Root exudate metabolites drive plant-soil feedbacks on growth and defense by shaping the rhizosphere microbiota

Abstract

By changing soil properties, plants can modify their growth environment. Although the soil microbiota is known to play a key role in the resulting plant-soil feedbacks, the proximal mechanisms underlying this phenomenon remain unknown. We found that benzoxazinoids, a class of defensive secondary metabolites that are released by roots of cereals such as wheat and maize, alter root-associated fungal and bacterial communities, decrease plant growth, increase jasmonate signaling and plant defenses, and suppress herbivore performance in the next plant generation. Complementation experiments demonstrate that the benzoxazinoid breakdown product 6-methoxy-benzoxazolin-2-one (MBOA), which accumulates in the soil during the conditioning phase, is both sufficient and necessary to trigger the observed phenotypic changes. Sterilization, fungal and bacterial profiling and complementation experiments reveal that MBOA acts indirectly by altering root-associated microbiota. Our results reveal a mechanism by which plants determine the composition of rhizosphere microbiota, plant performance and plant-herbivore interactions of the next generation.

Fig. 1

Benzoxazinoid release by maize roots is associated with changes in root-associated microbiota. a, b Benzoxazinoid (BX) profiles at the root surface (a, n = 14–15) and the soil cores (b, n = 5) of wild type (WT) B73 and bx1 mutant plants. For absolute quantities, refer to Supplementary Fig. 1. Stars indicate significant differences (two-sided Student’s t tests, *P < 0.05; ***P < 0.001). c, d Partial Canonical analysis of Principal Coordinates (CAP) of rhizosphere and root bacterial (c) and fungal (d) communities (n = 7–10). The CAP ordinations using Bray–Curtis distance were constrained for nesting the factor plant genotype within the factor sample type. Model, explained fraction of total variance and model significance are indicated above the plots. Axes report the proportions of total variation explained by the constrained axes. Non-conditioned bulk soil samples were included as negative controls. For unconstrained ordination and individual operational taxonomic units (OTUs), refer to Supplementary Fig. 1

Fig. 2

Benzoxazinoid soil conditioning increases plant defense and decreases plant growth. a Shoot biomass of 10-week old wild type (WT) B73 plants growing in soils previously conditioned by WT (BX+) or benzoxazinoid (BX)-deficient bx1 mutant plants (BX−) (n = 10). b, c Weight gain (b) and leaf damage (c) of Spodoptera frugiperda caterpillars (n = 10). d Changes in leaf phytohormones, defense marker genes, primary and secondary metabolites (n = 9–10). Stars indicate significant differences between conditioning treatments (two-sided Student’s t tests, *P < 0.05; **P < 0.01; ***P < 0.001). Arrows indicate metabolic markers which were used for phenotyping in subsequent experiments. DW, dry weight. L.O.D., below limit of detection

Fig. 3

Influence of soil conditioning of different benzoxazinoid mutants on plant and herbivore performance. a, b Average caterpillar weight gain (a) and shoot biomass (b) of the next generation of wild type (WT) B73 or W22 maize plants (“Response genotype”) growing in soils previously conditioned by corresponding WT (BX+) or benzoxazinoid (BX)-deficient bx-mutant plants (BX−) (“Conditioning genotype”, +SE, n = 10–15). Stars indicate significant differences between soil types (***P < 0.001, one-way ANOVA). Different letters indicate significant differences between individual soil types (P < 0.05, one-way ANOVA followed by multiple comparisons through FDR-corrected LSMeans). c Fold changes of leaf markers in the next generation of B73 or W22 plants growing in BX+ soil which were conditioned by corresponding WT (B73 or W22) plants relative to those growing in BX− soil which were conditioned by corresponding bx-mutant plants (n = 10). Stars indicate significant differences between soil types (ANOVA, FDR-corrected LSMeans, *P < 0.05; **P < 0.01; *** P < 0.001). For full datasets, refer to Supplementary Fig. 3. DW, dry weight. Cond., conditioning

Fig. 4

Application of MBOA restores benzoxazinoid-dependent plant-soil feedback effects. a, b Shoot biomass (a) and caterpillar growth rate (b) of wild type (WT) B73 plants growing in soils previously conditioned by WT (BX+) or benzoxazinoid (BX)-deficient bx1 mutant plants (BX−), as well as MBOA-complemented BX− soils (n = 15–19). Letters indicate significant differences between conditioning treatments (ANOVA, FDR-corrected LSMeans, P < 0.05). c Fold changes of leaf markers in WT B73 plants growing in BX+ soil or MBOA-complemented BX− soil relative to water-treated BX− soils (n = 15–19). For full datasets, refer to Supplementary Fig. 5. No significant differences were found between BX+ soil and BX− soil supplemented with MBOA (Supplementary Fig. 5). DW, dry weight. Cond., conditioning. Stars indicate significant differences between treatments and BX− soil (ANOVA, FDR-corrected LSMeans, *P < 0.05; **P < 0.01; ***P < 0.001)

Fig. 5

Benzoxazinoid-dependent changes in soil microbiota are necessary and sufficient to trigger changes in plant and herbivore performance. a, b Plant (a) and herbivore performance (b) of wild type (WT) B73 plants growing in soils previously conditioned by WT (BX+) or benzoxazinoid (BX)-deficient bx1 mutant plants (BX−). Soils were either left untreated, sterilized by X-ray or sterilized and complemented with microbial extracts from the respective non-sterilized soils. Stars indicate significant differences within soil treatments (ANOVA, FDR-corrected LSMeans, *P < 0.05; **P < 0.01; n = 6–10). c Redundancy analysis (RDA) of leaf markers (n = 6–10). Model, explained fraction of total variance and model significance are indicated above the plots. Axes report the proportions of constrained variation explained by the constrained axes. For corresponding unconstrained ordination and individual metabolite measures, refer to Supplementary Fig. 7 and Supplementary Tables 5 and 6. DW, dry weight. Cond., conditioning. n.s., no significant. Treat., treatment

Fig. 6

MBOA-induced effects on plant and herbivore performance depend on the soil microbiota. a, b Shoot biomass (a) and caterpillar growth rate (b) of wild type (WT) B73 plants growing in soils previously conditioned by benzoxazinoid (BX)-deficient bx1 mutant plants (BX−), which were complemented with water or MBOA with and without subsequent sterilization (+SE, n = 11). c Fold changes of leaf markers in WT B73 plants growing in BX− soils which were complemented with MBOA relative to water-treated BX− soils with and without subsequent sterilization (+SE, n = 11). For full datasets, refer to Supplementary Fig. 8. Cond. or C, conditioning. Treat or T, treatment; n.s., no significant; DW, dry weight. Stars indicate significant differences between complementation treatments within soil treatments (ANOVA, FDR-corrected LSMeans, *P < 0.05; **P < 0.01; ***P < 0.001)

Fig. 7

Benzoxazinoid-conditioning shapes soil microbiota in the roots and rhizosphere of the second plant generation. a, b Partial Canonical analysis of Principal Coordinates (CAP) of rhizosphere and root bacterial (a) and fungal (b) communities of wild type (WT) plants grown in soils previously conditioned by WT (BX+) or benzoxazinoid (BX)-deficient bx1 mutant plants (BX−) (n = 10). CAP ordinations using Bray–Curtis distance were constrained for the factors sample type and soil conditioning. Model, explained fraction of totasl variance and model significance are indicated above the plots. Axes report the proportions of total variation explained by the constrained axes. c, d Biplot with the bacterial Operational Taxonomic Unit (bOTU, c) and fungal OTU (fOTU, d) scores of the CAP showing the contribution of individual OTUs to the separation of the different treatments. For unconstrained ordination and individual OTUs, refer to Supplementary Fig. 9

Fig. 8

Proposed model for benzoxazinoid-dependent, microbiota-mediated plant performance