Impact of root-associated strains of three Paraburkholderia species on primary and secondary metabolism of Brassica oleracea

Paraburkholderia species promote Broccoli growth in a cultivar-dependent manner

Root tip inoculation of the two Broccoli cultivars with strains of three different Paraburkholderia species led to changes in leaf color (deep green leaves), shoot biomass, root biomass and root architecture (Fig. 1a). Percent change in biomass was used as a measure to assess the growth-promoting effects of the Paraburkholderia species in the two Broccoli cultivars. Two-way analysis of variance (ANOVA) was conducted to assess the influence of the two independent variables (strains of Paraburkholderia species and Broccoli cultivars) on both shoot and root biomass. The Paraburkholderia species included three levels (Pbg, Pbh, Pbt) and the Broccoli cultivars consisted of two levels (Coronado, Malibu). For shoots, all interactions, except Pbt-Malibu, resulted in significant increases in biomass relative to the non-treated control plants, while for roots all three Paraburkholderia species significantly increased the biomass in both Broccoli cultivars (Fig. 1b). In general, the relative impact of Paraburkholderia species was up to 3 times higher for root biomass than for shoot biomass (Fig. 1b). Two-way ANOVA showed highly significant interactions between the strains of Paraburkholderia species and Broccoli cultivars regarding the percent changes in shoot and root biomass (Supplementary Table S1). Overall, for cultivar Coronado the percent change in shoot biomass was about 40% compared to the control, and not significantly different between the different strains of Paraburkholderia species, whereas in cultivar Malibu the percent change in shoot biomass was significantly higher for Pbg (~ 70%) and Pbh (~ 90%) as compared to Pbt. Furthermore, inoculation with Pbh led to a significantly higher increase in shoot biomass in cultivar Malibu than in Coronado. Regarding the percent change in root biomass, only inoculation of Pbt showed significant differences between the two Broccoli cultivars. As indicated above, the shoot biomass of cultivar Malibu inoculated with Pbt was not significantly different from the control plants (Fig. 1b). Over a period of 11 days, both Pbg and Pbh-treated Broccoli cultivars showed significantly higher shoot and root biomass from 7 days post inoculation (dpi) onwards, while Pbt-treated plants showed higher shoot biomass in Coronado from 9 dpi onwards (Fig. 1c).

Relation between root colonization and plant growth promotion

The extent of root colonization of the strains of Paraburkholderia species was assessed for the two Broccoli cultivars at the early and late growth stages. The data was log-transformed, as they did not meet the ANOVA assumption for homogeneity of variance and normality. Three-way analysis of variance was conducted on the interaction effects of Paraburkholderia species strains, Broccoli cultivars and time after inoculation on root colonization. The Paraburkholderia species strains included three levels (Pbg, Pbh and Pbt), the Broccoli cultivars included two levels (Coronado, Malibu) and time after inoculation consisted of two levels (6 dpi, 11 dpi). There was a highly significant three-way interaction effect on root colonization (Supplementary Table S2). In general, Pbg showed significantly higher root colonization in both cultivars at both time points when compared to Pbh and Pbt (Fig. 2 and Supplementary Table S3). In addition, Pbg showed significantly higher root colonization in cultivar Coronado at 6 dpi (2.1 ± 0.1 × 10⁸ Cfu/mg roots) and at 11 dpi (8.1 ± 0.3 × 10⁷ Cfu/mg roots). In cultivar Malibu root colonization by Pbg was not significantly different between the two time points (1.0 ± 0.1 × 10⁸ Cfu/mg roots (6 dpi) and 1.0 ± 0.1 × 10⁸ Cfu/mg roots (11 dpi)). Pbt showed significantly lower root colonization in cultivar Malibu at both time points. Furthermore, Pbt on Malibu showed a significant decline in root colonization at the later time point (1.7 ± 0.2 × 10⁶ Cfu/mg roots (6 dpi) and 7.0 ± 0.4 × 10⁵ Cfu/mg roots (11 dpi)).

Paraburkholderia species altered primary and secondary metabolism of Broccoli shoot

Considering the extent of strain and cultivar-dependent variations in root colonization and plant growth promotion, we investigated the systemic effect of the bacterial strains on the shoot metabolome of the two Broccoli cultivars at 6 and 11 dpi. GC–MS and LC–MS-based non-targeted metabolomics analysis of shoot extracts were performed to profile the polar primary metabolites and semi-polar secondary metabolites, respectively. The data was subjected to ANOVA with correction for multiple testing (Benjamini-Hochberg) and metabolites that were significantly different (P < 0.05 and fold change > 2) between at least two treatments were used for multivariate analysis. Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA) were used to reduce the dimensionality of the data and explore specific patterns of change in metabolome in the different plant–rhizobacteria interactions.

Effects of Paraburkholderia on shoot primary metabolism

GC–MS-based non-targeted metabolomics demonstrated that out of the 138 polar metabolites detected, 68 (50%) were significantly different between at least two treatments. At 6 dpi, PCA indicated that the first three principal components (PCs) explained 62.8% of the total variance (Fig. 3a1). The first PC (PC1), explained 34.8% of the total variance and corresponded to the effect of the three Paraburkholderia treatments on the metabolome of both cultivars (Fig. 3b1, Clusters 1, 5, 8 and 9). Pbg had the greatest impact on shoot primary metabolism of both Broccoli cultivars, while inoculation with Pbh and Pbt resulted in changes in the shoot primary metabolome in a cultivar-dependent manner. Pbh had greater impact on shoot primary metabolome of Malibu, while Pbt had greater impact on shoot primary metabolome of Coronado (Fig. 3a1). The major changes in primary metabolism induced by Paraburkholderia included accumulation of sugars (Cluster 9) and depletion of amino acids (Cluster 5, phenylalanine, lysine and methionine) and phosphoenolpyruvate (PEP), a key intermediate in glycolysis and gluconeogenesis. Some of the representative metabolites in cluster 8 that showed accumulation in all interactions, except in the ineffective Pbt-Malibu interaction, include aspartic acid, mannonic acid and putrescine. The second principal component (PC2) explained 18.2% of the total variance and resulted from metabolites that showed variation between the two cultivars. Furthermore, treatment of the two cultivars with Pbg and Pbh widened the inherent variation in the level of some of the metabolites between the two Broccoli cultivars. (Fig. 3b1, Clusters 2 and 3). Amino acids such as glutamine, oxoproline (pyroglutamic acid), GABA (γ-aminobutyric acid) and isoleucine were intrinsically higher in Coronado than in Malibu.

At the later seedling growth stage (11 dpi), the initial inoculation of the roots of the two Broccoli cultivars with the strains of the Paraburkholderia species continued to have substantial impact on shoot primary metabolism. PCA showed that the first three principal components explained 72.1% of the total variance (Fig. 3a2). Here, the impact of all the three strains of the Paraburkholderia species on the Broccoli shoot metabolome was cultivar dependent and was greater in Malibu (Fig. 3a2). The first principal component (PC1) explained 44.2% of the total variance and resulted from metabolites that were accumulated (Fig. 3b2, Cluster 5) or reduced (Cluster 2) in the Paraburkholderia treatments. The Broccoli metabolites that decreased upon inoculation with the strains of the Paraburkholderia species encompassed amino acids such as lysine, phenylalanine, methionine, the non-proteinogenic amino acids ornithine and GABA, as well as PEP. In all plant–microbe combinations, except the ineffective partnership between Pbt-Malibu, PEP showed 11–14 fold decreases (Supplementary excel, Table S6). Sugars and other metabolites, including ascorbic acid and aspartic acid, represented the metabolites enhanced by the Paraburkholderia treatments when compared to the control plants (Cluster 5). Six days after treatment with Paraburkholderia, sugars showed greater abundance in cultivar Malibu than in cultivar Coronado (Fig. 3b2). However, at 11 dpi, sugars in Paraburkholderia-treated plants showed substantial depletion in cultivar Coronado as compared to 6 dpi (Fig. 3b1), whereas in cultivar Malibu, the temporal variation in the level of these sugars was less pronounced (Fig. 3b2, Cluster 5, Supplementary Figure S3 and S4). PC2, representing 19.4% of the total variance, was associated with metabolites in cluster 1 including glycine, that were depleted in all treatment combinations except in the controls and in the ineffective partnership between Pbt and Malibu (Fig. 3b2, Cluster 1). Oxoproline and some other metabolites in cluster 3 were intrinsically abundant in the shoots of cultivar Coronado.

Paraburkholderia impact on Broccoli primary metabolism is highly associated with soluble sugars

As sugars are the primary drivers of plant growth, we looked into their temporal dynamics, particularly related to sugar generation and utilization in the shoots of the two Broccoli cultivars treated with the strains of the Paraburkholderia species. The fold change in sugar level between Paraburkholderia treated and control plants at 6 dpi was used as a measure of sugar generation, while the fold change in sugar level of treated plants from 6 to 11 dpi was used as a measure of sugar utilization. In control plants, the sugar levels showed no significant difference between the two Broccoli cultivars at 6 dpi (supplementary Fig. S3). In contrast, treatment with the strains of the Paraburkholderia species showed substantial impact on the sugar generation in shoots of both Broccoli cultivars, resulting in significant increases in the level of fructose and its derivatives, glucose, sorbose, galactose and galactopyranose at 6 dpi. Moreover, the magnitude of sugar generation showed remarkable differences between the strains of the Paraburkholderia species-Broccoli cultivar combinations (Fig. 3c1). Pbg treatment resulted in the highest sugar generation when compared to Pbh and Pbt, and this ability was significantly higher in cultivar Malibu than in Coronado. The ineffective partnership between Pbt and Malibu had the least impact on sugar generation. Similarly, the utilization of sugar also showed noticeable differences among the strains of the Paraburkholderia species-Broccoli cultivar combinations. In Coronado, Pbg inoculation led to greater sugar utilization when compared to cultivar Malibu. The ineffective partnership between Pbt and Malibu showed reduced sugar utilization when compared to the effective partnership of Pbt with Coronado (Fig. 3c2).

Effects of Paraburkholderia on shoot secondary metabolism

From the 1,868 metabolites detected by LCMS, 1,386 (74%) were significantly different between at least two treatments. PCA of the metabolites at 6 dpi demonstrated distinct clustering of the samples based on the strains of the Paraburkholderia species-Broccoli cultivar combination (Fig. 4a1). Here PC1 explained 33.2% of the total variation and was associated with sample differences due to metabolites that were intrinsically more abundant in one of the two Broccoli cultivars (Clusters 3, 5 and 12, Fig. 4b1). Metabolites that were intrinsically more abundant in Coronado included aliphatic glucosinolates such as 2-methylbutyl glucosinolate and glucoiberverin as well as the aromatic glucosinolates glucotropaeolin and gluconasturtiin. The levels of 2-methylbutyl glucosinolate and glucoiberverin were 147 and 209 times higher in Coronado than in Malibu, respectively (Cluster 3). In Malibu, on the other hand, a number of phenolic compounds were intrinsically more abundant (Clusters 5 and 12).

The second principal component (PC2) explained 20.9% of the total variance and was associated with metabolites that were reduced (Fig. 4b1) Clusters 2 and 4) or induced (Clusters 7 and 11) by the Paraburkholderia treatments. Treatment of both cultivars with the Paraburkholderia species also widened the intrinsic cultivar variation in metabolites. Inoculation of Pbg had the greatest impact on the shoot secondary metabolome profile of both Broccoli cultivars, whereas the ineffective partnership between Pbt and Malibu had less pronounced impact on the shoot metabolome. Metabolites in cluster 2, comprising amino acids such as arginine, asparagine, tryptophan and N-acetylated glutamic acid/fucosamine, showed greater reduction in their abundance upon treatment with strains of the Paraburkholderia species when compared to the control. Cluster 4 encompassed metabolites that were more abundant in Malibu than in Coronado and included ascorbic acid ethyl ester, N-acetyl-tryptophan, and terpenoids putatively annotated as such as S-furanopetasitin and sonchuionoside C. The metabolites in clusters 7 and 11 were induced by all the strains of Paraburkholderia species and were dominated by phenolic compounds. In Malibu, inoculation of Pbg led to greater accumulation of flavonoids glycosides (i.e. kaempferol-di/tri-(feruloyl/coumaroyl glycosides and robinin), hydroxycinnamates (ferulic acid and its derivatives, caffeic acid derivatives such as chlorogenic acid) and indole-3-acetic-acid-O-glucuronide when compared to the other two Paraburkholderia species.

PC3 explained 4.9% of the total variance and was represented by Pbg-induced (Clusters 8 and 10) or Pbt-induced (Cluster 13) metabolites in both Broccoli cultivars. Pbg-enhanced metabolites in cluster 8 consisted of the flavonoid kaempferol 3-sophorotrioside, whereas Pbt-enhanced metabolites in cluster 13 included the hydroxycinnamate O-sinapoyl-beta-D-glucoside and resveratrol-sulfoglucoside, a stilbenoid.

Similarly, at 11dpi, inoculation with the strains of the Paraburkholderia species led to substantial changes in the shoot metabolite profiles of the two Broccoli cultivars (Fig. 4a2,b2). In the PCA, the first three PCs explained 51.1% of the total variance. The first PC, explaining 29.1% of the total variance is associated with metabolites that accumulated or were reduced in response to Paraburkholderia and the change in these groups of metabolites was more pronounced in Malibu cultivar (Fig. 4b2, Clusters 1, 2, 3, 4 and 5: up; 10 and 11: down). The induced metabolites in the above-mentioned clusters included flavonoids i.e. kaempferol-di/tri-glycosides (feruloyl/caffeoyl/coumaroyl), robinin, medicarpin-O-glucoside-malonate, as well as hydroxycinnamates, i.e. ferulic acid, caffeic acid and various derivatives of these metabolites. Furthermore, Paraburkholderia also induced coumarins such as eupatoriochromene and mahaleboside and mevalonate, a precursor of mevalonate pathways that goes into terpenoid biosynthesis. The reduced metabolites in both Broccoli cultivars included amino acids such as arginine, asparagine and N-acetylglutamic acid (Cluster 10). Meanwhile, metabolites in cluster 11 were also reduced by the Paraburkholderia treatment and these metabolites were intrinsically more abundant in cultivar Coronado. Some of the metabolites in cluster 11 included sulfur-containing metabolites such as 2-methylbutyl glucosinolate and glucoiberverin, derivatives of sulfurous amino acids including leucyl-cysteine and methionyl-isoleucine, as well as precursor or breakdown products of glucosinolates, for instance 6-methylthiohexanaldoxime and 3-methylsulfinylpropyl isothiocyanate.

The second PC (PC2) explained 23.8% of the total variance and was due to metabolites that were intrinsically more abundant in cultivar Malibu (Fig. 4b2, Clusters 6, 7, 8 and 9). Metabolites in clusters 6, 7, 8 and 9 showed significant reduction in all effective partnerships. Tryptophan, a building block for indolic glucosinolate and the growth hormone indole-3-acetic acid, N-acetylated amino acids including N-acetyl phenylalanine/tryptophan, terpenoids, i.e. S-furanopetasitin and sonchuionoside C, and sulforaphane, an isothiocyanate, are some of the metabolites in these clusters worth mentioning.

PC3 explained 6.2% of the total variance and was associated with yet unknown metabolites that showed Pbg specific alteration in both Broccoli cultivars (Clusters 13).

Paraburkholderia induces systemic resistance against the bacterial leaf pathogen Xanthomonas campestris in a cultivar-dependent manner

As shown above, the two Broccoli cultivars exhibited inherent differences in their shoot chemistry (Fig. 4). Furthermore, treatment of the plant roots with strains of the Paraburkholderia species led to substantial alteration of the shoot metabolome including metabolome signatures specific to the individual combination of Paraburkholderia species and Broccoli cultivar (Fig. 4). Based on this, we hypothesized that the inherent and induced differences in shoot chemistry between the two cultivars could contribute to a differential defense response against leaf pathogens. To address this hypothesis, treated and control plants of the two cultivars were challenged with two bacterial leaf pathogens, i.e. Xanthomonas campestris pv. armoraciae P4216 (Xca) and Xanthomonas campestris pv. campestris P4014 (Xcc).

The interaction effect of two independent variables (Paraburkholderia species and Broccoli cultivars) on disease severity of the two bacterial pathogens was assessed using beta regression. The strains of the Paraburkholderia species included three levels (Pbg, Pbh and Pbt) and the Broccoli cultivars consisted of two levels (Coronado and Malibu). There was a highly significant interaction effect of the strains of the Paraburkholderia species and Broccoli cultivars on disease severity on both Xanthomonas pathovars (Supplementary Table S4). No significant inherent variation in disease severity was observed between the two Broccoli cultivars when control plants were challenged with the two bacterial pathogens (Fig. 5). However, treatment of the roots with Paraburkholderia led to a clear reduction or enhancement of disease severity. For example, treatment with Pbg and Pbh enhanced disease severity by 18–28% in cultivar Coronado challenged by both bacterial pathogens, whereas Pbh and Pbt significantly reduced the disease severity in cultivar Malibu challenged by Xca (47% and 30%, respectively) and Xcc (33% and 28%, respectively) (Fig. 5).

Source: Ecology - nature.com