Apoplast Metabolomics Profiling Reveals Nitrogen-Dependent Modulation of Plant-Pathogen Interactions

Abstract

In the present study, we analysed the role played by the apoplast in the crosstalk between biotic and abiotic stress conditions. In particular, we studied the crosstalk between nitrogen (N) limitation and infection of the model plant Arabidopsis thaliana by Erwinia amylovora, an apoplastic bacterium. Our previous findings indicated that low N (LN) conditions increase E. amylovora in planta titres and expression of virulence factors. In this work, we extracted the apoplast wash fluids (AWF) from plants grown under low N or high N (HN) conditions and applied them to bacteria in vitro. We observed that LN-AWF induced stronger virulence gene expression than HN-AWF. Metabolomic analysis of both apoplast extracts revealed the presence of common metabolites; however, their proportions were distinct, indicating a direct effect of N availability on apoplast content. Interestingly, changes in the apoplast metabolite proportions were also observed early after bacterial infection, but only in plants grown under LN conditions. To evaluate the effect of single metabolites on virulence gene expression, we selected 43 metabolites and observed that 29 of them were activators, whereas two, GABA and citrate, acted as repressors. This study shows that environmental constraints, such as N availability, impact plant-pathogen interactions by altering the apoplastic content.

Keywords: Arabidopsis thaliana; Erwinia amylovora; hrp; apoplast; metabolome; multistress; nitrogen limitation; virulence.

FIGURE 1

LN‐AWF induces high Erwinia amylovora hrp expression in vitro and increases bacterial fitness in planta. (A) Apoplast wash fluid (AWF) and whole‐leaf extracts were prepared from Arabidopsis thaliana rosette leaves of plants supplied with 2 (Low N, LN) or 10 (High N, HN) mM NO₃ ⁻ E. amylovora cells were incubated for 6 h in vitro in rich medium (Luria Bertani, LB), inducing medium (IM), whole leaf extracts (LN‐L and HN‐L) or AWF (LN‐AWF and HN‐AWF). hrp gene expression was estimated using the hrp promoter‐β‐Gal reporter fusion. Expression values are given as Miller units as described in Zhang and Bremer (1995). Experiments were performed twice with similar results, the values presented in the graph correspond to the mean of both experiments (n = 6). Different letters indicate a significant difference according to the one‐way ANOVA test, post hoc Fisher's least significant difference (LSD) test, p < 0.01. (B, C) Bacterial titres of wild‐type E. amylovora in A. thaliana leaves supplied with 2 (Low N, LN) or 10 (High N, HN) mM NO₃ ⁻. Bacteria were incubated for 6 h in LB, IM, LN‐AWF or HN‐AWF prior to infection. Leaves were sampled 24 h post‐inoculation and colony‐forming units (CFU) were measured. (B) Plants grown in 2 mM NO₃ ⁻. (C) Plants grown in 10 mM NO₃. Two levels of significance threshold were considered according to the statistical test one‐way ANOVA; *p < 0.05, **p < 0.01. ns, non significant.

FIGURE 2

Gas chromatography–mass spectrometry analysis of metabolite content in low‐nitrogen apoplast wash fluid (LN‐AWF) and high‐nitrogen (HN)‐AWF. Heatmap of metabolites differently accumulated between LN‐AWF and HN‐AWF. The values correspond to the log₂ ratio of metabolites concentration in HN‐AWF versus LN‐AWF. (A) Apolar and fatty acids; (B) amino acids; (C) organic acids. Blue: more accumulated in HN‐AWF, Red: more accumulated in LN‐AWF, black: non significant. Only selected metabolites are shown.

FIGURE 3

Many metabolites present in the apoplast induce the expression of hrp genes. hrp gene expression was determined by lacZ activity in inducing medium without any carbon source (IM−C), rich medium (Luria Bertani, LB) and in IM complemented with a single carbon source (as indicated) at 5 mM. The lacZ reporter strain was grown in the corresponding medium at 28°C for 6 h, and β‐galactosidase activity was measured as described by Zhang and Bremer (1995). (A) We selected the carbon sources among the metabolites found more accumulated in LN‐AWF (Figure 1). Asterisks indicate a significant difference from the LB condition according to the one‐way ANOVA; (**p < 0.01). Values shown are the means and standard errors of three replicates. Three independent experiments were performed with similar results. (B) Heatmap of induction of hrp genes by a larger selection of metabolites, either accumulated in LN‐AWF (a), in HN‐AWF (b) or not significantly different in the two extracts.

FIGURE 4

Identification of repressors of Erwinia amylovora hrp gene expression. We used the ONPG test to determine hrp gene expression. E. amylovora was incubated 6H in inducing medium with no carbon source (IM−C) or in IM supplemented with two carbon sources, a potential repressor with a potential inducer. We tested two potential repressors, GABA (left column) and citrate (right column). Each repressor was tested in combination four different hrp inducers (black column). All metabolites were at the 5 mM concentration. hrp gene expression in the presence of the inducer and the repressor was compared to hrp expression in the presence of the inducer alone. Asterisks indicate significant differences according to one‐way ANOVA test (*p < 0.05; **p < 0.01; ns: p > 0.05).

FIGURE 5

Impact of Erwinia amylovora inoculation on metabolite levels in apoplast wash fluid (AWF). Gas chromatography–mass spectrometry (GC–MS) analysis of metabolite content reveals an effect of E. amylovora infection on apoplastic metabolome as soon as 6 h post‐inoculation (hpi) under low nitrogen (LN) conditions (A) and under both low and high N at 24 hpi (B). Metabolite content was analysed by GC–MS in AWF of Arabidopsis thaliana leaves 6 (A) and 24 (B) hpi following mock inoculation (black bars) or inoculation with E. amylovora (Ea; white bars). Plants were supplied with 2 (low N, LN) or 10 (high N, HN) mM NO₃ ⁻. Error bars represent standard error (n = 4). (A) Heatmap of selected metabolites that accumulate significantly differently between mock and infected tissue at 6 hpi. The data corresponds to the log₂ ratio of each metabolite content in mock‐inoculated versus infected leaves. Non‐significant metabolites are shown in grey. (B) Quantification of selected metabolites at 24 hpi. Asterisks indicate a significant difference between samples (water and infected samples at 6 or 24 hpi) according to t test (*p < 0.05).