding general rules and patterns. It appears difficult to link observations on the composition of AMF communities with the functional significance of AMF because AMF isolates of the same species can be functionally highly diverse (see above). Hence, in our opinion, it will be necessary to develop a method to use functional traits and functional groupings of AMF to characterize AMF communities. Experiments with plants and soil organisms have already shown that ecosystem properties can be successfully explained using functional traits and functional diversity (Diaz & Cabido, 2001; Heemsbergen et al., 2004). For example, Tilman et al. (1997) observed that functional composition and functional diversity rather than species diversity were the principal factors explaining effects of increased plant species diversity on plant productivity and nutrient acquisition. Functional diversity is defined here as ‘the value and range of those species and organismal traits that influence ecosystem functioning’. For further information regarding trait selection and measures of functional diversity in ecology, we refer to Lavorel & Garnier (2002), Cornelissen et al. (2003) and Petchey & Gaston (2006). A number of traits have been used to distinguish AMF types from each other. Some of these traits are useful for taxonomic purposes (spore morphology, ribosomal gene sequences), while other traits tell something about the functional significance of a particular AMF type for plant performance or for ecosystem functioning. We have identified 13 different functional traits of AMF that could be used to determine the functional diversity present within a specific AMF community (Table 1). Several of these functional traits could be used to predict the effect of AMF communities on plant performance and on specific ecosystem processes. This list is not all-embracing; other traits could also be added. Some of the functional traits listed in Table 1 are known to vary between different AMF and are of functional importance for plant growth and ecosystem functioning. For example, several studies have shown that P uptake by plants is related to the amount of external hyphae produced by a specific fungus. Plants colonized by AMF types that produce large amounts of external mycelium usually acquire more P from AMF compared with plants that are colonized by AMF that form low amounts of hyphae (Jakobsen et al., 1992). Similarly, plant communities grown with AMF types that produce an extensive mycorrhizal mycelium acquire more P compared with plant communities that are colonized by AMF producing few hyphae (van der Heijden et al., 1998, 2006). Hence, the hyphal length of AMF present in the soil could be used to make predictions about the mycorrhizal contribution to plant P acquisition and plant productivity (in case P is limiting plant growth). Arbuscular mycorrhizal fungi can also be responsible for a high proportion of total P uptake when there is no AMF effect on plant growth (Smith et al., 2004; van der Heijden et al., 2006). Hyphal length (together with AMF colonization) could, under these circumstances, be used to estimate the mycorrhizal contribution to P acquisition independent from plant productivity. Moreover, the formation and stability of soil aggregates is also related to hyphal length (Miller & Jastrow, 2000) pointing to the importance of this trait for ecosystem functioning. Other fungal traits listed in Table 1 are promising in that they could be used to characterize additional functional diversity present within AMF communities; for example, it has been shown that there is temporal and spatial variation in root and soil colonization by different AMF (Merryweather & Fitter, 1998; Smith et al., 2000). Some AMF appeared to be active in summer while other fungi were more abundant in autumn (Merryweather & Fitter, 1998). Plant productivity could be higher in communities where both fungal types are present, assuming that both fungal types are beneficial and supply nutrients to the plant. Hence, this trait could be important for sustaining plant productivity throughout the growing season. Moreover, some AMF acquire nutrients near the roots while other AMF forage further away and explore a different soil volume (Smith et al., 2000). This can lead to increased productivity because there could be complementary resource use. A similar phenomenon has been observed in plants with different rooting depth. Host preference and host range are two functional traits that are, potentially, very important. Plant diversity and plant productivity could be very dependent on fungal diversity in case AMF have a restricted host range or strong host preference, and when different fungi promote growth of different plant species (van der Heijden et al., 1998). It is, however, still unclear if there are many (uncultured) AMF with restricted host range or strong host preference and whether they are abundant. A recent suggestion that AMF identity rather than AMF diversity is more important in explaining effects of fungal diversity on plant productivity (Vogelsang et al., 2006) does not point to the importance of host preference. The functional significance of other traits is still unclear. Different AMF vary in a wide range of characters, including speed of root colonization (Hart & Reader, 2002b), amount of root colonization (Smith & Read, 1997), spore production (Bever et al., 2001), the frequency of hyphal fusions and the integrity of hyphal networks (Giovannetti et al., 1999; de la Providencia et al., 2005), formation of vesicles and auxiliary cells (Morton & Benny, 1990) and the physiological activities of nutrient uptake and transport pathways (Boddington & Dodd, 1999; Burleigh et al., 2002). The integrity and stability of hyphal networks might for example be important for disturbed or dry environments where AMF types with stable networks provide a better linkage with the soil and protect the vegetation against drought or disturbance. Moreover, it has been shown that some AMF types can acquire nitrogen (N) and transfer it to the plant (Johansen et al., 1992; Frey & Schüepp, 1993; Hodge et al., 2001). It is unclear whether all AMF are able to do this. Fungal N acquisition would be an important functional trait if there is variation among different AMF types and if AMF transfer significant amounts of N to the plant. Functional diversity of AMF is especially important if different AMF have different functions (e.g. one fungus provides drought resistance and another fungus is responsible for nutrient acquisition). Physiological studies that show that different fungi provide different services are extremely scarce, despite observations that fungal diversity increases plant productivity (van der Heijden et al., 1998; Gustafson & Casper, 2006; Lekberg et al., 2007) and plant diversity (van der Heijden et al., 1998; Vogelsang et al., 2006). Early work has shown that plants can be classified into different functional groups according to their ability to tolerate stress and disturbance (Grime, 1979). Perhaps such functional groupings are also useful to distinguish different fungal strategies: For example, some AMF types (e.g. Glomus intraradices and Glomus mosseae) appear to have a more ruderal lifestyle as they produce large amounts of spores and are also found in disturbed sites (Jansa et al., 2003; Öpik et al., 2006). Such AMF types could be classified as being disturbance tolerant. Other AMF types form mycelial networks in soil, produce few (or no) spores and are characteristic for undisturbed infertile ecosystems (Stukenbrock & Rosendahl, 2005; Öpik et al., 2006). Such AMF types might be adapted to nutrient-poor conditions and could perhaps be classified as stress tolerant. It is likely that most cultured AMF belong to the first group, pointing to the need to isolate more AMF types from undisturbed natural ecosystems. Several traits appear to have a taxonomic basis, especially at higher taxonomic resolution. Hyphal networks produced by members of the two major AMF suborders, the Glomeraceae and Gigasporacea vary. Hyphal networks of the Glomeraceae are thought to be better integrated compared with those of the Gigasporacea because there are many more hyphal fusions (de la Providencia et al., 2005; Voets et al., 2006). Moreover, members of the Glomeraceae are usually fast root colonizers, allocate a larger fraction of fungal biomass within the root (Hart & Reader, 2002b) and they form vesicles for storage of lipids (Morton & Benny, 1990). By contrast, members of the Gigasporacea lack vesicles (Morton & Benny, 1990) and they are thought to produce relatively more external mycelium (Hart & Reader, 2002b). These different fungal strategies suggest that AMF communities with a higher diversity of AMF genera or families contain larger functional diversity. Moreover, one fungal trait, the P uptake per unit hyphal length appeared to be AMF species specific (Munkvold et al., 2004) and this trait could perhaps be used to discriminate different AMF species. When using the functional traits discussed above it is also very important to consider that plants vary in their response to AMF and that some plant species are much more sensitive to changes in AMF community composition (van der Heijden et al., 1998). Hence, the response of plant communities to changes in AMF communities also depends on the identity of the plants present in a specific plant community. In order to identify functional groups of AMF it is necessary to develop a method to rapidly characterize functional AMF traits, for example by using functional genes present within AMF communities. However, currently, AMF genetics and the relation between AMF genes, their expression levels, and functional traits such as hyphal length and nutrient acquisition are not yet fully understood. For example, different AMF taxa contain different P-transporter genes (Benedetto et al., 2005; Harrison, 2005). Genetic variations in P-acquisition pathways (hyphal uptake, hyphal storage as polyphosphate, hyphal transport, and supply to the plant) could potentially explain why different AMF supply different amounts of P to plants (Boddington & Dodd, 1999; Harrison, 2005). It is, however, still unclear whether such functional genes (and their activities) can be used to explain the effects of AMF communities on plant nutrition and plant growth. This is an area that is still largely unexplored and where much progress can be made. Hopefully, the sequencing of Glomus intraradices, which is currently being performed, will increase our understanding of AMF genetics and direct us to suitable functional genes as biomarkers for AMF functional diversity. Moreover, enzyme tests have been successfully used to characterize functional diversity of ectomycorrhizal fungal communities present on root tips (Courty et al., 2005). Such approaches could be useful for AMF, if there is a way to apply this method in situ. At present, AMF identified in the field cannot be linked to their ecological function and their impact on plant growth. Therefore, the only way to link AMF community composition with functional traits of AMF is to make use of artificial ecosystems. In artificial ecosystems, the AMF community composition can be controlled, and AMF isolates can be selected that can be distinguished by molecular identification methods. The diversity of AMF communities can be manipulated, and the performance of the AMF community as a whole can be compared with the performance of each of the individual AMF isolates that compose the community. The functional diversity of the AMF communities can be calculated using the functional attribute diversity of the AMF types composing the communities (Petchey & Gaston, 2006) and using a selection of traits listed in Table 1. Subsequently, it can be tested whether specific ecosystem functions (e.g. productivity, nutrient acquisition and soil structure) are influenced by AMF functional diversity. The controlled environment of artificial ecosystems also requires that many important factors of the natural situation, such as interactions with other organisms are eliminated (Read, 2002). Another approach to connect AMF community composition and function is to manipulate entire AMF communities and compare changes in community composition and functional properties of the entire community. Examples could be treatments such as fertilization or use of the fungicide benomyl, but also more recent technologies such as the use of hyphal compartments (Schweiger & Jakobsen, 1999) and bioassay plants (Johnson et al., 2005). Although the natural situation is reflected as best as possible with these methods, they are indirect, and therefore sometimes less conclusive because other factors are simultaneously affected. To date, experiments investigating the effects of AMF diversity on plant growth or ecosystem functioning have been performed in one or two growing seasons, and usually include widespread, easily culturable ‘ruderal type’ fungi. Future experiments need to be performed for a longer, ecologically realistic, growth period and with fungi that are abundant in the ecosystem studied (and not necessarily easy to culture). Observations that adult plants and seedlings are colonized by different AMF communities (Husband et al., 2002), and that juveniles and adults sometimes benefit from different AMF (van der Heijden et al., 2006), suggest that AMF types may vary in ways that we still do not completely understand. In this respect it is also important to use stable isotope probing of DNA (Friedrich, 2006) or fatty acids (Olsson & Johnson, 2005) to analyse which fungi are physiological active, where they are active and in which type of ecosystems. Several experiments have now shown that AMF isolates are functionally diverse and that the composition of AMF communities has a large impact on plant performance, plant community structure and ecosystem functioning. At the same time an increasing number of studies have investigated AMF communities in plant roots in the field for a large number of plants and in a wide range of ecosystems. However, few studies have made a direct link between the composition of AMF communities in the field and the functional significance of such AMF communities for plant growth and ecosystem functioning. This is an area that needs much more attention, not only to better understand the ecology of plants and plant communities but also to monitor whether introductions of AMF into the field have been successful (e.g. to enhance agricultural production or increase the effectiveness of nature restoration projects). We have suggested a number of approaches that are helpful in this respect. In our opinion, defining functionally distinct AMF groups is essential if we want to understand fully the interactions between plant and AMF communities in agricultural and natural ecosystems. We have therefore identified a number of functional traits (see Table 1) that could be used to predict the effects of AMF communities on plant growth and ecosystem functioning. This, together with microcosm studies and novel technologies will help us to better understand the functional significance of AMF for plants and ecosystems. This research has been supported by a grant from the Netherlands Organisation for Scientific Research (NWO Vernieuwingsimpuls grant 016.001.023 awarded to MvdH). We thank the editor and the reviewers for constructive and helpful comments.
No takes yet. Share an insight, caveat, or question.
Heijden et al. (2007) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: