Identification using Morphology

Industry participants have long sought basic assurances from suppliers that exchanged goods and materials are properly identified and labeled in a manner that is useful for guaranteeing authenticity.⁴ Historically, this was a process that could be performed by examining whole plants and/or plant parts that were essentially intact after harvest. An experienced botanist or trained herbalist could use macromorphological features to confirm the identity of the botanical merchandise. To the extent possible, this traditional approach remains in practice,³ but changes in the type of botanical materials traded in international and domestic herbal commerce, a varying range of intermediate suppliers, and a decreasing population of herbal trade-related botanical taxonomists, have created a need for supplemental—and in some instances more precise—methods of identification.

The typical practicing taxonomist, such as those associated with university herbaria, natural history museums, botanical gardens, or public land stewardships, generally classify and apply names to plants with the assistance of reproductive characters, such as floral and fruit morphology.⁵ A high quality scientific botanical specimen is one that includes the date and location of collection, the collector’s name, and a representative sample of a plant’s intact leaves, stems, flowers or fruits, and often roots that have been carefully arranged between sheets of herbarium paper, and felt, pressed, and dried soon after collection. In some cases, quality specimens include mature fruits, bark, or other organs and parts that are useful to differentiate the material from its closest relatives. This approach has been maintained for centuries because of its effectiveness for providing taxonomists with a means to reliably identify a specimen. However, within the world’s herbarium collections, notable reserves of plant specimens are maintained that—although collected by experienced botanists—do not include adequate material critical for definitive taxonomic diagnosis. In any given herbarium, these specimens are generally classified to the extent possible (at the genus or family level) and placed in folders that are labeled “Indet,” an abbreviation for “indeterminate.”

The use of morphological characters for herbal identification in a commercial setting remains a viable and important approach.⁶ However, due to the fact that identification of many species requires relatively intact botanical specimens including reproductive structures, the method has limited utility across the broader herbal supply chain. Botanical material may lack key diagnostic characters at the time of harvest or the material may be acquired from primary or secondary producers in the form of partially processed, shredded or powdered vegetation, or as a processed liquid extract. In the case of cut and powdered botanical materials, the use of microscopy is often necessary. However, the availability of QC personnel who are adequately trained in microscopy is limited, as are adequate reference materials. [Editor’s note: To help meet this growing need, the American Herbal Pharmacopoeia will be publishing an extensive microscopy text soon.]

Identification using Phytochemistry

An alternative to the morphology-based method of identification uses a comparison of the sample’s phytochemistry to that of a reference standard created from botanical reference material of documented identity. In this widely used approach, well-established chemical separation techniques provide a profile that is characteristic of a species. Methods such as thin-layer chromatography (TLC) and high-performance liquid chromatography (HPLC) separate the constituent natural chemical compounds into distinctive profiles.^7,8 Other techniques such as mass spectroscopy (MS) and nuclear magnetic resonance (NMR) can provide distinct profiles without necessarily identifying individual phytochemical compounds.⁹ The presence of a specific profile indicating the levels of various compounds in comparison to a species’ reference standard provides a common basis for a confirmatory test of identification.²

Phytochemical approaches to botanical characterization are essential for investigating the presence and concentration of naturally-occurring chemical constituents in botanical materials and can also be used to help detect the presence (or absence) of adulterants and/or contaminants (e.g., heavy metals, pesticides, undisclosed pharmaceutical drugs, etc.). However, complete reliance on chemical profiles for herbal identification and QA/QC remains problematic; such methods have certain limitations.

The production and accumulation of secondary metabolites is a natural process in plants—a process that is dependent on the interaction of a plant’s genotype and its growing conditions. The genetic composition of an individual plant defines its ability to create specific compounds.^10,11 However, the same plant will produce different levels of secondary metabolites from year to year based on the season’s rainfall, temperature, nutrient availability, and other environmental factors.^12-14 The combined effect of these environmental and genetic variables creates a source of notable variation in the presence and concentration of natural compounds within members of a species, such as from plant to plant or population to population.^15-19 From a taxonomic perspective, the reliance on features (i.e., marker compounds) that may exhibit such wide variation creates a risk for applying the wrong name to a substrate or improperly authenticating botanical material.

Additionally, the use of marker compounds may be problematic in situations where the taxonomic range of the compound’s occurrence is not well understood. Related species potentially substituted for a given botanical may not have been thoroughly evaluated to determine if they possess similar chemistry. Investigation into the production of secondary compounds in plants is an active area of research among many members of the botanical community, but definitive data are not yet available on the full suite of natural compounds that each botanical species possesses. Therefore, use of marker compounds as a diagnostic indicator in the absence of definitive comparative data also may lead to inappropriate taxonomic diagnoses.^1,20

Identification using Nucleotide Sequences

The use of DNA sequence data for identity testing of plants has the potential to markedly improve QA and QC for herbal dietary ingredients.^21-24 Plants contain DNA in 3 cellular compartments: the nucleus, chloroplast, and mitochondrion. Specific stretches of DNA contained in these genomes, particularly in areas of the chloroplast and nuclear genomes, are useful for identification.

Over the last 10 to 15 years, various approaches using nucleotide data have been developed to help characterize variation within plant populations and plant species.^21,23 Many of these methods have relied on the use of DNA fragments (as opposed to specific DNA sequences) to characterize or define differences among individuals, populations, or species. Methods such as restriction fragment length polymorphisms (RFLP), randomly amplified polymorphic DNA (RAPD), or amplified fragment length polymorphisms (AFLP), are used to amplify numerous randomly located sections of an individual’s genome. These fragments are separated by gel electrophoresis to determine their length. DNA fragments of given lengths are the characters used to compare samples to each other or to reference materials. The overall genetic similarity between 2 samples is determined by the degree to which they share fragments of identical lengths, although it is normal for some variation to occur within and among species.

Without prior knowledge of the name of the substrate being tested and an existing reference library for fragment analysis, the broader applicability of DNA fragment-based analytical approaches to botanical identity testing is limited. Furthermore, these methods work best with samples consisting of fresh material comprised of a single species. In practice, many herbal preparations consist of more than one species and often contain degraded DNA that does not meet the rather stringent requirements for accurate and reliable analysis using fragment-based analytical approaches.

A more robust technique for botanical identification relies on the use of specific DNA sequences to uniquely characterize and differentiate plant material.²⁵ The specific sequence composition of select regions of DNA, as opposed to its overall length, is effective for identifying or discriminating among botanical species, varieties, populations, and even individual plants. Imagine trying to use the length of a telephone number to differentiate between several people in a phone list. However, when the individual numbers are used to dial a contact, each person can be specifically reached. The same model can be applied when trying to understand how individual sequences can be used to differentiate plants. In the case of DNA, 4 types of bases (adenine, thymine, guanine, and cytosine, symbolized by A, T, G, and C, respectively) constitute a plant’s DNA sequence in the same way that 10 numerals are available to compose a person’s telephone number. The arrangement or order of these bases, rather than the number of characters that make up the sequence, can accurately identify the plant material. Also, just as longer phone numbers have been adopted by the Bell system to accommodate a need to identify more callers, longer nucleotide sequences can be used to differentiate plants.

The term DNA barcode has been used to describe DNA sequences that can uniquely identify a species.^26,27 Enthusiastic supporters of DNA barcodes imagine a framework where all plants can be quickly identified using a well-defined set of DNA sequence data.²⁸ However, critics of DNA barcodes are concerned that this approach is too simplistic and fails to accurately classify all botanical species.^29,30 Often these debates become charged with theoretical implications that may never be tested in real-world applications. In the interest of side-stepping the DNA barcode debate, a practical solution is proposed for the identification of plants associated with herbal dietary supplements that employs the use of DNA sequence data within a broader molecular systematic context—a framework that avoids the potential pitfalls of simple DNA barcodes³¹ by using DNA sequence data in a more comprehensive phylogenetic system to identify targets and differentiate them from their closest relatives.^32,33

The fundamental utility of DNA data in molecular systematic applications, such as inferring evolutionary relationships or identity determination, relies on the basic principles of inheritance where, in most cases, exact copies of DNA are passed from one generation to the next. In rare instances, mutations or mistakes in DNA replication occur during gametogenesis, or the process that creates sex cells. Mutations can take many forms, including the substitution, deletion, or insertion of one or more nucleotides, as well as the repetition or inversion of 2 or more nucleotides at any given location. Generally speaking, a new mutation that arises in the sex cells of plants is passed on to offspring by sexual reproduction. If the mutation is not lethal, the offspring may mature and pass the mutation to its progeny. The rate at which any new mutation spreads through a population depends on the many forces associated with natural selection. Over time and many generations, a new mutation may spread to all members of a population, and much later still, occur in all members of a species.

The older a mutation is in an evolutionary sense, the deeper it will occur in a lineage of plants. For example, a mutation may be shared by all members of the same genus or family of plants. Discovering and noting these mutations helps to classify plants using molecular systematics. In this approach, plants that share mutations of the same origin are classified together. The more mutations that are shared, the more evolutionary history is in common among the organisms.

The goal of a DNA sequence-based approach for any identity related application in the herbal industry is to exploit the mutations that are shared across one species, variety, or population but are absent from their closest relatives. This task is relatively simple in lineages like ginkgo (Gingko biloba, Ginkgoaceae), where a single species represents an ancient lineage that has no closely related living relatives. However, the work becomes more difficult, but tractable, in closely related species with apparently shorter evolutionary histories, such as those observed in several genera of grasses and many other flowering plants.

The idea to include or exclude an individual plant from a particular taxonomic group based on its underlying DNA mutations is a common practice in biological systematics.^32-36 Performing the work of DNA-based analysis on herbal specimens is similar to putting this taxonomic approach into practice for industry stakeholders. The task of sorting through numerous vouchered specimens and choosing genes and nucleotide sequences that adequately delineate the taxonomic boundaries for each herb will take time and continued effort. Overall, this DNA-based systematic endeavor aims to adequately recognize important differences between plants of interest and their closest relatives, as well as exceed the established limits of DNA barcoding.³¹

Overview of the DNA-Based Approach

Methods for determining DNA sequence data from plants have been perfected and are now routinely performed in many laboratories around the world.²⁵ The general method includes 3 main steps: (1) Disruption of cell membranes, followed by isolation and purification of the nucleic acids; (2) Amplification of targeted sequence regions in the purified DNA using the polymerase chain reaction; and (3) Sanger sequencing of the amplified product. Once the DNA sequence data have been determined, identity testing can be performed by comparing the test sample sequence data to the sequence data from reference samples.

Hundreds of thousands of reference DNA sequences are maintained in the National Institutes of Health Genbank database (www.ncbi.nlm. nih.gov/Genbank). Nucleotide sequence data from unknown botanical samples can be quickly characterized by using the online comparative tool BLAST (the Basic Local Alignment Search Tool).³⁷ The BLAST program identifies sequences in the Genbank database that are most similar to the test sequence by applying a statistical measure of similarity between reference nucleotide data and the test data. Once the most similar Genbank sequences are identified, phylogenetic analysis can be performed using a combined dataset comprised of Genbank and test sample data. This process can quickly characterize unknown or unnamed samples by identifying well-characterized close relatives of the material. If further analysis of the unnamed sample is required, more DNA sequences can be added to the dataset using reference sequences from other members of the unnamed plant’s taxonomic group, which was defined by the initial round of testing.

Despite the unparalleled utility that Genbank offers for comparative analysis, a relatively small fraction of its data have imperfections stemming from occasional laboratory errors or improperly labeled specimens.³⁸ Depending on the specific implications of any final analysis, individual taxonomic diagnoses should be confirmed using properly verified data from vouchered reference specimens. (Readers should note that the Genbank application discussed here is for purposes of characterizing samples of unknown identity, such as potential adulterants, trace contaminants containing DNA [i.e., plants, animals, fungi, protozoa] or samples that simply became disconnected from their label.)

The Practical Utility of DNA in Herbal Samples

Nucleotide sequence-based approaches appear to successfully overcome many of the shortcomings associated with morphological and chemical identification techniques, including identification of both single element and mixture preparations. All living plant cells contain DNA and, except for extremely rare events, cells from different portions of a plant (i.e., roots, stems, and leaves) each contain DNA with the same sequences. As such, one plant cell, or a portion of a plant cell, contains adequate DNA for testing and analysis. Very small quantities of plant material are thus required for successful identification of substrates to determine taxonomy. Further, even after treatment in many harsh processing methods, fragments of DNA often are still available for analysis. Examples of successful DNA-based identification have been reported from fossils, pollen, botanical extracts, and refined oils.^39-42

A number of papers have been published detailing the successful application of DNA-based methods for confirming the identity or purity of herbal products. For instance, the botanical Hypericum perforatum (Clusiaceae), or St. Johns wort, is morphologically and chemically similar to other closely related Hypericum species, creating analytical challenges for its identification using morphological and phytochemical methods.¹ Howard et al. have shown that DNA-based methods can be used to effectively identify H. perforatum in herbal preparations.⁴³ DNA-based methods also have been used to effectively differentiate the economically important herbs Korean or Chinese ginseng (Panax ginseng, Araliaceae) and American ginseng (P. quinquefolius), which appear morphologically similar in powdered forms.⁴⁴ In support of purity determination for Chinese star anise (Illicium vernum, Illiciaceae), DNA sequence data has been shown as a useful tool to detect the presence of its known toxic adulterant Japanese star anise (I. anisatum).⁴⁵

Plant mixtures often represent particular challenges to morphological or chemical approaches of confirming identity. Assigning biochemical or morphological attributes that can be used reliably to determine individual botanical mixture components becomes increasingly difficult, if not impossible, as the number of plants contributing to a mixture increases.⁴⁶

DNA-based approaches are useful for deconstructing mixtures of 2 or more plants and could become useful for confirming the identity of components in blended herbal products. Recombinant DNA techniques that are applied in microbiology studies and environmental investigations can be applied to the identification of plant mixtures.⁴⁷ In these approaches, DNA is extracted from botanical mixtures in the same manner applied to single-element samples, and targeted regions of DNA are amplified from the extracted material. Prior to sequencing the amplified product, individual copies of DNA are combined with longer engineered fragments of DNA called vectors. The joined pieces of DNA are then inserted into bacteria that have been chemically treated and specifically prepared to receive vector DNA, a process called transformation. Generally, the engineered fragment of DNA will contain a gene that confers resistance to antibiotic drugs, e.g., kanamycin or ampicillin. Successfully transformed bacteria are plated onto a nutrient agar plate containing kanamycin or ampicillin. Bacteria that contain a vector (and the gene for antibiotic resistance) will grow, with each surviving colony representing one fragment of DNA amplified from the original sample mixture. The process is completed when the vector is purified from each bacterial colony and sequenced. Using this approach, botanical species present in any meaningful concentration can be detected in a mixture of 2 or more plants. The ratio of plants in each sample indicates how many colonies need to be screened to give a high likelihood of obtaining a sequence from each plant in the mixture.

The time and cost surrounding DNA-based analytical techniques have steadily decreased over the last 20 years. Procedures and nucleotide sequence data collection methods that once required days or weeks can now be completed in hours. Technological advances, such as robotics, have also increased levels of laboratory productivity, which allow fewer technicians to process more samples. Furthermore, the world’s schools of higher learning continue to produce botanists that are trained in molecular systematics and laboratory sciences—permitting competition to foster a continued trend toward faster and lower-cost DNA-based herbal testing services. Today, routine DNA-based analytical testing can be performed in a timely manner with service pricing in the neighborhood of morphological and phytochemical methods.

Identity Testing vs. Quality Determination

The availability of DNA-based identity testing procedures for plants used in herbal dietary supplements provides an expanded paradigm that may distinguish botanical ingredient identity from botanical ingredient quality. The preponderance of processed botanical material in the herbal supply chain precludes the broader, reliable application of morphology-based identity testing, and by default the herbal community relies largely on phytochemical measurements to ascertain sample identity. However, the scientific literature indicates that biochemical profiles are not ideal as a routine measure of identity, due to the inherent variation in the production of secondary compounds,^12,48 the possibility of diagnostic indicator compounds occurring in more than one botanical species,^1,20 as well as the occurrence of botanical mixtures in many herbal preparations.^46,49

The adoption and integration of DNA-based testing by the herbal dietary supplement community can provide an independent, sensitive, and specific means to ascertain the specific identity of botanical materials, addressing the FDA’s fundamental requirement for herbal product ingredient identification.

The opportunity to remove the ambiguity associated with botanical ingredient identity will allow members of the herbal dietary supplement community to more prominently recognize 2 distinct attributes of herbal ingredients: identity as well as quality. The combination of DNA-based identification (or, where possible, morphology-based authentication) complemented by measurements of biochemical compound concentration contained in a particular lot of raw material would be especially beneficial. These correlated but distinct measures will provide stakeholders with well-defined tools to discriminate and understand differences between physically similar sets of raw materials.

Matt Cimino, PhD, is a practicing molecular systematist and the DNA laboratory director at Stoney Forensic, Inc. in Chantilly, Virginia.

Acknowledgements

This work was supported by the National Institutes of Health (NIH) Office of Dietary Supplements (ODS) and the National Center for Complementary and Alternative Medicine (NCCAM) Grant R44AT1556. The author would like to thank the peer reviewers of this manuscript for their thoughtful comments.