Differentiation of Mitragyna speciosa, a narcotic plant, from allied Mitragyna species using DNA barcoding-high-resolution melting (Bar-HRM) analysis

Mitragyna speciosa (Korth.) Havil. [MS], or “kratom” in Thai, is the only narcotic species among the four species of Mitragyna in Thailand, which also include Mitragyna diversifolia (Wall. ex G. Don) Havil. [MD], Mitragyna hirsuta Havil. [MH], and Mitragyna rotundifolia (Roxb.) O. Kuntze [MR]. M. speciosa is a tropical tree belonging to the Rubiaceae family and has been prohibited by law in Thailand. However, it has been extensively covered in national and international news, as its abuse has become more popular. M. speciosa is a narcotic plant and has been used as an opium substitute and traditionally used for the treatment of chronic pain and various illnesses. Due to morphological disparities in the genus, the identification of plants in various forms, including fresh leaves, dried leaf powder, and finished products, is difficult. In this study, DNA barcoding combined with high-resolution melting (Bar-HRM) analysis was performed to differentiate M. speciosa from allied Mitragyna and to assess the capability of Bar-HRM assays to identify M. speciosa in suspected kratom or M. speciosa-containing samples. Bar-HRM analysis of PCR amplicons was based on the ITS2, rbcL, trnH-psbA, and matK DNA barcode regions. The melting profiles of ITS2 amplicons were clearly distinct, which enabled the authentication and differentiation of Mitragyna species from allied species. This study reveals that DNA barcoding coupled with HRM is an efficient tool with which to identify M. speciosa and M. speciosa-containing samples and ensure the safety and quality of traditional Thai herbal medicines.

considered illegal 16 . Many countries have banned kratom or implemented severe, strict action or penalties for its possession. However, there are increasing reports of people mixing kratom leaves with pharmaceutical drugs, Coca-Cola cocktails, cough syrup and strange ingredients such as mosquito coil ash 17 . The demand for kratom is not often met due to the restriction and unavailability of the species in the required quantity in areas convenient for kratom processing. As a result, M. speciosa is substituted with other plant species, including Mitragyna species, or adulterated. The other species may have similar or different morphological characters and may differ in their chemical profiles. Although of natural origin and having been used for many years, traditional medicines are not yet recognized officially in many countries due to concerns about their safety, quality and efficacy 18,19 . The major reasons for the increase in such concerns are intentional or inadvertent substitution and adulteration [20][21][22] . The adulterant and substituted species may have different or lower pharmacological action compared with that of their authentic counterparts. Even species within the same genus may exhibit differences in pharmacological action. This inadvertent substitution and adulteration can cause intoxication and even death 20,22 . Thus, quality assurance in terms of the identity of herbal drugs used in traditional medicines is vital. The WHO normal rules for conventions and practices on research and assessment of traditional medicines state that the initial phase in assuring the safety and efficacy of traditional medicines is correct identification 23 .
Taxonomic identification of tree species can be challenging. Plants of the same species may vary in their morphology according to their growing environmental conditions, their age, and time, and closely related species may exhibit similar morphologies 24 . In the past decade, molecular identification tools have been broadly used for plant identification. DNA barcodes, short sections of DNA sequences, have been proven as an alternative tool for the identification of medicinal plants. In our previous studies, it was shown that sequences from the nuclear internal transcribed spacer (ITS) region can be used to differentiate M. speciosa from related species by the polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) method 2 . However, the shortcomings of the PCR-RFLP approach are that it is time consuming and very limited in its ability to identify species when the samples are from incomplete specimens or damaged. More recently, several studies have shown that very closely related medicinal plant species can also be distinguished accurately by using DNA barcoding combined with high-resolution melting (Bar-HRM) analysis 21,[25][26][27] .
Therefore, in this study, we aimed to develop Bar-HRM analysis for the differentiation of M. speciosa from related species and the investigation of M. speciosa in suspected kratom samples. The generated Bar-HRM analysis profiles enable verification of the authenticity of narcotic plant species for law enforcement.

Results
DNA analysis and primer design. DNA was successfully extracted from all collected Mitragyna species and suspected kratom samples. The genomic DNA concentrations of the Mitragyna species and suspected kratom samples were quantified using NanoDrop measurements (Table S1). The suspected kratom DNA samples were inconsistent in quality and quantity. The total DNA concentrations of all the Mitragyna species and suspected kratom samples were provided in Table S1. rbcL, trnH-psbA intergenic spacer, matK, and ITS2 barcode sequences (Table 1) were amplified in all Mitragyna species and the suspected kratom samples. Positive PCR products with the expected length were confirmed with agarose gel electrophoresis. The trnH-psbA intergenic spacer was observed to have higher nucleotide variation (5.395%) than other regions, followed by ITS, matK, and rbcL (3.780%, 0.658% and 0.069%, respectively) ( Table 1).
Multiple sequences of M. speciosa were used to design HRM primers. The HRM primer pairs were designed for the flanking regions of each DNA barcode region, yielding HRM amplicons ranging from 71 to 110 bp ( Table 2; Fig. S1). The trnH-psbA intergenic spacer was observed to have the most variable sites (9 bp), which consisted of nine nucleotide insertion-deletion (indel) positions without single-nucleotide polymorphisms (SNPs), followed by ITS2 (1 bp), matK (0 bp) and rbcL (0 bp) ( Table 2). Highest variable characters (%) showed in psbA-trnH (8.82) by ITS2 (3.75), matK (1.39) and rbcL (1.39) ( Table 2). All Mitragyna species multiple sequence alignments were provided in supplementary figure (Fig. S2). However, when amplicons differ in just one or few nucleotides, they may present similar melting curve profiles with small shifts in T m ( Table 3). The GC content of the four HRM amplicons was calculated for the prediction of melting curve profiles. ITS2 had the highest average GC content at 63.27%, followed by rbcL, matK and the trnH-psbA intergenic spacer at 46.18%, 36.97%, and 35.35%, respectively ( Table 2). Bar-HRM analysis of four DNA loci of M. speciosa was performed using specific HRM primer pairs corresponding to the ITS2, matK, rbcL and trnH-psbA intergenic spacer barcode regions. All HRM primer pairs were designed with conserved sequences with a 100% match to the target sites, which www.nature.com/scientificreports/ facilitated primer annealing and elongation initiation of DNA polymerase. The matK and rbcL HRM primer pairs provided consistent amplicon sizes of 71 and 72 bp, respectively. The ITS2 and trnH-psbA intergenic spacer primers yielded variable-length amplicons of 79-80 and 101-110 bp, respectively ( Table 2).

Evaluation of HRM primer pairs for Mitragyna species. The four HRM primer pairs Mit-ITS2, Mit-
matK, Mit-rbcL and Mit-trnH-psbA were designed and used for HRM analysis of the four Mitragyna species (Fig. S1). The Bar-HRM procedure was conducted to identify Mitragyna species. All samples were amplified with four different primer pairs prior to defining the T m at the melting step in order to distinguish M. speciosa from related species (Table 3). The anticipated HRM amplicons from ITS2, matK, rbcL and trnH-psbA were 79-80 bp, 71 bp, 72 bp and 101-110 bp, respectively ( Table 2; Fig. S1). As shown in Fig. 1, the barcode regions of ITS2, rbcL, the trnH-psbA intergenic spacer and matK in M. speciosa presented similar melting curve profiles and therefore could be visually differentiated from the normalized melting curves of other Mitragyna species. The melting curve analysis revealed different melting peaks, which could be used to discriminate M. speciosa from the other three species. To obtain the best visualization of very small differences between individual melting curves, Bar-HRM analysis of all four regions was conducted using Bio-Rad Precision Melting software to acquire HRM profiles and improve visualization for differentiation. Precision Melting software was used to calculate the difference in melting curves with M. speciosa as a reference. All the final normalized melting curves and separation of the different melting curves of the four Mitragyna species are shown in Fig. 1A-H. The HRM curve analysis showed slight melting temperature shifts between PCRs for all the Mitragyna species ( Fig. 1; Table 3). However, the relative position and shape of the normalized melting curves were consistent compared to those of the differential melting curves (Fig. 1). Bar-HRM analysis of the four HRM primer pairs yielded a normalized plot and differential plot. The melting curves of the four Mitragyna species readily distinguished M. speciosa from the related species when using HRM analysis with four HRM primer pairs (Fig. 1). The melting profiles of the four Mitragyna plants could be separated into two clusters. All M. speciosa samples were collectively clustered (red line) (Fig. 1). In contrast, other allied species were clustered together (yellow, olive green and green lines) (Fig. 1). The differences in T m between M. speciosa and related species for ITS2, matK, rbcL and the trnH-psbA intergenic spacer were 1.3, 0.4, 0.8 and 0.3, respectively (Table 3). Although the four HRM primer pairs could be used to differentiate M. speciosa from related species, the ITS2 region, which provided the largest difference in T m , was selected for further investigation.
Investigation of M. speciosa in suspected kratom samples. According to the significant differences in the melting profiles of M. speciosa and other related Mitragyna species detecting using the Mit-ITS2 primer pair, the Bar-HRM approach was used to detect M. speciosa in suspected kratom samples (Table 4; Fig. 2A). In the Bar-HRM analysis of normalized (Fig. 2B) and differential curves using the M. speciosa curve as the reference (Fig. 2C), five out of six suspected kratom samples, namely, K-01, K-02, K-04, K-05 and K-06, had T m values ranging from 83.9 to 84.0 °C, which were similar to those of the references (Table 4). On the other hand, kratom sample K-03 had the differential curve with the highest distinct T m value of 85.2 °C, placing the sample in the non-M. speciosa cluster (Table 4; Fig. 2B,C). The melting curves of the six suspected kratom samples formed two clusters: the M. speciosa cluster (K-01, K-02, K-04, K-05 and K-06) and the non-M. speciosa cluster (K-03) (Fig. 2B,C). However, all HRM amplicons were sequenced and NCBI BLAST searched against the GenBank database for species confirmation. The NCBI BLAST results for suspected kratom samples K-01, K-02, K-04,  www.nature.com/scientificreports/ K-05 and K-06 indicated a very close match to M. speciosa, with the highest query coverage and maximum identity, and sample K-03 matched to non-M. speciosa (Table 4; Fig. S3).

Discussion
M. speciosa or kratom has a very long history of traditional usage in Southeast Asia. Because kratom is increasingly considered a less expensive and more readily available substitute drug 5 , it has reportedly gained more attention as an alternative to opium, at least in Thailand. Although kratom is a narcotic and banned in certain www.nature.com/scientificreports/ countries because of its opioid-like effects 5,6 , people living in rural areas continue to believe that kratom consumption is less harmful than the consumption of other banned drugs; in fact, its usage has not been reported to have any significant health threats, in contrast to opiate misuse 9,16,28,29 .
In the past decade, molecular analysis has become an acceptable tool for the authentication of medicinal plants. Many studies have revealed DNA barcode regions that can discriminate plants at the genus or species level 30 . In this study, four universal DNA barcode regions, rbcL, the psbA-trnH intergenic spacer, matK, and ITS2, were evaluated for their applicability in identifying four Mitragyna species. The ITS2 region yielded the highest interspecific divergence among the four species of Mitragyna and revealed the highest nucleotide variation among them (Table 1). Indeed, the ITS2 barcode region is one of the most variable regions in angiosperms and can be used for species differentiation 30,31 . Recently, Bar-HRM analysis has been successful in the identification of various medicinal plants 32 . The advantage of HRM is its capacity to screen variations in specific regions that would not be identified by Sanger sequencing; moreover, using HRM analysis of melting temperatures can allow any different nucleotides in a sequence to be detected 33    www.nature.com/scientificreports/ analysis. The optimal sequence length for obtaining accurate results from Bar-HRM analysis is 300 bp or less 34 . In our study, the lengths of the HRM amplicons were in the range of 71 to 110 bp, which is consistent with the findings of previous reports. High nucleotide consensus at HRM primer sites is critical for the annealing of primers and elongation of amplicons by DNA polymerase 35 . Our HRM primers designed for the four DNA regions showed 100% consensus in the flanking regions of each variable DNA site in the four Mitragyna species, which promoted the efficiency of DNA amplification. Further, we developed a Bar-HRM analysis method that can clearly differentiate M. speciosa from closely related species, including M. diversifolia, M. hirsuta and M. rotundifolia, and identified them directly in suspected kratom samples. In our study, the four newly designed HRM primer pairs (Fig. S1) were successfully used for the differentiation of M. speciosa from related species. Although all HRM analyses of the four regions displayed satisfactory discrimination of M. speciosa from allied Mitragyna species, the highest differential melting temperature (T m ) was observed in the Bar-HRM analysis of ITS2. The ITS2 region was used as a target for HRM analysis and has been demonstrated to be an effective tool for the detection and quantification of several plants, such as plants in the genus Sideritis 36 , nine herbal teas 37 , plants in the genus Artemisia 38 , two Ophiocordyceps species 39 , and edible plants 27 . The results from the Mit-rbcL and Mit-matK primer pairs showed a higher T m for rbcL HRM amplicons than for matK HRM amplicons. However, their amplicon sizes were very similar, and they showed the same type of nucleotide substitution (matK: C → T and rbcL: G → T). This study showed that the ITS2 amplicons had the highest GC content (63.27%), the trnH-psbA intergenic spacer amplicons had the lowest GC content (35.35%) ( Table 2), and nucleotide variations in these regions exhibited the largest difference in T m among the four Mitragyna species (Table 3). In our analysis, GC content was the most important metric for choosing candidate DNA barcode regions for combination with HRM analysis. A higher GC content will lead to a larger T m difference. Therefore, the GC content of nucleotide variations among species is the key factor for the authentication of plant species.
In the investigation of suspected kratom samples, five out of six samples (K-01, 02, 04, 05, and 06) clustered with M. speciosa, and only one suspected sample (K-03) was included in the non-M. speciosa cluster (Table 4; Fig. 2). Suspected kratom sample K-03 clustered with non-M. speciosa, consistent with the NCBI BLAST result (Table 4). This study proves the efficiency of Bar-HRM analysis for the identification of raw material and highly processed samples. The closed-tube system can be simultaneously performed without a post-PCR assay. Moreover, this method is suitable for routine analysis in the laboratory without the need for high expertise. Our research provides a valuable tool for the characterization of Mitragyna species that can be useful for quick one-step real-time PCR for the simultaneous identification of suspected kratom samples. Furthermore, this analysis is potentially useful for the identification and differentiation of closely related plant species in mixed-plant samples.

Conclusion
To the best of our knowledge, this is the first study of DNA barcoding coupled with HRM analysis for the differentiation of M. speciosa, a narcotic species, from closely related species and for the investigation of suspected kratom samples. ITS2 best reflected the relationships between the four Mitragyna species that we tested and can be used consistently to determine species identity. The Bar-HRM results from the Mit-ITS2 primer pair indicated that ITS2 can be used as an effective DNA barcode marker for Mitragyna species. Bar-HRM analysis provides a simple, sensitive, and reliable method for the investigation of M. speciosa and suspected kratom samples for routine analysis in forensic laboratories. These outcomes will aid in the authentication of M. speciosa in suspected kratom samples as well as deliver a new method with which to identify and differentiate suspected kratom samples, ensure safety and quality control for traditional medicine and identify narcotic plant species for law enforcement.  Table S2. All samples were identified by an expert taxonomist, Assoc. Prof. Thatree Phadungcharoen, Rangsit University, Thailand. Each voucher specimen was assigned a specific number and deposited in the Museum of Natural Medicine, Chulalongkorn University, Thailand. These authenticated plant specimens were further applied to perform Bar-HRM analysis using the nuclear region of the ITS and three chloroplast regions, namely, rbcL, the trnH-psbA intergenic spacer and matK. Five different forms of suspected kratom samples presented as M. speciosa, including fresh leaves (K-01), dried leaves (K-02 and K-03), powder (K-04), juice (K-05), and a cocktail (K-06), were collected from anonymous sources in the southern part of Thailand ( Fig. 2A; Table S3). www.nature.com/scientificreports/ (MP Biomedicals, USA). The quality and quantity of genomic DNA were determined using agarose gel electrophoresis and a NanoDrop spectrophotometer (Thermo Fisher, USA). All DNA was kept at − 20 °C prior to more in-depth analysis.

DNA isolation from authentic
Sequence analysis and HRM primer design. All sequences from the plastid DNA, including matK, rbcL and the trnH-psbA intergenic spacer, along with the ITS2 sequence of Mitragyna plants were retrieved from our previous study 2,40 (Table S2). Multiple alignments of sequences were performed using MEGA7 41 . Interspecific divergence was calculated using MEGA7 with the Kimura 2-parameter (K2P) distance model and pairwise deletion algorithm. Variable characters and GC content were calculated for all DNA barcode regions in the Mitragyna species (Table 1). All four DNA barcode regions were adopted to design the HRM primer pairs based on the consensus flanking region, which covered enough variable sites to allow the differentiation of narcotic species from nonnarcotic species. All the primers used in this were mentioned in Table S4. The expected sequences of HRM amplicons, nucleotide variation and average GC percentage were calculated for further analysis (Table 2).
Bar-HRM analysis. PCR  Testing of suspected kratom samples. Total genomic DNA was isolated from each suspected kratom sample and then subjected to HRM analysis with the Mit-ITS2 primer pair for the ITS2 region for identification via melting temperature (T m ). HRM amplicons of each sample were sequenced on an ABI 3730XL DNA Analyzer. NCBI BLAST analysis was performed against the GenBank database to authenticate the plant species.