Molecular Identification of Botanical Medicinal Materials

3.3 Molecular Identification of Botanical Medicinal Materials

Approximately 90% of medicinal materials recorded in the Pharmacopoeia of the People’s Republic of China (2010 edition) are derived from botanical sources. The huge international market of herbal medicinal materials suggests the importance of their correct identification. The pharmacological effects of herbal medicinal materials may vary among closely related species, subspecies, varieties, cultivars, and localities, not to mention the adulterants derived from distantly related species. Apart from conventional organoleptic and chemical methods, molecular approach provides an alternative and de fi nite method to identify these samples.

3.3.1 Discrimination at Inter-family and Inter-genus Levels

Adulteration of herbal materials by distantly related species from different families or genera is common. Molecular identification of these adulterants is relatively easy as their genetic makeups are quite different from the genuine species. DNA fingerprinting techniques usually show clear-cut results. For example, AP-PCR, RAPD and RFLP fi ngerprints of medicinal Panax species in family Araliaceae showed different patterns from the adulterants in families Nyctaginaceae, Phytolaccaceae, Campanulaceae, and Talinaceae [ 19, 45 ]. DNA sequencing is also useful to discriminate distantly related species. For example, trnL region is able to distinguish medicinal Stemona species in family Stemonaceae from adulterants in family Asparagaceae [ 71 ]. Similarly, trnL-F and trnH-psbA regions were used to distinguish Madouling derived from Aristolochia species (Aristolochiaceae) from the substitute derived from Cardiocrinum species (Liliceae) [ 20 ] . Identi fi cation of materials of different genera can also be achieved by DNA techniques. AP-PCR and RAPD were able to discriminate materials belonging to eight genera in family Asteraceae and identify the herbs Kudidan and Pugongying [ 23, 24 ]. Similarly, PCR-RFLP may be applied to differentiate four Codonopsis species (Campanulaceae) from two adulterants derived from Campanumoea and Platycodon species in family Campanulaceae [ 46 ]. DNA sequencing of ITS region was applied to distinguish 16 medicinal Dendrobium species from Pholidota species in the same family Orchidaceae [ 64 ] . Although DNA sequencing is useful to differentiate samples derived from distantly related species, such as at the family and genus levels, choosing a suitable DNA region is crucial. Some DNA regions, such as ITS and 5S, evolve rapidly and their sequence similarities at species level in some families are low. For example, the sequence similarity of ITS and 5S regions among Muxiang species (Asteraceae) and the toxic adulterants in Aristolochiaceae were only 56–58% and 20–30%, respectively [ 65 ]. Although such low similarity does not affect the differentiation of samples in different families, it may make sequence alignment and phylogenetic tree construction dif fi cult.

3.3.2 Discrimination at Inter- and Intra-species Levels

One of the major advantages of molecular identi fi cation is its high resolutio which allows differentiation samples at inter- or intra-species level. DNA fingerprinting, such as AP-PCR, RAPD, SCAR, DALP, and AFLP, readily differentiated closely related species of P. ginseng from P. notoginseng [ 18, 34, 38, 50 ]. DNA microarray with hybridization probes designed based on ITS and 5S sequences successfully detected several medicinal Dendrobium species [ 61, 62 ].

Choosing an appropriate DNA region with high variability and discrimination power is crucial for differentiation of closely related species by DNA sequencing.

For example, trnL is a relatively conserved region which could differentiate medicinal Stemona species (Stemonaceae) from adulterants derived from Asparagus species (Asparagaceae) but failed to discriminate the medicinal species (S. japonica , S. sessilifolia and S. tuberosa) and another closely related species S. parvi fl ora [ 71 ]. On the contrary, the ITS, 5S and trnH-psbA regions are highly varied regions which are commonly used for identification at species level. The ITS region is varied enough to discriminate all 16 medicinal Dendrobium species with inter-specific divergences ranging from 2 to 17% [ 64 ]. This region was also used to authenticate Baihuasheshecao derived from Hedyotis diffusa (Rubiaceae) and resolved all the 14 Hedyotis species studied [ 66 ] . In fact, the ITS-2 region is highly varied and found useful for discriminating most medicinal species and therefore has recently been proposed to be a DNA barcode for medicinal plants [ 13 ]. Although ITS shows high sequence variability among species and is the most frequently used region for species identi fi cation of herbal medicinal materials, the presence of multiple copies, which may be non-homogeneous, and the problem of secondary structure resulting in poor-quality sequence data are major drawbacks [ 76, 77 ]. Molecular cloning prior to DNA sequencing is necessary to solve these problems. Besides, fungal contamination is common in herbal medicinal materials and would interfere proper ampli fi cation of target ITS sequences by universal primers. Specially designed plant-speci fi c primers should be used in such conditions. The 5S region is a highly varied region and frequently used for species and subspecies differentiation. It readily discriminated Swertia mussotii from S. chirayita , S. franchetiana, and S. wolfgangiana with interspecific divergences ranged from 31 to 65% [ 70 ]. It also differentiated Dangshen derived from Codonopsis pilosula and C. pilosula var. modesta with intra-speci fi c similarity of 95–98%, respectively, and interspecific similarity ranged from 70 to 73% [ 69 ] .

In our experience, however, the sequence of 5S region is sometimes too varied, making it dif fi cult for sequence alignment. Moreover, this region has multiple copies and molecular cloning prior to sequencing is essential. TrnH-psbA region is a complementary DNA barcoding region showing the highest amplification successful rate and discrimination rate among 9 tested loci [ 15, 78 ]. It is used to identify 19 Aconitum species with an average inter-specific similarity of 85% [ 68 ]. The two closely related medicinal species, A. carmichaeli and A. kusnezoffii, were clearly distinguished by a 56 bp sequence inversion in their trnH-psbA sequences. A disadvantage of the trnH-psbA region is the presence of poly-A structure which reduces the successful rate of DNA sequencing. Besides, sequence alignment may be dif fi cult due to the frequent presence of nucleotide insertion and deletion. In spite of the highly discriminative ability at species level, trnHpsbA could not resolve the relationship between Cardiocrinum giganteum and its variety C. giganteum var. yunnanense , but the trnL-F region could [ 20 ] . This example demonstrated that there is no single universal locus suitable for differentiating all taxa at different levels. Searching for a suitable region that suits the purpose is not avoidable.

3.3.3 Discrimination among Cultivars and Geographical Culture Origins

Herbal medicinal materials derived from various cultivars or collected from different geographical origins may be traced using molecular techniques. For example, the herb Huajuhong is derived from Citrus grandis or its cultivar C. grandis “Tomentosa.” ISSR fi ngerprinting using six primers generated 57 DNA fragments which readily differentiated four samples of Citrus grandis from 15 samples of C. grandis “Tomentosa.” Although there were a few nucleotide substitutions in their ITS sequences, cladistic analysis showed that ITS was unable to differentiate these cultivars as they could not form distinct clusters [ 40 ] . AP-PCR fingerprints of Dangshen collected from different geographical origins in China showed that samples from Sichuan and Hubei generated a characteristic fragment of 0.8 kb using the primer OPC-02. Specific fragments of 1.15, 0.63, and 1.15 kb were obtained in samples from Shanxi, Sichuan, and Gansu, respectively, using the primer OPC-04. The AP-PCR primer OPC-05 ampli fied specific fragments of 1.25 and 1.6 kb for samples from Gansu, while a 0.9 kb fragment is characteristic in Hubei samples [ 25 ] .