Among the many DNA-based methods with potential application in the authentication of medicinal plant materials, DNA barcoding has attracted particular attention recently and has been touted by some as a superior authentication method. In order to assess the utility of DNA barcoding in the quality control of raw materials, extracts and finished herbal medicinal products, we examined four universal barcodes (matK, rbcL, trnH-psbA and ITS2) in 17 authentic raw materials and in extracts and tablets made from those raw materials. Authentic herbarium specimens were also included. Samples were chemically profiled and marker compounds identified by UPLC-MS. DNA was extracted with the NucleoSpin® 96 Plant II kit and each genetic target amplified by PCR, and Sanger sequenced. Results were compared to sequences in GenBank using BLAST. For raw materials, the rate of successful amplification and sequencing varied for the four targets but was low overall (highest for rbcL and trnH-psbA, 25% and 33% respectively). Identification to species was achieved for 18% (trnH-psbA), 15% (rbcL), 5% (matK) and 0% (ITS2) across all samples. Of dried materials, 53% were correctly identified by at least one barcode. Importantly, only one extract yielded amplifiable DNA; 20 extracts did not and could not be identified by any of the four barcodes. In contrast, the phytochemical integrity of all samples was confirmed by UPLC-MS.
The results demonstrate that DNA barcoding using universal barcode sequences, which are relatively long, is not suitable for routine authentication of raw materials, extracts or finished herbal medicinal products, most likely due to DNA degradation arising from drying and processing. The use of shorter, species-specific sequences may yield better results. In order to avoid confusion, we recommended the term 'DNA barcoding' be applied only when universal (not species-specific) target sequences are used, in line with the original definition of DNA barcoding [1].