Combining DNA Barcoding and HPLC Fingerprints to Trace Species of an Important Traditional Chinese Medicine Fritillariae Bulbus

Fritillariae Bulbus is a precious Chinese herbal medicine that is grown at high elevation and used to relieve coughs, remove phlegm, and nourish the lungs. Historically, Fritillariae Bulbus has been divided into two odourless crude drugs: Fritillariae Cirrhosae Bulbus and Fritillariae Thunbergii Bulbus. However, now the Chinese Pharmacopoeia has described five Fritillariae Bulbus—the new additions include Fritillariae Pallidiflorae Bulbus, Fritillariae Ussuriensis Bulbus, and Fritillariae Hupehensis Bulbus. Because the morphology of dried Fritillariae Bulbus is similar, it is difficult to accurately identify the different types of Fritillariae Bulbus. In the current study, we develop a method combining DNA barcoding and high-performance liquid chromatography (HPLC) to help distinguish Fritillariae Cirrhosae Bulbus from other Fritillariae Bulbus and guarantee species traceability of the five types of Fritillariae Bulbus. We report on the validation of an integrated analysis method for plant species identification using DNA barcoding that is based on genetic distance, identification efficiency, inter- and intra-specific variation, calculated nearest distance, neighbour-joining tree and barcoding gap. Our results show that the DNA barcoding data successfully identified the five Fritillariae Bulbus by internal transcribed spacer region (ITS) and ITS2, with the ability to distinguish the species origin of these Fritillariae Bulbus. ITS2 can serve as a potentially useful DNA barcode for the Fritillaria species. Additionally, the effective chemical constituents are identified by HPLC combined with a chemical identification method to classify Fritillaria. The HPLC fingerprint data and HCA (hierarchical clustering analysis) show that Fritillariae Cirrhosae Bulbus is clearly different from Fritillariae Thunbergii Bulbus and Fritillariae Hupehensis Bulbus, but there is no difference between Fritillariae Cirrhosae Bulbus, Fritillariae Ussuriensis Bulbus, and Fritillariae Pallidiflorae Bulbus. These results show that DNA barcoding and HPLC fingerprinting can discriminate between the five Fritillariae Bulbus types and trace species to identify related species that are genetically similar.


Efficiency of PCR Amplification and Sequencing
The three loci-ITS, ITS2, and psbA-trnH-all showed PCR amplification and high sequencing efficiency (100%). The locus matK showed no PCR amplification and a sequencing efficiency of 0%. Moreover, the effective sequence ratios of ITS, ITS2, and psbA-trnH were the same (100%).

Genetic Divergence Determination
The lengths of the aligned sequence (base pairs) and variable sites for the ITS and ITS2 regions were 610/46 and 235/18, respectively. The results are shown in Figures 1 and 2. Six parameters were used to characterise inter-and intraspecific divergence. The intraspecific distance and interspecific distance of ITS were 0.0000-0.0145 and 0.0020-0.0052, respectively, and the average distance was 0.0024. For ITS2, the intraspecific distance was 0.0000-0.0085, the interspecific distance was 0.0021-0.0053, and the average distance was 0.0030. The results indicated that ITS2 had the highest interspecific divergence, followed by ITS. Meanwhile, ITS exhibited lower interspecific divergence compared with the other regions. At the intraspecific level, ITS showed the lowest divergence, while ITS2 displayed the highest variation level. Basic local alignment search tool (BLAST) analyses were performed on the 88 sequences. The results indicated that compared with the same sequence in NCBI, query cover was over 99%.

Neighbour-Joining (NJ) Tree Identification
The NJ tree constructed from 44 ITS sequences showed that Fritillariae Cirrhosae Bulbus could be clustered into one group, Fritillariae Thunbergii Bulbus and Fritillariae Hupehensis Bulbus could be clustered into one group, Fritillariae Pallidiflorae Bulbus could be clustered into one group, and Fritillariae Ussuriensis Bulbus could be clustered into one group. The results are shown in Figure 3.

Neighbour-Joining (NJ) Tree Identification
The NJ tree constructed from 44 ITS sequences showed that Fritillariae Cirrhosae Bulbus could be clustered into one group, Fritillariae Thunbergii Bulbus and Fritillariae Hupehensis Bulbus could be clustered into one group, Fritillariae Pallidiflorae Bulbus could be clustered into one group, and Fritillariae Ussuriensis Bulbus could be clustered into one group. The results are shown in Figure 3.

Cluster Analysis by High-Performance Liquid Chromatography
To validate the results of the HPLC fingerprint analysis and further elucidate the resemblance relationship among the samples, hierarchical clustering analysis (HCA) was applied using the Statistical Product and Service Solutions (SPSS) 19.0 software and the Unscrambler X 10.0 software. HCA results showed that Fritillaria cirrhosae was clustered into a group of circle samples with Fritillaria ussuriensis as black marker. The results are shown in Figure 4. Fritillariae Thunbergii Bulbus and Fritillariae Hupehensis Bulbus were clearly distributed on both the sides by HPLC. Fritillariae Pallidiflorae Bulbus were dispersed in the same side of Fritillariae Cirrhosae Bulbus, and HPLC could not distinguish Fritillariae Cirrhosae Bulbus from other Fritillaria. The alkaloid content is shown in Table S3. The chemical constituents of Fritillaria were similar. The results were consistent with the results of fingerprint analysis by HPLC. Therefore, HCA can also help to

Cluster Analysis by High-Performance Liquid Chromatography
To validate the results of the HPLC fingerprint analysis and further elucidate the resemblance relationship among the samples, hierarchical clustering analysis (HCA) was applied using the Statistical Product and Service Solutions (SPSS) 19.0 software and the Unscrambler X 10.0 software. HCA results showed that Fritillaria cirrhosae was clustered into a group of circle samples with Fritillaria ussuriensis as black marker. The results are shown in Figure 4. Fritillariae Thunbergii Bulbus and Fritillariae Hupehensis Bulbus were clearly distributed on both the sides by HPLC. Fritillariae Pallidiflorae Bulbus were dispersed in the same side of Fritillariae Cirrhosae Bulbus, and HPLC could not distinguish Fritillariae Cirrhosae Bulbus from other Fritillaria. The alkaloid content is shown in Table S3. The chemical constituents of Fritillaria were similar. The results were consistent with the results of fingerprint analysis by HPLC. Therefore, HCA can also help to distinguish Fritillaria plants, but it is not enough for identification. a principal component diagram. All samples were divided into three parts. SPSS 19.0 was used for systematic clustering, and the clustering method of inter-group connection was adopted. The measurement standard was the squared Euclidean distance, and the effective percentage of samples was 100 percent. Chromatographic data showed that Fritillariae Thunbergii Bulbus, Fritillariae Hupehensis Bulbus, Fritillariae Ussuriensis Bulbus, and Fritillariae Pallidiflorae Bulbus had similar fingerprints. The Fritillariae Cirrhosae Bulbus samples were clustered into one branch, and the counterfeit products were clearly on different branches. Fritillariae Cirrhosae Bulbus and Fritillariae Thunbergii Bulbus had similar fingerprints. These results showed that HPLC cannot distinguish the two kinds of traditional Chinese medicines. The results are shown in Figure 5.  A total of 44 Fritillariae Bulbus medicines were analysed by chromatographic fingerprints and a principal component diagram. All samples were divided into three parts. SPSS 19.0 was used for systematic clustering, and the clustering method of inter-group connection was adopted. The measurement standard was the squared Euclidean distance, and the effective percentage of samples was 100 percent. Chromatographic data showed that Fritillariae Thunbergii Bulbus, Fritillariae Hupehensis Bulbus, Fritillariae Ussuriensis Bulbus, and Fritillariae Pallidiflorae Bulbus had similar fingerprints. The Fritillariae Cirrhosae Bulbus samples were clustered into one branch, and the counterfeit products were clearly on different branches. Fritillariae Cirrhosae Bulbus and Fritillariae Thunbergii Bulbus had similar fingerprints. These results showed that HPLC cannot distinguish the two kinds of traditional Chinese medicines. The results are shown in Figure 5.

Species Resources and DNA Barcodes of Fritillariae Bulbus
There are many studies on the Fritillariae Cirrhosae Bulbus plant. However, for the high circulation of traditional Chinese medicine on the market, DNA barcodes identification of Fritillariae Bulbus should be studied.

Species Resources and DNA Barcodes of Fritillariae Bulbus
There are many studies on the Fritillariae Cirrhosae Bulbus plant. However, for the high circulation of traditional Chinese medicine on the market, DNA barcodes identification of Fritillariae Bulbus should be studied. Fritillariae Cirrhosae Bulbus in the 2015 edition of Chinese Pharmacopoeia can be used for medicinal purposes, and Fritillaria species have similar morphologies. Wild Fritillariae Bulbus grows in high altitude areas and is difficult to harvest, requiring a large amount of manpower and material resources. It is very important to identify Fritillariae Bulbus effectively. It is difficult to distinguish Fritillariae Cirrhosae Bulbus from other Fritillariae Bulbus medicinal materials with traditional methods. Powder identification is also a common method for the identification of traditional Chinese medicines. Fritillariae Cirrhosae Bulbus and other Fritillariae Bulbus medicinal materials are white, tasteless powders. The classification of Fritillariae Cirrhosae Bulbus medicinal materials has always been controversial, and there are differences among different types of Fritillariae Cirrhosae Bulbus medicinal materials in the Chinese Pharmacopoeia. DNA barcode identification may be useful in the classification of Fritillaria parent sources. The ITS and ITS2 gene regions have been written into the Chinese Pharmacopoeia and have been widely used to identify plants of the Fabaceae and Compositae families as well as other plants that are traditionally used in Chinese medicines. ITS and ITS2 have good amplification efficiency for the identification of Fritillariae Cirrhosae Bulbus. PsbA-trnH also exhibits sufficient amplification efficiency in Fritillariae Cirrhosae Bulbus, but it is difficult to distinguish Fritillariae Cirrhosae Bulbus and its counterfeits accurately. The locus matK cannot be efficiently amplified in Fritillariae Cirrhosae Bulbus.

Identification of Fritillariae Bulbus
ITS and ITS2 can clearly distinguish the five Fritillariae Bulbus types. In the NJ tree, Fritillariae Bulbus can be clearly divided into different branches. The same results can be obtained by using HPLC to identify Fritillariae Bulbus. Fritillariae Cirrhosae Bulbus (six dry bulbs from Liliaceae, In summary, we have established new chemical and molecular analysis methods for discriminating the five Fritillariae Bulbus. The results revealed that DNA barcoding overcomes the limitations of HPLC fingerprinting for differentiating close genetic relationships and similar chemical-composition species to guarantee an accurate and scientific confirmation of herbal identities in medicinal materials from multiple sources.

Sampling
In total, 44 samples representing 10 Fritillaria species were collected from different areas of China. All of the plant species were identified by Lijuan Zhang. Detailed information about the materials used and sequences obtained in the study are provided in Table S1.

DNA Extraction, Amplification, and Sequencing
Genomic DNA was extracted from dried Bulbus samples using a Plant Genomic DNA Extraction Kit (Tiangen Biotech, Beijing, China) according to the manufacturer's protocol. The quality of the extracted DNA samples was verified to ensure their suitability for subsequent amplification. The four candidate barcodes regions (ITS, matK, psbA-trnH, and ITS2) were tested for their feasibility for Fritillaria barcoding.
The PCR products were detected by 1.2% agarose gel electrophoresis and visualized under ultraviolet light. DNA extraction, amplification, and sequencing were carried out with identical procedures and conditions described above. Forty-four sequences from five Fritillariae Bulbus were submitted in Genbank (Table S2).

Sequence Analysis
DNA sequences were aligned with NCBI standard sequences by using the software Clustal W [35] and MegaX and assembled with CodonCode Aligner version 3.7.1 (CodonCode, USA). Removal of the primer region and low-quality region, manual correction, and stitching were performed. In addition, all of the ITS sequences in our study were delimited based on hidden Markov model (HMM) annotation methods. The average intra-and inter-specific distances were calculated to evaluate the intraspecies variation and interspecific divergence [36,37]. Cluster analyses were conducted using a neighbour-joining tree based on the pairwise Kimura 2-parameter model between individuals.

Preparation of the Reference Solution
Two milligrams of each alkaloid reference substance, peimisine (1), verticine (2), verticinone (3), and imperialine (4), were precisely weighed and placed in a 2 mL flask. A known amount of methanol was added to the vial, and the mixture was shaken to ensure uniformity of the prepared mixed reference solution.

Preparation of Test Solution
Approximately two grams of medicinal powder (sieve 4) was precisely weighed, placed in a 100 mL flask, and soaked in a 4 mL aqueous ammonia solution for 2 hours. The solution was then mixed with 60 mL of a mixed solution of trichloromethane-methanol (4/1) and homogenized and refluxed for 3 hours in a water bath at 80 • C before cooling and filtering. The filtrate was placed in an evaporating dish and allowed to dry. The sample was reconstituted in methanol and transferred to a 2 mL volumetric flask. Methanol was added to a known amount, and the mixture was shaken. The solution was filtered through a 0.22 µm microporous membrane filter before HPLC analysis.

Methodological Investigation
To determine the precision, three solutions of reference substances with different concentrations were sampled in triplicate. The retention time (RT) and peak area (PA) of each peak were measured, and RSD values were calculated. The retention times and peak areas of fritillarin, sibelin, fritillarin B, and fritillarin A had RSD values ranging from 0.12% to 1.09% and from 0.68% to 3.84%, respectively.
Reproducibility was established by extracting six samples of Fritillariae Thunbergii Bulbus and Fritillariae Cirrhosae Bulbus from the same batch according to the abovementioned methods. The peak area of each reference substance was determined in the range of 0.85-2.75% according to the chromatographic conditions described above, which showed that the method had good reproducibility.
In the stability experiment, the sample solutions were analysed at 0, 2, 4, 8, 12, 24, and 36 hours. The peak area was recorded. The RSD values of the corresponding concentration of the peak area of each reference substance ranged from 0.71% to 3.47%, indicating that the sample was stable for at least 36 hours.
Sample determination: a total of 44 batches of 10 Fritillaria medicinal materials were sampled and analysed according to the preparation method of the solution. Sample analysis was carried out according to the chromatographic conditions described above. The sample injection volume was 1 µL, and chromatographic charts were recorded for 10 min. Meanwhile, the reference solution was injected to determine the retention time of alkaloids. Representative HPLC chromatograms of 10 Fritillaria medicinal materials are shown in Figures S1-S10.