Metabolic profiling revealed the organ‐specific distribution differences of tannins and flavonols in pecan

Abstract Carya illinoinensis is rich in phenolic metabolites such as tannins and flavonols, but both the composition and the distribution of these nutritional constituents in most pecan organs were still unclear. In this experiment, a comprehensive qualification and quantification of phenolic metabolites in eight organs of pecan were conducted for the first time. Ninety‐seven phenolic metabolites were identified, in which twelve were identified for the first time in pecan, including a series of ellagitannins with high molecular weight. Hydrolysable tannin was the dominant kind of phenolic metabolites in pecan. The metabolic profiles of tannins in pecan were extended. Thirty‐three phenolic metabolites were quantified, among them the highest content was ellagic acid pentose in testa. From this experiment, we can see that the distribution of phenolic metabolites in pecan was organ‐specific, tannins tend to accumulate in pecan testa with both diverse structures and high contents, while flavonols tend to accumulate in organs such as branch, bark, or leaf. Among all organs, testa contained the highest content of phenolics, which might play important roles in protecting pecan kernel from diseases and insects. A massive phenolic metabolites' matrix in different pecan organs was built in this experiment, which should be useful for related researches in the future and help provide a theoretical basis for using these organs as functional foods.

The most important part of pecan is by no means the edible part of the fruit-the kernel. Pecan fruit consists of pericarp and the seed (nut). The seed includes the inner flesh kernel and the brown pellicle wrapped around the flesh which was called testa. The epicarp and mesocarp were undivided in pecan and together they formed the shuck, while the shell is the endocarp. Besides kernel, other organs of pecan trees also have different values and can be used. Pecan shell is the by-product of food industry. Pecan shells aqueous extract can protect mice from oxidative damage induced by cigarette smoke exposure and reduced the locomotor activity and anxiety symptoms induced by smoking withdrawal (Reckziegel et al., 2011), it also has hepatoprotective activity against ethanol-induced liver damage (Müller et al., 2013). The oxidative properties of margarines supplemented with pecan nutshell extracts, rosemary extract, and butylated hydroxytoluene (BHT) were investigated, while pecan nutshell extracts had equally effects with the other two antioxidants and may be considered as a natural product replacement for the synthetic antioxidant BHT (Engler Ribeiro, de Britto Policarpi, Dal Bo, Barbetta, & Block, 2017). The antiproliferative and antitumor activity of pecan shells and their relationship with phenolics were also investigated (de la Rosa et al., 2014;Hilbig, Policarpi, et al., 2018). Chopped pecan shells were used for making tea in Brazil and were thought to have the diuretic and digestive effects (Prado et al., 2014;Engler Ribeiro et al., 2017). Flavonols in pecan bark and leaf were reported to have antidiabetic and hepatoprotective actives (Abdallah, Salama, Abd-elrahman, & El-Maraghy, 2011;Gad, Ayoub, & Al-Azizi, 2007). Many pecan organs besides kernel have the potential of being used in food or health food industry. The bioactives of phenolics in different pecan organs had attracted interests and had been investigated, but the composition and the distribution of these phenolic constituents in most pecan organs were still unclear.
Similar to walnut, pecan also belongs to waste-heavy materials for that about 70% of the fruit weights are shells and shucks (Han et al., 2018). Harvesting pecan nuts produced a lot of shucks, while cracking pecan nuts produced huge amount of shells. Pecan trees tend to produce excessive male inflorescence, which easily led to the over consume of tree nutrition. Remove over-bearing male inflorescence can protect the tree and will produce by-products.
Pruning and grafting will also produce by-products such as leaves, branches, and sometimes barks. Active phenolics were contained in these by-products, but their compositions were not well understood. A comprehensive metabolic investigation of phenolics in various organs of pecan is needed for better utilization of each part. On the other side, the distribution of secondary metabolites in plants is organ-specific. They usually share the same upstream metabolic pathway. And then, due to the differences in enzymes activity or type of each organ, these metabolites tend to have different synthetic and catabolic rates or have different chemical modification reactions such as methylation or glycosylation, which led to diverse structures to play diverse functions. So, a comprehensive metabolic investigation of phenolics in different pecan organs is also needed for better understanding of the organ-specific distribution of these active metabolites in pecan.
By far, the phenolics in pecan were mostly studied in kernels (Gong & Pegg, 2017;Jia et al., 2018;Robbins et al., 2015;Robbins, Ma, Wells, Greenspan, & Pegg, 2014;de la Rosa et al., 2011), while studies in shell (Hilbig, Alves, et al., 2018), leaf (Gad et al., 2007;Ishak, Ahmed, Abd-Alla, & Saleh, 1980;Lei et al., 2018), and bark (Abdallah et al., 2011) were much less. There is no report of the rest of the organs. The ultra-high-performance liquid chromatography coupled with hybrid linear ion trap and Orbitrap mass spectrometer (UHPLC-LTQ-Orbitrap MS n ) is the newest mass technic which has many advantages such as high sensitives and accuracies, while using ultra-high-performance liquid chromatography coupled with triple quadrupole mass spectrometer (UHPLC-QQQ-MS n ) can get a better result of quantity. In this paper, we used these methods of both rapid identification and accuracy quantification to get a comprehensive survey of the distribution of phenolic metabolites in various pecan organs.

| Materials and chemicals
Eight different organs of pecan, including kernel (without testa), testa, shell, shuck, leaf, branch, bark, and flower (male inflorescence), were collected in October 2018 (except the flower were in May 2019) at the scientific orchards of the Institute of Botany, Jiangsu Province and Chinese Academy of Sciences. Samples were collected from healthy adult trees of cultivar "Pawnee." Six biological replicates were prepared for each organ, and each biological replicate contained samples from three trees. After transported back to the laboratory, the testa was peeled manually from the surface of kernels and was stored at −70°C, so as samples of other parts. Before use, all samples were powdered and homogenized in liquid nitrogen with mortars and pestles.

| Sample extraction
Samples were extracted according to the methods of Regueiro et al. (2014) and Robbins et al. (2014) with slight modifications. The whole experiment was carried out in a dark room illuminated by a red light. Pecan samples (500 mg) and 1 ml mixed solution of methanol/water/acetic acid (70/29.5/0.5, v/v/v) were placed together in a 10-ml centrifuge tube, ultrasonic extracted for 5 min on ice bath, and centrifuged at 8,000 g for 10 min at 4°C (Hettich, Andreas Hettich GmbH & Co. KG). The supernatants were transferred into clean centrifuge tubes. Then, the mixed solutions were added again into the tubes with the kernel residues and ultrasonic extracted again with the same method. The supernatants were combined, and 1 ml n-hexane was added, vortexed for 1 min, centrifuged at 8,000 g for 5 min at 4°C for the purpose of defat. This defat process was also repeated again. The methanol layers were collected, combined and stored at −20°C until further analysis.

| UHPLC-QQQ-MS n quantification
The quantification of phenolic metabolites was performed on the Waters Acquity UPLC system (Waters, Corp.) using the same column and chromatographic conditions as used in previous qualification.

| Total phenolic content
The total phenolic content was measured according to the methods of our previous report (Jia et al., 2018). Briefly, methanol extract (10 μl) was mixed with 2 ml of 7.5% (w/v) sodium carbonate and 2.5 ml of 10% (v/v) Folin-Ciocalteu regent, and put in water bath at 50°C for 15 min in the dark. The reaction solutions were cooled to room temperature, and then, the absorbance was measured at 760 nm using UV spectrophotometer (Shimadzu UV-2100, Shimadzu Corporation). Ellagic acid was used as standard reference, and the results were expressed as milligrams of ellagic acid equivalents (EAE) per gram of defatted kernel weight (mg EAE g −1 ). The bland solution was made by pure methanol and treated along with samples under the same protocol. All samples were measured in triplicates on three biological replicates.

| 2, 2-Diphenyl-1-picrylhydrazyl (DPPH) free radical scavenging assay
The antioxidant capacities were firstly measured by the DPPH free radical scavenging assay according to the methods of previous report (Jia et al., 2018). Briefly, 4 ml of DPPH radical solution (39.43 mg DPPH in 1 L methanol) was mixed with 10 μl of methanol extract and kept in dark for 30 min. Then, absorbance was measured at 515 nm with UV spectrophotometer. The bland solution was made by pure methanol and treated along with samples under the same protocol.
All samples were measured in triplicates on three biological replicates. Absorbance of blank was subtracted from each sample. Trolox was used as standard reference, and the results were expressed as μmol trolox equivalents (TE) per gram of defatted kernel weight (μmol TE g −1 ).
The mixture was kept under dark for at least 16 hr and diluted with ethanol to absorbance of 0.70 ± 0.05 at 734 nm with UV spectrophotometer before use. Then, 40 μl methanol extracts of samples were mixed with 2 ml ABTS˙+ solution, and the absorbance was measured at 734 nm after rested still for 6 min. The bland solutions were made by pure methanol and treated along with samples under the same protocol. All samples were measured in triplicates on three biological replicates. Absorbance of blank was subtracted from each sample. Trolox was used as standard reference, and the results were expressed as μmol trolox equivalents per gram of defatted kernel weight (μmol TE g −1 ).

| Method validation and statistical analysis
Commercial standard compounds were used to validate the linearity, precision, repeatability, and stability of the optimized method.
At least six appropriate concentrations of standard solution in duplicate were prepared to obtain the linearity. The LOD (limit of detection) and LOQ (limit of qualification) for each analyte were acquired at a S/N of 3 and 10, respectively. Six injections of mixed standard solutions, including the highest, middle, and lowest concentration of linearity, were used to assess the precision and reproducibility.
Six independent replicates were studied for each sample solution. The MS data were processed by the Xcalibur (version 4.1) software, and PCA was performed by the SPSS (version 18.0) software.

| Method validation
The liquid chromatographic method had been optimized, and the final parameters were chosen under the consideration of both separation resolution and analytical efficiency. Commercial standard compounds were used to verify the optimized analytical method (Table S1). Results showed that the method had good linearity over wide ranges with regression coefficients (r 2 ) over .99. The LOD and the LOQ were 0.0009 μg/ml and 0.0029 μg/ml for catechin, 0.0008 μg/ml and 0.0026 μg/ml for ellagic acid, 0.0008 μg/ml and 0.0028 μg/ml for quercetin, respectively. The precision, repeatability, and stability of the method were also studies using standard compounds (Table S2). The low relative standard deviations (RSDs) demonstrated that this method is precise, repeatable, and stable and can be used in the following experiment.

| Identification of phenolic metabolites in different pecan organs
A total of 97 phenolic metabolites have been identified from 8 pecan organs, including 57 hydrolysable tannins, 19 condensed tannins, 1 complex tannin, 18 flavonols, and 2 other compounds ( Figure 1a, Table 1). There were significant differences in the numbers of metabolites identified in different pecan organs (Table 1). The number of phenolic compounds identified from leaves was largest among all organs, followed by testa, while the least phenolic compounds were identified from shells. This result showed the difference in the complexity of phenolic compounds among different pecan organs, similar result can be seen in the principal component analysis (PCA) too ( Figure 1b). The samples of testa and leaf were far from the others, which showed the specificity of phenolic compounds in these organs. The samples of kernel, shell, branch, and bark were close to each other, which showed the similarity of composition between them. The phenolic compositions of kernel were quite different from those of testa. Of the 97 phenolic metabolites, 30 were found in all organs, 34 in only one organ, and 21 in only two organs (Table 1).
On the basis of pedunculagin/casuariin isomer (bis-HHDP-glucose, m/z 783), metabolites with one and two galloyl substituents were found at m/z 933 (peaks 17, 40, 48, and 54) and 1,085 (peaks 78 and 82) and were assigned as glansrin C and eucalbanin A/cornusiin B isomer. Glansrin C and eucalbanin A/cornusiin B isomer had been detected from walnut before (Regueiro et al., 2014), and this is the first report of these two tannins in pecan. be pterocarinin A. This compound had been found in walnut before (Regueiro et al., 2014), while this is the first report in pecan.

F I G U R E 5
Deduced metabolic pathway of part of the hydrolysable tannins in pecan organs. Different colors of metabolites represented different structure categories: brown, ellagitannins; blue, hydrolysable tannins mixed with both HHDP and galloyl groups; pink, complex tannin. K, kernel; T, testa; SL, shell; SK, shuck; BK, bark; BH, branch; L, leaf; F, flower (male inflorescence) ellagitannins were reported the first time in pecan. Through this study, the tannin profile and each route of tannin metabolic pathway of pecan had been expanded ( Figure 5), which laid a foundation for future researches and uses.

| Identification of condensed tannins and their distribution
A series of condensed tannins were identified in these organs.

| Identification of flavonols and their distribution
All flavonoid metabolites identified in this experiment were flavonols, and most of them were eluted in the later part of elutes. There were three types of flavonol aglycones, quercetin, azaleatin, and caryatin, which contained their identical fragment ions at m/z 301, 315, and 329 in the MS 2 spectrum, respectively. One tricky problem about flavonol identification in pecan is they have similar MW with ellagic acid, methyl ellagic acid, and dimethyl ellagic acid derivatives.
Luckily, the high-accuracy MW can solve these problems. arabinoside, diglucoside, and rutinoside (Ishak et al., 1980), and their structures were elucidated with NMR method, so this is the first report of azaleatin galactose in pecan, and this is also the first report of azaleatin glucoside in pecan bark.
The high-accuracy MW also helped us to identify peak 81 (exper- were assigned as quercetin galloyl hexoside instead of ellagic acid galloyl hexoside. A series of quercetin glycosides had been isolated from pecan leaves and elucidated with NMR method, including glucoside, rhamnoside, arabinoside, galactoside, and galloyl galactoside (Abdallah et al., 2011;Gad et al., 2007;Ishak et al., 1980).  (Ishak et al., 1980). Peak 86 had the same molecular formula but different structure, the fragment ion at m/z 301 suggested it contained a quercetin aglycone and a rhamnoside; therefore, it was assigned as quercetin rhamnoside, which had been reported in pecan leaves (Gad et al., 2007).

| Identification of other phenolic compounds and their distribution
The structure of peaks 28 and 41 ([M−H] − , 1,207) was much special, they were identified as complex tannins, which means a flavonolbased motif linked with ellagitannins through C-C bond (Okuda et al., 2009). By analysis of the fragmentation pattern, they were identified as stenophyllanin A/B isomer, which contained two HHDP, one galloyl and one epicatechin groups. This is the first report of this complex tannin in pecan. Brevifolin carboxylic acid (peak 39) and neochlorogenic acid (peak 16) were also assigned according to their high-accuracy MWs, fragmentation patterns and compared with literature reports (Regueiro et al., 2014). Neochlorogenic acid was found only in pecan leaf in this experiment, and this is the first report of this compound in pecan.

| Content
In order to get more accurate result, we use the UHPLC-QQQ-MS n equipment with MRM mode to carry out the quantitative analysis ( acid pentose (9.65 µg/g), and ellagic acid rhamnoside (4.80 µg/g), but ellagic acid pentose was one of the top five phenolics in all organs.
The total content of condensed tannins was much lower than that of hydrolysable tannins, their content was varying from 190.71 µg/g to 1.79 µg/g, highest content appeared in testa, while lowest in flower.
Combined with previous qualification results, we can see that tannins tend to accumulate in pecan testa with both diverse structures and high contents.
The distribution of flavonols was much different with tannins, they were the prevailing phenolics in branch (1,223.18 µg/g), bark (667.80 µg/g), and leaf (308.64 µg/g), but rare in testa (10.04 µg/g) and kernel (3.29 µg/g). This result consisted with previously reports while most flavonols were isolated and identified firstly from pecan branch, bark or leaves (Abdallah et al., 2011;Gad et al., 2007;Ishak et al., 1980). The highest content among flavonols appeared at 1,156.38 µg/g of caryatin in branch, followed by 665.17 µg/g of caryatin in bark and 250.46 µg/g of quercetin rhamnoside in leaf.
Combined with previous results, we can see that flavonols tend to accumulate in organs such as branch, bark, or leaf in pecan with both diverse structures and high contents.
The total content of all phenolics quantified with LCMS was highest in testa, so were the TPC and antioxidant capacities (Table 2), all these results indicated that pecan testa contained high concentration of phenolics. Such high concentration of phenolics plays important roles in protecting pecan kernel from invasion of diseases and insects. The total content of phenolics quantified with LCMS was lowest in shuck, so was the TPC. The TPC and antioxidant capacities of pecan shell were also relatively high, which is consistent with previous results. At present, the research on antioxidant capacities of pecan is limited to kernel and shell (Biomhoff et al., 2006;de la Rosa et al., 2011de la Rosa et al., , 2014Prado et al., 2013Prado et al., , 2014Flores-Cordova et al., 2017;Hilbig, Alves, et al., 2018;Jia et al., 2018;Robbins et al., 2015;Villarreal-Lozoya et al., 2007;Wu et al., 2004). Although there were some differences among different research results due to the differences of producing areas, varieties, and extraction methods, the antioxidant capacities of shell were higher than that of kernel in all reports, about 5-7 times of that of kernel. In this experiment, we separated the testa from kernel, which provided a more accurate result of different pecan organs.

| CON CLUS IONS
The composition and distribution of phenolic metabolites in eight different pecan organs were analyzed for the first time. A rapid qualitative method of LTQ-Orbitrap MS and an accurate quantita- be discriminated more clearly and accurately. Because the previous researches were mainly focused on pecan kernels, so this is the first report of many phenolic metabolites in other organs. The contents of thirty-three phenolics were determined under MRM mode.
Combined with the previous qualification, the compositions of phenolic metabolites in different pecan organs were more clearly.
A massive phenolic metabolites' matrix in different pecan organs was built in this experiment, which should be useful for related researches in the future and help provide a theoretical basis for using these organs as functional foods.