A phosphorus threshold for mycoheterotrophic plants in tropical forests

Abstract

The majority of terrestrial plants associate with arbuscular mycorrhizal (AM) fungi, which typically facilitate the uptake of limiting mineral nutrients by plants in exchange for plant carbon. However, hundreds of non-photosynthetic plant species-mycoheterotrophs-depend entirely on AM fungi for carbon as well as mineral nutrition. Mycoheterotrophs can provide insight into the operation and regulation of AM fungal relationships, but little is known about the factors, fungal or otherwise, that affect mycoheterotroph abundance and distribution. In a lowland tropical forest in Panama, we conducted the first systematic investigation into the influence of abiotic factors on the abundance and distribution of mycoheterotrophs, to ask whether the availability of nitrogen and phosphorus altered the occurrence of mycoheterotrophs and their AM fungal partners. Across a natural fertility gradient spanning the isthmus of Panama, and also in a long-term nutrient-addition experiment, mycoheterotrophs were entirely absent when soil exchangeable phosphate concentrations exceeded 2 mg P kg^-1 Experimental phosphorus addition reduced the abundance of AM fungi, and also reduced the abundance of the specific AM fungal taxa required by the mycoheterotrophs, suggesting that the phosphorus sensitivity of mycoheterotrophs is underpinned by the phosphorus sensitivity of their AM fungal hosts. The soil phosphorus concentration of 2 mg P kg^-1 also corresponds to a marked shift in tree community composition and soil phosphatase activity across the fertility gradient, suggesting that our findings have broad ecological significance.

Keywords: arbuscular mycorrhizal fungi; epiparisitism; mycoheterotroph; phosphorus; tropical forest.

Figure 1.

The mycoheterotrophs (a) Voyria tenella and (b) V. corymbosa in a lowland tropical forest in Panama. The root system of the mycoheterotroph V. tenella is intensely colonized by arbuscular mycorrhizal (AM) fungi (d). In (c)(i), fungal material is visible as the light-coloured ring surrounding the central vasculature. In (c)(ii), fungal material (hyphae and coils) is rendered in red and plant material is not shown. In (d), plant material is displayed in grey, and fungal material in red. The same image stack is displayed in (d)(i–iv) with the plant material made increasingly transparent. Confocal micrographs (c)(ii) and (d)(i–iv) were obtained by differential staining of plant and fungal tissues, shown as three-dimensional projections (AMIRA). (a) and (b) Courtesy of Christian Ziegler. (a,b) Scale bar, 20 mm; (c) scale bar, 1 mm; (d) scale bar, 100 µm.

Figure 2.

(a) Numbers of the mycoheterotroph Voyria tenella sharply decline with increasing soil exchangeable phosphorus (P) across a naturally occurring gradient in lowland tropical forests in Panama. The solid line depicts the fitted response of a GLM with negative binomial errors (n = 37). Red dashed lines indicate the 95% CI. The blue shaded region represents the concentrations of soil exchangeable phosphorus found in +P plots in the nearby factorial nutrient-addition experiment (electronic supplementary material, figure S1). (b) The abundant mycoheterotroph V. tenella (i) and less common congener V. corymbosa (ii) are eliminated by phosphorus (P) addition in a long-term factorial nutrient-addition experiment in a lowland tropical forest in Panama. The figure contrasts 16 no-P plots (control, N, K, NK treatments) with 16 +P plots (P, NP, KP, NPK treatments). Values are fitted responses of a GLM with negative binomial errors and show 95% confidence intervals. The effects of individual fertilization treatments on numbers of V. tenella and V. corymbosa are presented in electronic supplementary material, figure S3.

Figure 3.

The relative abundance—in the soil—of the AM fungal taxa most strongly associated with (a) Voyria tenella and (b) Voyria corymbosa are reduced but not eliminated in +P treatments in a long-term factorial nutrient-addition experiment in a lowland tropical forest in Panama. (i) Upper bars represent the relative abundance of AM fungal taxa in the roots of V. tenella and V. corymbosa (averaged across control, N, K, NK treatments; n = 16). (ii) Lower bars illustrate the effect of experimental phosphorus addition on the proportional abundance of AM fungal taxa in the soil. The figure contrasts 16 no-P plots (control, N, K, NK treatments) with eight +P plots (P, NP treatments). Values are fitted responses of GLM with negative binomial errors and show 95% confidence intervals. Significant effects of phosphorus addition are asterisked. Data are based on read counts from 454-sequencing. See electronic supplementary material, figure S4 for the effects of individual fertilization treatments on the relative abundance of AM fungal taxa in the soil.

Figure 4.

The AM fungal partners of mycoheterotrophs Voyria tenella and Voyria corymbosa under unfertilized conditions were present both in the soil and in the roots of autotrophic plants when fertilized with phosphorus, although the relative abundance of AM fungal taxa shifted in response to phosphorus addition. Interactions between mycoheterotrophs, AM fungi and autotrophs in unfertilized control plots are represented in (a), and in phosphorus-addition plots in (b). The latter displays potential linkages (indicated by shaded blue region) between mycoheterotrophs and AM fungi based on their partners in unfertilized control plots; there were no mycoheterotrophs actually found in phosphorus-addition plots. Values are based on the mean of four unfertilized control plots and four phosphorus-addition plots. The widths of bars representing AM fungal OTUs (ii) are scaled to the relative abundance of OTUs in soil communities in control (a)(ii) and phosphorus-addition (b)(ii) treatments, respectively. The thickness of the linkages is scaled to reflect the proportion of the AM fungal community constituting the linkages. AM fungal OTUs found in the roots of either species of Voyria and their linkages are depicted in colour (see legend), and those AM fungal OTUs not interacting with mycoheterotrophs in grey. Full list of OTU codes are given in electronic supplementary material, table S3. Only the 50 most abundant AM fungal OTUs (in the soil) are plotted, and AM fungal OTUs making up less than 1% of the total number of sequences in a sample type are omitted for clarity. Data are based on read counts from 454-sequencing. V. ten, Voyria tenella; V. cor, Voyria corymbosa; ALSB, Alseis blackiana; DESP, Desmopsis panamensis; HEIC, Heisteria concinna; SIMA, Simarouba amara; SORA, Sorocea affinis; TET2, Tetragastris panamensis; VIR1, Virola sebifera.