Article pubs.acs.org/JAFC
Cite This: J. Agric. Food Chem. 2018, 66, 10317−10326
Comprehensive Metabolomics Analysis of Mandarins (Citrus reticulata) as a Tool for Variety, Rootstock, and Grove Discrimination Shi Feng,#,‡,∥ Liying Niu,†,‡,∥ Joon Hyuk Suh,‡ Wei-Lun Hung,‡ and Yu Wang*,#,‡
J. Agric. Food Chem. 2018.66:10317-10326. Downloaded from pubs.acs.org by UNIV OF LOUISIANA AT LAFAYETTE on 10/04/18. For personal use only.
#
Department of Food Science and Human Nutrition, University of Florida, 572 Newell Drive, Gainesville, Florida 32611, United States ‡ Citrus Research and Education Center, Food Science and Human Nutrition, University of Florida, 700 Experiment Station Road, Lake Alfred, Florida 33850, United States † Institute of Farm Product Processing, Jiangsu Academy of Agricultural Sciences, Nanjing 210014, People’s Republic of China S Supporting Information *
ABSTRACT: The metabolite profile responsible for the quality of mandarin fruit is influenced by preharvest factors including genotype, rootstock, grove location, etc. In this paper, mandarin varieties were discriminated using metabolomics. Additionally, effects on metabolic profiles due to grove location and rootstock differences were also investigated. Results revealed that mandarin varieties could be differentiated using the metabolite profile, while the compositions of flavonoids have the potential for variety differentiation. With regard to fruits of the same variety, grove location might determine the overall profile of metabolites, whereas rootstock possibly affected composition of secondary metabolites. Pathway enrichment analysis demonstrated that biosynthesis pathways of terpenoids and steroids involving limonene and linalool were highly influenced by variety diversity. Moreover, the flavonoid biosynthesis pathway, involving hesperetin, naringenin, eriodictyol, and taxifolin, was indicated to have a close relationship with rootstock differentiation. This study provides useful and important information with depth for breeding and optimizing preharvest practices. KEYWORDS: metabolomics, mandarin, variety discrimination, rootstock, grove location, pathway enrichment analysis
■
INTRODUCTION The delicate flavor of mandarin is responsible for the popularity of this fresh fruit that is largely consumed worldwide. In the citrus fruit family, mandarin (Citrus reticulata) constitutes a large, distinctive, and highly varied group. According to Tanaka’s classification,1 mandarin can be principally divided into natural subgroups including Satsuma mandarin (Citrus unshiu Marcovitch), King mandarin (Citrus nobilis Loureiro), Mediterranean mandarin (Citrus deliciosa Tenore), common mandarins (Citrus reticulata Blanco), small-fruited mandarin (Citrus indica, Citrus tachibana and Citrus reshni),2 as well as mandarin hybrids including tangor (Citrus reticulata × Citrus sinensis, i.e. mandarin × orange hybrid) and tangelo (Citrus paradisi × Citrus reticulata, i.e. grapefruit hybrid × mandarin).3 Different varieties of mandarin fruit present diverse qualities, such as appearance, nutritional value, and flavor profile. Besides cultivar genotype, other preharvest factors including rootstock, environmental conditions (depending on the location of grove) and cultural practices also play an important role determining fruit quality.4−7 To provide information with depth for breeding and optimizing preharvest practices, changes of metabolites related to these factors need to be studied using metabolomics. Metabolomics is increasingly applied to a wide range of disciplines including plant science, medical science, environmental science, etc.8−10 It can provide detailed information regarding similarities and differences in the metabolite composition among samples. The most common techniques used in these metabolomics studies are mass spectrometry (MS) based. © 2018 American Chemical Society
However, no single analytic methodology is ideal for all metabolites. Typically, a combination of techniques is necessary to analyze the majority of metabolites of differing polarity and varying molecular weights. Therefore, a comprehensive metabolomics strategy using multi-MS techniques (GC-MS, GC-MS/ MS, and LC-MS/MS) is a valuable tool to display genetically or environmentally induced variations in metabolite compositions. Some metabolites of mandarin fruits are closely related to the flavor profile. Sugars and organic acids are essential primary metabolites involved in the fundamental metabolic pathways in mandarin. Additionally, they contribute sweetness and sourness to the gustatory perception of mandarin, whereas the secondary metabolites, flavonoids, contribute bitterness.11−13 Meanwhile, volatile compounds dominate the olfactory perception of mandarin, even demonstrated as key factors that characterize mandarin flavor during sensory evaluation.14 Although showing no direct influence on mandarin flavor, two other primary metabolites, amino acids and fatty acids, are major nutrients from which flavor-related volatiles are derived.15 The comprehensive metabolic profiling of mandarins will have great potential to facilitate a variety or preharvest practices for discrimination based on its close correlation with flavor attributes and the modification of flavor attributes through genomic and phenomic tools (e.g., linking Received: Revised: Accepted: Published: 10317
July 20, 2018 September 10, 2018 September 12, 2018 September 12, 2018 DOI: 10.1021/acs.jafc.8b03877 J. Agric. Food Chem. 2018, 66, 10317−10326
Article
Journal of Agricultural and Food Chemistry
semiquantified based on the concentration of the internal standard. In the case of unavailable commercial reference compounds, compounds were tentatively identified using the NIST library and comparing retention indices with the literature and online databases (Flavornet, The Pherobase, PubChem and LRI & Odor Database). Determination of Sugars, Organic Acids, and Amino Acids. The analytical methods were modified from Cerdán-Calero et al.16 Ten milliliters of juice was centrifuged (2000g, 5 min, 4 °C), and then the supernatant was filtered through a 0.22 μm nylon filter. Ten microliters of filtered juice was mixed with 40 μL of adonitol (1 mg/mL in methanol, internal standard) and then exhaustively dried using a SpeedVac evaporator (Thermo Scientific, Waltham, MA). The dried residue was next mixed with 30 μL of methoxyamine hydrochloride (20 mg/mL in pyridine) and vortexed for 2 h at room temperature. Afterward, 80 μL of N-methyl-N-(trimethylsily)trifluoracetamide were added into the sample, and the mixture was vortexed for an additional 30 min. GC-MS analysis of the derivatized sample was carried out using an Agilent 7890 GC coupled with an Agilent 5975C MS (Santa Clara, CA) in EI mode with 70 eV. The sample solution (1 μL) was injected at 230 °C in split mode (10:1) onto an Rxi-5 MS capillary column (30 m × 0.25 mm, 0.25 μm film thickness). The oven temperature was programmed at 70 °C for 5 min, increased to 270 °C at 4 °C/min, and then ramped to 320 °C (5 min hold) at 20 °C/min. The MS was set to scan from m/z 60 to 650. Helium was used as carrier gas with a flow rate of 1.1 mL/min. Identification and semiquantification of target compounds were conducted as described above. Determination of Flavonoids and Limonoids. One hundred milligrams of crushed dried flesh was mixed with 1 mL of methanol (containing 0.1% formic acid) and 10 μL of catechin (1 mg/mL in methanol, internal standard) and then vortexed for a continuous 15 min. After centrifugation (3000g, 5 min, 4 °C), the supernatant was collected and the solvent of the supernatant was removed using a SpeedVac evaporator. The dried residue was reconstituted in 500 μL of methanol and then passed through a C18 solid phase extraction cartridge (Varian Inc., Harbor City, CA). The eluate was rinsed out using 500 μL of methanol and collected for LC-MS analysis. The LC-MS analysis was performed following a previously reported procedure.19 The SRM transitions, collision energy, RF lens and retention time of the analytes, and internal standard are provided in Table S1. Determination of Fatty Acids. Twenty milligrams of dried fruit was added to a 15 mL glass tube and mixed with 20 μL of 5% butylated hydroxytoluene methanol solution. The sample was then extracted with 2 mL of tert-butyl methyl ether (MTBE)/methanol solution (2:1, v/v) containing an internal standard (heneicosanoic acid, C21:0) by vortexing for 30 min. After centrifugation at 1500g, 4 °C for 10 min, supernatant was transferred into a glass tube and dried under a stream of nitrogen gas. The dried residue was redissolved in 1 mL of hexane. To derivatize fatty acids to fatty acid methyl esters (FAMEs), 1 mL of 1 M potassium hydroxide methanol solution was added to each sample. After shaking for 30 min, the supernatant (1 μL) was analyzed using GC-MS/MS (TSQ 8000, Thermo Fisher Scientific, San Jose, CA) equipped with an EI ion source. The separation was performed on an HP-88 Column (100 m × 0.25 mm, 0.20 μm film thickness, Agilent Technologies, Santa Clara, CA). The injection volume was 1 μL, using a split ratio of 20:1. The oven temperature was programed from 100 to 170 °C (28 min hold) at 10 °C/min, and then ramped to 210 °C (10 min hold) at 2 °C/min. The temperatures of the GC injector and transfer line were set at 250 and 240 °C, respectively. The MS was operated with timed-selective reaction monitoring (SRM). The ion source temperature was set at 280 °C, and EI energy was 70 eV. SRM transitions were optimized as follows: m/z 74 → 43 for palmitic acid, stearic acid, arachidic acid, palmitoleic acid, oleic acid and vaccenic acid, m/z 81 → 79 for linoleic acid, and m/z 79 → 77 for α-linolenic acid. Statistical Analysis. Principle component analysis (PCA) and partial least-squares-discriminant analysis (PLS-DA) were performed using SIMCA (Version 15.0, Umetrics, Umeå, Sweden). One-way analysis of variance (ANOVA) (P < 0.05) was performed using SPSS v.20.0 (IBM Corp., Armonk, NY) to estimate significant differences
phenotypic traits to the genetic information for particular genotypes). Diverse compositions of metabolites have been reported among mandarin genotypes, rootstocks, microenvironment, etc.16−18 Nevertheless, no comprehensive investigation on the metabolite profile has been reported regarding the unique fruit quality of mandarin cultivars. In the present study, we defined and compared four mandarin varieties, targeting both selected primary and secondary metabolites via various mass-spectrometry-based metabolomics approaches. On the basis of the integrated analytical results, the biomarkers responsible for discriminating mandarin variety, rootstock, and grove location were identified, and finally the biomarkers involved in metabolic pathways closely related to the variety differentiation and rootstock discrimination were discussed.
■
MATERIALS AND METHODS
Fruit Samples. Clementines (Citrus clementine hort. ex Tanaka) were obtained from three commercial groves and labeled according to their sources, i.e., Clem-H from Halos (Los Angeles, CA), Clem-C from Cuties (Pasadena, CA) and Clem-B from Bee Sweet Citrus (Fowler, CA). Satsumas (Citrus unshiu Marcovitch) and Pages (Citrus tangelo) were obtained from Bee Sweet Citrus. Daisy (Citrus reticulata Blanco) fruits were obtained from Lindcove Research and Extension Center, California. They were harvested from three rootstocks, i.e., Daisy/RUBT (Rubidoux trifoliate orange), Daisy/CARR (Carrizo citrange) and Daisy/SHRL (Schaub rough lemon). All the fruits were stored at −20 °C before analysis. Chemicals. Reference compounds (Tables 3−6) were purchased from Sigma-Aldrich (St. Louis, MO), Indofine Chemical Company, Inc. (Hillsborough Township, NJ), Yuanye Biotechnology Co., Ltd. (Shanghai, China), ChromaDex (Irvine, CA), and Nu-Chek Prep, Inc. (Elysian, MN). Internal standards (4-heptadecanone, adonitol, catechin, and heneicosanoic acid) and n-alkane standards were purchased from Sigma-Aldrich and Nu-Chek Prep, Inc. Fruit Sample Preparation. Fruit juice was hand-squeezed with an electric juicer prior to determination of volatiles, sugars, organic acids, and amino acids. For the analyses of fatty acids, flavonoids, and limonoids, fruit flesh was cut into small pieces and freeze-dried at −52 °C, 0.11 mbar for 6 days (FreeZone 2.5 L, benchtop freeze-dry system, Labconco, Fort Scott, KS). For each variety, three fruits were sampled and used to prepare for the fruit juice/flesh while each juice/flesh sample was analyzed in triplicate. Determination of Volatile Components. For each extraction, 5 mL of hand-squeezed juice was transferred into a 40 mL headspace vial and capped with a PTFE/silicon septum cap. Sodium chloride (1.8 g) and a magnetic stir bar were added to the juice sample to facilitate the volatile extraction while 5 μL of 4-heptadecanone (0.4 mg/mL in methanol) was added as an internal standard. The juice sample was incubated in a water bath (40 °C, 20 min) and then extracted by a 50/30 μm DVB/CAR/PDMS fiber (Supelco Inc., Bellefonte, PA) for 30 min. The volatile analysis was carried out on a Clarus 680 gas chromatograph (GC) (PerkinElmer, Inc., Waltham, MA) equipped with a Clarus SQ 8T mass spectrometry (MS) detector. Separation was achieved using a TR-FFAP capillary column (30 m × 0.25 mm, 0.25 μm film thickness, Thermo Scientific, Bellefonte, PA). The oven temperature was initially set at 40 °C (2 min hold) and then ramped at 5 °C/min to 230 °C (10 min hold). The temperature of the GC injector was set at 250 °C. Mass spectra in the electron impact mode (MS-EI) were applied at 70 eV ionization energy. The MS was set to scan the m/z range from 50 to 300 in the positive mode. Helium was used as the carrier gas at a flow rate of 1.1 mL/min. A series of n-alkanes was used to determine linear retention indices (RI) of each volatile. Identification of volatile compounds was achieved by comparison with reference compounds through comparing the (1) retention indices on the FFAP capillary column and (2) mass spectra in EI modes. Relative concentrations of target compounds were 10318
DOI: 10.1021/acs.jafc.8b03877 J. Agric. Food Chem. 2018, 66, 10317−10326
Article
Journal of Agricultural and Food Chemistry
conditions, such as elevation and soil depth.30,31 Furthermore, our results could also be supported by the work of Robards et al.,32 in which flavonoids were implied to be ideally suited to differentiate citrus varieties. Biomarkers for Mandarin Variety Discrimination and Metabolic Network Analysis Related to Variety Discrimination. To identify variables responsible for the variety differentiation, PLS-DA was performed on individual varieties based on the results of all metabolites (Figure 1a). The plot
among data. After the test of homogeneity of variances, analytes that showed homoscedasticity (p > 0.05) were analyzed by ANOVA and Tukey while Welch and Dunnett T3 (posthoc) were applied to analytes that showed heteroscedasticity (p < 0.05). The metabolic pathway enrichment analysis was implemented using MBRole (http://csbg.cnb.csic.es/mbrole2/index.php). The purpose of metabolic pathway enrichment analysis is to identify coordinately changed Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways using metabolite data. “Biosynthesis of terpenoids and steroids (map01062)” and “Flavonoid biosynthesis (map00941)” from KEGG database (http://www.genome.jp/kegg/) were used as references for pathway mapping.
■
RESULTS AND DISCUSSION Metabolite Profiling of Mandarins. Four varieties of mandarin fruits were used in metabolite studies, including Clementine, Satsuma, Page, and Daisy. Clementine fruits were collected from three different groves (grafted on the same rootstock), whereas Daisy mandarin fruits were harvested from three different rootstocks in the same grove. Clementine mandarins have been demonstrated to be a hybrid of Willowleaf mandarin and sweet orange.20,21 Page mandarin is a hybrid between Minneola tangelo and Clementine mandarin22 while Daisy mandarin is considered to be a hybrid between Fortune mandarin and Fremont mandarin.23 From a recent study, the inferred parents of Satsuma mandarin were proposed as Kishu mandarin and Kunenbo mandarin.24 In this study, both primary metabolites (sugars, organic acids, amino acids, and fatty acids) and secondary metabolites (volatile compounds, flavonoids, and limonoids) were profiled in mandarin fruits. GC-MS was applied to the analyses of volatile components, sugars, organic acids, and amino acids, while flavonoids, limonoids, and fatty acids were analyzed using LCMS/MS and GC-MS/MS, respectively. A total of 104 metabolites were identified, including nine sugars (sugar alcohol and sugar acids included), six organic acids, 11 amino acids, 8 fatty acids, 41 volatiles, and 29 flavonoids and limonoids. All analytes were semiquantified using corresponding internal standards for comparing the relative changes (Tables 3−6).19,25,26 Discrimination of Metabolic Profiles among Different Varieties of Mandarin. To examine the variety differentiation among mandarin cultivars, PCA was applied to individual varieties based on metabolite profiling data. From PCA, mandarins were clearly distinguished by varieties, explained by 45.0% on the first two principal components (PC1 and PC2) (Supporting Information Figure S1a). In the PCA score plot, Daisy cultivars were grouped closely together although they were harvested from different rootstocks. Clementine fruits from three locations in California showed some separation. However, they could still be grouped together and distinguished from other varieties. This was consistent with our hypothesis. In fact, metabolomic analysis has been successfully applied to various fruits and vegetables, including citrus fruits,27 avocado,28 and pepper,29 to illustrate variety discrimination. To look more closely at the metabolite profiles of mandarin fruits, multiple PCAs were applied to individual varieties based on the results of the primary and second metabolites. Results indicated that primary metabolites failed to discriminate mandarin fruits by variety (Supporting Information Figure S1b), whereas second metabolites, specifically flavonoids, showed potential to group mandarins by variety (Supporting Information Figure S1c). This is consistent with previous studies that demonstrated the sugar, organic acid, and amino acid contents of mandarin fruits varied greatly depending upon growth
Figure 1. (a) PLS-DA score plot of mandarins. The result of the PLS-DA clearly distinguished all mandarin varieties. (b) PLS-DA loading plot of mandarins. Limonene was demonstrated to be an essential compound for distinguishing four mandarin varieties.
was validated by a 999x permutation test (y-intercepts of 0.23 to −0.122 for Q2 value). The result of the PLS-DA clearly distinguished all mandarin varieties. Meanwhile, in the loading plot of the PLS-DA (Figure 1b), limonene was demonstrated to be an essential compound for distinguishing four mandarin varieties. On the basis of the PLS-DA results, significant metabolites were identified by considering the variable importance for projection (VIP). The metabolites with a VIP value over 1.0 were considered as potential markers for distinguishing different mandarin varieties. Meanwhile, one-way ANOVA (p < 0.05) was performed to estimate significant differences of metabolite contents among mandarin cultivars (file 1 in Supporting Information). Combining the results of VIP values and ANOVA, 18 metabolite markers were selected (Table 1). Pathway enrichment analysis using these markers (Table 1) indicated possible biological pathways related to these metabolites (Table 2). The metabolic pathway, biosynthesis of terpenoids and steroids (BTS) (Figure 2a), showed high matched/total percentages with low p-values (
