Authentication of the botanical origin of Western herbal products using Cimicifuga and Vitex products as examples

Because the regulation of herbal products as dietary supplements in the USA is less strict than for OTC drugs in Europe, the substitution and adulteration of black cohosh dietary supplements with related Cimicifuga plants has frequently occurred in North America [5‐7]. Additional reports have described hepatotoxicity associated with the use of black cohosh products, especially from products derived from the wrong plant parts instead of the roots and rhizomes or different Cimicifuga species [8‐12]. The Ministry of Health, Labour and Welfare of Japan as well as other national ministries have updated their notifications about the risk of adverse liver reactions caused by the use of products containing black cohosh.

One possible reason why Cimicifuga plants native to Asia have been found in dietary supplements in the USA is that Chinese crude drugs were mistakenly recognized as black cohosh and were mixed with it. Cimicifuga plants are also used as traditional medicines in East Asia including Japan; the rhizome of C. dahurica (Turcz.) Maxim., C. foetida L. and C. heracleifolia Kom. is defined in the Chinese Pharmacopoeia (CP) as a crude drug named Cimicifuga Rhizome with heat-clearing and detoxifying effects, whereas the Japanese Pharmacopoeia (JP) describes C. simplex (DC.) Wormsk. in addition to the above three species as the botanical origin of Cimicifuga Rhizome (or “升麻” in Japanese) [13]. It is illegal in Japan for any health foods to contain materials or extracts prepared from C. simplex, C. dahurica, C. foetida or C. heracleifolia instead of C. racemosa, because Cimicifuga Rhizome is a pharmaceutical ingredient used exclusively for drugs.

Because of these problems, we developed genetic and chemical methods to simultaneously discriminate C. racemosa from other related species for the evaluation of black cohosh products that are commercially available in Japan.

Molecular biological techniques are useful and powerful methods to detect genetic differences among closely related plant species [14‐17]. For Cimicifuga plants, a random amplified fragment length polymorphism fingerprinting method was developed to identify the four sympatric species distributed in eastern North America (C. racemosa, C. americana Michx., C. pachypoda Elliott and C. rubifolia Kearney) [18], and a multiplex amplification refractory mutation system (ARMS) method was developed to identify four Asian species (C. dahurica, C. foetida, C. heracleifolia, and C. simplex) [19]. To the best of our knowledge, there has been no report about the simultaneous identification of both North American and Asian species. To establish a DNA analytical method that can discriminate C. racemosa from possible contaminated species, we prepared six Cimicifuga plants, identified each species by sequencing, and attempted to develop polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) and ARMS methods [20].

Eight referential Cimicifuga plants were kindly provided by the Research Center for Medicinal Plant Resources, National Institute of Biomedical Innovation, Health and Nutrition, and were identified as six species (one C. racemosa, three C. simplex and one each of C. japonica (Thunb.) Spreng, C, dahurica, C. heracleifolia and C. rubifolia) by DNA sequencing on the ITS and trnL−F regions (Table 2). Based on these sequences and sequences obtained from the DNA databases, we first developed a novel PCR-RFLP method for the identification of C. racemosa targeting on its unique BstBI restriction site (5ʹ TT*CGAA 3′) on the ITS1 region (Fig. 1). When the 441 bp of a DNA fragment on the ITS1 region was amplified by PCR from referential plant samples and digested with BstBI, two unique fragments of 134 and 307 bp were observed only from the referential C. racemosa plants (Fig. 2).

Although we found this to be a convenient and useful method to identify species of plant samples and to detect the presence of C. racemosa in black cohosh products, we could not detect contaminated or adulterated products by this method because undigested fragments remained. Therefore, we next developed an ARMS method for discriminating C. racemosa from other related species, using two types of specific sense primers targeting the nucleotide substitution at position 61 on the trnL gene (Fig. 3). The species-specific and the common fragments (284 bp) were successfully observed from C. racemosa and the other five species when the PCR was performed with DNAs from referential plants using the specific primer sets for C. racemosa and those for others, respectively (Fig. 4). Our results indicated that the ARMS analysis was a more accurate method than the PCR-RFLP method to identify C. racemosa from the related species, although the ARMS method takes a longer time.

We then applied the ARMS method to test the original plants of eight black cohosh products available on the Japanese market, and the results demonstrated that three of the eight products were not derived from C. racemosa (Fig. 5; Table S1). We also performed TLC and HPLC analyses for detecting two marker compounds—actein (1) and 23-epi-26-deoxyactein (2)—(Fig. 6) from black cohosh products, and we found that the three of the eight products did not contain marker compounds, as was also shown by the results of the ARMS analysis (data not shown). These TLC and HPLC results completely followed the results of the ARMS analysis and strongly demonstrated that the ARMS analysis is a useful method for authenticating the botanical source of black cohosh products.

Although the ARMS analysis was a very precise and straightforward way to evaluate the botanical origin of black cohosh products as described above, it had a disadvantage in that hardly any DNA was obtained from the products consisting of dry extracts. On the other hand, chemical analyses using marker compounds were also difficult to use for evaluating the botanical origin because most of the metabolites from closely related plant species overlap. The EP sets an identification test for black cohosh to confirm the presence of actein (1) and 23-epi-26-deoxyactein (2) by TLC. However, since these compounds are present in other Cimicifuga species as well as C. racemosa, the EP sets three additional TLC tests for substitution and adulteration to compare the chromatogram obtained from the reference solution with each chromatogram from C. americana, C. foetida, and the mixture of C. dahurica and C. heracleifolia [21].

In the course of our studies on the botanical origins of black cohosh products, we found and reported a simple way to detect substitution and adulteration [22]. We prepared 25 black cohosh products including six OTC drugs purchased in Europe (Table 3: BC-1 to -19, BC-M1 to –M6) and five Cimicifuga Rhizome crude drugs of JP grade (Table 3: CR-1 to -5), and we genetically identified their origins by DNA sequencing and analyzed their constituents by liquid chromatography-mass spectrometry (LC/MS). The sequencing analysis of the ITS region revealed that five of the 10 tested health food products containing powdered plant material contained undeclared C. dahurica, and that Cimicifuga Rhizome products were derived from C. dahurica except for CR-3, which contained both C. dahurica and C. heracleifolia (Table 3).

When these 15 samples were analyzed by LC/MS for the detection of actein (1) and 23-epi-26-deoxyactein (2), they showed three chromatographic patterns (I, II and III) based on the extracted ion chromatograms (XIC) at m/z 735.4 corresponding to actein (Fig. 7). All of the black cohosh products derived from C. racemosa (BC-1, -3, -4, -9, and -15) definitely contained marker compounds and showed pattern I. Of the 10 products that contained plant materials of C. dahurica, all of the Cimicifuga Rhizome samples (BC-6 and BC-12) showed pattern II, whereas the chromatograms of the other black cohosh samples (BC-7, -16, and -19) showed pattern III.

Pattern III appears to overlap with patterns I and II, and the products of pattern III consisted of powdered materials and dry extracts; it was thus strongly suggested that BC-7, -16 and -19 consisted of powdered materials of C. dahurica and an extract prepared from C. racemosa. Taking these results together with our above-described findings, we speculated that patterns I, II, III represent their botanical origins as C. racemosa, C. dahurica and the above-described mixture, respectively. We therefore established a simple method to detect substitution and adulteration of black cohosh products by monitoring a specific mass range on LC/MS.

We then analyzed the other black cohosh samples (which consisted only of dry extracts) by LC/MS to evaluate their botanical origins, and we found that 13 products including OTC drugs correctly used the dry extract of C. racemosa, and we observed that the XIC from BC-5 and BC-18 were out of synchronization with all three patterns (Table 3). Our finding that all the OTC drugs of EP grade were classified in the pattern I group strongly supports the usefulness of this method. The use of our new method could thus be a possible solution to the problem of adulteration of black cohosh products available in Japan.

The dried ripe fruit of Vitex agnus-castus L. is defined under the names of agnus castus fruit (ACF) and chaste tree, respectively, in the EP and USP−NF. ACF is used to treat symptoms of premenstrual syndrome and climacteric disorder, and it is becoming popular in Japan as the active ingredient of OTC drugs and health food products. On the other hand, the fruit of its congenic species, V. rotundifolia L. f. and V. trifolia L., has been used as a crude drug named shrub chaste tree fruit (SCTF) (蔓荊子 in Japanese) for its pain relief, sedative and anti-inflammatory effects.

Although the EP sets an identification test for ACF to confirm the presence of aucubin and agnuside by TLC as well as a purity test by macroscopic analysis based on the size of the fruit to distinguish V. agnus-castus from other Vitex species [23], the chemical components and morphological characteristics of Vitex species are too similar to discriminate between them. Considering the internationalization of the supply chain of herbal plants, this situation could lead to misidentification of the botanical origin of these crude drugs, resulting in unexpected adverse reactions as well as a decline in their pharmaceutical efficacy. We therefore attempted to identify the marker compounds that distinguish ACF and SCTF with the goal of establishing quick methods to evaluate their origins.

We first prepared crude drugs of ACF of EP grade and SCTF of Japanese standards for non-pharmacopoeial crude drugs (non-JP) grade to confirm their origins by DNA sequencing, and we applied them in a differential analysis using LC/MS data [24]. The sequencing analysis of the LFY/FLO gene revealed that two samples of ACF available in Germany and five samples of SCTF in Japan were derived from V. agnus-castus and V. rotundifolia or its mixture with V. trifolia, respectively (Table S2). To identify candidate marker compounds for ACF against SCTF, we performed a high-resolution LC/MS analysis of crude drug samples in an automated differential analysis, and we obtained 23 peaks as candidates. We then filtered these candidates on the basis of each formula and intensity, and seven peaks were finally picked up as characteristic compounds for ACF.

The chemical formulas calculated based on the accurate mass and MS fragmentation pattern of these peaks were collated with the phytochemical metabolite database, and we selected a putative diterpene with a novel structure—ID 1552 (RT 19.02 min, m/z 335.2228, formula C₂₀H₃₁O₄ for [M-H]⁻)—as the first target to identify the chemical structure (Fig. 8). An ACF health food product was used as a source material of peak ID 1552. The repeated column chromatography and preparative TLC of a CHCl₃ extract of the ACF health food product afforded compound 3/3a as a mixture. Compound 3/3a was obtained as a white powder. The ¹H-NMR and ¹³C-NMR spectroscopic data of 3/3a (Table 4) were very similar to those of known compounds, leucasperols A and B [25], and the complete structures of 3/3a were determined on the basis of the spectroscopic data from HMBC, COSY, NOESY and differential NOE NMR experiments (Fig. 9). Since there have been no reports on the isolation of compounds bearing this structure other than leucasperols A and B from Leucas aspera L., we proposed the trivial names chastol and epichastol for compounds 3 and 3a, respectively.

We analyzed additional ACF products by LC/MS to verify the usefulness of 3/3a as marker compounds, and we successfully confirmed their detection (Fig. 10; Table S3). Since avoiding contamination from closely related species is a significant requirement for pharmaceuticals of natural origin, this information will be valuable for the quality control of both ACF and SCTF products from the viewpoint of regulatory science.

Metabolomics have recently been adopted for the classification of variables and the identification of marker compounds based on the characteristics of their metabolic profiles in phytochemical investigations. For example, chemometric analyses such as a principal component analysis, a partial least-squares discriminant analysis, and an orthogonal partial least-squares discriminant analysis (OPLS-DA) are frequently used as multivariate tools for discrimination analyses [26‐29]. In order to identify marker compounds for SCTF against ACF, we analyzed SCTF and ACF products by LC/MS to apply to multivariate analyses [30].

When the LC/MS data set was subjected to OPLS-DA, the clusters of SCTF and ACF were clearly separated on the score plots, and 28 ions were picked up as the most influential variables that contribute to the segregation of SCTF and ACF samples on the S-plots (Fig. 11; Table S4). We then filtered these candidates on the basis of each formula and intensity, and seven peaks were finally picked up as specific compounds for SCTF. The chemical formulas calculated based on the accurate mass and MS fragmentation pattern of these peaks were collated with the phytochemical metabolite database, and we selected a putative triterpene—ID 6436 (RT 4.59 min, m/z 663.3920, formula C₄₀H₅₅O₈ for [M-H]⁻)—as the first target to identify the chemical structure.

The SCTF crude drug was extracted with CHCl₃ and fractionated using normal and reversed-phase column chromatography and prep-HPLC, monitoring the marker compound on LC/MS as a guide to obtain compound 4. Compound 4 was isolated as an amorphous colorless powder, and its molecular formula was determined as C₄₀H₅₆O₈ by high-resolution electrospray ionization mass spectrometry, with 13 degrees of unsaturation. Its complete structure was determined as 3-O-trans-feruloyl tormentic acid on the basis of spectroscopic data from ¹H-NMR, ¹³C-NMR, and HMQC and HMBC analyses (Tables 5, 6; Fig. 12); this is the first time that this compound has been isolated from a Vitex plant.

We also analyzed additional SCTF and ACF products by LC/MS to evaluate the usefulness of compound 4 as a marker compound for SCTF and successfully verified its specificity (Fig. 13; Table S5). The limit of detection of compound 4 in SCTF raw materials was 0.02 mg/ml. These results demonstrated that monitoring a marker compound for SCTF—3-O-trans-feruloyl tormentic acid—should help prevent the use of adulterated materials in ACF products, and would also be a useful means of identifying the origin of raw materials in SCTF products.

With the changes in cause-specific morbidity and the increasing interest in self-medication, Western herbs are becoming popular in Japan. Because herbal products and their ingredients contain thousands of constituents, the authentication of original species from herbal products is the first and most important task to ensure their quality and thus their safety and efficacy. This review described the importance of quality control in the regulation of herbal products, using Cimicifuga and Vitex products as examples. There are many other Western herbs being developed for OTC drugs, foods with functional claims and health foods, and the proper use of the correct origin for each herb is necessary. It is important to continue to work on the quality control of herbal products from the viewpoint of regulatory science.