Reproducing plant microbiome research reveals site and time as key drivers of apple tree phyllosphere bacterial communities

Abstract

Manipulating plant microbiomes is foreseen as a key biocontrol avenue to tackle the accelerating challenges of global change in agriculture. Several recent studies have identified the spatiotemporal dynamics of phyllosphere microbial communities, stressing the need to understand plant microbiome drivers to design efficient biocontrol interventions. Yet, these works are often performed on small sample counts, rarely provide sufficient information on the relative impact of time or local environment, and are seldom repeated to assess reproducibility. To address these limits, we performed a longitudinal sampling across multiple orchards of contrasting agricultural practices to study the ecological drivers of phyllosphere bacterial communities of apple tree (Malus domestica, Borkh.). We sampled up to eight apple cultivars at six orchards (three conventional, three organic) in the Eastern Townships (Canada) in 2022 and 2023. In contrast with common cross-sectional microbiome studies, our work builds on a two-year sampling design, thus allowing for the evaluation of the reproducibility of previous plant microbiome research. Our results support previous findings indicating that site and time are major drivers of apple tree bacterial community structure, yet their relative influence vary across the two sampling years. In addition, our data showed that leaf and flower bacterial alpha diversity is lower at organic sites compared to conventional sites. Overall, this study provides a comprehensive longitudinal multi-site study design highlighting the value of assessing reproducibility in plant microbiome studies and paving the way for future research in this field.

Keywords: Agricultural practices; Apple tree; Bacterial communities.; Longitudinal; Phyllosphere; Plant microbiome; Reproducibility; Spatiotemporal dynamics.

Fig. 1

Study design and principal coordinate analysis (PCoA) ordinations of variation in bacterial community composition in 2023. (A) Map indicating the six sites. (B) Table indicating the cultivars sampled at each site. (C) PCoA on May samples (flowers + leaves) across sites (n = 157). Colors indicate plant compartment (pink = flower, green = leaf). (D) PCoA on leaf samples (n = 292) showing the variation through time. Point color indicate sampling time (pink = May, green = July, yellow = August). (E) PCoA on leaf samples (n = 292) showing the variation across time and sites. Conventional sites are colored in shades of orange while organic sites in shades of blue. Ordinations are based on Bray-Curtis dissimilarities among samples. Ellipses indicate 95% confidence intervals. Shapes indicate agricultural practices (square = conventional, circle = organic).

Fig. 2

Variation in bacterial community composition across sites of different agricultural practices. (A) PCoA showing the cultivars Honeycrisp and Spartan (n = 108). (B) PCoA showing variation in community composition across time. Ordinations are based on Bray-Curtis dissimilarities among samples. Ellipses indicate 95% confidence intervals. Colors indicate agricultural practices (yellow = conventional; blue = organic). Combinations of point shapes and colors indicate sites (square: yellow = A, blue = C; circle: yellow = B1, blue = B2; diamond: yellow = D1, blue = D2). (C) Top 20 most relatively abundant genera at all sites depending on agricultural practices and sampling time.

Fig. 3

Alpha diversity indices across time and sites for cultivars Honeycrisp and Spartan. (A) Plots are separated according to the six sites and indicate change in alpha diversity across time. (B) Plots are separated according to the agricultural practice and indicate change in alpha diversity across time. (C) Plots indicates the change in alpha diversity between practices at different time points. Asterisks indicate the results of a post-hoc Dunn’s test (^*p ≤ 0.05; ^**p ≤ 0.01; ^***p ≤ 0.001, ^**** p ≤ 0.0001).

Fig. 4

Effect of agricultural practice on flower bacterial community composition and diversity. (A) PCoA illustrating the variation within flower samples. Combinations of point shape and color (yellow = conventional, blue = organic) indicate sites (yellow circle = A, yellow square = B1, blue square = C). Ordination based on Bray-Curtis dissimilarities among samples. Ellipses indicate 95% confidence intervals for each practice. (B) Alpha diversity index of conventional and organic samples. Asterisks indicate the results of a post-hoc Dunn’s test (^*p ≤ 0.05; ^**p ≤ 0.01; ^***p ≤ 0.001, ^****p ≤ 0.0001). (C) Relative abundance of the top 20 genera found in flower samples for both conventional and organic sites.

Fig. 5

Bacterial community composition and diversity across cultivars of different susceptibility levels to fire blight at sites A, B1, and C. PCoAs performed on (A) leaf (n = 162) and (B) flower (n = 53) samples. Ordinations are based on Bray-Curtis dissimilarities among samples. Ellipses indicate 95% confidence intervals. (C) Alpha diversity through time for leaf and flower samples. Asterisks indicate the results of a post-hoc Dunn’s test (^*p ≤ 0.05; ^**p ≤ 0.01; ^***p ≤ 0.001, ^****p ≤ 0.0001).

Fig. 6

Bacterial community variation across two sampling years. PCoAs showing the effect of (A) sampling year and (B) sampling time. (C) Relative abundance of top 20 genera across the two sampling years for each time point. (D) Shannon alpha diversity index across time and sites for both years.