Bacterial Strains and Culture Conditions
The Xanthomonas oryzae pv. oryzae (Xoo) strain PXO99A was kindly provided by Prof. G. L. Qian (Nanjing Agricultural University) [37]. Xoo was cultivated at 28℃ on nutrient agar (NA) medium in plates or in nutrient broth (NB) medium in flasks. NA medium consists of 10 g of sucrose, 5 g of peptone, 3 g of beef extract, 1 g of yeast powder, and 15 g of agar powder per liter of distilled water. NB medium contained the same components but lacked agar powder.
Chemicals
Resveratrol (1) was isolated from Smilax china (purity > 95%). Oxyresveratrol (2) and piceatannol (3) were purchased from Hangzhou Great Forest Biomedical Ltd. (purity > 95%) (Zhejiang, China). Six derivatives of the above three stilbenoids (1–3), (di-hydro-resveratrol (4), di-hydro-oxyresveratrol (5), di-hydro-piceatannol (6), tri-methyl-resveratrol (7), tetra-methyl-oxyresveratrol (8), and tetra-methyl-piceatannol (9)) were synthesized in our lab. All other chemicals used in this study were purchased from Sigma Chemical (St. Louis, MO, USA).
Minimum Inhibitory Concentrations (MIC)
MICs of compounds (1–9) were determined by following the two-fold dilution method [38]. In brief, overnight cultures of Xoo (1%, v/v) were resuspended in fresh nutrient agar (NB) medium in the presence of the samples (0.001–0.1 mg/mL for compounds (1–6) and 0.01–1 mg/mL for compounds (7–9)) in 96-well plates, then incubated at 28 °C and 180 rpm for 18 h. The MICs of compounds (1–9) against Xoo are shown in Table 1.
Antibacterial investigation of compounds (1-9) against Xoo
The antibacterial activity of compounds (1–9) against Xoo was investigated as per previous research, with some modifications [39]. Briefly, overnight cultures of Xoo (1%, v/v) were resuspended in the fresh NB medium supplemented with samples at concentration gradients in test tubes, then incubated at 28 °C and 180 rpm for 18 h. The concentrations of compounds (1–6) were 0, 2.5, 5, 10, 25, 50, and 100 μg/mL, and compounds (7–9) were 0, 50, 100, 200, 500, and 1000 μg/mL. The same amount of DMSO was used as the control. Then cultures were analyzed for antibacterial activity at 600 nm by a spectrophotometer (BioTek, Vermont, USA). Percentage inhibition was calculated as follows:
Inhibition % = [1− (Ai/A0)] × 100
where Ai is the OD600 of the cultures with compounds (1–9) and A0 is the OD600 of the control culture.
Bacterial growth measurement
The effects of resveratrol on the Xoo growth were measured by following the previous methods, with some modifications [40]. Briefly, overnight culture of Xoo (1%, v/v) were resuspended in the fresh NB medium supplemented with resveratrol at different concentrations (0, 2.5, 5, 10, 25, 50, and 100 μg/mL), then incubated at 28 °C and 180 rpm. The same amount of DMSO was used as the control. The OD600 values of the culture were measured every 2 h for up to 24 h by a microplate reader (BioTek, Vermont, USA). The Xoo growth was evaluated by plotting the values of OD600 against time. The results were shown in Fig. S1.
Transmission electron microscope of the Xoo flagella
Xoo flagella were detected by transmission electron microscopy (TEM) following the methods, with some modifications [22]. Each Xoo overnight culture was diluted into fresh NB medium at 1% (v/v), then incubated at 28 °C and 180 rpm for 18 h. Resveratrol was added at 11.67 μg/mL to ensure 50% effect on growth. The suspension was deposited onto grids, then stained with 2% uranyl acetate for 30 s and dried for 10 min at room temperature, with the flagella then observed by TEM (Tecnai 12, Philips, Holland).
Extraction of Xoo metabolite
Metabolites of Xoo were extracted according to our previous study, with some modifications [10]. Each Xoo overnight culture was diluted into 30 mL of fresh NB medium at 1% (v/v) in Erlenmeyer flasks, then incubated at 28 °C and 180 rpm for 18 h (OD600 ≈ 0.82). Resveratrol was added at 11.67 μg/mL to ensure 50% effect on growth (OD600 ≈ 0.41). To ensure bacterial equality, two Erlenmeyer flasks of culture were combined to obtain one treated group sample. The same amount of DMSO was used in the control group. Fourteen biological replicates were used for the treatment and control groups, respectively. After incubation, the cell culture was chilled by brief incubation on ice. The cell pellet was obtained by centrifugation at 12,000 rpm for 15 min at 4 °C. Subsequently, the cell pellet was washed three times with phosphate-buffered saline (PBS), and then transferred to a 10-mL microtube equipped with 3.8 mL of precooled methanol/water (1/0.9, v/v), and stored at 4 °C until use. Mixtures were then extracted with a homogenizer for 5 min on ice, with 2 mL of chloroform added. After vortexing, place the mixtures on ice for 10 min, then centrifuged at 12,000 rpm for 15 min at 4 °C. Subsequently, the supernatants were transferred to new centrifugal tubes and treated under vacuum with a Speed-Vac Concentrator (Thermo SAVANT, SC110A-230) to completely remove methanol. The supernatants were stored at -80 °C overnight, and then lyophilized in a freeze drier. All samples were stored at -80 °C for further analysis.
Nuclear Magnetic Resonance (NMR) Measurements
According to published methods [10], the lyophilized extracts were dissolved in 600 μL of 99.8% D2O PBS buffer (pH 7.4) equipped with 0.05% (w/v) sodium 3-(trimethylsilyl) propionate-2,2,3,3-d4 (TSP). After vortexing, the mixtures were centrifuged at 12000rpm for 15 min to discard sediments. The supernatants were transferred to new NMR tubes for 1H-NMR analysis.
The 1H-NMR spectra of samples were recorded on a Bruker AVANCE III 500 MHz NMR spectrometer at 298 K. D2O was used for field frequency locking, TSP was used as the chemical shift reference (1H, 0.00 ppm). A transverse relaxation-edited Carr-Purcell-Meiboom-Gill (CPMG) sequence [90(τ-180-τ) nacquisition] with a total spin-echo delay (2 nτ) of 40 ms was used to suppress the signals of proteins. 1H-NMR spectra were measured with 128 scans in 32 K data points with a spectral width of 10, 000 Hz. The spectra were Fourier transformed after multiplying the free induction decays (FIDs) by an exponential weighting function corresponding to a line-broadening of 0.5 Hz.
Data Preprocessing and Peak Assignments
Before analysis, the 1H-NMR spectrum was manually phased and baseline corrected using Bruker Topspin 3.0 software (Bruker GmbH, Karlsruhe, Germany) and referenced to TSP at 0.0 ppm. Subsequently, the ASCII format files were obtained by the convert of MestReNova (Version 8.0.1, Mestrelab Research SL). And then these files were read into R software (http://cran.rproject.org) for multivariate analysis. The spectra between 0.2 and 10 ppm were segmented with an average binning of 0.005 ppm. The regions influenced by the residual resonance of water was cut off between 4.4 and 5.175 ppm. Then, all spectra were conducted probabilistic quotient normalization and mean-centered before multivariate statistical analysis.
The NMR resonances were assigned by querying metabolomics databases, including the Human Metabolome Database (HMDB, http://www.hmdb.ca) and Madison-Qingdao Metabolomics Consortium Database (MMCD, http://mmcd.nmrfam.wisc.edu), in conjunction with the Chenomx NMR suite 7.5 (Chenomx Inc., Edmonton, Canada) and statistical total correlation spectroscopy (STOCSY) [10].
Multivariate Data Analysis
Multivariate statistical analysis was applied to the NMR data, and included principal component analysis (PCA) and supervised orthogonal signal correction partial least-squares discriminant analysis (OSC-PLS-DA). Unsupervised PCA was first used to reduce the dimensionality of the imported NMR data, and some new latent variables, principal components, were obtained, with such components smaller than variables before transformation. And then filter out irrelevant effects and maximize the discrimination of inter group differences by supervised OSC-PLS-DA. The OSC was applied prior to PLS-DA to filter out unrelated variables not concerning class discrimination to minimize the influence of unrelated signals.
The quality of the OSC-PLS-DA model was evaluated by repeated two-fold cross-validation. The R2 and Q2 parameters reflected the prediction ability and the goodness-of-fit of the constructed models. In order to the further validate the supervised model, a permutation test (2000 times) was performed [41]. And color-coded loading plots were constructed to reveal variables that contributed to group separation. The fold-change values of metabolites and their associated p-values corrected by the Benjamini and Hochberg-adjusted method were calculated and visualized in colored tables [42]. In addition, receiver operating characteristic (ROC) curves were used to verify the classifier performance of the established OSC-PLS-DA models after 200 times repeated two-fold cross-validation [10].
Univariate Statistical Analysis
Univariate analyses, including nonparametric Mann-Whitney tests and the parametric Student’s t-tests [10], which were used to detect difference in crucial metabolites between groups. The fold-change values of the identified metabolites as well as p-values between groups were calculated. Then false discovery rate was controlled by adjusting p-values according to the Benjamini-Hochberg method when proceeding with multiple comparisons.
Statistical Analysis
All experiments were run in triplicate, and experimental results were expressed as means ± standard deviation or averages. Data were analyzed by one-way analysis of variance (ANOVA) and Duncan’s multiple range test were performed using SPSS version 17.0 (SPSS Inc., Chicago, IL, USA) statistical software with a significant P value of p < 0.05.
Availability of data and materials
All data generated or analysed of this study are described in this paper.