Ectomycorrhiza and the open nitrogen cycle in an afrotropical rainforest


The evolution of the ectomycorrhizal symbiosis raises an important challenge for ecologists. How did ectomycorrhiza manage to find a niche where it could outcompete the highly successful, and much older, arbuscular mycorrhizal symbiosis? To what extent can we use current ecosystems where ectomycorrhizal and arbuscular mycorrhizal plants co-occur in order to understand their differential performance? Tropical ecosystems provide major opportunities in this regard. Some tropical forests are located on very old and highly weathered substrates, where phosphorus (P) is the limiting element, constituting a habitat ideally suited for arbuscular mycorrhizal trees. But some forests, like those of the Guineo-Congolian region of western Africa, are dominated by ectomycorrhizal trees. The functioning of the ectomycorrhizal symbiosis in these forests is still largely unknown. In this issue of New PhytologistTedersoo et al. (pp. 832–843) describe enzyme activities and stable isotope patterns of ectomycorrhizas in a Gabon rainforest that shed some light on the role of these fungi in accessing nutrients from organic sources. They determined potential enzyme activities (of enzymes that are involved in the breakdown of organic material or the release of nutrients from organic sources) and stable isotopes (13C and 15N) of individual ectomycorrhizal root tips that were assigned to different fungal species based on molecular sequence analysis. As both parameters (enzyme activities, stable isotopes) reflect aspects of nutrient acquisition, a reasonable hypothesis would be that both would be correlated. The forest, Crystal Mountains National Park in north-western Gabon, which has an unique flora due to its status as a Pleistocene forest refuge, is very rich in ectomycorrhizal fungi, with 101 ectomycorrhizal fungal species revealed on 287 root tips in this study. Ample opportunities for fungal niche differentiation should therefore exist as well.

Their study clearly showed that potential enzyme activities and 15N natural abundance of individual root tips were not correlated. Lack of correlation between both parameters raises the question whether both indicate the same aspects of mineral nutrition or uptake from organic sources. Lack of correlation seems bad news, even though microbial soil ecologists have learned to accept that different assessments of soil microbial communities regularly do not correlate. Lack of correlation can both be due to inherent methodological problems in these assays but also be explained (away) by noting that both parameters reflect different aspects of that mineral nutrition. Currently, we cannot confidently conclude which happens to be the case, although there is no doubt that both methods have limitations.

‘Lack of correlation seems bad news, even though microbial soil ecologists have learned to accept that different assessments of soil microbial communities regularly do not correlate.’

Tedersoo et al. note that their enzyme assay indicates potential activity. Buée et al. (2005) noted that the enzyme assay measures activity in a buffer solution rather than at the water potential of the soil. This issue of water availability is important because of the question to what extent the rate-limiting step for these enzymes in the soil is constituted by enzyme levels or by substrate availability. Under field conditions substrate availability has been shown to be the rate-limiting step for phosphatase (Tinker & Nye, 2000) and protease (Mooshammer et al., 2012). In the study by Tedersoo et al., the enzyme involved in acquisition of organic nitrogen (leucine aminopeptidase) did not show significantly higher levels in ectomycorrhizal root tips compared with nonmycorrhizal roots, suggesting that acquisition of organic nitrogen (N) by ectomycorrhizal fungi was unimportant in this forest. Several data further supported the conclusion that the benefit of being ectomycorrhizal in these forests is not due to enhanced performance in acquisition of organic N: the δ15N signal was higher in ectomycorrhizal trees than in the arbuscular mycorrhizal trees (in contrast to studies from boreal and arctic ecosystems where N is the limiting nutrient and where ectomycorrhizal plants have a lower δ15N signal than arbuscular mycorrhizal plants) and the small difference in δ15N between sporocarps of saprotrophic and ectomycorrhizal fungi (in contrast to earlier studies that suggest that both guilds can be separated by their 15N (and 13C) signals). Their data therefore show an open and rapid N cycle. But an open N cycle (characterized by high inputs and high N losses, due to the fact that N is not a limiting nutrient) is not characteristic for all tropical forests that are dominated by ectomycorrhizal trees. Observations that the recalcitrant litter of ectomycorrhizal trees slows down the N cycle, and that such trees gain competitive superiority due to the ability of their ectomycorrhizal fungi to access organic N sources have also been reported (Brearley et al., 2003; McGuire et al., 2010). Surprisingly, extremely high δ15N signals were observed in sporocarps of species of Hygrocybe and unidentified members of the Gomphales, suggesting either new trophic modes or new mechanisms of isotope fractionation. These fungi also had lower δ13C signatures than both ectomycorrhizal and saprotrophic fungi and these low values are currently unexplained as well.

If not N, what about P? Acid phosphatase was almost three times higher in ectomycorrhizal than in nonmycorrhizal root tips. It was also significantly higher in ectomycorrhizas from the /tomentella-thelephora lineage than the /russula-lactarius lineage, the two major lineages of tropical ectomycorrhizas (et al., 2012). But such a strong uncoupling of organic N and organic P acquisition is surprising in the face of the strong stoichiometric constraints (Sinsabaugh et al., 2009). The rapid litter breakdown rate in the studied forest would also argue a major role for organic P acquisition. For the six further enzymes that Tedersoo et al. assessed, and that are involved in carbon (C) cycling, a further question needs to be asked – to what extent are such enzymes functional? This question pertains to the hotly debated issues to what extent ectomycorrhizas exhibit substantial saprotrophic activity, being able to gain C by decomposing organic materials. However, recent suggestions that C compounds, not nutrient limitation, limit breakdown rates in tropical forests (Hättenschwiler et al., 2011) may necessitate reconsidering the role of these enzymes in the ectomycorrhizosphere.

Despite significant differences in enzyme activities and δ15N signatures between ectomycorrhizal root tips of different fungi (both in terms of lineage and exploration type), such differences did not or hardly translate into differences between different ectomycorrhizal tree species. There were very few differences in enzyme activities between tree species, and equally stable isotope signatures showed few differences, not only within the guild of ectomycorrhizal trees but also between ectomycorrhizal and arbuscular mycorrhizal trees. Small functional differences between ectomycorrhizal tree species may not be very surprising, considering the low host specificity and high ability to form common mycorrhizal networks of afrotropical ectomycorrhizal fungi (Diédhiou et al., 2010; Tedersoo et al., 2011); however it raises the issue about the functional relevance of differences between fungal lineages, if trees average out such differences.

Here we have an unsolved paradox of scale: while the study of Tedersoo et al. shows the benefits of ever more detail by analyzing the signals of individual root tips, we are still a long way from explaining why individual fungal species (or at least the 101 species assessed in this study) do really matter in understanding the niche that ectomycorrhizal trees occupy in this habitat with its open N cycle.