There is a growing interest amongst community ecologists in functional traits. Response traits determine membership in communities. Effect traits influence ecosystem function. One goal of community ecology is to predict the effect of environmental change on ecosystem function. Environmental change can directly and indirectly affect ecosystem function. Indirect effects are mediated through shifts in community structure. It is difficult to predict how environmental change will affect ecosystem function via the indirect route when the change in effect trait distribution is not predictable from the change in response trait distribution. When response traits function as effect traits, however, it becomes possible to predict the indirect effect of environmental change on ecosystem function. Here we illustrate four examples in which key attributes of ectomycorrhizal fungi function as both response and effect traits. While plant ecologists have discussed response and effect traits in the context of community structuring and ecosystem function, this approach has not been applied to ectomycorrhizal fungi. This is unfortunate because of the large effects of ectomycorrhizal fungi on ecosystem function. We hope to stimulate further research in this area in the hope of better predicting the ecosystem- and landscape-level effects of the fungi as influenced by changing environmental conditions.
The study of community ecology is essential from the standpoint of land management, conservation and, therefore, policy. One major area of interest in community ecology concerns the environment as a determinant of the structure of the biological community. Another important area of interest concerns the effect of community structure on ecosystem function, especially what has been termed ‘biodiversity effects on ecosystem functions’ (Hillebrand & Matthiessen, 2009). For both there is at least an implicit concern for and, in many cases, an explicit interest in the various traits possessed by organisms because traits both determine suitability to habitat (and, therefore, membership in communities), and influence ecosystem function (Fig. 1a). Traits possessed by organisms that influence their response to the environment and, therefore, influence their suitability in a particular community are referred to as response traits (Lavorel & Garnier, 2002). An example of a response trait is the ability to grow at low water potentials, which is known to vary among fungal taxa (Coleman et al., 1989) and which will likely influence the composition of communities in arid habitats. Traits that influence ecosystem function are referred to as effect traits (Lavorel & Garnier, 2002). An example of an effect trait is the rate at which mycorrhizal fungal litter decomposes, which is known to vary among fungal taxa (Koide & Malcolm, 2009) and which can affect the size of the detritus pool and the rate of nutrient and carbon cycling.
One of the goals of community ecology is to predict the effects of disturbance (environmental change) on ecosystem function. There are both direct effects of environmental change on ecosystem function and indirect effects mediated by a shift in community structure (Fig. 1a). Consider a disturbance such as a drying climate, which can influence nutrient cycling in two ways. First, decomposition of litter and the cycling of nutrients will slow as a direct consequence of drying (Moore, 1986). Second, drying will likely have an impact on ectomycorrhizal fungal community structure (Swaty et al., 2004) and thus have an indirect effect on nutrient cycling. Only species possessing traits that help them cope with increasing water shortage are expected to persist in the community. If this shift in community structure results in an increase in species producing fungal litter that is relatively recalcitrant, there will be a further reduction in nutrient cycling.
The direct and indirect effects of environmental change on ecosystem function may be difficult to distinguish unless the timescales for direct and indirect effects are different. For example, if drying is imposed faster than communities change in response to drying, then the direct effects of drying can be assessed in the absence of a community shift. If environmental change and community shift occurs simultaneously, it becomes quite difficult to distinguish between direct and indirect effect of environmental change on ecosystem function, particularly if response traits are unlinked from effect traits (Fig. 1a). In that case, as the environment filters out species with unsuitable response traits, one is not able to predict the effect of this filtering on the effect traits possessed by the species in the community.
However, it is possible for certain traits to function as both response traits and effect traits (Fig. 1b). Our goal is to demonstrate for ectomycorrhizal fungi that such dual response–effect traits have disproportionate consequences for ecosystem function because the response traits favored by the environment are the very same effect traits that influence ecosystem function and, due to ecological filtering, they are naturally found in high frequency in the community. Such traits are thus among those that should be given special consideration in ecological studies. They permit us to distinguish between direct and indirect effects of environmental change on ecosystem function because the indirect effects become predictable. While the literature on functional traits is relatively well developed for plants (see Lavorel & Garnier, 2002), regrettably it has not yet become well developed for the ectomycorrhizal fungi (but see Bueé et al., 2007; Lilleskov et al., 2011) even though such fungi may have large effects on ecosystem function (Read et al., 2004).
Among the potentially important ectomycorrhizal fungal response traits influencing their abundance in various communities are preference for N source, exploration morphotype and the deposition of melanin in cell walls. We suggest herein how such traits are also effect traits as they influence ecosystem function, primarily carbon (C) and nitrogen (N) cycling. In particular we highlight four examples in which the traits possessed by ectomycorrhizal fungi that help to determine their membership in particular communities are the very same traits that may have large effects on ecosystem function. The four examples are those for which we could assemble the most complete relationship between environment, trait, community structure and ecosystem function based on the available literature and our own data. We hope that these examples will serve to stimulate a healthy discussion about and further research into the links among environmental conditions, community structure, ecosystem function and species attributes that serve as both response and effect traits. We consider this to be an important yet poorly investigated intersection of research areas with implications for conservation, restoration and management in the midst of alterations of land use and climate.
Example 1: Variation in N source, preference for N source, N and C cycling
The availability of organic N, ammonium and nitrate in the soil solution are influenced by the rates of mineralization, microbial immobilization, nitrification and denitrification, which are themselves influenced by factors such as pH, water content, temperature and litter quality (Hobbie, 1992; Maag & Vinther, 1996; Persson et al., 2000). Thus, habitats differ in the kinds of N sources available to both plants and their ectomycorrhizal fungi.
Plant species exhibit a range of capacities to absorb the various sources of N available in the soil solution (Gessler et al., 1998), and it is not unreasonable to hypothesize that natural selection has promoted adaptations that allow plants to efficiently utilize N sources that are frequently available to them (Falkengren-Grerup, 1995). Thus, arctic species that experience higher concentrations of organic N sources might be expected to utilize organic forms of N more frequently than temperate species from habitats with rapid rates of N mineralization (Kielland, 1994).
It stands to reason that, with respect to obtaining N, ectomycorrhizal fungi face similar selection pressures and ecological filtering to plants. It is certainly true that ectomycorrhizal fungi vary in their capacity to absorb the various forms of N that are available in nature; some isolates grow best on mineral sources of N while others grow better on organic forms (Abuzinadah & Read, 1986). Most species appear to prefer ammonium over nitrate, while others exhibit little preference (Finlay et al., 1992). However, while it is abundantly clear that the addition of mineral N to the environment can have marked effects on the structure of ectomycorrhizal fungal communities (Lilleskov et al., 2002a), it is less clear that this is in any way related to variation in the ability to utilize the various N sources available to them in their habitats.
Nevertheless, patterns similar to those observed for plants have been observed for the ectomycorrhizal fungi (Tibbett et al., 1998; Taylor et al., 2000; Lilleskov et al., 2002b). The study of Tibbett et al. (1998) concerned the abilities of isolates of Hebeloma from temperate and arctic habitats to acquire N from organic and inorganic sources of N. The isolates from the arctic sites, where N mineralization is slow, were more frequently able to acquire N from organic sources than were the isolates from the temperate sites, where N mineralization is more rapid. Taylor et al. (2000) documented communities of ectomycorrhizal fungi along an N deposition gradient from Scandinavia to central Europe. Lilleskov et al. (2002b) studied ectomycorrhizal fungal communities along a relatively short atmospheric N deposition gradient caused by a fertilizer manufacturing plant in Alaska. In these studies it was first demonstrated that the structure of the ectomycorrhizal fungal communities varied significantly along the N availability gradients. It was further shown that fungal isolates from sites on the high end of the mineral N concentration gradient were less capable of utilizing protein as an N source than isolates from sites on the low end of the gradient. These three case studies provide evidence that, just as for plants, ecological filtering or natural selection may favor physiological capacities in the fungi that allow them to utilize N sources that are frequently available to them.
The consequences of this, in terms of ecosystem function, are potentially large. Although ectomycorrhizal fungi do not simply pass N on to their hosts after they have absorbed it from the environment under all circumstances (Finlay et al., 1992), many ectomycorrhizal fungi do transfer to their hosts substantial quantities of N (Abuzinadah et al., 1986). The consequence of the capacity to specialize on N sources that are most available in the habitat is that a greater fraction of the total ecosystem N will be controlled within the plant–fungus symbiosis than would be possible if the fungi were less capable of acquiring the available N sources. Less of the N would theoretically be found in the saprotrophic microbial pool or be subject to leaching or denitrification. To the extent that N limits plant productivity, ecosystem productivity would be increased. Moreover, because the cycling of N is tightly coupled with that of C, specialization by ectomycorrhizal fungi on the type of N that is most available in the environment may lead to greater soil C sequestration via enhanced photosynthesis and greater soil C deposition by the mycorrhizal fungi (Orwin et al., 2011).
Example 2: Rooting density, fungal exploration strategy, C and N cycling
Fine roots are the primary source of C for ectomycorrhizal fungi. Therefore, the ability to colonize new roots is essential to their survival. However, early successional communities are often characterized by low rooting density (Peay et al., 2011) and quite probably, a patchy distribution of roots. Because ectomycorrhizal fungi vary markedly in the distance they can grow from their colonized root, in early successional communities ectomycorrhizal fungal species with long-distance exploration morphotypes (Agerer, 2001) may be favored over those with shorter outgrowths from colonized roots. This hypothesis was tested by Peay et al. (2011), who found that communities with low to intermediate rooting densities had a high frequency of long-distance exploration types while communities with high rooting density had a high frequency of short- and medium-distance exploration types.
The basic unit of construction of mycelial fungi (as opposed to yeasts) is the hypha, a linear structure comprising a chain of individual cells. Consequently, single hyphae are slender, on the order of a few microns thick. Such hyphae are obviously capable of transporting materials over some distance but, because of their small diameter, long-distance transport of materials is inherently inefficient. They are also slender enough to be easily broken by the slightest movement of soil and are highly vulnerable to fungivorous arthropods (Johnson et al., 2005). However, some fungal species, particularly some members of the Basidiomycota, are capable of aggregating many hyphae into what have been called cords or rhizomorphs. These are the characteristic structures of species that possess the long-distance exploration morphotype. Cords or rhizomorphs comprise both a symplast of living, individual hyphae and a (sometimes large) apoplast. Materials are assumed to move passively down concentration and hydrostatic pressure gradients in both symplast and apoplast but, within the symplast, active transport mechanisms are also possible (Cairney, 1992). By virtue of their large diameter, cords and rhizomorphs are assumed to be inherently more efficient in moving materials over relatively large distances than individual hyphae (Cairney, 1992) and, of course, they are far more resistant to breakage. Much more has been learned about the nature and ecological significance of the cords and rhizomorphs of saprotrophic fungi (Boddy, 1999) than those of ectomycorrhizal fungi. Nevertheless, in terms of acquiring and transporting resources such as C, N and P, the two types of fungi are no different. Both explore their substrates to locate sources of limiting resources, absorb the resources, and transport them to regions of the mycelium undergoing growth or experiencing high physiological activity. Thus, saprotrophic fungal cords and ectomycorrhizal fungal cords function similarly (Cairney, 1992).
There are many, potentially large, ecosystem-level functional consequences of a shift in community structure from a preponderance of species with shorter exploration morphotypes to species with longer exploration morphotypes. As shown in example 1, more effective N uptake would result in a greater fraction of the total ecosystem N being controlled within the plant–fungus symbiosis than would otherwise be possible, and less of the N would theoretically be found in the saprotrophic microbial pool or be subject to leaching or denitrification. Greater ecosystem productivity would result from transfer of N from fungi to host plant.
The relatively large amount of C allocated to ectomycorrhizal fungi by their hosts (c. 20% of total photosynthesis, Hobbie, 2006), and the composition of their cell walls (Koide & Malcolm, 2009) makes ectomycorrhizal fungal litter and its decomposition products a significant soil pool of both C and N (Clemmensen et al., 2013). Weigt et al. (2012) have noted that species producing cords or rhizomorphs tend to produce larger pools of biomass than those that produce mainly shorter distance exploration types. They estimated that, on average, long-distance exploration types contributed c. 15 times more biomass in their extraradical mycelium than short-distance exploration types. Long-distance exploration types may also contribute disproportionately to the pool of neromass (litter). Upon death, rhizomorphs and cords are likely to turn over more slowly than individual hyphae (Treseder et al., 2005), possibly because they possess far less surface area per unit weight than individual hyphae (Koide & Malcolm, 2009) or because of their hydrophobicity (Unestam & Sun, 1995). Thus, rooting density, which may influence the complement of exploration morphotypes in the community, may indirectly control the size and nature of the C and N pools in the soil.
Rhizomorphic or cord-forming ectomycorrhizal fungal species may also influence soil organic matter pools in another, indirect way. The withdrawal of water from a volume of soil by cord-forming fungi (Brownlee et al., 1983) may influence the rate of litter decomposition. In wet weather the extraction of water by ectomycorrhizal fungi and transfer of this water to the host is assumed to be largely inconsequential insofar as litter decomposition is concerned. Indeed, under moist conditions, the limited saprotrophic activity of some ectomycorrhizal fungi (Koide et al., 2006) may actually contribute to decomposition directly. However, Koide & Wu (2003) suggested that in dry weather the extraction of water from litter by ectomycorrhizal fungi may retard its decomposition by decreasing water availability to decomposer microorganisms. This could increase soil C sequestration in the form of litter and reduce the rates of nutrient cycling, nutrient leaching and denitrification. Whether a reduced nutrient cycling rate affects ecosystem productivity depends, of course, on the relative constraints of water and nutrient availability.
Example 3: Elevated atmospheric CO2 concentration, fungal exploration strategy, N and C cycling
Elevated atmospheric CO2 concentration may increase ecosystem productivity (Lloyd & Farquhar, 1996). This can be accompanied by a significant shift in the structure of ectomycorrhizal fungal communities, frequently expressed in higher abundances of species with long-distance exploration morphotypes (Fransson, 2012 and references therein).
The mechanism by which elevated atmospheric CO2 concentration and the accompanying enhanced productivity increase the abundances of ectomycorrhizal fungal species with long-distance exploration morphotypes is not entirely clear. Some have reasoned that this occurs because such species, which produce and maintain large structures in the soil, require greater amounts of C than do short-distance exploration types (Godbold et al., 1997). That remains a viable hypothesis but, unfortunately, we do not yet know whether long-distance exploration species actually require more C for their production and maintenance than short-distance exploration types. Carbon demand is determined by the rate of fungal production, among other things, but the size of the standing crop of fungal tissues is not necessarily an indication of rate of production because of possible variation in the rate of turnover of tissues. Indeed, some evidence is consistent with the hypothesis that short-distance morphotypes require more C than long-distance types (Wallander et al., 2010).
As a consequence of elevated atmospheric CO2 concentration, N limitations may become more severe in temperate forests while P limitations may increase in tropical forests (Lloyd & Farquhar, 1996 and references therein). Thus, an alternative hypothesis is that plants preferentially associate with ectomycorrhizal fungi possessing the long-distance exploration morphotypes under conditions of elevated atmospheric CO2 concentration in order to increase their capacity to access N or P. Depending on root density, the exploitation of volumes of soil at some distance from the root via ectomycorrhizal fungi with a long-distance exploration strategy may allow more effective N capture. In addition, long-distance types may also improve the effectiveness of N capture when nutrients are not uniformly distributed in the soil. Nutrients are frequently patchily distributed because of the heterogeneous nature of litter, the formation of soil horizons, and spatial variation in mineralogy, microtopography, animal activity and the location of individual plants (Stark, 1994). The patchiness may occur at a physical scale that could hamper their acquisition by nonmycorrhizal roots or roots colonized by short-distance exploration types (Fitter, 1994). Long-distance exploration types may be particularly useful in simultaneously acquiring resources from multiple resource patches.
We hypothesize, therefore, that a shift in community structure resulting in an increased abundance of species possessing the long-distance morphotype (rhizomorphic or cord-forming species) as a consequence of elevated CO2 concentrations will potentially have large direct and indirect effects on N and C cycling as detailed in example 2. However, greater C allocation to ectomycorrhizal fungal biomass under elevated CO2 could actually result, under some circumstances, in increased N immobilization and thus reduced N availability to the host (Näsholm et al., 2013).
Example 4: Water stress, cell wall melanization, C and N cycling
The pigment melanin may play an essential role in protecting fungal cells from a range of stressors including UV radiation (Singaravelan et al., 2008), high concentrations of heavy metals (Gadd & Derome, 1988) and reduced water potential (Kogej et al., 2006). Melanin is not a single compound but a group of complex polymers composed of phenolic and indolic monomers. These are found in animals, plants, bacteria and fungi (Butler & Day, 1998). In fungi, melanin is deposited in the cell wall where it forms complexes with other cell wall components such as chitin, β-glucans and proteins. While they are nonessential in primary metabolic pathways, melanins may still serve essential survival functions, particularly under stressful environmental conditions (Butler & Day, 1998).
Ectomycorrhizal fungi vary greatly both in the concentration of melanin in their cell walls and in the type of melanin they produce. There are four known classes of fungal melanins based on their biochemical precursors including γ-glutaminyl-3,4-dihydroxybenzene (GDHB) melanin, dihydroxyphenylalanine (DOPA) melanin, dihydroxynaphthalene (DHN) melanin and catechol melanin (Butler & Day, 1998). The type of melanin produced by fungi seems to be largely governed by their phylogeny: Basidiomycetes produce GDHB and DOPA melanins; Ascomycetes produce primarily DHN melanin but may also produce DOPA and catechol melanins (Butler & Day, 1998).
The heavily melanized and ubiquitous ascomycetous ectomycorrhizal fungus, Cenococcum geophilum Moug. & Fr., is known for its ability to tolerate greater water stress than many other ectomycorrhizal fungal species (Mexal & Reid, 1973; Pigott, 2006) and there is reason to believe that this drought tolerance is, in part, the result of heavy melanization of its cell walls (C. Fernandez & R. Koide, unpublished data). Water stress, in fact, appears to result in the dominance of Cenococcum geophilum in many arid and seasonally water-stressed communities (Querejeta et al., 2009).
While water stress may predispose some communities to comprise a large proportion of melanized fungi, melanization may, in turn, influence decomposition. The melanization of the cell walls of C. geophilum, for example, is thought to make it highly resistant to decomposition (Malik & Haider, 1982). Litter resulting from the death of ectomycorrhizal fungi represents a large input into forest ecosystems (Clemmensen et al., 2013). Thus, communities with C. geophilum as a dominant member of the ectomycorrhizal fungal community may contain an unexpectedly high concentration of C stored in the form of highly recalcitrant fungal litter or partial decomposition products. Indeed, Meyer (1964) noted that a large proportion of C. geophilum ectomycorrhizas appearing to be dead accumulated in the soil and may be an important contributor to stable soil C. This has been supported using minirhizotron imaging to document the turnover rate of C. geophilum ectomycorrhizas, which were found to have a median persistence that was 4–10 times longer than ectomycorrhizas produced by other ectomycorrhizal fungal species. A large proportion of the C. geophilum ectomycorrhizas was also found to be dead using a vitality stain (Fernandez et al., 2013). In addition to the fungal tissues that directly colonize root tips, the extramatrical mycelium and sclerotia (hard, resting structures) produced by C. geophilum are also likely to be large and stable pools of C in forest soils. Fogel & Hunt (1983) found that melanized hyphae accounted up to 68% of the total hyphal mass found in the soil of Douglas fir stands and speculated that accumulation is caused by high resistance to decay, and Dahlberg et al. (1997) found that sclerotia of C. geophilum may account for a surprisingly large fraction of total belowground ectomycorrhizal fungal biomass and necromass. Sclerotia of C. geophilum have been found to range in age from 200 to 1200 yr old in mineral soil horizons using 14C dating and were found to be older than the humic substances found in the same horizon (Watanabe et al., 2007). Even more remarkable are the reports suggesting that what were once thought to be C spherules formed from the impact of a comet c. 12 000 yr ago are in fact undecomposed sclerotia of C. geophilum (Scott et al., 2010)! This indicates that communities in which melanized ectomycorrhizal fungi are dominant may contain an unexpectedly high concentration of C and N sequestered in fungal tissues.
While plant ecologists have discussed, for some time, the utility in examining response and effect traits in the context of community structuring and ecosystem function, this approach has not yet been applied widely to the ectomycorrhizal fungi. This is unfortunate because of the large effects ectomycorrhizal fungi have on ecosystem function. We have called attention to four examples in which the traits possessed by ectomycorrhizal fungi, as they interact with the environmental conditions of the habitat, determine their membership in a particular community. Furthermore, we have suggested ways in which the same traits may significantly influence ecosystem function, particularly C and N cycling. In other words, some response traits are effect traits. The coincidence of response and effect traits is important because a response trait favored by the environment and, therefore, found in high frequency in the community, is the very same effect trait that influences ecosystem function. This makes it easier to predict the indirect effects of environmental change, mediated through a community shift, on ecosystem function. Traits that are both response and effect traits thus have disproportionate consequences in communities and should be given special consideration in ecological studies. Other ectomycorrhizal fungal traits that may be of particular interest to consider as dual response and effect traits include those having to do with host preference, dispersion (animal vs wind, aboveground vs belowground fruiting, etc.), the ability to access resources from litter and hydrophobicity. We have avoided discussion of the linkage of traits, including tradeoffs among trait states, primarily because there are so few studies that address them. This is certainly an area of research that deserves more attention. It would be interesting to determine, for example, whether the degree of melanization is negatively correlated with growth rate. If that were true a high degree of melanization might be disadvantageous to these fungi under continuously high water availability. In any case, we hope that this contribution will stimulate further research at the fascinating intersection of community structuring and ecosystem function. In our view, research into these dual, response and effect traits will enable us to better manage natural resources because we will be more able to predict the ecosystem- and landscape-level effects of the fungi as influenced by changing environmental conditions.
We thank the anonymous reviewers and M-A. Selosse for their valuable comments on previous versions of this manuscript. We gratefully acknowledge funding from the US National Science Foundation, US Department of Agriculture, the Alfred P. Sloan Foundation, the Department of Plant Science of the Pennsylvania State University, USA and the College of Life Sciences of Brigham Young University, USA.