Analysis of 2-(2-Phenylethyl)chromones by UPLC-ESI-QTOF-MS and Multivariate Statistical Methods in Wild and Cultivated Agarwood

Agarwood is the fragrant resinous material mainly formed from species of Aquilaria. 2-(2-phenylethyl)chromones, especially the highly oxidized 5,6,7,8-tetrahydro-2-(2-phenylethyl)chromones, are the main representative compounds from agarwood. It is important to determine whether agarwood in trade is from cultivated trees or natural trees in the Convention on the International Trade in Endangered Species (CITES). We characterized the 2-(2-phenylethyl)chromones in agarwood by ultra-performance liquid chromatography coupled with electrospray ionization mass spectrometry (UPLC–ESI-QTOF-MS) and differentiated wild from cultivated agarwood by metabolomic analysis. A total of 141 chromones including 50 potentially new compounds were evaluated as belonging to four structural classes (unoxidized 2-(2-phenylethyl)chromones, 5,6,7,8-tetrahydro-2-(2-phenylethyl)-chromones, bi-2-(2-phenylethyl)chromones, and tri-2-(2-phenylethyl)chromones). The metabolic difference between wild and cultivated agarwood was analyzed by component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA). Fourteen markers of metabolisms in wild and cultivated agarwood were constructed (e.g., 6,7-dimethoxy-2-(2-phenylethyl)chromone, 6,8-dihydroxy-2-(2-phenylethyl)chromone, 6-methoxy-2-(2-phenylethyl)chromone, etc.). These results indicated that UPLC–ESI-QTOF-MS-based metabonomics analysis in agarwood may be useful for distinguishing wild agarwood from cultivated agarwood.


Introduction
Agarwood, also known as aloeswood or eaglewood in several regions, chen xiang (China), jinkoh or kanankoh (Japan), gaharu or kalabak (Malaysia and Indonesia), krissna (Thailand and Lao), agar (India), and oud (Middle East), is a fragrant resinous heartwood obtained from certain trees belonging to Malvales Thymelaeaceae [1,2]. It is widely accepted that agarwood was created as a response of the tree to various forms of injury, including natural injuries and artificial injuries [3][4][5]. The supply of agarwood (wild sources) is far lower than the market demand because only a few source trees in nature can actually produce agarwood, and its production is slow. Due to the exploitation of this tree, all Aquilaria spp. were listed on Appendix II of the Convention on the International Trade in Endangered Species (CITES) in 2004. Nine Aquilaria species-A. crassna (critically endangered), A. rostrata (threatened), A. banaensae, A. beccariana, A. cimingiana, A. hirta, A. malaccensis, A. microcarpa, and A. sinensis (vulnerable)-were listed on the International Union for Conservation of Nature (IUCN)

Peak
RT (
Peaks 20, 27, 28, 32, 35, and 42, showed a similar molecular ion at m/z 267. Peaks 20, 27, and 35 were identified as qinanone D isomers [13], all of which are substituted by hydroxyl in the benzyl moieties, although it was not possible to distinguish the position of hydroxyl because they showed the same fragmentation pattern. Peaks 28, 32, and 42 were identified as 7-hydroxy-2-(2-phenylethyl) chromone isomers, all of which are substituted by hydroxyl in the chromone moieties [15].
Peaks 20, 27, 28, 32, 35, and 42, showed a similar molecular ion at m/z 267. Peaks 20, 27, and 35 were identified as qinanone D isomers [13], all of which are substituted by hydroxyl in the benzyl moieties, although it was not possible to distinguish the position of hydroxyl because they showed the same fragmentation pattern. Peaks 28, 32, and 42 were identified as 7-hydroxy-2-(2-phenylethyl)chromone isomers, all of which are substituted by hydroxyl in the chromone moieties [15].
The molecular formula C18H19ClO6 was assigned to compound 83 and 86. These compounds were tentatively identified as 8-chloro-5,6,7-trihydroxy-2-(4-methoxyphenethyl)-5,6,7,8-tetrahydrochromene and its isomers [8], one of which was found in the extracts of agarwood for the first time. Peak 66 was identified as agarotetrol [19]. The MS/MS of this compound displayed the fragment ions at m/z 301, 283, 255, and 227, in accordance with the successive losses of waters (301, 283) as well as subsequent two CO (255,227). Fragment ion at m/z 164 corresponded to the loss of C 7 H 7 (91) from the fragment ion of 255. Additionally, another agarotetrol isomer was detected in the extract (compound 73) that matched with aquilarone B [19].

Multivariate PCA and OPLS-DA Analysis of UPLC-MS Data
The unsupervised PCA score plot explained 71.0% of the total variance (R2). The agarwood samples were separated into a wild group (w) and two cultivated groups (a,c) ( Figure 6). Next, all data sets of the wild and cultivated agarwood metabolite profiles were analyzed by supervised multivariate statistics and orthogonal partial least squared discriminant analysis (OPLS-DA). The S-plot easily visualized the variables that changed most significantly at the top or the bottom of the plot (Figure 7). The variables (peaks 13, 41, 42, 48, 50, 57, 58, 67, 69, 71, and 94) on the top showed the most upregulated metabolites in wild agarwood, and the ones (peaks 110, 113, and 123) on the bottom, the most downregulated. Among them, 14 metabolites were selected as biomarkers from the variables important in the projection (VIP) >1.0 in the statistical analysis. The variables were putatively assigned and presented in Table 3, which can serve as biomarkers between wild and cultivated agarwood. The main compounds were 6,7-dimethoxy-2-(2-phenylethyl)chromone, 6,8-dihydroxy-2-(2-phenylethyl)chromone, 6-methoxy-2-(2-phenylethyl)chromone, 6-methoxy-2-[2-(4-methoxyphenyl)ethyl]chromone, and others, as much different for the two types of agarwood samples. These biomarkers could significantly discriminate differences between groups. The results thus obtained are reliable and can be used for, for example, metabolic or biosynthetic pathway analysis. putatively assigned and presented in Table 3, which can serve as biomarkers between wild and cultivated agarwood. The main compounds were 6,7-dimethoxy-2-(2-phenylethyl)chromone, 6,8-dihydroxy-2-(2-phenylethyl)chromone, 6-methoxy-2-(2-phenylethyl)chromone, 6-methoxy-2-[2-(4-methoxyphenyl) ethyl]chromone, and others, as much different for the two types of agarwood samples. These biomarkers could significantly discriminate differences between groups. The results thus obtained are reliable and can be used for, for example, metabolic or biosynthetic pathway analysis.

Chemicals and Materials
Acetonitrile (Merck HPLC grade, Darmstadt, Germany) and formic acid (Sigma-Aldrich, MS Grade, St. Louis, MO, USA) were used. Ultrapure water was deionized and purified by the Milli-Q purification system (Millipore, Bedford, MA, USA). Other analytical grade reagents and chemicals were purchased from Guangzhou Chemical Reagent Factory (Guangzhou, China).

Plant Material
Twenty-one agarwood samples were selected and analyzed. All plantation-cultivated (via the whole-tree agarwood-Inducing technique (Agar-Wit, Zhongshan, China) and others) and wild-type agarwood (A. sinensis) was collected from Zhongshan, Chaozhou (   were used as follows: ESI voltage, 4500 V; nebulizer gas, 55; auxiliary gas, 55; curtain gas, 35; turbo gas temperature, 500˝C; declustering potential, 100 V; collision energy, 10 eV. The samples were acquired with an IDA (Information-Dependent Acquisition) method, which can automatically select candidate ions for further product ion study. The TOF mass scan range was operated from m/z 100 to 2000, and the product ion scan range was m/z 50 to 2000. The collision energy (CE) was set from 45 eV. The MS calibration was set using the Calibrant Delivery System (CDS) by direct injection at a flow rate of 300 µL¨min´1.

UPLC-MS Data Processing and Multivariate Analysis
Peak detection, alignment, and identification were processed using PeakView v2.0 and MarkerView v1.2.1 software (Applied Biosystems/MDS Sciex, Foster City, CA, USA). Statistical analyses with the 1066ˆ21 matrix were carried out by SIMCA-P (version 13.0, Umetrics, Umea, Sweden). Unsupervised principal component analysis (PCA) was used for obtaining similarities or latent differences between groups. Orthogonal projection to latent structure discriminant analysis (OPLS-DA) was then implemented to detect maximum information from the data set and to distinguish the metabolites induced by different groups. Potential biomarkers were identified by analyzing the VIP and S-plot.

Conclusions
The UPLC-ESI/QTOF/MS-based metabolite-profiling approach enabled the tentative identification of 141 metabolites in agarwood extract on the basis of their accurate masses and MS/MS spectra in positive ion mode together with the reported data. The method applied combined the excellent separation effect of a small-particle-size C18-column (1.9 µm) and an IDA method; as such, the high resolution enabled the separation of a great number of isomers, with the high sensitivity, mass accuracy, and detections of the isotopic pattern combined with QTOF/MS for both molecular and fragment ions. We summarized the MS characterization of 2-(2-phenylethyl)chromones and identified these compounds for distinguishing wild from cultivated eaglewood. However, due to the fact that most of the compounds have the same molecular formula and similar fragment ions, some chromone isomers cannot be distinguished. This is the first time that a detailed study of these chromones has been carried out by UPLC-QTOF-MS analysis. At the same time, 2-(2-phenylethyl)chromones were used to distinguish wild from cultivated agarwood; 14 2-(2-phenylethyl)chromones were selected as biomarkers. The cultivars can be distinguished from the wild by their pattern. The present result was in line with some other findings that the ions of 267, 281, 283, 311, 313, 341, and 349 may be able to be differentiated into cultivated or wild [14,25] agarwood. It is interesting to note that there were 44 potentially new bi-2-(2-phenylethyl)chromones that had not been detected in previous work, and some of them can serve as biomarkers between wild and cultivated agarwood. The proposition for two types of agarwood needs to be further confirmed with a broader set of samples in future metabolic studies.