Lineage overwhelms environmental conditions in determining rhizosphere bacterial community structure in a cosmopolitan invasive plant

Abstract

Plant-microbe interactions play crucial roles in species invasions but are rarely investigated at the intraspecific level. Here, we study these interactions in three lineages of a globally distributed plant, Phragmites australis. We use field surveys and a common garden experiment to analyze bacterial communities in the rhizosphere of P. australis stands from native, introduced, and Gulf lineages to determine lineage-specific controls on rhizosphere bacteria. We show that within-lineage bacterial communities are similar, but are distinct among lineages, which is consistent with our results in a complementary common garden experiment. Introduced P. australis rhizosphere bacterial communities have lower abundances of pathways involved in antimicrobial biosynthesis and degradation, suggesting a lower exposure to enemy attack than native and Gulf lineages. However, lineage and not rhizosphere bacterial communities dictate individual plant growth in the common garden experiment. We conclude that lineage is crucial for determination of both rhizosphere bacterial communities and plant fitness.Environmental factors often outweigh host heritable factors in structuring host-associated microbiomes. Here, Bowen et al. show that host lineage is crucial for determination of rhizosphere bacterial communities in Phragmites australis, a globally distributed invasive plant.

Fig. 1

Differences in microbial community structure among lineages and as a function of distance between sampling locations. a Principal coordinates analysis of Bray–Curtis dissimilarity for the microbial communities associated with Native (green), Gulf (blue) and Introduced (orange) lineages of Phragmites australis, collected from across the United States (Supplementary Fig. 1). b Bray–Curtis dissimilarity plotted as a function of the distance (km) between sampling locations. Regression statistics are found in Supplementary Table 2. n = 5 replicates per P. australis population (eight Introduced populations, five Gulf populations, eight Native populations)

Fig. 2

Core microbiome analysis for each lineage. Abundance of taxa belonging to the core microbiome of different Phragmites australis lineages. Colored bars indicate the taxon was present in 100% of the samples for that lineage, whereas gray bars indicate that it was only present in some samples

Fig. 3

Microbial community structure of different lineages of Phragmites australis grown in a common garden experiment. a Principal coordinates analysis of Bray–Curtis dissimilarity for the entire (open symbols) and active (closed symbols) microbial community, along with the pre-incubation community (purple). b Principal coordinates analysis of the Bray–Curtis dissimilarity values for just the active microbial community, indicating the importance of population (PERMANOVA, F _59,14 = 20.56, p < 0.001), and lineage (F _71,2 = 19.79, p < 0.001)

Fig. 4

Abundance of KEGG pathways potentially associated with plant–microbe interactions in the rhizosphere soils of Introduced, Gulf, and Native lineages. In each of the three pathways, data are the mean ± standard error of the mean and different letters denote significantly different means within each pathway. The abundance of each pathway was significantly lower in the Introduced lineage rhizosphere microbial communites than in those communities associated with the Gulf and Native lineages. Toxin/metal detoxification: ANOVA, F _2,75 = 47.56, p < 0.001; antimicrobial biosynthesis: ANOVA, F _2,75 = 35.23, p < 0.001; antimicrobial degradation: ANOVA, F _2,75 = 29.33, p < 0.001)

Fig. 5

Structural Equation Model fitted with range-standardized coefficients. Solid lines indicate that the driver influences the likelihood of the model via a χ ² Likelihood ratio test. Dashed paths indicate no detectable influence of the driver (p > 0.05). Standardized coefficients are presented for each path. Letters a and b denote groupings via post-hoc tests. χ ² test statistics can be found in Supplementary Tables 5 and 6