ARTICLE pubs.acs.org/JAFC
Chemical and Genetic Assessment of Variability in Commercial Radix Astragali (Astragalus spp.) by Ion Trap LC-MS and Nuclear Ribosomal DNA Barcoding Sequence Analyses Wei-Lie Xiao,†,‡ Timothy J. Motley,§ Uchenna J. Unachukwu,‡ Clara Bik San Lau,|| Bei Jiang, ^ Feng Hong,† Ping-Chung Leung,|| Qing-Feng Wang,§,# Philip O. Livingston,† Barrie R. Cassileth,† and Edward J. Kennelly*,‡ †
Integrative Medicine Service, Memorial Sloan-Kettering Cancer Center, New York, New York, USA Department of Biological Sciences, Old Dominion University, Norfolk, Virginia, USA Institute of Chinese Medicine, the Chinese University of Hong Kong, Hong Kong, China ^ College of Pharmacy, Dali University, Snowman Road, Xiaguan, Dali 671000, Yunnan, P. R. China ‡ Lehman College and the Graduate Center, the City University of New York, Bronx, New York, USA # Wuhan Botanical Institute, the Chinese Academy of Sciences, Wuhan 430072, P. R. China
)
§
ABSTRACT: Radix Astragali (Huangqi) has been demonstrated to have a wide range of immunopotentiating effects and has been used as an adjuvant medicine during cancer therapy. Identity issues in the collection of Radix Astragali exist because many sympatric species of Astragalus occur in the northern regions of China. In order to assess the quality, purity, and uniformity of commercial Radix Astragali, 44 samples were purchased from herbal stores in Hong Kong and New York City. The main constituents, including four isoflavonoids and three saponins, were quantitatively determined by liquid chromatography mass spectrometry (LC-MS). There was significant sample-to-sample variability in the amounts of the saponins and isoflavonoids measured. Furthermore, DNA barcoding utilizing the variable nuclear ITS spacer regions of the 44 purchased Radix Astragali samples were sequenced, aligned and compared. Eight polymorphic point mutations were identified which separated the Radix Astragali samples into three groups. These results indicate that the chemical and genetic variability that exists among Radix Astragali medicinal products is still a consistency and quality issue for this herbal. Two-way ANOVA analysis showed significant effects on the contents of the seven tested compounds when both phylogenetic and geographic (i.e., point of purchase) factors were considered. Therefore, chemical profiles determined by LC-MS and DNA profiles in ITS spacer domains could serve as barcode markers for quality control of Radix Astragali. KEYWORDS: Radix Astragali, Astragalus, DNA barcoding, LC-MS
’ INTRODUCTION Radix Astragali, known as Huangqi in China, is prepared from the dried roots of Astragalus membranaceus (Fisch) Bunge and Astragalus mongholicus Bunge. It is one of the most commonly used traditional Chinese herbal drugs, and was reported to have immunostimulant, hepatoprotective, antidiabetic, diuretic, and sedative effects.1 In spite of the recent taxonomic revision combining several species into A. mongholicus, many local collectors and botanists still recognize A. membranaceus and A. mongholicus as separate species.2 The two species traditionally have been separated by the presence of more dense hairs on the latter species and by differences in their ovaries and pods. Astragalus membranaceus is the more commonly used species for Radix Astragali. Both species grow in the northeast, north, and northwest of China as well as in Mongolia, and A. membranaceus extends into Korea.2 The major bioactive compounds found in Radix Astragali are isoflavonoids, saponins and polysaccharides, which have various biological activities.3 Recent research on pharmacological properties and clinical applications has demonstrated that Radix Astragali has a wide range of immunopotentiating effects and could be used as an adjuvant medicine during cancer therapy.4-6 r 2011 American Chemical Society
These findings have increased the interest and demand for Radix Astragali. Nowadays, various Radix Astragali preparations are commercially available not only in China as a TCM component but also in the United States as a dietary supplement. A popular product, the raw dried root is commonly used because of its relatively low price. Many analytical methods including HPLC and LC-MS have been used for quality control of Radix Astragali crude drugs, monitoring the levels of selected active isoflavonoids or saponins.7-12 In addition, Radix Astragali is graded by its root appearance in the market; the longer and thicker roots are regarded as better quality.13 Previous studies have focused on Astragalus samples collected from local herbal stores or collected directly from the field7,10,11,14-17 or crude drugs (e.g., oral solutions, injections, concentrated granules and tablets).8,9,18,19 In these cases, the sampling was limited, and more intensive sampling is required to get a clearer picture about product consistency and quality of Radix Astragali products. Received: July 26, 2010 Revised: January 21, 2011 Accepted: January 25, 2011 Published: February 15, 2011 1548
dx.doi.org/10.1021/jf1028174 | J. Agric. Food Chem. 2011, 59, 1548–1556
Journal of Agricultural and Food Chemistry In order to compare the quality of Radix Astragali in herbal markets, we have developed a validated LC-MS method through simultaneous determination of isoflavonoids and saponins, including calycosin-7-O-β-D-glucoside, ononin, calycosin, formononetin, and astragalosides I, II, and IV; 44 commercial raw roots from herbal markets in New York City and Hong Kong were analyzed. DNA analysis was used to validate sample identity and determine genetic variation in botanicals.7,20-24 Similar genetic testing has already proven to be useful for studies of species used in Radix Astragali.7,21,24 We conducted a genetic study to identify a stable marker system to further compare variation of samples found in the Chinese and American markets and determine if genetic variation is congruent with the chemical variation occurring among samples.
’ MATERIALS AND METHODS Chemicals and Reagents. HPLC grade acetonitrile was from J. T. Baker (Phillipsburg, NJ, USA); HPLC grade methanol was from E. Merck (Darmstadt, Germany); distilled water was further purified by Milli-Q system (Millipore, Milford, MA, USA). Sephadex LH-20 (25100 μm) was manufactured by Pharmacia Fine Chemicals (Piscataway, NJ, USA), and reversed-phase C18 silica gel (40 μm) was obtained from J. T. Baker (Phillipsburg, NJ, USA). Polyvinylidene difluoride (PVDF) syringe filters with a pore size of 0.45 μm were purchased from National Scientific Co. (Duluth, GA, USA). Total genomic DNA was extracted from dried roots using the DNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA) and β-mercaptoethanol molecular grade (EMD Biosciences). PCR reaction mixes were 8 μL of ddH2O, 1 μL of DNA, 12.5 μL of GoTaq Green Master Mix (Promega, Madison, WI, USA), 0.25 μL of BSA (EMD Biosciences), 1.25 μL of DMSO (EMD Biosciences), and 1.0 μL of each 20 μM primer (Integrated DNA Technologies). PCR products were then purified and cleaned following the manufacturer's protocol with a QIAquick PCR purification Kit (Qiagen), and automated sequencing was accomplished with Big Dye chemistry (Applied Biosystems Inc. Foster City, CA, USA). Radix Astragali Products. The Radix Astragali samples (dried roots) were purchased in 2009 from herbal stores in the New York City and Hong Kong City areas. There are 44 products in total, among them 22 from New York and another 22 from Hong Kong. The samples were identified T.J. Motley on the basis of morphology and a DNA sequence blast search comparison against the NCBI GenBank DNA database. The blast search of the 44 samples was found to have 99-100% comparable genetic identities to available published A. membranaceus sequences (N = 7). The next most similar sequence was A. propinqus (n = 1) with 98% genetic similarity. This verifies that all purchased samples were in the Radix Austragali species complex, and likely represent collections of A. membranaceus. The root specimens were vouchered and are deposited under A. membranaceus in the ODU herbarium; the New York samples (NY 1-22) were deposited under T. J. Motley collection #3237, and Hong Kong samples (HK 1-22) were deposited under T. J. Motley collection #3238. All samples were dried by lyophilization prior to the chemical and DNA analysis. Standards and Controls. Calycosin-7-O-β-D-glucoside (1) was isolated from one of Radix Astragali samples by the following procedures. Dried Radix Astragali samples (5.0 g) were extracted with 95% aqueous ethanol at room temperature (3 100 mL). After the EtOH was removed in vacuo, the residue (1.52 g) was separated over reversedphase C18 eluting with MeOH-H2O (1:4, 2:3, 1:1, 3:1, and 0:1) to give five fractions (I-V). Fraction II (250.0 mg) was further separated by Sephadex LH-20 eluting with methanol to give four subfractions (AD). Calycosin-7-O-β-D-glucoside (1) (10.2 mg) was obtained from subfraction B by recrystallization. The structure of 1 was determined by 1D and 2D NMR spectra, and the purity was more than 98% as determined
ARTICLE
Figure 1. Chemical structures of seven standards: 1, calycosin-7-O-β-Dglucoside; 2, ononin; 3, calycosin; 4, astragaloside IV; 5, astragaloside II; 6, formononetin; 7, astragaloside I. by ion trap LC-MS. Ononin (2) (purity >98.0%) and calycosin (3) (purity >98.0%) were purchased from ChromaDex (Irvine, CA, USA). Astragaloside IV (4) (purity >98.0%), astragaloside II (5) (purity >98.0%), formononetin (6) (purity >98.0%), and astragaloside I (7) (purity >98.0%) were obtained from Zhongxin Innova Laboratories (Tianjing, China) (Figure 1). Instrumentation. 1H NMR (300 MHz) and 13C NMR (75 MHz) spectra were recorded using a Bruker AVANCE 300 MHz NMR spectrometer. The NMR spectra were obtained in CDCl3, with chemical shifts expressed in δ and coupling constants (J) in hertz. 2D NMR (HSQC, HMBC, 1H,1H COSY, and ROESY) spectra were also obtained from a Bruker AVANCE 300 MHz NMR spectrometer. LC-MS was performed on a Thermo Finnigan LCQ mass spectrometer (San Jose, CA, USA) in the positive mode with a Waters 2690 separations module. The instrument was equipped with an atmospheric pressure chemical ionization (APCI) source and controlled by Xcalibur software. The discharge current was set to 5 μA. The vaporizer and capillary temperatures were set to 450 and 200 C, respectively. Nitrogen was used as the sheath gas and auxiliary gas at flow rates of 80 and 10 units, respectively. A mass range of 200-2000 amu was scanned. Separations were carried out on a Waters 2690 HPLC equipped with a Phenomenex Hydro C18 column (4.6 250 mm, 5 mm) at ambient temperature. The mobile phase consisted of water (A) and MeCN (B) with a flow rate of 1 mL/min. The mobile phase composition began with 0% B, followed by a linear increase to 10% B in 5 min, 10%-55% B in 30 min, and 55%-100% B in 10 min. Tissue disruption for DNA extraction was accomplished using a Mini Beadbeater 96 (GlenMills Inc., Clifton, NJ, USA) with zirconia/silica 1 mm glass beads. PCR was accomplished with GenAmp 2720 thermal cycler (Applied Biosystems Inc.), amplification products visualized on DigiDoc-it imaging system (UVP, Upland, CA, USA) and dried down for sequencing in a Savant DNA 120 vacuum concentrator (Thermo Fisher Scientific, Waltham, MA, USA). Automated sequencing was done using a 3730XL capillary DNA sequencer (Applied Biosystems Inc.). Preparation of Standards for LC-MS Analysis. Stock solutions (1 mg/mL) for the seven standards were prepared by dissolving individual standards in HPLC grade MeOH. Working standard solutions containing each of the 7 standards were prepared by diluting the stock solutions with methanol to a series of proper concentrations. The solutions were brought to room temperature and an aliquot of 10 μL was injected into LC-MS for analysis. Sample Preparation for LC-MS Analysis. Samples of Radix Astragali (0.5 g) were ground and extracted with 95% EtOH (15 mL) by ultrasonication for 30 min under ambient temperature, and centrifuged at 3000 rpm for 10 min. After the supernatant was decanted, the residue was further extracted according to the same method two more times. The combined solutions were evaporated to dryness under vacuum at 25-30 C. The residue was reconstituted with 10 mL of MeOH in a 1549
dx.doi.org/10.1021/jf1028174 |J. Agric. Food Chem. 2011, 59, 1548–1556
Journal of Agricultural and Food Chemistry
ARTICLE
Figure 2. Representative total ion currency (TIC) chromatograms (NY 2 and 20 from group 1, NY 21 and HK 6 from group 2, and NY 7 and 11, and HK 13 from group 3) and MS spectra of the labeled seven peaks from TIC chromatogram of sample NY 2 detected by positive APCI-MS. volumetric flask and filtered through a 0.45 μm membrane before injection. The sample 10 μL was injected to LC-MS for the determination of isoflavonoids and astragalosides. Validation of Analytical Method. The sample NY 2 was selected randomly for the analytical method development. This LC-MS method was used for quantitative analysis and validated with respect to linearity, recovery, and sensitivity. Calibration curves were obtained by plotting the peak area versus the concentration of the standard. Each calibration curve was established on five data points covering the concentrations of 1.00-100.0 μg/mL for standards 1, 6 and 7, 0.50-50.0 μg/mL for standards 2, 3 and 4, 1.00-50.0 μg/mL for standard 5, respectively. Limits of detection (LOD) and limits of quantification (LOQ) were calculated as signal-to-noise ratios with minimal values of 3:1 and 10:1, respectively. Intraday variations were determined with three injections of a standard mixture solution over a 1 day period. Interday variations were evaluated by performing three injections of standard mixture solutions for
three consecutive days. The percentage relative standard deviation (RSD, %) of the retention time (Rt) and peak area (pA) were taken as the measures of precision. Recovery test was used to evaluate accuracy of this method. The recovery for the extraction of the Radix Astragali samples was achieved by adding known amounts of the standards 1-7 to the ground roots of Radix Astragali (0.25 g) prior to extraction. The recovery was determined by subtracting the values obtained for the control matrix preparation from the sample added with standards, divided by the amount of standards, and expressed as percentages. DNA Extraction. Total genomic DNA was extracted from dried root samples of Radix Astragali following the manufacturer’s protocol for the DNeasy Plant Mini Prep. Tissues were lysed for 15 s prior to extraction. DNA Amplification. DNA was amplified using the polymerase chain reaction (PCR).25 PCR reactions were performed using 8 μL of ddH2O, 1 μL of DNA, 12.5 μL of GoTaq Green Master Mix, 0.25 μL of BSA, 1.25 μL of DMSO, 1.0 μL of each 20 μM primer, and 1 μL of 1550
dx.doi.org/10.1021/jf1028174 |J. Agric. Food Chem. 2011, 59, 1548–1556
Journal of Agricultural and Food Chemistry
ARTICLE
Table 1. Concentration of Seven Well-Characterized Standards (mg/g) in the Different Radix Astragali Samples (n = 3) standards samples 1 (calycosin-7-O-β-D-glucoside)
2 (ononin)
3 (calycosin)
4 (astragaloside IV) 5 (astragaloside II) 6 (formononetin) 7 (astragaloside I)
Group 1a NY 1 NY 2 NY 8 NY 13 NY 14 NY 20 HK 1
0.169 ( 0.002 0.575 ( 0.006 0.449 ( 0.003 1.030 ( 0.016 0.607 ( 0.013 0.201 ( 0.001 0.114 ( 0.004d
0.061 ( 0.002 0.185 ( 0.004 0.133 ( 0.006 0.238 ( 0.001 0.211 ( 0.001 0.058 ( 0.001 0.018 ( 0.001d
0.107 ( 0.003 0.218 ( 0.001 0.043 ( 0.001 0.232 ( 0.002 0.031 ( 0.0004 0.030 ( 0.001 0.066 ( 0.001
NY 5 NY 18 NY 21 HK 6 HK 8 HK 15 HK 20
0.925 ( 0.009 0.282 ( 0.008 0.218 ( 0.009 1.595 ( 0.033b 0.381 ( 0.011 0.456 ( 0.004 0.179 ( 0.003
0.333 ( 0.018 0.110 ( 0.001 0.133 ( 0.004 0.693 ( 0.007b 0.211 ( 0.002 0.156 ( 0.001 0.115 ( 0.002
0.042 ( 0.002 0.017 ( 0.001 0.011 ( 0.001d 0.096 ( 0.001 0.068 ( 0.002 0.018 ( 0.001 0.079 ( 0.003
0.056 ( 0.002 0.052 ( 0.001 NDc 0.047 ( 0.001 ND 0.141 ( 0.001b ND
0.509 ( 0.004 0.165 ( 0 0.002 0.081 ( 0.0002 0.050 ( 0.002 0.391 ( 0.018 0.121 ( 0.0002 0.023 ( 0.001d
0.667 ( 0.005 0.092 ( 0.001 0.136 ( 0.003 0.164 ( 0.001 0.109 ( 0.004 0.045 ( 0.001d 0.148 ( 0.002
1.699 ( 0.006b 0.345 ( 0.010 0.246 ( 0.009 0.192 ( 0.001 0.486 ( 0.006 0.316 ( 0.001 0.138 ( 0.002
0.667 ( 0.009b 0.098 ( 0.003 0.031 ( 0.001 0.069 ( 0.002 0.043 ( 0.002 0.049 ( 0.004 0.050 ( 0..001
0.257 ( 0.001 0.239 ( 0.001 0.128 ( 0.001 0.487 ( 0.002 0.431 ( 0.008 0.154 ( 0.003 0.605 ( 0.004
0.804 ( 0.009 0.264 ( 0.011 0.134 ( 0.001 0.260 ( 0.008 0.233 ( 0.001 0.129 ( 0.005 0.304 ( 0.002
0.080 ( 0.001 0.074 ( 0.003 0.056 ( 0.001 0.068 ( 0.001 0.066 ( 0.003 0.090 ( 0.001 0.166 ( 0.004 0.052 ( 0.001 0.062 ( 0.001 0.072 ( 0.001 0.051 ( 0.001 0.053 ( 0.003 0.032 ( 0.001 0.040 ( 0.0004 0.040 ( 0.002 0.049 ( 0.001 0.049 ( 0.001 0.048 ( 0.001 0.040 ( 0.002 0.066 ( 0.001 0.194 ( 0.004 0.095 ( 0.003 0.137 ( 0.003 0.049 ( 0.001 0.165 ( 0.001 0.080 ( 0.001 0.045 ( 0.001 0.041 ( 0.001 0.051 ( 0.0002 0.045 ( 0.001
0.114 ( 0.002 0.051 ( 0.001 0.100 ( 0.006 0.193 ( 0.005 0.146 ( 0.001 0.211 ( 0.008 0.131 ( 0.001 0.051 ( 0.003 0.128 ( 0.002 0.361 ( 0.001 0.104 ( 0.003 0.275 ( 0.001 0.054 ( 0.001 0.420 ( 0.004 0.611 ( 0.010 1.211 ( 0.012 0.714 ( 0.007 1.023 ( 0.006 1.122 ( 0.001 0.943 ( 0.002 0.439 ( 0.009 0.705 ( 0.002 0.379 ( 0.002 0.186 ( 0.002 0.527 ( 0.001 0.489 ( 0.003 2.284 ( 0.090b 1.888 ( 0.034 0.381 ( 0.011 0.286 ( 0.003
0.350 ( 0.001 0.244 ( 0.001 0.166 ( 0.002 0.138 ( 0.001 0.205 ( 0.005 0.393 ( 0.007 0.571 ( 0.009 0.105 ( 0.007 0.226 ( 0.002 0.177 ( 0.002 0.170 ( 0.001 0.191 ( 0.003 0.089 ( 0.004d 0.178 ( 0.001 0.188 ( 0.007 0.120 ( 0.001 0.105 ( 0.004 0.310 ( 0.001 0.138 ( 0.003 0.260 ( 0.001 0.790 ( 0.020 0.551 ( 0.008 0.645 ( 0.004 0.142 ( 0.002 0.451 ( 0.004 0.340 ( 0.014 0.094 ( 0.002 0.112 ( 0.002 0.134 ( 0.002 0.112 ( 0.001
Group 2 0.037 ( 0.001 ND ND ND ND ND ND
Group 3 NY 3 NY 4 NY 6 NY 7 NY 9 NY 10 NY 11 NY 12 NY 15 NY 16 NY 17 NY 19 NY 22 HK 2 HK 3 HK 4 HK 5 HK 7 HK 9 HK 10 HK 11 HK 12 HK 13 HK 14 HK 16 HK 17 HK 18 HK 19 HK 21 HK 22
0.325 ( 0.007 0.290 ( 0.006 0.198 ( 0.001 0.815 ( 0.012 0.235 ( 0.013 0.380 ( 0.007 0.563 ( 0.007 0.391 ( 0.003 0.471 ( 0.003 1.299 ( 0.051 0.384 ( 0.001 0.436 ( 0.015 0.158 ( 0.002 0.130 ( 0.003 0.211 ( 0.010 0.664 ( 0.003 0.829 ( 0.005 0.640 ( 0.002 0.244 ( 0.001 0.498 ( 0.008 0.393 ( 0.004 0.664 ( 0.007 0.614 ( 0.001 0.427 ( 0.007 0.783 ( 0.013 0.263 ( 0.005 0.586 ( 0.004 0.127 ( 0.001 0.884 ( 0.003 0.625 ( 0.014
0.097 ( 0.001 0.038 ( 0.001 0.078 ( 0.004 0.026 ( 0.001 0.049 ( 0.001 0.072 ( 0.003 0.229 ( 0.010 0.079 ( 0.002 0.095 ( 0.002 ND 0.065 ( 0.002 0.103 ( 0.005 0.202 ( 0.005 0.029 ( 0.002 0.156 ( 0.002 0.015 ( 0.001 0.164 ( 0.002 0.037 ( 0.001 0.334 ( 0.006 0.033 ( 0.001 0.089 ( 0.001 0.032 ( 0.002 0.127 ( 0.004 0.022 ( 0.001 0.067 ( 0.001 0.017 ( 0.0005 0.072 ( 0.003 0.068 ( 0.002 0.078 ( 0.001 0.123 ( 0.006 0.230 ( 0.002 0.238 ( 0.010 0.380 ( 0.001 0.104 ( 0.003 0.144 ( 0.001 0.324 ( 0.005 0.081 ( 0.002 0.249 ( 0.004 0.169 ( 0.007 0.247 ( 0.002 0.241 ( 0.006 0.051 ( 0.001 0.345 ( 0.002 0.157 ( 0.002 0.227 ( 0.005 0.124 ( 0.002 0.183 ( 0.002 0.052 ( 0.001 0.440 ( 0.006 0.090 ( 0.001 0.267 ( 0.002 0.034 ( 0.002 0.167 ( 0.001 0.631 ( 0.001b 0.080 ( 0.003 0.468 ( 0.004 0.315 ( 0.002 0.086 ( 0.002 0.193 ( 0.001 0.089 ( 0.002
ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND 0.041 ( 0.001 ND 0.060 ( 0.001 0.053 ( 0.001 ND ND ND 0.017 ( 0.0004d ND 0.046 ( 0.001
a
Group numbers refer to the groups from phylogenetic relationship of the tested samples as shown in Figure 3. b The values in bold are the highest concentration of each standard among all the tested samples. c ND means under the quantitative levels. d The values in bold and italic are the lowest concentration of each standard among the detectable samples. 1551
dx.doi.org/10.1021/jf1028174 |J. Agric. Food Chem. 2011, 59, 1548–1556
Journal of Agricultural and Food Chemistry
ARTICLE
Table 3. Two-Way ANOVA Fit Model of Concentrations of Compounds 1-7 with the Interaction between Sample Source and Phylogenetic Grouping Added as Model Effects phylogenetic
sample source
phylogenetic groups and
compd
grouping only
only
sample source
1
0.0542
0.6810
0.0301*a
2
0.0001*
0.2552
0.0048*
3
0.1393
0.0952
0.0248*
4 5
0.0431* 0.0395*
0.0119* 0.0001*
0.0002* 0.0012*
6
0.0580
0.0040*
0.0152*
7
0.4092
0.0263*
0.0341*
a
Asterisk denotes significant effects on the variances in the concentrations of compounds tested; N = 44, samples were run in triplicate.
Figure 3. Maximum Parsimony tree containing 44 Radix Astragali and other species of Astragalus from GenBank. Numbers above branches are bootstrap support values; numbers below branches of 3 Radix Astragali clades show number of genetic changes that differentiate each group or individuals in group 3.
Table 2. Mean Distribution and Analysis of Variance (ANOVA) Results for Significant Differences of Compounds 1-7 among Sample Source compd
HK samplesa
NY samplesa
F ratio
Prob > F
1
0.514 ( 0.33
0.473 ( 0.30
0.557
2
0.218 ( 0.15
0.148 ( 0.08
11.972
