Editor’s note: Less than 24 hours after the New York Times broke the news of New York Attorney General Eric T. Schneiderman’s investigation of herbal supplements on February 3, 2015, herbal science experts, natural products industry trade associations, analytical experts, and others issued press releases, member advisories, and letters to the editor questioning the NY AG-commissioned scientist’s use of DNA barcoding to authenticate several store-brand herbal supplements from major US retailers. According to Schneiderman’s office, the analysis revealed a lack of the labeled herbs or the presence of unlisted ingredients in a majority of the products. However, as numerous analytical and regulatory experts have noted in the weeks since the announcement, DNA barcoding is a developing technology that is neither appropriate for the authentication of botanical extracts nor approved by the US Food and Drug Administration (FDA) for such purposes. In the weeks following the NY AG’s press release, two white papers from expert authors were published outlining the limitations of DNA-barcoding for botanicals. ABC Chief Science Officer Stefan Gafner, PhD — whose doctorate is in the field of pharmacognosy and natural products chemistry — summarizes and explains these documents below.

White Paper by Four Analytical Experts Evaluates Capabilities and Limitations of DNA Barcoding for Testing of Botanical Dietary Supplements

On March 11, 2015, a white paper was published on the capabilities and limitations of DNA barcode testing for the identification of botanical dietary supplement ingredients in finished products.¹ The document indicates that the conclusions from the DNA barcoding tests commissioned by the NY AG are not justified.

The white paper is authored by four highly respected experts in the fields of DNA and other analytical methods for botanical dietary supplements: Danica T. Harbaugh Reynaud, PhD, CEO of AuthenTechnologies LLC, in Berkeley, California, a leading laboratory in DNA authentication of botanicals; Brent D. Mishler, PhD, professor in the Department of Integrative Biology at the University of California at Berkeley; James Neal-Kababick, Director of Flora Research in Grants Pass, Oregon, a leading laboratory for botanical dietary supplement analysis; and Paula N. Brown, PhD, director of the Natural Health and Food Products Research Group at the British Columbia Institute of Technology in Burnaby, British Columbia. The document was underwritten by four dietary supplement trade organizations: the American Herbal Products Association (AHPA), the Consumer Healthcare Products Association (CHPA), the Council for Responsible Nutrition (CRN), and the United Natural Products Alliance (UNPA).

As stated on the first page of the document, the purpose of the paper is to provide an “expert assessment of the capabilities and limitations of DNA barcoding for botanical dietary supplement authentication to inform assessments of the applicability and accuracy of DNA test results.”

The focus of the white paper — clearly written as a critique of Schneiderman’s test results² — is less on the capabilities of DNA barcoding, but it does provide a good overview of the limitations and possible pitfalls, which are often overlooked in papers published in the scientific literature (e.g., the now largely discredited report from 2013 on the authentication of herbal ingredients in dietary supplements authored by Steven Newmaster, PhD, et al. at the University of Guelph,³ which prompted the NY AG to look into the quality of botanical dietary supplements). It is important to understand that DNA barcoding is only one of many DNA-based methods for identification of herbal ingredients, and as such, the comments made in the paper are applicable only to DNA barcoding and not to other, more specific genetic methods used for identification of plant materials.

The white paper first explains the principles underlying DNA barcoding and the main experimental steps involved in the technique. One of the crucial steps in using genetic methods for identification of plant species is to find the gene region(s) that contain enough variation to allow the species of interest to be distinguished from other species. The choice of the best regions for DNA barcoding is still a matter of debate, and, according to a recent review of the peer-reviewed literature, a total of 17 different regions are reportedly used by scientists for authentication of plant material.⁴ After extraction from the sample, the DNA is amplified, which means that a large number of copies of the chosen gene regions are obtained using polymerase chain reaction (PCR). The resulting genes are sequenced — a process that determines the order in which the four components of DNA (adenine, cytosine, guanine, and thymine) are aligned — and compared to reference sequences from authentic botanical material, or to sequences in publicly available databases like GenBank.

The authors stress that the applicability of DNA barcoding has to be evaluated on a case-by-case basis and comment that “failure to appreciate the effects of extraction on botanical materials indicates the general lack of understanding by scientists performing DNA testing on finished products and the importance of the distinction between raw material and processed materials, especially extracts.”

There is a relationship between the level of processing of herbal dietary supplements and the length of the DNA fragments; the more processed the ingredients are, the more fragmented the original DNA becomes. As a result, DNA barcoding of highly processed botanical materials often results in findings indicating the absence of any DNA, or the presence of DNA from excipients or fillers.

The paper continues with a number of key aspects to be looked at when considering DNA-based methods for herbal ingredient identification. The identification of the appropriate genes to be amplified has to be carefully evaluated. Also important is the inclusion of control samples: appropriate positive and negative controls must be used for each of the steps in the testing process to avoid false positive or false negative results. Another key aspect is the use of authentic reference materials. Ideally, the resulting DNA sequences should be compared to sequences of authenticated fresh or dried whole plant material. Comparison with publicly available databases, particularly GenBank, carries the risk of identification errors due to potentially erroneous or missing entries in the databases. Finally, the choice of the right algorithm for the analysis of the DNA sequence is crucial. The most reliable results are obtained when specific alogorithms are used that consider species-specific DNA features rather than those that are based on the similarity of the match with sequences from public databases. As with any other analytical method, an appropriate validation of the DNA barcoding method needs to be performed.

A few pages in the document are devoted to answering questions with regard to the analysis performed by James A. Schulte II, PhD, on the samples collected by the NY AG. Since the NY AG has not released any details on the testing method, most of the questions raised cannot be answered with certainty, but based on the little information that is available, the authors conclude that the results of the NY AG’s investigations are invalid. It remains to be seen if the results obtained from chemical testing by independent analytical laboratories will confirm the authors’ suspicions that the cease-and-desist letters were based on erroneous results.

While it is impossible to give a comprehensive overview on such a complex topic as genetic authentication of botanical ingredients in a short paper, the authors have been able to explain many of the caveats of using DNA barcoding in a way that is accessible to the reader, and they have given ample evidence as to why DNA barcoding is generally not appropriate for highly processed botanical ingredients such as herbal extracts. When in doubt, researchers may prefer to rely on well-established, appropriate, and scientifically valid analytical methods, e.g., those provided in official pharmacopeias and other authoritative compendia and/or scientific journals.

NPA White Paper by Former FDA Regulators Stresses Importance of Validation by Chemical Analyses and Highlights Issues with the Detection of Processed Plant DNA

The Natural Products Association released a white paper on March 17, 2015, questioning DNA barcoding as a valid method to determine the identity of herbal dietary supplement ingredients.⁵

The document was written by Daniel Fabricant, PhD, CEO of NPA, and Corey Hilmas, MD, PhD, NPA’s senior vice president of scientific and regulatory affairs. Dr. Fabricant is the former director of the FDA’s Division of Dietary Supplement Programs and Dr. Hilmas served as chief of the FDA’s Dietary Supplement Regulation Implementation Branch.

The NPA publication came less than a week after the publication of the previously described white paper commissioned by AHPA, CHPA, CRN, and UNPA.

Drs. Fabricant and Hilmas start with a brief summary of the investigation into herbal dietary supplements by the NY AG and a call for increased transparency and the release of more details on the testing used as a basis for the cease-and-desist letters sent to GNC, Target, Walgreens, and Walmart. This is followed by an executive summary listing a number of key limitations of the DNA barcoding technology: e.g., the lack of high-quality DNA in many botanical extracts, which may prevent the ingredients listed on the dietary supplement labels to be detected; the inability to quantify the amounts of the alleged contaminants; the possibility for cross-contamination; and the absence of a large library with DNA sequences from properly authenticated and vouchered plant material.

The white paper then goes into the details of DNA barcoding and the various steps involved along the process. The authors stress the importance of removing certain secondary plant metabolites ahead of the gene amplification, since these metabolites can interfere with the PCR used to amplify the DNA. One paragraph is devoted to the selection of the gene regions (loci) that are best suited to distinguish the various plant species from each other. The authors provide some historical background about the loci selection process, and they conclude that the combination of two loci suggested by the Consortium for the Barcode of Life (CBOL) discriminates only to species level at a probability of 72%. They also note that the species delimitation (i.e., the process by which boundaries among species are determined) for plants is not always very straightforward due to hybridization within many plant genera, which is not the case for the animal kingdom.

A particular concern with DNA barcoding methods is the use of publicly available DNA sequence databases, due in part to the lack of vouchered specimens linking the DNA sequence to the specimen. (Other shortcomings of public DNA databases not mentioned in the paper include the occurrence of errors in specimen records and missing sequences from loci that prevent the use of the two-loci system proposed by CBOL.) Based on the data released by the NY AG, the authors write that “all we can conclude at this point is that [Dr. Schulte] used publicly available DNA sequences that did not involve voucher specimen, leading to erroneous results.”

A short paragraph is dedicated to the impact of processing on the plant DNA, explaining that the process of extraction, as well as the solvents used in the process, impact the quality and amount of DNA that can be detected. Other processes that impact DNA quality are excessive heat treatments, or submission of the material to UV light, often leading to only small DNA fragments remaining. These fragments are shorter in length than what is required to generate a useful barcode. Finally, the authors question the expertise of Dr. Schulte and the accreditation of his laboratory.

The NPA white paper ends with a discussion of appropriate test methods for botanical dietary supplements. In the opinion of the authors, chromatographic methods are the gold standard for the identification of botanicals — in particular the approach where chemical fingerprints are obtained by high-performance liquid chromatography combined with mass spectrometric detection (HPLC-MS) — and they suggest that DNA barcoding should never be used without a chemical fingerprint method for botanical ingredient authentication.

The white paper raises the same questions about the applicability of DNA barcoding to herbal extracts as the white paper written by Harbaugh Reynaud et al., but there is some additional information in the NPA paper, e.g., the potential for interference of secondary metabolites with the PCR and the historical perspective on the development of DNA barcoding. Some of the details are quite technical, which may make it difficult to follow for readers who are unfamiliar with DNA-based methods. In addition, there are a few statements that are either unclear or potentially confusing, e.g., “moreover, if the study actually was scientifically valid and fit for purpose, the science would appear to be inconsistent with the dietary supplement current Good Manufacturing Practices,” or “chromatographic methods are the gold standard as stand-alone methods for the identification of botanicals in diverse matrices such as finished dietary supplement products. These tests produce a chemical fingerprint, corresponding to masses, which are characteristic of the phytochemical profile for a particular species.”

Overall, the authors conclude that DNA barcoding is not an appropriate choice to determine the identity of herbal extracts in dietary supplements. Or, as they put it, “in the end, if the method is unable to detect DNA, all it can conclude is that it was unable to detect a DNA barcode for a particular botanical extract.”

References

1. Harbaugh Reynaud DT, Mischler BD, Neal-Kababick J, Brown, PN. The capabilities and limitations of DNA barcoding of dietary supplements [white paper]. March 2015. Available at: www.crnusa.org/NYAG/The-Capabilities-and-Limitations-of-DNA-Testing-FINAL-3-10-2015.pdf. Accessed March 20, 2015.
2. New York State Office of the Attorney General. A.G. Schneiderman Asks Major Retailers To Halt Sales Of Certain Herbal Supplements As DNA Tests Fail To Detect Plant Materials Listed On Majority Of Products Tested. [press release]. February 3, 2015. Available at: www.ag.ny.gov/press-release/ag-schneiderman-asks-major-retailers-halt-sales-certain-herbal-supplements-dna-tests. Accessed March 11, 2015.
3. Newmaster SG, Grguric M, Shanmughanandhan D, Ramalingam S, Ragupathy S. DNA barcoding detects contamination and substitution. BMC Medicine. 2013;11:222. doi:10.1186/1741-7015-11-222. Available at: www.biomedcentral.com/content/pdf/1741-7015-11-222.pdf. Accessed March 16, 2015.
4. Techen N, Parveen I, Pan Z, Khan IA. DNA barcoding of medicinal plant material for identification. Curr Opin Biotechnol. 2014;25:103-110. Abstract available at: www.sciencedirect.com/science/article/pii/S0958166913006721. Accessed March 16, 2015.
5. Fabricant DS, Hilmas CJ. DNA barcoding for botanical authentication [white paper]. March 2015. Available at: www.npainfo.org/App_Themes/NPA/docs/regulatoryLegislative/White%20Paper/NPAWhitePaper_DNABarcoding.pdf. Accessed March 20, 2015.