Bar-HRM: a reliable and fast method for species identification of ginseng (Panax ginseng, Panax notoginseng, Talinum paniculatum and Phytolacca Americana)

Plants samples and DNA extraction

Plant tissues including two Phytolacca species (P. americana and P. japonica), three Talinum species (T. fruticosum, T. paniculata and T. triangulare) and two Panax species (P. ginseng and P. notoginseng) were obtained from Queen Sirikit Botanic Garden (QSBG) (Table 1). The plant tissues were ground with liquid nitrogen. DNA from all samples was extracted using the Nucleospin Plant® II kit (Macherey-Nagel, Düren, Germany) according to the manufacturer’s instructions.

Literature search

Research literature involved ginseng studies was accessed through the Web of Science Core Collection. Only publications indexed in the Science Citation Index Expanded were included in the search. Each document’s type, year of publication and corresponding author’s country were recorded. The literature search results were visualized using an open source data visualization framework called “RAWGraphs” (Mauri et al., 2017).

Data mining for sequence analyses

The DNA sequences of plant species in genus Phytolacca, Talinum and Panax were searched and extracted from GenBank on National Center for Biotechnology Information (NCBI) website using the keyword “name of each genus and each chosen barcode region” (ITS, matK, rbcL, trnL and trnH-psbA). MEGA 6 program was used for sequence alignment and analysis (Tamura et al., 2013). The following characteristics were recorded: average GC content, conserved and variable site (%).

Fragment amplification and HRM analysis

Real-time PCR amplification and DNA melting with fluorescence measurements were performed on a Rotor-Gene Q HRM system (Qiagen, Hilden, Germany). The total volume of 20 µL reaction mixture contained 20 ng genomic DNA, 10 µL of MeltDoctor™ HRM Master Mix (Applied Biosystems, Foster City, CA, USA), 0.2 µL of 10 mM forward primers and reverse primers (Table 2). Conditions are as follows; 95 °C for 5 min followed by 35 cycles of 95 °C for 30 s, 57 °C for 30 s, and 72 °C for 20 s. The temperature increased from 60 to 95 °C, at 0.1 °C/s.

Literature search

A total of 2,724 published articles containing the word ‘ginseng’ were found when performing the literature search (July 2019), and totaling 601 other references types (reviews, proceedings, papers, meetings, abstracts, etc.) identified by our database searches (Thomson Reuters Web of Science). About 71% (1,927 articles) of the ginseng published articles were found to be about plants in Panax genus. Authors from South Korea have the highest contribution compared with those from other counties. Second and third to Korea were China and USA (Fig. 1). We furthered our database searches with the words ‘method’ and ‘technique’ and found that only 39 articles from 1,927 Panax published articles focusing on the method or technique used for species authentication/identification/discrimination.

Data mining and in silico analyses

GenBank accessions were collected to assemble DNA barcode sequences of Panax, Phytolacca and Talinum. Data was present for most regions of the three selected genus, except for psbA-trnH and trnL of Phytolacca and trnL of Talinum. The total number of Panax sequences collected for the respective regions are as follows: 301, 196, 126, 86 and 4 species for ITS2, matK, psbA-trnH, rbcL and trnL, respectively. The total number of Phytolacca sequences collected for the respective regions are as follows: 6, 19, and 18 species for ITS2, matK, and rbcL, respectively. The total number of Talinum sequences collected for the respective regions are as follows: eight, 14, six, and eight species for ITS2, matK, psbA-trnH and rbcL, respectively.

Sequence length, GC content and variation within sequences lead to different T_m values and melting profiles which are main focus in High Resolution Melting (HRM) analysis. Therefore, all collected sequences were then analyzed using MEGA6 for average GC content (%), conserved and variable site (%). As can be seen from Table 2, the analyzed ITS2 fragment from all three plant groups was found to have a higher nucleotide variation (84.48% in Panax species, 8.44% in Phytolacca species and 21.55% in Talinum species) than other regions. The nucleotide variation within amplicons of Panax species was found to be as follows: ITS2 >psbA-trnH >matK >rbcL >trnL, Phytolacca species was found to be as follows: ITS2 >matK >rbcL, and Talinum species was found to be as follows: ITS2 >matK >psbA-trnH >rbcL (Table 2). It is suggested that in this study trnL is least suitable for ginseng species discrimination in terms of both lack of DNA data and nucleotide variation. In contrast, ITS2 poses great potential for this study. Variation in melting profiles for the different markers could also predict from an average %GC content of amplicons. The ITS2 region had the highest average %GC content in all three plant groups, with 62.4% in Panax, 60.8% in Phytolacca, and 72.2% in Talinum (Table 3). Based on these results, it was predicted that the ITS2 primer pair would be the best marker choice for HRM analyses with the target species.

Fragment amplification for HRM analysis

There was inconsistency in the lengths of psbA-trnH fragment obtained from PCR amplification due to high indel in the region. As amplicon length is one of main factors affecting HRM analysis, the psbA-trnH was not included here. Although matK has been proposed as one of standard plant barcodes in terms of species identification, in this study HRM had a low success rate in PCR amplification with the matK primers. Therefore, we did not choose the matK for analysis here. Three primer sets including ITS2, rbcL and trnL were selected. HRM analyses was carried out in triplicate on each of the seven species including two Phytolacca species, three Talinum species and two Panax species to establish the melting profiles for each primer pair. The analysis is presented in T_m value of each species and the melting profiles of amplicons from each region are illustrated in Figs. 2–4. In this study, we also tested the hypothesis that Bar-HRM can discriminate true ginseng (P. ginseng and P. notoginseng) and other two plant species from other genus that looks very similar to the true ginseng: Thai ginseng (T. paniculata) and poisonous species (P. americana) (Fig. 5).

The HRM analysis result of rbcL (Figs. 2A–2C), trnL (Figs. 3A–3C) and ITS2 (Figs. 4A–4C) regions are shown as melting curves. The rbcL primers generated a unique melting curve for each Talinum and Panax species (Figs. 2B and 2C), whereas similar melting curves from two samples were observed in the Phytolacca species (Fig. 2A). This suggests the rbcL has adequate ability to discriminate the tested Talinum and Panax species but not the Phytolacca species. The melting curves of P. americana and P. japonica generated from trnL primers were nearly the same (Fig. 3A). Similarly, the shapes of the melting curves of P. ginseng and P. notoginseng were nearly identical to each other (Fig. 3C). The individual melting curves from trnL analysis were reproducibly obtained from each of the three different Talinum species (Fig. 3B). In contrast, all ITS2 amplicons from the different species of Phytolacca, Talinum and Panax species yielded distinctive HRM profiles (Figs. 4A–4C).

We then used only ITS2 primers in order to confirm its ability in discriminating the true ginseng (Panax species) from plant species that closely resemble the Panax ginseng in root form T. paniculata and P. americana. The melting profiles of ITS2 amplicons were shown in Fig. 5. HRM analysis based on ITS2 region can detect differences among samples and thus the two species that resemble true ginseng can be discriminated. A sequence alignment of the tested species was performed to justify the nucleotide differences between them. Within the fragment amplified by the ITS2 primers pair (285 bp in length), there were 110 variable sites found. The result here is consistent with our prediction that indicated ITS2 would be efficient in identifying the tested plant’s species. As many distributors claim false plants are part of the true ginseng family, one wonders how we can be sure about which plants the ginseng products are derived from exactly. From our results, Bar-HRM using ITS2 primers pose a great potential to authenticate the real ginseng.