Assessment of Genetic and Chemical Variability in Curcumae Longae Rhizoma (Curcuma longa) Based on DNA Barcoding Markers and HPLC Fingerprints

Abstract

Curcumae Longae Rhizoma (Curcuma longa L.) is an important traditional Chinese medicine with multiple beneficial effects. To elucidate the genetic and chemical differences among Curcumae Longae Rhizoma samples, three DNA barcoding markers (rbcL, matK, and ITS-LSU D1/D3) and HPLC fingerprinting were employed in this study. The discriminatory power of rbcL and matK was low, as they only detected one sequence type that showed 100% similarity with more than 20 congeneric species in the Barcode of Life Data Systems (BOLD) database. In contrast, ITS-LSU D1/D3 showed sufficient discriminatory power to precisely identify all of the market samples as C. longa L. in a BLAST search as well as differentiate each sample based on 2–10 ITS-LSU D1/D3 haplotypes with intragenomic variability (mean p-distance: 0.7%, range: 0–2.6%; mean number of differences: 9.6 sites, range: 0–38 sites). HPLC fingerprinting of 13 commercial samples showed a similarity that ranged from 0.769 to 0.996, indicating that the sample quality varied. A cluster analysis based on 5 common peak areas from the HPLC chromatogram resulted in two groups. Group I included 9 samples with a relatively high chemical content, and group II contained 4 samples with a low chemical content. A Mantel test revealed a low correlation (r=0.1721, p=0.047) between genetic and chemical differences. Our findings suggest that the integrated approach of ITS-LSU D1/D3 DNA barcoding and HPLC fingerprinting provides a comprehensive, precise, and convenient method to clarify the genetic and chemical differences in Curcumae Longae Rhizoma.

Curcumae Longae Rhizoma, also known as Jiang Huang or turmeric, is an important traditional Chinese medicine and food spice prepared from the dry rhizome of Curcuma longa L. (Zingiberaceae). It has been widely used in traditional Chinese medicine to promote Qi circulation, dissolve blood sludge, induce menstruation, and relieve pain.¹⁾ In the food industry, Curcumae Longae Rhizoma is commonly used as a spice and food coloring agent.²⁾ It is widely produced in South China, including the Sichuan, Guangxi, Guangdong, Yunnan, and Zhejiang Provinces. Due to highly similar morphological traits, Curcumae Longae Rhizoma is frequently confused with herbs of the same genus, such as Curcumae Radix (also known as Yujin), which is the dry tuber of C. longa L., C. phaeocaulis VAL., C. wenyujin Y. H. CHEN et C. F. LING, or C. kwangsiensis S. G. LEE et C. F. LING, and Curcumae Rhizoma (also known as Ezhu), which is the dry rhizome of C. phaeocaulis VAL., C. wenyujin Y. H. CHEN et C. F. LING, or C. kwangsiensis S. G. LEE et C. F. LING.^1,3) The chemical composition of turmeric rhizomes often vary between varieties, medicinal parts, geological locations, cultivation conditions, and post-harvest processing,^4,5) which can affect the quality, stability, and homogeneity of medical materials. Thus, it is necessary to investigate the genetic and chemical differences in Curcumae Longae Rhizoma for quality control.

DNA barcoding is a species identification system that uses short DNA sequences of a standardized gene region. In 2009, rbcL and matK were proposed as plant core barcodes by the Plant Working Group of Consortium for the Barcode of Life (CBOL)⁶⁾; however, the discriminatory power of these regions was insufficient. Thus, other regions, such as ITS2⁷⁾ and ITS,⁸⁾ were proposed. Previous studies tested the ability of DNA barcoding to identify Curcuma species.^9–12) This research yielded sequence data that are stored in GenBank and can be used as a reference library. To improve the efficiency of rbcL and matK, we used the nuclear ribosome fragment ITS-LSU D1/D3 (Fig. 1) as a DNA barcoding marker because this fragment contains several quickly evolving regions, including internal transcribed spacers (ITS1 and ITS2) and part of the 5′ end of a large subunit (D1, D2, and D3).^13,14)

Fig. 1. Schematic Representation of the ITS-LSU D1/D3 Fragment Amplification Strategy

The ITS-LSU D1/D3 is partial of the ribosomal DNA, which generally shows multiple tandem repeats in the chromosome, containing the small subunit (SSU), the internal transcribed spacer 1 (ITS1), the 5.8 S subunit, the internal transcribed spacer 2 (ITS2), and the large subunit (LSU). Arrows indicate the locations and directions of forward primer DUAN3 (positions at the 3′ ends of SSU) and reverse primer DUAN6 (positions at the middle of LSU).

Many analytical methods, including twice development TLC,¹⁵⁾ HPLC,¹⁶⁾ GC-MS,^17,18) capillary electrophoresis (CE),¹⁹⁾ and NMR,²⁰⁾ are used to distinguish the rhizomes or radixes of curcuma species based on their chemical composition. With high efficiency and rapid separation, HPLC fingerprinting has been widely used for species identification and quality control for traditional Chinese medicines,^21,22) including Curcumae Longae Rhizoma^23–25) and ginger.²²⁾ The optimization of HPLC conditions enabled the simultaneous detection of the major chemical compounds in Curcumae Longae Rhizoma, including curcuminoids (curcumin (CUR), desmethoxycurcumin (DMC), bisdesmethoxycurcumin (BDMC)) and volatile constituents.²⁵⁾

To clarify the genetic and chemical differences of Curcumae Longae Rhizoma in the traditional Chinese medicine market, we developed a comprehensive, precise, and simple method that combines DNA barcoding with HPLC fingerprinting. The sequences of three loci (rbcL, matK, and ITS-LSU D1/D3) and HPLC profiles were obtained for each commercial sample, and cluster analyses were performed on these genetic and chemical data.

MATERIALS AND METHODS

Collection of Plant Materials

In this study, 13 commercial samples of Curcumae Longae Rhizoma [the dry rhizome of C. longa L. (Zingiberaceae)] were collected from medicinal markets and drug stores (Table S1). An authentic sample (CL10) of C. longa L. was provided by Mr. Yushi Ye from the South China Botanical Garden of the Chinese Academy of Sciences (Guangzhou, Guangdong Province, China). The corresponding voucher specimens were verified by Prof. Min Ye (Peking University, China) and deposited in the State Key Laboratory of Natural and Biomimetic Drugs, School of Pharmaceutical Sciences at Peking University in Beijing, China.

Generation of DNA Barcoding Markers

Genomic DNA was extracted from 30 mg of dried and powdered rhizomes or leaves using a Plant Genomic DNA Kit (Tiangen Biotech Co., Beijing, China) according to the manufacturer’s instructions. Three DNA barcoding markers (rbcL, matK, and ITS-LSU D1/D3) were amplified by PCR using a set of universal primers for each locus^26–28) (Table 1). PCR reactions were performed on a T100™ Thermal Cycler (BioRad, Hercules, CA, U.S.A.) in a 25 µL reaction volume containing 1 µL genomic DNA, 0.5 µL each forward/reverse primer (10 µM), and 23 µL PCR Gold Mix TSE101-1 (Beijing Tsingke Biotech Co., Beijing, China). The PCR products were separated in a 1% agarose gel by electrophoresis and visualized under UV light. The PCR products of rbcL and matK were directly subjected to sequencing using the same PCR primers. The PCR products of ITS-LSU D1/D3 were purified using a DNA Gel Recovery Kit (Tsingke Biotech Co.) and cloned into Escherichia coli DH5α using a pClone007 Blunt Simple Vector Kit (Tsingke Biotech Co.). Ten clones for each sample were sent to Tsingke Biotech Co. for DNA sequencing using sequencing primers M13F (5′-TGT AAA ACG ACG GCC AGT-3′) and M13R (5′-CAG GAA ACA GCT ATG ACC-3′). He obtained sequence files were imported into Sequence Scanner software version 1.0 (Applied Biosystems, Foster, CA, U.S.A.) to assess sequence quality. Traces with a trace score lower than 20 or a contiguous read length lower than 300 bp were considered as low-quality sequences and discarded. High-quality sequence traces were trimmed, assembled, manually edited using the SeqMan program (DNAStar Inc., Madison, WI, U.S.A.), and deposited in GenBank with the following accession numbers: kx832045 for rbcL, kx832044 for matK, and kx831971-kx832025 for ITS-LSU D1/D3 (Table S1). Sequences from fungal contamination or fungal endophytes were excluded.

Table 1. Primers and PCR Reaction Conditions

Region	Primer name	Sequence (5′→3′)	Reaction conditions	Reference
ITS-LSU D1/D3	Duan3	GTCGTAACAAGGTTTCCGTA	95°C 3 min; 95°C 30 s, 55°C 30 s, 72°C 1 min, 35 cycles; 72°C 5 min	Duan et al., 2009
	Duan6	GGCATAGTTCACCATCTTTC
rbcL	rbcLa-F	ATGTCACCACAAACAGAGACTAAAGC	95°C 3 min; 95°C 30 s, 55°C 30 s, 72°C 30 s, 35 cycles; 72°C 5 min	Kress et al., 2009
	rbcLa-R	GTAAAATCAAGTCCACCRCG
matK	390F	CGATCTATTCATTCAATATTTC	95°C 3 min; 95°C 30 s, 52°C 40 s, 72°C 1 min, 45 cycles; 72°C 5 min	Cuénoud et al., 2002
	1326R	TCTAGCACACGAAAGTCGAAGT

Molecular Identification Using DNA Barcoding

Sequences were aligned in MEGA 6.0²⁹⁾ using the Muscle algorithm. The intragenomic variability of the multiple ITS-LSU D1/D3 sequences within individual and overall samples was analyzed. The number of haplotypes, variable sites, and indel sites were determined using DnaSP 5.10.01 software.³⁰⁾ Using MEGA 6.0 software, the p-distance and number of differences were calculated, and cluster analyses were conducted using a neighbor-joining tree based on pairwise p-distances between individuals. Species identification for rbcL and matK sequences were performed using the Barcode of Life Data Systems (BOLD) (http://v3.boldsystems.org/index.php/databases). To identify species with our ITS-LSU D1/D3 sequences or ITS (extracted from the corresponding ITS-LSU D1/D3 sequences), we established a reference database and performed a local BLAST search. The reference database included 10 ITS-LSU D1/D3 sequences from the reference sample CL10 of C. longa L. The sequences were verified and incorporated with 2946 sequences mined from GenBank (http://www.ncbi.nlm.nih.gov/genbank/) using the key words “internal transcribed spacer” and “zingiberaceae.”

HPLC/UV Analysis

To prepare sample solution, 0.01 g of powdered Curcumae Longae Rhizoma was extracted with 10 mL of methanol at room temperature with sonication for 30 min. To prepare standard solutions, reference standards of BDMC, DMC and CUR were mixed and dissolved in methanol to make the mixed standard solution (each 0.1 mg/mL) and then diluted continuously to construct the standard curves. All solutions were filtered through a 0.22-µm filter membrane before injection to the HPLC system.

For HPLC/UV analysis, chromatographic separation was completed using an Agilent 1260 HPLC instrument (Agilent, Waldbronn, Germany) equipped with a C18 column (250×4.6 mm, 5 µm). The mobile phase consisted of acetonitrile (A) and water containing 0.1% (v/v) formic acid (B) with the following gradient program: 0 min: 20% A, 80% B; 33 min: 80% A, 20% B; 38 min: 90% A, 10% B; 45 min: 90% A, 10% B; 46 min: 20% A, 80% B; 53 min: 20% A, 80% B. The flow rate was 1.0 mL/min, and the column temperature was 30°C. To monitor multiple chemical compounds simultaneously, the detection wavelengths were set at 250, 254, 280, 320, 370, and 430 nm. Chromatographic data were accessed and processed using the Agilent OpenLAB CDS Chemstation Edition C.01.05 (Agilent, Waldbronn, Germany). A similarity analysis was completed using Chromatographic Fingerprint Similarity Evaluation Software (2012. 130723 version) from the Chinese Pharmacopoeia Commission. A pairwise Euclidean distance matrix was generated from the peak areas of the 5 common peaks in the HPLC chromatograms, and a cluster analysis was performed using the nearest distance method in MATLAB 7.0.

HPLC/MS Analysis

HPLC/MS analysis was performed on an Agilent series 1100 HPLC instrument coupled with an LCQ Advantage ion-trap mass spectrometer (Thermo Fisher, U.S.A.) via an electrospray ionization interface. The HPLC conditions were the same as those for HPLC/UV analysis. The HPLC eluent was introduced into the electrospary ionization source in a post-column splitting ratio of 4 : 1. It was operated in the positive ion mode: source voltage 4.0 kV; sheath gas (N2) 40 arbitrary units; auxiliary gas (N2) 10 units; capillary temperature 320°C; capillary voltage 30 V; and tube lens offset voltage 20 V. Data were analyzed using Xcalibur™ 1.4 software (Thermo Fisher).

Relationship between Genetic Distance and Chemical Distance

To determine how well the chemical composition corresponded to the genetic background, we compared the p-distance matrix of ITS-LSU D1/D3 sequences with the Euclidean distance matrix of the HPLC data using the vegan Mantel function in R with 999 permutations.³¹⁾

RESULTS

Identifying Commercial Samples Using rbcL and matK Sequences

The rbcL and matK regions were PCR amplified and sequenced with a 100% success rate, resulting in 14 rbcL and matK sequences each. A previous study²⁾ showed that the primer pair 390F/1326R did not work for matK gene amplification in turmeric. In contrast, we found that this primer pair produced PCR products if additional PCR cycles (45) were completed (Table 1). The sequence alignment showed only one sequence for rbcL (527 bp) and matK (873 bp) in all of the samples. A unique species identification using the rbcL (KX832045) and matK (KX832044) sequences is difficult because the BOLD results showed that the rbcL and matK haplotypes were shared by 21 and 27 species, respectively (Tables S2, S3). The rbcL sequence showed 100% similarity with 85 accessions from 21 species in 5 genera and 4 families. Out of these 85 accessions, approximately 36.5% were from C. longa and 11.8% were from C. amada. The rbcL sequence was mainly detected in Curcuma spp. (95.3%); however, it was also detected in other genera (e.g., Alpinia nigra), classes (e.g., Cryptolepis buchananii and Barleria montana), and phyla (e.g., Cycas rumphii). The matK sequence showed 100% similarity with 45 accessions from 27 species (except sp. JLS 71485 and sp. SS-0909401) in the Curcuma genus. Out of these 45 accessions, 11.1% were from C. zedoaria and 8.9% were from C. longa. These results indicate that the plant core barcodes rbcL and matK have insufficient discriminatory power for identifying Curcumae Longae Rhizoma.

Identifying Commercial Samples Using ITS-LSU D1/D3 Sequences

The ITS-LSU D1/D3 fragment was PCR amplified, cloned, and successfully sequenced for all samples. A total of 73 high-quality ITS-LSU D1/D3 sequences with a sequence length ranging from 1464 to 1492 bp were obtained. The sequences contained a total of 62 haplotypes with 2–10 haplotypes in each specimen (Table 2). We also obtained 56 sequences from fungus contamination and endophytes, which were excluded from further analyses. Sequence difference in ITS-LSU D1/D3 region among samples was shown as Fig. S1. An analysis of intragenomic variability in the samples revealed 85 variable sites (including 50 singletons and 35 parsimonies) and 72 indel sites. The mean p-distance was 0.7% (range: 0–2.6%), and the mean number of differences was 9.6 sites (range: 0–38 sites). A cluster analysis showed that each specimen was located on a unique branch, and CL8 was isolated from the remaining 13 samples, i.e., CL8 was not a close relative (Fig. 2). These results suggest that the ITS-LSU D1/D3 region has sufficient intragenomic variability to distinguish C. longa L. samples of different origins.

Table 2. Summary of Intragenomic Variability of ITS-LSU D1/D3 Sequences in 14 Curcuma longa L. Samples

Samples	No. of sequences	Haplotypes	Sequence length (bp)	Variable sites (singletons)	Variable sites (parsimonys)	Total variable sites	Indel sites	% mean p-distance (range)	Mean No. of difference (range)
CL1	3	3	1477–1486	12	0	12	27	0.6 (0.1–0.8)	8.0 (1.0–12.0)
CL2	4	4	1483–1491	6	0	6	14	0.2 (0.0–0.5)	3.0 (0.0–6.0)
CL3	2	2	1468–1485	10	0	10	25	0.6 (0.6–0.6)	9.0 (9.0–9.0)
CL4	5	4	1470–1486	2	3	5	22	0.2 (0.0–0.3)	2.6 (0.0–4.0)
CL7	2	2	1469–1478	6	0	6	25	0.4 (0.4–0.4)	6.0 (6.0–6.0)
CL8	2	2	1472–1474	34	0	34	36	2.3 (2.3–2.3)	34.0 (34.0–34.0)
CL9	3	3	1478–1492	15	0	15	27	0.7 (0.5–0.8)	10.0 (7.0–11.0)
CL10	10	10	1467–1486	20	14	34	56	0.7 (0.1–1.4)	10.6 (1.0–21.0)
CL11	8	8	1465–1486	14	4	18	38	0.4 (0.1–0.8)	5.2 (1.0–9.0)
CL12	7	7	1469–1486	14	7	21	51	0.5 (0.1–1.0)	7.5 (0.0–15.0)
CL13	8	7	1477–1488	7	12	19	37	0.5 (0.0–0.9)	7.5 (0.0–13.0)
CL14	5	5	1469–1490	12	9	21	53	0.7 (0.3–1.2)	10.2 (3.0–16.0)
CL15	8	8	1464–1489	6	16	22	59	0.6 (0.1–1.3)	8.9 (2.0–17.0)
CL16	6	6	1464–1486	14	8	22	49	0.6 (0.3–0.9)	9.0 (4.0–13.0)
Overall	73	62	1464–1492	50	35	85	72	0.7 (0.0–2.6)	9.6 (0.0–38.0)

Fig. 2. Cluster Analysis of 14 Curcuma longa L. Samples Using the Neighbor–Joining Method

The intragenomic ITS-LSU D1/D3 sequences of each sample were assigned as one group. The evolutionary divergence was estimated by the p-distance from averaging all sequence pairs between groups. The scale bar indicates 0.1% differences. The analysis included 73 nucleotide sequences.

The taxonomy for Curcuma spp. remains controversial. Therefore, we used an authenticated sample of C. longa L. from the South China Botanical Garden as a reference plant. A BLAST search was used to verify our reference samples of C. longa L. (CL10). Three out of 10 ITS LSU D1-D3 sequences (KX831974, KX831979, and KX831982) showed >99% similarity with the GenBank sequences of C. longa isolate Zn_45.5 (KF304500, KF304501, KF304502, KF304504, KF304505, and KF304506), C. longa voucher 360 (KM226976), C. longa voucher 361 (KM226977), C. haritha isolate 118.1 (KJ872082 and KJ872083), and C. amada isolate Zn_44.1 (KF304458, KF304456, and KF304455), indicating that these samples were congeneric species. Therefore, the ITS-LSU D1/D3 sequences of CL10 were incorporated into the ITS sequence file downloaded from GenBank, which served as the reference database. Next, we used different ribosome fragments (ITS-LSU D1/D3 and ITS) to identify the botanical origin of the commercial samples. The BLAST (Expected value <0.00001) results of both ITS-LSU D1/D3 and ITS consistently assigned our reference plant of C. longa L. CL10 as one of the top hits for each query sequence (identity for ITS-LSU D1/D3 and ITS ranged from 97.64–100% and 96.37–100%, respectively) (Table S4). Thus, we identified all of the market samples as C. longa L. and submitted their ITS-LSU D1/D3 sequences into GenBank.

Chemical Discrimination Using HPLC Fingerprinting

Methanol extracts of the 13 commercial samples were analyzed using HPLC. All of the compounds had different UV absorptivities at different UV wavelengths. A wavelength of 250 nm was chosen for generating chromatograms because it showed better signal responses for the detection of all analyte peaks, whereas 430 nm was selected for detecting CUR, DMC, and BDMC because it resulted in higher signal intensities. Using a gradient solvent system, the chemical profile, consisting of clearly separated peaks, was obtained during a 40-min run for each sample (Fig. 3, Fig. S2). Five common peaks (labeled as 1–5) were detected in all samples at retention times of 21.8, 22.54, 23.2, 32.8, and 36.1 min. Three common peaks, 1, 2, and 3 were identified as BDMC, DMC and CUR, respectively, by comparing with reference standards. The other common peaks were tentatively identified as arturmerone (m/z 217.0, common peak 4), α-Turmerone (m/z 219.0, common peak 5) and β-Turmerone (m/z 219.0, common peak 5′) by HPLC/MS/MS analysis (Fig. 4). To reveal the chemical difference among samples, similarity, peak areas and the content of curcuminoids were determined from the HPLC fingerprints. The similarities of the fingerprints ranged from 0.769 to 0.996 when each sample was compared with the reference chromatogram. The peak areas of the 5 investigated chemical compounds were different significantly with a maximum/minimum value ranged from 6.6–161.1-fold (Table 3). The contents of the three curcuminoids in each sample were determined by reference standards (Fig. S3). As shown in Table 4, the content of BDMC, DMC and CUR were 0.428–21.921, 0.316–16.190, and 1.133–34.961 mg/g, respectively. BDMC, DMC and CUR were markedly varied among samples with a maximum/minimum value of 51.2, 51.2, and 30.9-fold (Table 4), respectively.

Fig. 3. HPLC Profiles (250 nm) for 13 Market Samples of Curcumae Longae Rhizoma (Curcuma longa L.)

The chemical compounds from the 5 labeled peaks were selected for the UPGMA clustering analysis. Standards: 1. bisdemethoxycurcumin (BDMC); 2. demethoxycurcumin (DMC); 3. curcumin (CUR).

Fig. 4. Mass Spectra of the Chromatographic Peaks of Curcumae Longae Rhizoma (Curcuma longa L.)

Peak 4: Arturmerone; Peak 5: α-Turmerone; Peak 5′: β-Turmerone.

Table 3. The Peak Areas of the 5 Common Peaks in the HPLC Chromatograms of 13 Market Samples of Curcumae Longae Rhizoma (Curcuma longa L.)

Voucher No.	Peak 1	Peak 2	Peak 3	Peak 4	Peak 5	Total
CL1	5131.7	8068.4	13157.5	2880.2	1908.9	31146.7
CL2	6195.3	4361.2	8004.7	1632.8	1157.9	21351.9
CL3	10281	5976.3	10666.8	2171.9	1280.7	30376.7
CL4	8315.4	5494.7	12113.5	1707.8	1087.1	28718.5
CL7	4153.8	5798.5	16525.1	2185.6	962.7	29625.7
CL8	133.7	227.4	887.9	451.9	225.3	1926.2
CL9	6655.4	8026.4	12474.4	1861.7	1162.4	30180.3
CL11	2970.9	3139.2	8349.1	1308.4	518.1	16285.7
CL12	6253.3	7057.3	12082.2	1853.4	1111.4	28357.6
CL13	152.0	478	1707.3	599.8	237.3	3174.4
CL14	63.8	374.5	476.2	439.4	103.2	1457.1
CL15	125.7	1051.9	1435.9	449.9	191.6	3255.0
CL16	3492.9	5043.4	8653.7	1547	44.8	18781.8
Max/min	161.1	35.5	34.7	6.6	42.6	21.4

The peak area data for peaks 1–3 were detected at 430 nm, and the peak area data for peaks 4–5 were detected at 250 nm.

Table 4. Contents (mg/g) of Bisdemethoxycurcumin (BDMC), Demethoxycurcumin (DMC) and Curcumin (CUR) in 13 Market Samples of Curcumae Longae Rhizoma (Curcuma longa L.)

Voucher No.	BDMC	DMC	CUR	Total
CL9	14.295	16.105	26.425	56.825
CL3	21.921	11.954	22.613	56.488
CL7	9.032	11.594	34.961	55.587
CL1	11.089	16.190	27.863	55.142
CL4	17.786	10.979	25.662	54.427
CL12	13.449	14.143	25.596	53.188
CL2	13.327	8.684	17.001	39.012
CL16	7.642	10.066	18.369	36.077
CL11	6.544	6.211	17.727	30.482
CL15	0.559	1.985	3.156	5.700
CL13	0.614	0.823	3.728	5.165
CL8	0.576	0.316	2.000	2.892
CL14	0.428	0.613	1.133	2.174
Max/min	51.2	51.2	30.9	26.0

To investigate weather the HPLC fingerprint profiles can discriminate the samples with different content of compounds, a pairwise Euclidean distance matrix was generated from the peak area for the 5 common peaks in the HPLC chromatogram, and a cluster analysis was completed using the nearest distance method. The results revealed that two main groups (Fig. 5). Group I included 9 samples (CL1–4, CL7, CL9, CL11, CL12, CL16) with relatively high chemical content (total curcuminoids: 30.482–56.825 mg/g; curcumin: 17.001–34.961 mg/g), including all of the samples of Guangxi province (CL1–3, CL7, CL12), three samples of Sichuan province (CL9, CL11, CL16), and one sample of Yunnan province (CL4). Group II consisted of 4 samples (CL8, CL13–15) of Sichuan province with relatively low chemical content (total curcuminoids: 2.174–5.700 mg/g; curcumin: 1.133–3.728 mg/g). The clustering results indicated that the chemical composition of the samples was similar within groups but markedly different between groups.

Fig. 5. Cluster Analysis for 13 Market Samples of Curcumae Longae Rhizoma (Curcuma longa L.) Using the Nearest Distance Method

A pairwise Euclidean distance matrix was generated from the peak areas of the 5 common peaks in the HPLC chromatograms.

Relationship between Genetic Distance and Chemical Distance

The Mantel test is often used to compare two sets of dissimilarities or distances, i.e., measure the correlation between two dissimilarity/distance matrices. In this study, we employed the Mantel test to assess how well the chemical composition corresponded to the genetic background of samples. First, the ITS-LSU D1/D3 sequences of each sample were grouped, imported, and aligned in MEGA software. The p-distances between groups were calculated to create a p-distance matrix. Second, the peak area of the 5 common peaks for each sample were extracted, and the pairwise Euclidean distances were calculated in MATLAB to create a Euclidean distance matrix. We then compared the p-distance matrix of ITS-LSU D1/D3 sequences with the Euclidean distance matrix of the HPLC data using a Mantel permutation test (999 times) in R. The Mantel test showed a low correlation (r=0.1721, p=0.047) between genetic distance and HPLC distance, suggesting that chemical composition differences do not correspond well with genetic variations among the Curcumae Longae Rhizoma samples.

DISCUSSION

In the present study, 13 commercial samples of Curcumae Longae Rhizoma and an authenticated sample of C. longa L. were identified and distinguished using three DNA barcoding markers (rbcL, matK and ITS-LSU D1/D3). We found that the rbcL and matK markers could not assign unique species for all of the market samples of Curcumae Longae Rhizoma (Tables S2 and S3), whereas ITS-LSU D1/D3 sequences assigned them as C. longa L. and precisely distinguished each individual sample (Table S4). Although the rbcL and matK markers have been widely used as core barcodes for identifying land plant species by the CBOL since 2009,⁶⁾ their discriminatory power is limited.³²⁾ Therefore, a new DNA barcode was suggested. Our results highlight the high discriminatory power of ITS-LSU D1/D3 as a novel DNA barcode. ITS-LSU D1/D3 contains three well-documented divergent regions (ITS1, ITS2, and LSU D1/D3), and ITS2,⁷⁾ ITS,⁸⁾ and LSU D1/D3¹⁴⁾ have been proposed as barcoding markers. In addition, there are multiple haplotypes with intragenomic variability (Table 2) that may provide sufficient discriminatory power for the barcoding of each species and the discrimination of herbs with different origins.

Several previous studies identified a single sequence of ITS1 or ITS2 for each individual of C. longa by directly sequencing PCR products.^9,12) Using clone sequencing for ITS-LSU D1/D3, we identified 2–10 haplotypes within each sample. Our results are consistent with other studies^10,11) reported different ITS sequences among specimens. Therefore, our findings support the use of clone sequencing rather than the direct sequencing of PCR products to obtain accurate sequence data for ribosomal DNA (rDNA) fragments (ITS, ITS2, ITS-LSU D1/D3).

We developed a method for the HPLC fingerprinting of Curcumae Longae Rhizoma using a gradient solvent system with detection at 250 and 430 nm. This method allows for the simultaneous discrimination of the 5 main chemical compounds in these specimens. In addition, the cluster analysis based on the 5 common peak areas clearly distinguished low-content samples. Previous studies reported that the content of curcumin was different significantly within different sources^24,33) or different lines.⁴⁾ Similarly, we showed that the content of curcuminoids varied among samples, especially in samples of Sichuan province which included four samples (CL8, 13–15) with relatively low content of curcumin (Table 4). Although it is qualified according to the requirement of Chinese Pharmacopeia (curcumin content >1.0%), these samples could not be considered as the desirable medical materials with high quality. The various in content of curcuminoids may make it difficult to control the purity, stability and consistency in Traditional Chinese Medicine. Thus, the HPLC fingerprinting method developed in this study may be useful for quality control of Curcumae Longae Rhizoma.

Many researches have used DNA barcoding or HPLC analysis for identifying the species of C. longa L. or its medical materials of Curcumae Longae Rhizoma (turmeric).^{2,9,12,16,33,34)} However, little is known about the relationship of genetic and chemical differences by using integrated approach. In this study, we found a low correlation (r=0.1721, p=0.047) between genetic distance and HPLC distance using the Mantel test. This result indicates that chemical composition differences do not correspond well with genetic variations among Curcumae Longae Rhizoma samples. Our findings were similar to He et al.³⁵⁾ who reported that the relationships (for Coptis spp.) inferred by chemical taxonomy are different to molecular phylogeny. The phytochemical profile of herbs was affected by several factors, such as genetic compositon, growing stage, ambient environmental condition, post-harvest processing, etc.³⁶⁾ Thus, it is necessary to investigate the chemical composition of herbs and identify species based on their genetic background.

Our findings highlight the combined use of HPLC fingerprinting and DNA barcoding with ITS-LSU D1/D3 to assess the quality of Curcumae Longae Rhizoma sold in the medicinal market. ITS-LSU D1/D3 showed high discriminatory power for species identification. HPLC fingerprinting is a simple, sensitive, and accurate method to determine the chemical composition of herb samples. Thus, these combined systems can precisely identify the genetic and chemical differences within herbs and may be useful for the screening of high-quality medical materials.

Acknowledgments

This work was supported by the Department of Education of Guangdong Province [Grant No.: Guangdong higher education letter (2013) 113] and provincial key platform and key scientific research projects in colleges and universities of Guangdong in 2016 (Grant No.: 2016KTSCX133. Combining DNA barcoding and HPLC fingerprint for authentication of traditional Chinese medicine: a case study of Curcumae Longae Rhizoma). We thank Professor Jingping Liao and Mr. Yushi Ye of the South China Botanical Garden at the Chinese Academy of Sciences for their kind help with sample collection.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

The online version of this article contains supplementary materials.

REFERENCES

• 1) Chinese Pharmacopoeia Commission. Pharmacopoeia of the People’s Republic of China. Vol. 1, China Medical Science and Technology Press, Beijing, pp. 247–248 (2010).
• 2) Parvathy VA, Swetha VP, Sheeja TE, Sasikumar B. Detection of plant-based adulterants in turmeric powder using DNA barcoding. Pharm. Biol., 53, 1774–1779 (2015).
• 3) Zhang XJ, Qiu JF, Guo LP, Wang Y, Li P, Yang FQ, Su HX, Wan JB. Discrimination of multi-origin Chinese herbal medicines using gas chromatography-mass spectrometry-based fatty acid profiling. Molecules, 18, 15329–15343 (2013).
• 4) Hayakawa H, Minaniya Y, Ito K, Yamamoto Y, Fukuda T. Difference of curcumin content in Curcuma longa L. (Zingiberaceae) caused by hybridization with other Curcuma species. Am. J. Plant Sci., 2, 111–119 (2011).
• 5) Li S, Yuan W, Deng G, Wang P, Yang P, Aggarwal BB. Chemical composition and product quality control of turmeric (Curcuma longa L.). Pharm. Crop., 2, 28–54 (2011).
• 6) Hollingsworth PM, Forrest LL, Spouge JL, Hajibabaei M, Ratnasingham S, van der Bank M, Chase MW, Cowan RS, Erickson DL, Fazekas AJ, Graham SW, James KE, Kim K-J, Kress WJ, Schneider H, van AlphenStahl J, Barrett SCH, van den Berg C, Bogarin D, Burgess KS, Cameron KM, Carine M, Chacon J, Clark A, Clarkson JJ, Conrad F, Devey DS, Ford CS, Hedderson TAJ, Hollingsworth ML, Husband BC, Kelly LJ, Kesanakurti PR, Kim JS, Kim Y-D, Lahaye R, Lee H-L, Long DG, Madrinan S, Maurin O, Meusnier I, Newmaster SG, Park C-W, Percy DM, Petersen G, Richardson JE, Salazar GA, Savolainen V, Seberg O, Wilkinson MJ, Yi D-K, Little DP, CBOL Plant Working Group. A DNA barcode for land plants. Proc. Natl. Acad. Sci. U.S.A., 106, 12794–12797 (2009).
• 7) Chen S, Yao H, Han J, Liu C, Song J, Shi L, Zhu Y, Ma X, Gao T, Pang X, Luo K, Li Y, Li X, Jia X, Lin Y, Leon C. Validation of the its2 region as a novel DNA barcode for identifying medicinal plant species. PLoS ONE, 5, e8613 (2010).
• 8) China Plant BOL Group. Comparative analysis of a large dataset indicates that internal transcribed spacer (its) should be incorporated into the core barcode for seed plants. Proc. Natl. Acad. Sci. U.S.A., 108, 19641–19646 (2011).
• 9) Shi LC, Zhang J, Han JP, Song JY, Yao H, Zhu YJ, Li JC, Wang ZZ, Xiao W, Lin YL, Xie CX, Qian ZZ, Chen SL. Testing the potential of proposed DNA barcodes for species identification of zingiberaceae. J. Syst. Evol., 49, 261–266 (2011).
• 10) Záveská E, Fér T, Šída, Otakar, Krak K, Marhold K, Leongškorničková J. Phylogeny of Curcuma (Zingiberaceae) based on plastid and nuclear sequences: Proposal of the new subgenus Ecomata. Taxon, 61, 747–763 (2012).
• 11) Vinitha MR, Kumar US, Aishwarya K, Sabu M, Thomas G. Prospects for discriminating Zingiberaceae species in India using DNA barcodes. J. Integr. Plant Biol., 56, 760–773 (2014).
• 12) Chen J, Zhao JT, Erickson DL, Xia N, Kress WJ. Testing DNA barcodes in closely related species of Curcuma, (Zingiberaceae) from myanmar and china. Mol. Ecol. Resour., 15, 337–348 (2015).
• 13) Kuzoff RK, Sweere JA, Soltis DE, Soltis PS, Zimmer EA. The phylogenetic potential of entire 26 s rDNA sequences in plants. Mol. Biol. Evol., 15, 251–263 (1998).
• 14) Huang Q, Duan Z, Yang J, Ma X, Zhan R, Xu H, Chen W. SNP typing for germplasm identification of amomum villosum lour. based on DNA barcoding markers. PLOS ONE, 9, e114940 (2014).
• 15) Zhang JS, Guan J, Yang FQ, Liu HG, Cheng XJ, Li SP. Qualitative and quantitative analysis of four species of Curcuma Rhizomes using twice development thin layer chromatography. J. Pharm. Biomed. Anal., 48, 1024–1028 (2008).
• 16) Yang FQ, Wang YT, Li SP. Simultaneous determination of 11 characteristic components in three species of curcuma, rhizomes using pressurized liquid extraction and high-performance liquid chromatography. J. Chromatogr. A, 1134, 226–231 (2006).
• 17) Yang FQ, Li SP, Zhao J, Lao SC, Wang YT. Optimization of GC-MS conditions based on resolution and stability of analytes for simultaneous determination of nine sesquiterpenoids in three species of Curcuma Rhizomes. J. Pharm. Biomed. Anal., 43, 73–82 (2007).
• 18) Xiang Z, Wang X, Cai X, Zeng S. Metabolomics study on quality control and discrimination of three Curcuma species based on gas chromatograph-mass spectrometry. Phytochem. Anal., 22, 411–418 (2011).
• 19) Lechtenberg M, Quandt B, Nahrstedt A. Quantitative determination of curcuminoids in curcuma, rhizomes and rapid differentiation of Curcuma domestica, val. and Curcuma xanthorrhiza, roxb. by capillary electrophoresis. Phytochem. Anal., 15, 152–158 (2004).
• 20) Jung Y, Lee J, Kim HK, Moon BC, Ji Y, Ryu dH, Hwang GS. Metabolite profiling of curcuma species grown in different regions using 1 h NMR spectroscopy and multivariate analysis. Analyst, 137, 5597–5606 (2012).
• 21) Xu C, Yang B, Zhu W, Li X, Tian J, Zhang L. Characterisation of polyphenol constituents of linderae aggregate, leaves using HPLC fingerprint analysis and their antioxidant activities. Food Chem., 186, 83–89 (2015).
• 22) Yudthavorasit S, Wongravee K, Leepipatpiboon N. Characteristic fingerprint based on gingerol derivative analysis for discrimination of ginger (zingiber officinale) according to geographical origin using HPLC-DAD combined with chemometrics. Food Chem., 158, 101–111 (2014).
• 23) Tian M, Li M. Study on HPLC fingerprint of Rhizoma Curcumae longae. Res. Prac. Chin. Med., 4, 22–24 (2009).
• 24) Duan X. Optimized extraction, quantitative evaluation and antioxidation in vitro of curcumins from herbal Curcuma species. Master dissertation. Sichuan Agricultural University, Chengdu. pp. 12–29 (2013).
• 25) Wang Y. The comparative research on chemical constituents of curcumae radix, curcumae longae rhizoma and curcumae rhizoma. Master dissertation. Chong qing University, Chong qing. pp. 13–34 (2014).
• 26) Duan Z, Huang Q, Yang J, Diao L, Zhan R, He R, Xu H, Yan P, Chen W. DNA extraction method research for DNA barcode analysis of Chinese medicinal materials. Tradit. Chin. Drug Res. Pharmacol., 20, 480–484 (2009).
• 27) Kress WJ, Erickson DL, Jones FA, Swenson NG, Perez R, Sanjur O, Bermingham E. Plant DNA barcodes and a community phylogeny of a tropical forest dynamics plot in panama. Proc. Natl. Acad. Sci. U.S.A., 106, 18621–18626 (2009).
• 28) Cuénoud P, Savolainen V, Chatrou LW, Powell M, Grayer RJ, Chase MW. Molecular phylogenetics of caryophyllales based on nuclear 18s rDNA and plastid rbcl, atpb, and matk DNA sequences. Am. J. Bot., 89, 132–144 (2002).
• 29) Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: Molecular Evolutionary Genetics Analysis version 6.0. Mol. Biol. Evol., 30, 2725–2729 (2013).
• 30) Librado P, Rozas J. Dnasp v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics, 25, 1451–1452 (2009).
• 31) Oksanen J, Blanchet FG. Kindt R, Legendre P, O’Hara RB, Simpson GL, Solymos P, Stevens MHH, Szoecs E, Wagner H. “Vegan: Community Ecology Package. R package version 2.3–5.”: ‹http://cran.r-project.org/web/packages/vegan/›, 2016.
• 32) Hollingsworth PM, Graham SW, Little DP. Choosing and using a plant DNA barcode. PLoS ONE, 6, e19254 (2011).
• 33) Li R, Xiang C, Ye M, Li HF, Zhang X, Guo D. Qualitative and quantitative analysis of curcuminoids in herbal medicines derived from Curcuma species. Food Chem., 126, 1890–1895 (2011).
• 34) Rafi M, Wulansari L, Heryanto R, Darusman LK, Lim LW, Takeuchi T. Curcuminoid’s content and fingerprint analysis for authentication and discrimination of Curcuma xanthorrhiza, from Curcuma longa, by high-performance liquid chromatography-diode array detector. Food Anal. Method., 8, 2185–2193 (2015).
• 35) He Y, Hou P, Fan G, Arain S, Peng C. Comprehensive analyses of molecular phylogeny and main alkaloids for coptis, (ranunculaceae) species identification. Biochem. Syst. Ecol., 56, 88–94 (2014).
• 36) Dhami N, Mishra AD. Phytochemical variation: how to resolve the quality controversies of herbal medicinal products? J. Herb. Med., 5, 118–127 (2015).