Tissue collection and processing.
Branch and single leaf assay. For the branch and single leaf assays, Valencia sweet orange (Citrus sinensis) grafted onto Swingle rootstock was propagated in a greenhouse and inoculated with CLas using infectious ACP. Five ACP adults enclosed in a nylon mesh drawstring bag were applied and confined to newly emerged leaves for 14 days. Trees were then treated with insecticides, imidacloprid, and carbaryl, to eliminate ACP, and the trees were kept free of ACP until the time of sampling. For branch analyses, all leaves for a given time point were detached via razor blades, flash-frozen, and held in liquid nitrogen during sampling. Petioles were removed and stored at -80°C for CLas quantitation via qPCR. For metabolome analysis, ⅛ in (3.175 mm) diameter punches were dispensed into the wells of 96 well plates containing 500 µL 50% ethanol. Tissues were lysed by repeat freeze-thaw cycles between − 80°C and room temperature. The resulting extracts were filtered through 0.22 um filter plates via centrifugation before analysis via LC-MS. Leaf punches were collected in the same manner for the single-leaf assay. Collection of plant material comply with relevant institutional, national, and international guidelines and legislation
Field collected tissues. Stems, roots, and leaves from 50 trees (n = 150) were collected from 5 different citrus Florida citrus groves in 2016 34. In 2017, stems, roots, and leaves were collected from 80 trees located in 7 different orchards (n = 240). Each tree was divided into 4 quadrants (North, South, East, and West), and stems with attached leaves were collected from each of the quadrants and pooled. Topsoil from two sides of the tree and approximately 1.5 feet away from the base of the trunk near the irrigation line was removed, and the feeder roots near the irrigation line were sampled, shaken to remove soil, and sealed in a plastic bag. Gloves were changed, and clippers and shovels were sterilized with 30% household bleach between each tree that was sampled. All samples were immediately placed on ice for transit to the laboratory, where they were placed at 4˚C and processed within 24 hours.
MS data acquisition. The tissue extracts were prepared in 100% ethanol, spiked with 1µM sulfadimethoxine internal standard, and analyzed with UltiMate 3000 UPLC system (Thermo Scientific) using a Kinetex™ 1.7 µm C18 reversed-phase UHPLC column (50 X 2.1 mm) and Maxis Q-TOF mass spectrometer (Bruker Daltonics) equipped with ESI source. The column was equilibrated with 2% solvent B (98% acetonitrile, 0.1% formic acid in LC-MS grade water with solvent A as 0.1% formic acid in water), followed by a linear gradient from 2% B to 10% B in 0.2 min and then to 100% B at 12 min, held at 100% B for 2 min. Following each run, the column was equilibrated at 2% B for 1 min at a flow rate of 0.5 mL/min. MS spectra were acquired in positive ion mode in the range of 80-2000 m/z. A mixture of 10 µg/mL of each sulfamethazine, sulfamethizole, sulfachloropyridazine sulfadimethoxine, amitriptyline, and coumarin-314 was run at the beginning and the end of each batch (one 96-well plate). An external calibration with ESI-L Low Concentration Tuning Mix (Agilent Technologies) was performed prior to data collection, and internal calibrant Hexakis(1H,1H,3H-tertrafluoropropoxy)phosphazene was used throughout the runs. The capillary voltage of 4500 V, nebulizer gas pressure (nitrogen) of 1.4 bar, ion source temperature of 180°C, and dry gas flow of 4 L/min, were used. For acquiring MS/MS fragmentation, the 7 most intense ions per MS1 were selected. A stepping function was used to fragment ions at 50%, 100%, 150%, and 200% of the CID, with a timing of 25% for each step. Similarly, basic stepping of collision RF of 250 to 1500 Vpp with a timing of 25% for each step and transfer time stepping of 50, 75, 100, and 150 µs with a timing of 25% for each step was employed. MS/MS active exclusion parameter was set to 5 and released after 30 seconds. The mass of internal calibrant was excluded from the MS/MS list using a mass range of m/z 921.5–924.5. The data were deposited in the MassIVE online repository and are available under the IDs: MSV000082967 (2D leaf mapping); MSV000082962 (3D branch mapping); MSV000082963 and MSV000085416 (field study).
MS data analysis. The collected HPLC-MS raw data files were first converted from Bruker’s d to mzXML format and then processed with the open-source MZmine2 software [https://www.ncbi.nlm.nih.gov/pubmed/20650010?dopt=Abstract]. crop filtering with a retention time (RT) range of 0 to 14 min chromatograms. Mass detection was performed with a signal threshold of 1E3 and a 0.04-s minimum peak width. The mass tolerance was set to 20 ppm, and the maximum allowed retention time deviation was set to 5 s. For chromatographic deconvolution, the local minimum search algorithm with a 30% chromatographic threshold, minimum RT range of .6 sec, minimum relative height of 1%, minimum absolute height of 5E2, the minimum ratio of peak top/edge 2, and peak duration range of 0.04 - min was used. After isotope peak removal, the peak lists of all samples were aligned within the corresponding retention time and mass tolerances. Gap filling was performed on the aligned peak list using the peak finder module with 1% intensity, 10-ppm m/z tolerance, and 0.05-min RT tolerance, respectively. After the creation and export of a feature matrix containing the feature retention times, exact mass, and peak areas of the corresponding extracted ion chromatograms, the sample metadata was added to the feature matrix metadata of the samples.
All of the peaks that were present in any of the blanks with a signal-to-noise ratio (S/N) below 3:1 were removed from the final feature table.
Data pretreatment and statistical analysis. The data pretreatment and following statistical analysis were carried out with the MetaboAnalyst platform 55. The feature tables generated with MZmine were filtered to remove features with near-constant, very small values and values with low repeatability using the interquartile range (IQR) estimate. A detailed description of the methodology is given in 56. The samples were normalized using quantile normalization. The data were further scaled by mean centering and divided by standard deviation for each feature.
Principal component analysis (PCA) and partial least-squares discriminant analysis (PLS-DA) 57 were used to explore and visualize variance within data and differences among experimental categories. Random forests (RF) 58 supervised analysis was used to further verify the validity of determined discriminating features.
Molecular networking. The molecular network was created using the online workflow at GNPS platform (gnps.ucsd.edu) 1718. The data were clustered with MS-Cluster with a parent mass tolerance of 0.1 Da and an MS/MS fragment ion tolerance of 0.1 Da to create consensus spectra. The consensus spectra that contained less than 3 spectra were discarded. A network was then created where edges were filtered to have a cosine score above 0.65 and more than 4 matched peaks. The edges between two nodes were kept in the network if and only if each of the nodes appeared in each other's respective top 10 most similar nodes. The spectra in the network were then searched against GNPS's spectral libraries. All matches kept between network spectra and library spectra were required to have a score above 0.7 and at least 5 matched peaks. The molecular networks and the parameters used are available at the links below:
2D leaf mapping: https://gnps.ucsd.edu/ProteoSAFe/status.jsp?task=105598d2d782412ca0c988bfe933c032
3D branch mapping: https://gnps.ucsd.edu/ProteoSAFe/status.jsp?task=e2f1a1e367a7450fb940fe7cda04dc19
Field study: https://gnps.ucsd.edu/ProteoSAFe/status.jsp?task=2b0b1e554bc14d1fba321257b0ffc827
Field study, feature-based molecular network 59: https://gnps.ucsd.edu/ProteoSAFe/status.jsp?task=4ac8aff08af5433292143047c8cc5e90
2D/3D visualization. The procedure for the creation and visualization of 3D models is described in detail previously 13. Briefly, the 2D images and 3D model of the sampled plants were created and the coordinates for sampled spots were selected according to described protocol 13. The abundances of detected metabolites were normalized and autoscaled 55, and the coordinates for spots corresponding to each sample were inserted into the feature tables along with the spot size. The 2D or 3D models were drag-and-dropped into the ‘ili website in the browser (https://ili.embl.de/), followed by the feature table with coordinates. All figures were generated with the “Jet” color map. Spot size, opacity, and border opacity were adjusted for optimal visualization.
Synthesis of feruloyl putrescine hydrochloride. As the feruloyl putrescine was not available commercially, we have synthesized this compound for further compound identification validation and disc assay testing. N-Boc putrescine (1 equiv, 4.25 mmol, 0.91 mL) and ferulic acid (1.06 equiv, 4.5 mmol, 880 mg) were combined in anhydrous CH2Cl2 (32.5 mL) with stirring and the solution was cooled to 0°C. Then, a solution of N,N′-dicyclohexylcarbodiimide (DCC, 1.7 equiv, 7.2 mmol, 1.49 g) in anhydrous CH2Cl2 was added dropwise. The reaction mixture was allowed to warm to room temperature and stirred at for 2 days. Then, the mixture was filtered to remove precipitated dicyclohexylurea and the filtrate was concentrated in vacuo. The crude residue was purified over silica gel using 2–4% MeOH in CH2Cl2 to afford N-Boc feruloyl putrescine as a light yellow solid in 78% yield (1.29g). For Boc deprotection, trifluoroacetic acid (TFA, 15 mL) was added to a stirred solution of N-Boc feruloyl putrescine in CH2Cl2 (75 mL) under an inert atmosphere of Ar. The reaction was allowed to stir at room temperature for 40 min, then the solvent was removed in vacuo. The residue was dissolved in methanol (15 mL) with HCl (25 mL) and evaporated in vacuo to yield feruloyl putrescine hydrochloride as a light yellow solid (75% overall yield). 1H and 13C NMR data were consistent with those previously reported 60. HRMS (ESI) exact mass calculated for [M + H]+ (C14H21N2O3) is m/z 265.1547.
Simulations. Simulations were performed using the CLas Ishi-1 M15 metabolic model9. Standard biomass constraints were maintained to predict the overall CLas growth rate. All model simulations were performed using the Gurobi Optimizer v.5.6.3 solver (Gurobi Optimization) in the COBRA toolbox 61 for MATLAB (MathWorks). We simulated the maximal growth rate of CLas using flux-balance analysis. The main metabolic compounds affecting CLas growth were identified using the metabolome data, that was anthranilate, ferulate, glutamate, N-acetyl-L-ornithine, ornithine, putrescine, tyrosine, and vanillin and sensitivity analysis, looking for metabolite-specific growth responses was performed by varying uptake rates among 1x10− 12, 1x10− 10, 1x10− 8, 1x10− 6, 1x10− 4, 1x10− 2, 1x10− 1, 1x100, 1x101, 1x102, 1x103. Additionally, we performed a sensitivity analysis, which deployed a phenotypic phase plane that facilitates the observation of effects on the CLas growth by varying a particular constraint, in this case putrescine and ferulic acid. Predicted growth rates were compared with experimental results.
Microbial culture assays. Citrus metabolites were assayed using a previously developed disc-diffusion assay for L. crescens 27. Briefly, L. crescens liquid cultures were incorporated to a soft agar (0.8%) overlay and applied to a 20 mL solid agar (1.5%). L. crescens strain BT-1 ]was maintained and grown exclusively on the previously described bBM7 + 1.0 methyl-β-cyclodextrin. To these overlaid plates were applied autoclave-sterilized 6 mm paper discs (Whatman, NJ, USA), previously loaded with 35 µL of a given metabolite solution and dried in a sterile biosafety cabinet. Once discs were applied, plates were sealed and stored upside down in a 28 ℃ incubator for 6 days to allow a clear zone of inhibition development for measurement. 35 mg/mL solutions of putrescine (Fisher Scientific, Waltham, MA, USA), and feruloylputrescine (synthesis described above) were prepared using sterile water.
The efficacy of bioflavonoids and ferulic acid in vitro CLas-hairy root assay.
The in vitro CLas-hairy roots assay was performed according to the previously described protocol 28. CLas-citrus hairy roots were generated using CLas-infected citrus plant tissues, and the diagnosis of CLas was confirmed by quantitative PCR (qPCR). CLas-hairy roots were surface sterilized, and ~ 100 mg was transferred into multi-well plates containing Gamborg’s B-5 medium with 1% sucrose. Different concentrations of bioflavonoids (Horbaach, https://horbaach.com/products/citrus-bioflavonoids-complex-1500mg-300-vegetarian-caplets) and ferulic acid (Fisher Scientific, Catalog No. ICN10168505): 125, 250, 500, and 1000 ppm/mL, were added, vacuum infiltrated and incubated on a rotator shaker at 50 rpm in the dark at 25°C for 72 h. The experiments were carried out with six biological replicates, positive control (oxytetracycline hydrochloride), untreated CLas hairy roots, and an equal concentration of ethanol solvent used to dissolve the bioflavonoids and ferulic acid as negative controls. After the treatments, tissue samples were treated with PMAxx dye (propidium monoazide, Biotium, Fremont, CA) to inactivate dead CLas bacterial DNA. Further, total DNA was extracted, and viable bacterial titer was estimated by qPCR analysis using primers specific to the CLas gene encoding the Ribonucleotide reductase β-subunit (nrdB, RNR-F/RNR-R) 62 and the relative CLas titers were estimated and plotted relative to untreated using the 2 − ΔΔCt method. After normalization of target Ct with an endogenous reference gene (Ct') glyceraldehyde3-phosphate dehydrogenase 2 (GAPC2)63 to correct for DNA template concentration differences among the samples, it was plotted relative to untreated controls.
Primers used in this study
RNR-F | CATGCTCCATGAAGCTACCC | 62 |
RNR-R | GGAGCATTTAACCCCACGAA |
GAPC2-F | GAGGAGATCCCATGGGCAAA | 63 |
GAPC2-R | AAGAGGAGCTAGGCAGTTGG |