Elevational dynamics in soil fungal communities
The results presented here show that fungi with different life strategies seem to be favored by different environmental conditions, as indicated by functional differences among the elevational zones. Even in functional groups with no significant differences in richness, we observed strong differences in composition as a function of elevation. Similar elevational differences in richness and composition also have been observed in various functional groups of plants and animals (Cardelús et al . 2006; McCain 2009; McCain and Grytnes 2010; Guoet al. 2013).
Our results suggest the importance of vegetation type for species richness and community composition of functional groups of fungi associated with plants. For ECM fungi, richness correlates positively with taxonomic richness and density of ECM host plants both on global and regional scales (Tedersoo et al. 2014). Similarly, the data presented here suggest that ECM fungal richness mirrors host richness and density along elevation gradients in both the Neo- and Palaeotropics, as inferred from previous studies on tree community richness and composition of the sampled forest types (Aiba and Kitayama 1999; Malizia et al . 2012). This also results in important differences in the elevational patterns of ECM fungi between the Neo- and Palaeotropics. In the Neotropics, upper montane forests tend to harbor the highest density of ECM trees, i.e., with Alnus acuminata as the principal ECM host in the Yungas in Argentina, andAlnus acuminata and several Quercus spp. in the mountain range spanning southern Costa Rica and northern Panama (Malizia et al . 2012; Kappelle 2016; Wicaksono et al . 2017; Nouhra et al. 2018). These findings are in agreement with sporocarp-based studies in other areas of the wet Neotropics that reported highest richness of ECM fungi in upper montane forests (Mueller et al. 2006; Gómez-Hernández et al. 2012). In Borneo, in non-ultramafic soils, richness and density of ECM trees are high in lowland and lower montane forests and decline in upper montane forests, which is reflected here and is discussed in detail by Geml et al . (2017). At high elevations, ECM hosts such as Leptospermum recurvum tend to be dominant only in ultramafic soils (Aiba and Kitayama 1999), which were not included in this study. The above patterns in tropical mountains differ from those observed in temperate mountains, where ECM fungal richness tends to decrease monotonically with increasing elevation (Bahram et al. 2012; Nouhra et al. 2012; but see Bowman and Arnold 2018). The greater diversity and abundance of ECM fungi in montane forests likely contribute to the accumulation of soil OM through the continuous addition of fixed C to the rhizosphere, as in ECM-dominated boreal forests (Clemmensen et al. 2013), and as a result of negative interactions with saprotrophic fungi that inhibit decomposition (Gadgil and Gadgil 1971; Fernandez and Kennedy 2016).
Richness of plant pathogenic and saprotrophic fungi generally correlates positively with temperature as reported by a wide range of studies from landscape-level to global scales (Gómez-Hernández et al. 2012, Tedersoo et al . 2014, Geml 2019). In our study, richness (both proportional and absolute) of wood decomposers, and to some extent plant pathogenic fungi, was predictably higher at lower elevations, but richness patterns of saprotrophs were less predictable. In all three regions, the lowest proportional richness of saprotrophs always coincided with the highest proportional richness of ECM fungi, possibly as a result of competitive interactions mentioned above, i.e. “the Gadgil effect”. Regardless of differences in richness, strong community structure as a function of elevation, observed in all functional groups, suggests habitat specificity and elevational turnover of species within functional guilds as well. We note that some of the fungi identified under the functional guilds of plant pathogens and saprotrophs may prove to be endophytes, raising an interesting direction for future study. Our work suggests that given the ecological and phylogenetic overlap of endophytic, saprotrophic, and pathogenic taxa in terms of ITS sequences, endophytes will broadly follow the patterns described here.
We observed several consistent patterns regarding the distribution of taxonomic groups across functional guilds, suggesting a certain level of phylogenetic conservatism with respect to environmental niches. For example, the consistently higher richness of Sordariomycetes, particularly Hypocreales, Sordariales and Xylariales, at lower elevations is apparent in various functional groups and is in agreement with higher host and substrate diversity for these fungi, which include plant pathogens as well as saprotrophs. Unlike in boreal and temperate forests, Sordariomycetes are the dominant class of plant endophytic fungi in tropical lowland forests (Arnold and Lutzoni 2007), in agreement with their prevalence in soil at low elevations. In a similar manner, the higher richness of Leotiomycetes, particularly Helotiales, in montane cloud forests is apparent in several functional groups, such as root-associated fungi, plant pathogens, saprotrophs, and wood decomposers. Helotiales appear to be the most diverse order of ascomycetes in arctic tundra ecosystems (Semenova et al. 2015), and the above trend confirms that many Helotiales taxa thrive in relatively colder climates. The fact that habitat preference can already be observed at the level of taxonomic orders, often irrespective of functional guild, suggests shared physiological constraints and environmental optimum for relatively closely related taxa.