Bacterial Pathogen Infection Triggers Magic Spot Nucleotide Signaling in Arabidopsis thaliana Chloroplasts through Specific RelA/SpoT Homologues

Magic spot nucleotides (p)ppGpp are important signaling molecules in bacteria and plants. In the latter, RelA-SpoT homologue (RSH) enzymes are responsible for (p)ppGpp turnover. Profiling of (p)ppGpp is more difficult in plants than in bacteria due to lower concentrations and more severe matrix effects. Here, we report that capillary electrophoresis mass spectrometry (CE-MS) can be deployed to study (p)ppGpp abundance and identity in Arabidopsis thaliana. This goal is achieved by combining a titanium dioxide extraction protocol and pre-spiking with chemically synthesized stable isotope-labeled internal reference compounds. The high sensitivity and separation efficiency of CE-MS enables monitoring of changes in (p)ppGpp levels in A. thaliana upon infection with the pathogen Pseudomonas syringae pv. tomato (PstDC3000). We observed a significant increase of ppGpp post infection that is also stimulated by the flagellin peptide flg22 only. This increase depends on functional flg22 receptor FLS2 and its interacting kinase BAK1 indicating that pathogen-associated molecular pattern (PAMP) receptor-mediated signaling controls ppGpp levels. Transcript analyses showed an upregulation of RSH2 upon flg22 treatment and both RSH2 and RSH3 after PstDC3000 infection. Arabidopsis mutants deficient in RSH2 and RSH3 activity display no ppGpp accumulation upon infection and flg22 treatment, supporting the involvement of these synthases in PAMP-triggered innate immune responses to pathogens within the chloroplast.


Analyte Assignment
Pre-spiking of 15 N labeled ppGpp enabled the assignment of decomposition products during TiO2 enrichment. The following figure S4 shows how ppGpp partially decomposed to two isobaric analytes, which we assign as ppGp with a phosphate ester either in the 2´or 3´ position, designated as ppGp (2´); ppGp (3´). Given the same ratio of unlabeled and labeled analytes, we conclude that there is no natural ppGp in plants.

Figure S1
: ppGpp and decomposition product signals after TiO2 extraction of ppGpp from wild type plants (Col-0) with heavy isotope standard pre-spiking. Red trace: ppGpp of plant origin; orange trace: decomposition products derived from plant ppGpp during extraction. Grey traces: heavy ppGpp and ppGp as a result of concomitant heavy ppGpp degradation.
In order to unambiguously assign the decomposition products, we generated a mixture of heavy ppGp (2´and 3´) by basic treatment of 15 N labeled ppGpp, which leads to cyclophosphate formation to the required reference compounds. Figure S5 shows the clean conversion of the ppGpp reference to the desired mixture by removal of one phosphate unit (Pi). The coupled 31 P NMR also validates the regioselectivity. Figure S2: ppGpp hydrolysis leads to a regioisomeric mixture of ppGp (2´and 3´) as shown by 31 P NMR spectroscopy.
The obtained mixture of ppGp (2´) and ppGp (3´) was analyzed by CE-MS, also including a synthetic sample of pGpp, which is another isobaric analyte. Indeed, the CE-MS protocol is capable to separate ppGp (2´) and ppGp (3´) and pGpp, as shown in figure S6. This underlines that our assignment of the TiO2 decomposition products is correct.

Limit of Detection, Limit of Quantitation, Linearity
LOD and LOQ were determined via spiking of extracted plant material (150 mg FW) with heavy 15 N ppGpp standard (three biological repeats) on the QQQ System. LOD was defined as a signal to noise ratio (S/N) of above 3:1 and LOQ as S/N higher than 10:1. LOD was determined as 10 nM and LOQ was determined as 30 nM. One representative chromatogram of LOD and LOQ is shown in figure S5 (LOD) and figure S6 (LOQ), respectively. Automatic noise selection (see figure S5 and S6, red) was artificial, so noise was selected manually (see figure S5 and S6, blue) and calculated by data analysis software.