Arbuscular mycorrhizal interactions of mycoheterotrophic Thismia are more specialized than in autotrophic plants

Abstract

In general, plants and arbuscular mycorrhizal (AM) fungi exchange photosynthetically fixed carbon for soil nutrients, but occasionally nonphotosynthetic plants obtain carbon from AM fungi. The interactions of these mycoheterotrophic plants with AM fungi are suggested to be more specialized than those of green plants, although direct comparisons are lacking. We investigated the mycorrhizal interactions of both green and mycoheterotrophic plants. We used next-generation DNA sequencing to compare the AM communities from roots of five closely related mycoheterotrophic species of Thismia (Thismiaceae), roots of surrounding green plants, and soil, sampled over the entire temperate distribution of Thismia in Australia and New Zealand. We observed that the fungal communities of mycoheterotrophic and green plants are phylogenetically more similar within than between these groups of plants, suggesting a specific association pattern according to plant trophic mode. Moreover, mycoheterotrophic plants follow a more restricted association with their fungal partners in terms of phylogenetic diversity when compared with green plants, targeting more clustered lineages of fungi, independent of geographic origin. Our findings demonstrate that these mycoheterotrophic plants target more narrow lineages of fungi than green plants, despite the larger fungal pool available in the soil, and thus they are more specialized towards mycorrhizal fungi than autotrophic plants.

Keywords: Thismia; arbuscular mycorrhizal (AM) fungi; habitat filtering; mycoheterotrophy; phylogenetic niche conservatism; specificity.

Figure 1

Highest likelihood tree (LnL = −10519.28) showing the phylogenetic relationships among the Glomeromycota operational taxonomic units (OTUs) found in all the samples, including several reference sequences. The colored circles indicate the presence of the fungal OTUs according to plant group (mycoheterotrophic, yellow; autotrophic, green) and the pool of fungal OTUs present in the soil (blue). Mycoheterotrophic plants of the genus Thismia are associated with fungi in the Glomerales family (one subclade: Rhizophagus/Sclerocystis sp.); and green plants are also associated with fungi in the Glomerales family (two subclades: Rhizophagus/Sclerocystis sp. and Glomus sp.). The soil also harbors fungi from the Glomerales family, and also from the Diversisporales and Archaeosporales families within the Glomeromycota phylum.

Figure 2

Nonmetric multidimensional scaling plot (metaMDS) showing an ordination of the fungal community dissimilarities (Comdist) among all the samples. The fungal community dissimilarities are calculated based on the average phylogenetic distance between each fungal operational taxonomic unit (OUT) in one sample and the total OTUs in the other sample. Each symbol represents the Comdist of the fungal communities including all the OTUs found in each species per site. Permutational MANOVA (Adonis) showed significant fungal community dissimilarity between mycoheterotrophic (MH) Thismia plants, green plants, and soil (F = 25.4; R ² = 0.486; P = 0.001).

Figure 3

Fungal community structure based on the net relatedness index (NRI) for each species per site. The graph represents the fungal communities’ phylogenetic dispersion patterns as explained by the ‘type’ of material (mycoheterotrophic (MH) plants, green plants and soil). Negative NRI values indicate that the fungal communities are overdispersed in the phylogenetic tree, while positive NRI values indicate phylogenetic clustering. The NRI was significantly different in MH plants compared with green plants and soil. MH plants harbor more phylogenetically clustered AM fungal communities in their roots than green plants and the soil. Green plants also have significantly more clustered fungal communities than the soil. The mixed‐effects model estimates with 95% confidence intervals are shown. See Supporting Information Table S2 for statistical details.