Metabolite Analyses

For each clone, leaves (located in the face of bunches) were collected throughout the plot from five healthy (without visual foliar symptoms, as control, annotated C) and five Esca-diseased vines showing partial apoplexy symptoms (annotated D). For each healthy vine, two leaves were collected and pooled (10 leaves collected in total). For Esca-diseased vines, 10 leaves were collected from both symptomatic (three leaves in total) and asymptomatic shoots (seven leaves in total) of the same vines, and respectively, annotated Ds+ and Ds- (Figure ​(Figure1).1). Ds+ leaves were still green without apparent necrosis. Sampling was performed at the end of July, just at the beginning of symptom appearance. Samples were immediately frozen in liquid nitrogen and stored at -80°C until use. They were ground into a fine powder in liquid nitrogen before use.

Sampling scheme and pictures. For each clone, leaves were collected from healthy (without visual symptoms, as control, annotated C; left side of the figure) and Esca-diseased vines (^∗). For Esca-diseased vines, leaf samples were collected from both asymptomatic and symptomatic shoots of the same vines, respectively, annotated Ds– and Ds+ (right side of the figure).

Pigments were extracted from leaf powder (30 ± 5 mg) with 5 mL of 80% (v/v) acetone supplemented with 0.5% (w/v) CaCO₃ overnight at 4°C under continuous agitation. After centrifugation (10,000 × g for 10 min at 4°C), the supernatant was collected and absorbance was measured at 470, 647, and 663 nm. Pigment concentrations were then calculated according to Lichtenthaler method (Lichtenthaler, 1987).

Samples preparation, analysis, and data processing were performed as previously described (Fiehn et al., 2006, 2008; Krzyzaniak et al., 2018). Briefly, leaf samples (50 ± 10 mg) were resuspended in 1 mL of frozen water:acetonitrile:isopropanol (2:3:3) containing ribitol at 4 μg/mL and extracted for 10 min at 4°C with shaking at 1,400 rpm. After centrifugation (20,000 × g, 5 min), 100 μl supernatant were collected and dried for 5 h in a SpeedVac vacuum centrifuge. Samples were TMS-derivatized and analyzed using an Agilent 7890B gas chromatograph coupled to an Agilent 5977B mass spectrometer. For processing, data files were converted in NetCDF format and analyzed with AMDIS¹. A home retention indices/mass spectra library built from the NIST, Golm², and Fiehn databases and standard compounds were used for metabolite identification. Peak areas were also determined with the Targetlynx software (Waters) after conversion of the NetCDF file in Masslynx format. AMDIS, Target Lynx in splitless and split 30 mode data were compiled into a single Excel file for comparison (Supplementary Table S1). After blank mean substraction, peak areas were normalized to ribitol and fresh weight (in μg/mg fresh weight).

Some of the collected samples (5 C, 3 Ds- and 2 Ds+ ones of each clone) were prepared by overnight extraction in methanol (0.1 g/mL). Analysis of phenolics was performed by HPLC using a Beckman System Gold chromatography system equipped with a diode array detector Model 168 and a Beckman 507 sample injector equipped with a 20 μL sample loop as described by Krzyzaniak et al. (2018). Phenolics were separated on a Kinetex C18 column (4.6 × 100 mm, 2.6 μm, Phenomenex) at a flow rate of 1 mL/min and a mixture of solvent A (1.5% phosphoric acid in MilliQ water) and solvent B (100% acetonitrile) as mobile phase. Phenolics were eluted within 30 min with a linear gradient from 0 to 40% solvent B. Retention times were 6.43 min for trans-caffeoyltartaric acid, 7.58 min for trans-coumaroyl-tartaric acid, 11.89, 12.07, and 12.19 min for quercetin glycosides (quercetin-3-O-glucoside, rutin, and quercetin-3-O-galactoside, respectively), and 13.02 min for kaempferol-3-O-glucoside. Identification of the different phenolics was performed by comparison of their retention times and UV-vis spectra with those of reference compounds. Amounts of each phenolic compound occurring in the leaves were averaged from peak areas at 310 nm.

Qualitative variations of fluorescent compounds were measured in the same set of samples (5 C, 3 Ds- and 2 Ds+ replicates of each clone) analyzed by HPLC. Grounded leaves (200 mg) were mixed with methanol at 0.1 g/mL, centrifugated (3,000 × g, 5 min, room temperature), and the supernatant was then diluted in methanol (1:300). Methanolic extracts were analyzed with a Horiba Aqualog Horiba^® spectrofluorimeter using a 1-cm pathlength quartz cuvette. Excitation-Emission Matrices (EEMs) were acquired from 600 to 225 nm (3 nm steps for excitation) and from 211 to 617 nm (3.34 nm steps for emission) wavelengths.

Data were mathematically corrected in order to minimize inner filter effects, withdraw Rayleigh scattering and normalized to a Starna 1-ppm quinine sulfate reference cell. PARAFAC modeling was carried out using the drEEM tutorial (Murphy et al., 2013) accompanying Matlab code. The PARAFAC model was built with a number of components sufficient enough to best fit the variability of the EEM dataset and to validate through the core consistency and split half validation procedure. After model validation, F_max values for each PARAFAC component were obtained in order to represent the original EEM intensities.