Introduction

Traditional Chinese medicine (TCM), encompassing Chinese herbal medicine, continues to gain international recognition. According to the National Bureau of Statistics of China, the turnover of the Chinese herbal medicine market in 2019 reached 165.3 billion yuan for the domestic market and $6.175 billion for the international side1. Apiaceae, a family of flowering plants, is recognized as a significant resource for TCM, comprising 65 genera and 262 species2. Among them, Notopterygii Rhizoma et Radix (NReR), known as "Qiang-huo" in Chinese, holds a long history dating back to the Han Dynasty, approximately 2000 years ago3. According to the Pharmacopoeia of the People's Republic of China, NReR is derived from the roots and rhizomes of Notopterygium incisum Ting ex H. T. Chang or N. franchetii H. de Boissieu4. It encompasses a complex array of chemical constituents, including volatile oils and terpenes, coumarins, sugars, glycosides, phenolic acids, polyalkynes, and alkaloids5. Modern pharmacological studies have demonstrated its anti-inflammatory, antibacterial, antioxidant, antiarrhythmic, anticancer, antipyretic, and analgesic activities6. Currently, NReR serves as a raw material for over two hundred types of Chinese (Tibetan) patent medicines.

Notopterygium incisum and N. franchetii were listed as national third-class protected plants as early as 1987, and they were successively included as "Near-Threatened" species in China's Red List of Biodiversity and China Species Red List. In the past few decades, excessive excavation and habitat destruction lead the wild resources of NReR to be drastically reduced7,8,9. The scarcity of resources and the increase of market demand have driven the price up, which motivated adulteration intentionally. Reports indicate that Angelica sylvestris L., Pleurospermum rivulorum (Diels) Hiroe, Polygonum cuspidatum Sieb et Zucc, and Sanguisorba officinalis L. are frequently sold as NReR in the medicinal market10,11,12. These species share similar organoleptic characteristics but differ in chemical constituents compared to NReR13,14,15,16,17. Traditional methods used to authenticate NReR and its adulterants, such as macroscopy11,18,19, microscopy20, and chemical profiles21,22, can provide certain recognition and differentiation to some extent. But, these methods are prone to geographical variations, growth stages, and storage conditions, which may affect identification accuracy23. Therefore, it is necessary to establish a simple and accurate identification method to distinguish NReR from adulterations.

DNA barcode technology is currently used as an effective tool to identify species. This method provides a large amount of genetic information with high accuracy and objectivity, and it can standardize and automate the identification process, establishing an easy-to-use application system in a short time23. One potential DNA barcode for identifying medicinal plants and their close relatives is the internal transcribed spacer 2 (ITS2), which has attracted attention due to its unique advantages such as being short and conducive to amplifying degraded samples24,25. While a few studies have examined the molecular identification of NReR and its adulterants using DNA barcoding26,27,28, the composition of commercial NReR in the Chinese medicinal market needs to be sorted out to ensure the subsequent formulation of quality control standards and clinic safety for NReR.

Thus, the aim of this study is to investigate whether ITS2 is a valuable marker for identifying genuine NReR from its adulterants and to gain insight into the composition of commercial NReR in China.

Materials and methods

Plant materials

Eighteen ITS sequences available for genuine NReR (N. incisum and N. franchetii) were first downloaded and screened from NCBI GenBank as reference. Moreover, 168 commercial crude drug samples under the name of NReR were collected from herbal markets, pharmacies, and online shops in 23 provinces and municipalities of China. Voucher specimens are deposited in Herbarium of Kunming Medical University. Detailed information is presented in the Table S1. All methods of experimental research on plants were performed in accordance with the relevant institutional, national, and international guidelines and legislation.

DNA extraction, polymerase chain reaction (PCR) amplification, and sequencing

The surface of all herbal materials was cleaned with 75% ethanol to avoid fungal DNA contamination. About 50 mg of the materials were cut into pieces, added with 10% polyvinylpyrrolidone (PVP), and then ground with a FastPrep bead mill (Retsch MM400, Germany). Total genomic DNA was extracted using the modified CTAB procedure of Doyle and Doyle29 or using the Plant Genomic DNA Kit (Tiangen Biotech, Beijing, China). Agarose gel electrophoresis showed slight smearing in some DNA samples, indicating partial degradation. The universal primers ITS-S2F (5′-ATGCGATACTTGGTGTGAAT-3′) and ITS-S3R (5′-GACGCTTCTCCAGACTACAAT-3′) were used to amplify the complete ITS2 region24. The PCR reaction conditions were the same as described previously30. PCR products purifying and sequencing were completed by Tsingke Biotechnology Co., Ltd (Beijing, China).

Data analysis

The sequences of genuine and commercial NReR were assembled using the MAFFT v.731, and manually adjusted where necessary using the BioEdit32. The assembled sequences were annotated and trimmed to obtain the complete ITS2 region based on a Hidden Markov Model (HMM)33. The genetic distances were calculated using MEGA v.734 according to Kimura 2-parameter (K2P) model34. Barcoding gaps comparing the distributions of the pairwise intra- and inter-specific distances with distance intervals of 0.002 were estimated in Microsoft Excel 2016. The true NReR presenting a minimum inter-specific distance value higher than their maximum intra-specific distance were considered successfully discriminated from potential adulterant plant species35. Wilcoxon two-sample tests were performed as described previously24,36. Haplotype matrix was generated by DNAsp v.637. BLASTN and the nearest distance methods were both used to evaluate the species authentication efficacy38. Sequences were uploaded onto NCBI database with a minimum identity cut of 99% for a top match according to the BLAST program (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Neighbor-joining (NJ) tree was constructed based on haplotypes, performing 1,000 bootstrap replicates in MEGA v.734.

Results

Amplification, sequencing and sequence characteristics

Genomic DNA was extracted from a total of 168 commercial "Qiang-huo" products, out of which 48 samples failed to amplify due to severe DNA degradation. The length of 138 combined sequences ranges from 226 to 256 bp. The GC content of the sequences shows a mean value of 58.3% with a range of 53.2% to 68.2%. The aligned length of 263 bp exhibits 143 variable sites, a rate of 54.4% (Table 1). These findings suggest that sequences for the sampled "Qiang-huo" were relatively variable.

Table 1 ITS2 sequence characters of samples.
Full size table

Assessment of barcoding gap

The average interspecific distance between N. incisum and N. franchetii was 0.039. The interspecific distance between N. incisum and the adulterant species ranges from 0.037 (N. oviforme) to 0.659 (Broussonetia papyrifera). Notopterygium franchetii shows a similar interspecific distance with the adulterants, with the maximum interspecific distance being 0.699 from B. papyrifera and the minimum being 0.036 from N. oviforme. The intraspecific genetic distance within N. incisum (0.005) and N. franchetii (0.005) was both smaller than interspecific distance between NReR and adulterants (Fig. 1). Our results show that the intra- and inter-specific variation of ITS2 had distinct gaps (Fig. 2). Additionally, Wilcoxon’s two-sample tests reveals that the mean of the inter-specific divergences was significantly higher than that of the corresponding intra-specific variations (p < 0.001, Table 2).

Figure 1
figure 1

Genetic distances from genuine NReR to its adulterants.

Full size image
Figure 2
figure 2

Relative distribution of inter-specific divergences and intra-specific variations for ITS2 sequences. The colored bars in each box represent inter-specific (above) and intra-specific (below) genetic distances.

Full size image
Table 2 Wilcoxon two-sample tests for distribution of intra- vs. inter-specific divergences.
Full size table

Evaluation of species authentication capability of ITS2

The BLASTN method exhibits a 100% success rate in identifying the tested commercial samples (Table 3). These samples consist of nine species, namely N. incisum, N. franchetii, N. oviforme, Levisticum officinale, A. amurensis, Ostericum scaberulum, B. papyrifera, Haplosphaera himalayensis, and Heracleum fargesii. Each identification result was supported by best hit of accessions obtained from the NCBI database (Table S1). A few samples initially identified as O. scaberulum, B. papyrifera, Ha. himalayensis and He. fargesii were subsequently confirmed through additional sampling and sequencing.

Table 3 Comparison of authentication efficiency for ITS2 using different methods.
Full size table

A total of 25 haplotypes were generated from the ITS2 sequences of genuine NReR and 120 commercial samples (Fig. S1). Combining these haplotypes with the BLASTN results, N. incisum was assigned to Haps_1-7, N. franchetii to Haps_8-15, L. officinale to Hap_17, A. amurensis to Hap_18, O. scaberulum to Haps_19-20, Ha. himalayensis to Haps_21-22, O. scaberulum to Hap_23, He. fargesii to Hap_24, and B. papyrifera to Hap_25 (Table 4). With the exception of N. incisum, the NJ tree analysis reveals that haplotypes representing different species formed isolated clades. While the haplotypes representing N. incisum does not form a monophyletic group on the NJ tree, the sequences to be identified as potential authentic species were clustered together with haplotypes representing genuine species (Fig. 3). Hence, the NJ tree method also exhibits a 100% success rate for NReR identification (Table 3).

Table 4 Information on the haplotypes associated with commercial NReR samples. The species identity of the samples was determined by BLAST queries on NCBI. Sequences of the joining (NJ) tree of NReR and its adulterantsgenuine NReR downloaded and screened from GenBank are highlighted in bold.
Full size table
Figure 3
figure 3

Neighbor joining (NJ) tree of NReR and its adulterants constructed based on the haplotypes. Detailed information about the haplotypes associated with each species is shown in Table 4. Bootstrap values are shown (≥ 50%) next to the branch.

Full size image

Survey of commercial NReR in the Chinese medicine markets

This study represents the most comprehensive nationwide sampling of commercial “Qiang-huo” to date, comprising a total of 168 samples obtained from 23 provinces. Based on their external morphology and odor characteristics, these samples proved challenging to distinguish from one another (Fig. 4A). Of 168 samples, except 48 failed to be amplified, molecular identification results show that 65 samples (54.2%) were identified as authentic “Qiang-huo”, while 55 samples (45.8%) were identified as adulterants (Fig. 4B). Further identification using the BLASTN method reveals that the adulterants belonged to seven different species, namely L. officinale (17 samples), A. amurensis (13 samples), O. scaberulum (five samples), B. papyrifera (L.) Vent. (two samples), Ha. himalayensis (two samples) and He. fargesii (one sample) (Table S1). Levisticum officinale, the most widely sold adulterant, was found in medicinal markets of nine provinces, followed by N. oviforme, A. amurensis and O. scaberulum found in eight, six and two provinces, respectively. Notably, B. papyrifera and Ha. himalayensis were only detected in Yunnan and Xizang, respectively. Regarding spatial distribution, samples from Guizhou, Hunan, Jiangsu, Qinghai, and Zhejiang were all confirmed as genuine NReR, while in Chongqing, Liaoning, Inner Mongolia, Shaanxi, and Xizang, no authentic NReR was detected. In ten provinces, including Hubei, Jilin, Jiangxi, and others, only one adulterant was discovered. Similarly, within the six provinces, such as Sichuan, Guangxi, Gansu, and others, two distinct types of adulterants were observed. The scenario in Yunnan and Anhui is more complicated, as three distinct types of adulterants were found (Tables S1, S2).

Figure 4
figure 4

(A) Morphology of commercial “Qiang-huo” (SC06: Notopterygium incisum, AH08: N. franchetii, AH16: Angelica amurensis, AH18: Levisticum officinale, AH19: N. oviforme, SC16: Heracleum fargesii, XZ02: Haplosphaera himalayensis, YN06: Ostericum scaberulum, YN14: Broussonetia papyrifera), (B) The composition and proportion of the commercial “Qiang-huo” products identified by barcode ITS2.

Full size image

Discussion

In recent years, DNA molecular identification technology has emerged as a robust tool for TCMs identification. This technology stands out for its ease of operation, cost-effectiveness, and high accuracy. In 2009 at the 3rd World DNA Barcode Conference, it was announced that the matK and rbcL markers are the core sequences of plant DNA barcodes, with ITS and trnH-psbA as complementary sequences39. The longer length of these markers has shown some weakness in amplification, sequencing and alignment, which is exacerbated if the materials are highly processed and have degraded DNAs40,41. Unlike Western herbs, most Chinese medicinal herbs are subjected to traditional processing procedures to increase their potency, minimize negative effects, and change their medicinal properties for a particular clinical use before they are released into dispensaries, practitioners, and the market42. According to the Chinese pharmacopoeia, the medicinal herbs are typically cleaned, cut, dried, and then processed, including stir-frying, charring, steaming, boiling, and calcining4,43. Raw materials processing methods can cause DNA degradation, posing challenges in obtaining standard DNA barcode sequences for samples. The use of mini-DNA barcoding technology, which focuses on shorter yet more efficiently amplified sequences through PCR, can partially overcome this limitation24,41,44,45.

ITS2, a mini-barcode spanning 160 to 330 bp in length, has emerged as the predominant marker for identifying plant medicinal materials24. Its growing popularity can be attributed to its ease of amplification and remarkable discriminatory capabilities across different taxonomic levels24. In this study, ITS2 performs well, with a higher amplification rate of 71.4%. Unsuccessfully PCR amplified NReR samples could be attributed to the high temperature drying process, causing DNA degradation. Both nucleotide signature (NS) and genome skimming metagenomics (GSM) emerge as promising solutions for fragmented and degraded plant materials identification46. NS consists of distinct nucleotide sequences that are exclusive to a specific taxonomic group. Previous studies, such as those on American Ginseng47, Cistanches Herba48, and Pinelliae Rhizoma49, have successfully demonstrated the efficacy of nucleotide signatures in identifying medicinal materials. GSM is the low-coverage shotgun sequencing of total DNA. When this approach is applied on herbal products, sequencing library is built without PCR amplification of barcode regions, circumventing the limitations of PCR in conventional DNA barcoding, such as DNA degradation during product manufacturing and PCR bias because of primer mismatch, etc. GSM produces millions of reads in a single run. After quality control, reads could then be clustered into operational taxonomic units (OUTs) based on similarity at defined threshold (usually 99–100%). Representative consensus sequences from each cluster would then be subject to taxonomic assignment, usually by alignment-based software like BLAST or k-mer based methods like Kraken46,50. As its sequencing cost is decreasing year by year, GSM technology will be a prospective method for the identification of TCMs herbs.

The validity of DNA barcoding relies heavily on the availability of a precise reference database, as it serves as the cornerstone of DNA barcoding. In this study, we constructed a comprehensive ITS2 matrix encompassing genuine NReR and their commonly found adulterants for the first time. This matrix will play a crucial role in both monitoring NReR and exploring potential substitute sources. The matrix includes two authentic species of NReR (N. incisum and N. franchetii, comprising 16 haplotypes) and seven confused species (nine haplotypes) (Table 4; Fig. S1). Out of the 120 samples sold as "Qiang-huo", only 54.2% were identified as authentic NReR, while the rest were identified as adulterants, including L. officinale, N. oviforme, A. amurensis, O. scaberulum, He. fargesii, Ha. himalayensis, and B. papyrifera (Fig. 4B). These findings highlight the complexity of the "Qiang-huo" market, with the presence of previously unreported species. According to the Chinese Pharmacopeia, only N. incisum and N. franchetii are listed as sources of NReR, and the former one exhibits superior quality and efficacy51. Our analysis revealed that N. incisum accounted for two-thirds of the authentic products, further indicating a preference for N. incisum in the market (Fig. 4B; Table S1). The NJ tree showed that adulterants of NReR were distantly related to N. incisum and N. franchetii (Fig. 3). Levisticum officinale was introduced to China in 1957, and used as a substitute for the traditional Chinese medicine “dang gui”, the roots of A. amurensis, He. fargesii and O. scaberulum52. The chemical and pharmacological analysis results of L. officinale and A. amurensis dramatically differs from those of NReR53,54,55,56,57,58. Notopterygium oviforme and Ha. himalayensis were grouped together in a strongly supported clade (bootstrap support = 99), closely related to authentic NReR, exhibiting genetic distances of 0.059–0.084 and 0.036–0.037, respectively (Figs. 1, 3). While N. oviforme has been traditionally regarded as a regional substitute to NReR, albeit with inferior quality59, additional research is urgently needed to explore the chemical constituents and pharmacological efficacy of both N. oviforme and Ha. himalayensis. This investigation aims to ascertain their potential as viable substitutes for NReR. Broussonetia papyrifera may represent a contaminant in the Yunnan samples, as its morphology significantly differs from that of NReR, and we also detected genuine NReR in these samples.

Conclusions

In this study, the origin plants of commercial “Qiang-huo” in the market was clarified, and the reference matrix of NReR and its adulterants was successfully established. This achievement is crucial for the future industrial development of NReR. However, it is important to acknowledge that the amplification efficiency of the ITS2 region is not always optimal, which is a challenge encountered in DNA barcoding identification of many medicinal materials. Therefore, alternative approaches such as NS and GSM appear to be promising solutions for overcoming this issue. These cutting-edge methodologies have the potential to revolutionize the field, providing more comprehensive and accurate identification results for medicinal materials.