Interactions with fungi vary among Tripsacum dactyloides genotypes from across a precipitation gradient

Abstract

Plant-associated microbes, specifically fungal endophytes, augment the ability of many grasses to adapt to extreme environmental conditions. Tripsacum dactyloides (Eastern gamagrass) is a perennial, drought-tolerant grass native to the tallgrass prairies of the central USA. The extent to which the microbiome of T. dactyloides contributes to its drought tolerance is unknown. Ninety-seven genotypes of T. dactyloides were collected from native populations across an east-west precipitation gradient in Kansas, Oklahoma and Texas, and then grown together in a common garden for over 20 years. Root and leaf samples were visually examined for fungal density. Because fungal endophytes confer drought-tolerant capabilities to their host plants, we expected to find higher densities of fungal endophytes in plants from western, drier regions, compared to plants from eastern, wetter regions. Results confirmed a negative correlation between endophyte densities in roots and precipitation at the genotype's original location (r = -0.21 P = 0.04). Our analyses reveal that the host genotype's origin along the precipitation gradient predicts the absolute abundance of symbionts in the root, but not the relative abundances of particular organisms or the overall community composition. Overall, these results demonstrate that genetic variation for plant-microbe interactions can reflect historical environment, and reinforce the importance of considering plant genotype in conservation and restoration work in tallgrass prairie ecosystems.

Keywords: Endophytes; Tripsacum dactyloides; fungi; local adaptation; perennial; plant microbiome; precipitation; stress gradient hypothesis; tallgrass prairie.

Figure 1.

Map depicting the original collection sites of the T. dactyloides genotypes. The collection sites span the east-west precipitation gradient, where MAP (mean annual precipitation) is negatively correlated with longitude (Fig. 2). The site of the common garden in Woodward, Oklahoma, is denoted by a magenta star.

Figure 2.

Across the area where our genotypes originated, precipitation is negatively correlated with longitude (P = 9.86e−11, r = 0.6). Each point is the original provenance of one genotype used in this study.

Figure 3.

A magnified view of an endophyte-positive T. dactyloides leaf. The fungal hyphae have been dyed and are shown in dark blue.

Figure 4.

(A) Root endophyte density as a function of MAP (millimetres) at each genotype’s site of origin (P = 0.043). Leaf endophyte density as a function of MAP (millimetres) at each genotype’s site of origin (P = 0.62). We define density as the percentage of fields of vision that were endophyte positive (see ‘Methods’ section). Each dot represents an individual sample, and the lines are the regression lines of best fit. The grey shading represents the 95 % confidence interval.

Figure 5.

Venn diagram portraying the number of taxa that were unique to and shared between soil rhizosphere and root endosphere.

Figure 6.

Relative abundances of bacterial phyla (A) and classes (B) found in 48 samples of T. dactyloides. We have separated the root endosphere sequences (first panels) and the soil rhizosphere samples (second panels).

Figure 7.

Relative abundances of fungal orders (panel A) and genera (panel B) in 23 samples of T. dactyloides. We have separated the root endosphere sequences and the rhizosphere samples.

Figure 8.

CAP ordination for CLR-transformed bacterial (A) and fungal (B) samples, constrained by sample type (i.e. root endosphere vs. rhizosphere). Neither bacterial nor fungal community composition differed between root and rhizosphere (P > 0.05).

Figure 9.

CAP ordination for CLR-transformed bacterial (A) and fungal (B) samples, constrained by precipitation at the genotype’s original location. Root and rhizosphere bacterial or fungal communities did not co-vary with genotypes’ provenances (P > 0.05) across the precipitation gradient.