Analysis of Flavonoids in Rhamnus davurica and Its Antiproliferative Activities

Rhamnus davurica Pall. (R. davurica) has been used as a traditional medicinal herb for many years in China and abroad. It has been well documented as a rich source of flavonoids with diversified structures, which in turn results in far-ranging biological activities, such as anti-inflammation, anticancer, antibacterial and antioxidant activities. In order to further correlate their anticancer potentials with the phytochemical components, the fingerprint profile of R. davurica herb from Dongbei was firstly investigated using HPLC-ESI-MS/MS. Thirty two peaks were detected and identified, 14 of which were found in R. davurica for the first time in this work. Furthermore, a total of 23 peaks were resolved as flavonoids, which are the major components found in R. davurica. Meanwhile, the antiproliferative activities against human cancer cells of HT-29 and SGC-7901 in vitro exhibited distinct inhibitory effects with IC50 values at 24.96 ± 0.74 and 89.53 ± 4.11 μg/mL, respectively. Finally, the general toxicity against L-O2 cells displayed a much higher IC50 at 229.19 ± 8.52 μg/mL, which suggested very low or no toxicity on hepatic cell viability. The current study revealed for the first time the correlations between the flavonoids of R. davurica with their antiproliferative activities, which indicated that the fingerprint profile of flavonoids and their anticancer activities could provide valuable information on the quality control for herbal medicines and their derived natural remedies from this valuable medicinal plant.


Introduction
Flavonoids are one of the most ubiquitous polyphenolic secondary metabolites existing in natural products, which are widely found in plants and their corresponding derived foods. In plants, the basic chemical skeleton of flavonoids is biosynthesized by a series of condensation reactions between three malonyl residues (A ring) and hydroxycinnamic acid (B ring), namely a C6-C3-C6 skeleton [1]. These secondary metabolites could be classified into several types of structures according to the variations in the aglycones, sugar moieties and intersaccharide linkage [2]. Flavonoids play important roles in plant ecology and physiology, and many plants being rich in flavonoids have been used as herbal medicines or functional foods for thousands of years [3]. Recent studies showed that flavonoids and their glycoconjugates have exhibited remarkable pharmacological and biological properties, including anti-inflammatory, antioxidant, antitumor, antiallergic, antibacterial, antiarteriosclerotic and antiestrogenic activities [4][5][6]. To date, more than 8000 natural flavonoids have been identified with diversified activities, and it has been reported that their discrepant biological activities are mainly due

Optimization of Chromatographic Conditions and HPLC Fingerprint Profile of R. davurica (Dongbei)
A number of previous studies were carried out to optimize the HPLC analytical conditions for the separation of diverse flavonoids in the Rhamnus extracts [11]. For the systematic investigations, the HPLC method was optimized to reduce run-time while maintaining the chromatographic separation efficiency and peak shape. Besides, acetonitrile (ACN) was selected as the organic phase due to its lower background and higher sensitivity than that of methanol. Furthermore, formic acid (0.1%, v/v) was applied as an additive in the mobile phase, which could facilitate better resolution by reducing the peak tailing [24,25]. Due to the structural properties of the flavonoid compounds, the characteristic wavelength of 360 nm was chosen for the detection [26]. At last, the final optimized conditions were obtained within 35 min by using a Waters Sunfire RP-C18 column and a binary elution gradient of ACN/water (0.1% FA).
After extracted three times from the barks of R. davurica, the components of flavonoids were further enriched by a polyamide column and then analyzed with the optimized HPLC conditions (Section 3.5.1). The HPLC fingerprint profile of the phytochemical components from R. davurica at 360 nm is shown in Figure 1. In more detail, a total of 32 peaks were resolved under the optimal chromatographic conditions, which are subsequently summarized in Table 1 according to their corresponding peak numbers.
Molecules 2016, 21,1275 3 of 14 the HPLC method was optimized to reduce run-time while maintaining the chromatographic separation efficiency and peak shape. Besides, acetonitrile (ACN) was selected as the organic phase due to its lower background and higher sensitivity than that of methanol. Furthermore, formic acid (0.1%, v/v) was applied as an additive in the mobile phase, which could facilitate better resolution by reducing the peak tailing [24,25]. Due to the structural properties of the flavonoid compounds, the characteristic wavelength of 360 nm was chosen for the detection [26]. At last, the final optimized conditions were obtained within 35 min by using a Waters Sunfire RP-C18 column and a binary elution gradient of ACN/water (0.1% FA). After extracted three times from the barks of R. davurica, the components of flavonoids were further enriched by a polyamide column and then analyzed with the optimized HPLC conditions (Section 3.5.1). The HPLC fingerprint profile of the phytochemical components from R. davurica at 360 nm is shown in Figure 1. In more detail, a total of 32 peaks were resolved under the optimal chromatographic conditions, which are subsequently summarized in Table 1 according to their corresponding peak numbers.  Table 1.

Structural Identifications of Flavonoids Using HPLC-ESI-MS/MS Analysis
According to the HPLC-UV chromatogram of R. davurica shown in Figure 1, 32 components were detected and are summarized in Table 1. Structural identifications and characterizations were successfully carried out based on the comparisons of their ESI-MS/MS data with the corresponding standards and fragmentation pathways reported in the previous literature. Based on the identifications, the corresponding chemical structures present in R. davurica are shown in Figure 2. Twenty three compounds out of 32 are flavonoids and their derivatives, which consisted of the major components of R. davurica. Considering that the different types of aglycones, the sites of glycosylation and other substituents and the corresponding fragmentation pathways all contribute to the structural diversities, those detected components were further divided into four groups in order to simplify the MS/MS illustrations, including flavone aglycones, flavonoid glycosides, anthraquinones and others. The ESI-MS/MS analysis of those components was conducted in negative ion mode, and their retention times (Rt), calculated molecular masses and MS/MS data are shown in Table 1, respectively. The identifications, interpretations and verifications of these compounds are discussed thereafter in more detail. Figure 1. The HPLC-UV chromatogram of R. davurica (Dongbei) at 360 nm using the optimized analytical method. The peak numbers in this figure correspond to those used in Table 1.

Structural Identifications of Flavonoids Using HPLC-ESI-MS/MS Analysis
According to the HPLC-UV chromatogram of R. davurica shown in Figure 1, 32 components were detected and are summarized in Table 1. Structural identifications and characterizations were successfully carried out based on the comparisons of their ESI-MS/MS data with the corresponding standards and fragmentation pathways reported in the previous literature. Based on the identifications, the corresponding chemical structures present in R. davurica are shown in Figure 2. Twenty three compounds out of 32 are flavonoids and their derivatives, which consisted of the major components of R. davurica. Considering that the different types of aglycones, the sites of glycosylation and other substituents and the corresponding fragmentation pathways all contribute to the structural diversities, those detected components were further divided into four groups in order to simplify the MS/MS illustrations, including flavone aglycones, flavonoid glycosides, anthraquinones and others. The ESI-MS/MS analysis of those components was conducted in negative ion mode, and their retention times (Rt), calculated molecular masses and MS/MS data are shown in Table 1, respectively. The identifications, interpretations and verifications of these compounds are discussed thereafter in more detail.    Table 1.  Table 1.

Identification of Flavone Aglycones
Due to the characteristic parent structure of 2-phenyl chromone, most of the fragment ions for the flavone skeleton were observed in the ESI-MS/MS spectra, and further fragment ions could be more informative for structural elucidation. In this way, Peaks 9, 16, 18, 19, 21, 22, 23, 24, 25, 26 and 29 were tentatively identified as aglycones because of their similar fragmentation features of the ESI-MS/MS spectra, as shown in Table 1 Tentatively, the precursor ion at m/z 551 was formed by two sakuranetin with the further neutral losses of H 2 O and H 2 . Hence, Peak 28 was determined to be the sakuranetin dimer, which was found in R. davurica for the first time in this study. However, previous studies revealed similar carbon-carbon interflavanoid linkage types, such as 3,8'-coupled dimer of sakuranetin [29], chamaejasmin and 3"-epidiphysin [30].  [11,31]. Peak 32 presented the [M − H] − ion at m/z 359, and its MS/MS fragment ions at m/z 285, 267, 241 and 223 were similar to that of Peak 18, which suggested Peak 32 as a luteolin derivative.

Identification of Flavonoid Glycosides
According to Table 1, flavonoid conjugates account for a large proportion of the components detected in the extract of R. davurica. For a better illustration, we classified them into two groups: C-glycosides and O-glycosides, which were identified by comparing their MS/MS spectra with those of corresponding standards and the literature reported previously.
As for C-glycosides from R. davurica, the MS/MS spectra and proposed fragmentation pathways are shown in Figure 3

Identification of Flavonoid Glycosides
According to Table 1, flavonoid conjugates account for a large proportion of the components detected in the extract of R. davurica. For a better illustration, we classified them into two groups: C-glycosides and O-glycosides, which were identified by comparing their MS/MS spectra with those of corresponding standards and the literature reported previously.
As for C-glycosides from R. davurica, the MS/MS spectra and proposed fragmentation pathways are shown in Figure 3 [3]. The two glucosides were also identified in R. davurica for the first time in this work.   [2], and both of them were found in R. davurica for the first time in this study. Furthermore, Peak 6 was validated by the corresponding standard. Peak 5 gave the deprotonated molecular ion at m/z 461 with a [Y 0 ] − at m/z 299 (the loss of a hexose moiety), suggesting that it was a diosmetin monosaccharide. Meanwhile, the presence of the ion at m/z 284 was yielded due to the loss of the methyl moiety, and its glycosylation took place at C-7. Therefore, Peak 5 was tentatively identified as diosmetin 7-O-glucoside [4] [3]. The two glucosides were also identified in R. davurica for the first time in this work.  (Table 1), Peak 30 was identified as question in accordance with a C 16 H 12 O 5 formula, considering the retention times, LC profile with C18 column and the similar CID fragmentation pathways reported for the two isomers [36].

Others
Peak 12 showed a precursor ion at m/z 195 and fragment ions at m/z 167, 152, 136 and 108, which were similar to the MS/MS data of Iodolactone, and was thus tentatively deduced as an iodolactone derivative [37]. As for Peak 20, the [M − H] − ion at m/z 169 gave a series of product ions at m/z 151, 125, 107, 83 and 57 due to the consecutive neutral loss of H 2 O (18 Da), C 2 H 2 (26 Da), CO (28 Da), C 2 (24 Da) and C 2 H 2 (26 Da). Hence, Peak 20 was identified as oxireno [4,5]cyclopenta [1,2-c]pyran, which was similar to catalpol aglycone [38]. Moreover, the two compounds above were firstly found in R. davurica. Table 2 shows the results of method validation for the three representative standards, of which the calibration curves exhibited good linearity at 360 nm (R 2 ≥ 0.9990). In total, 32 components can be classified into five groups based on the structural identifications above. Hence, the individual component contents in R. davurica were quantified by using the corresponding standards for calibration for each group (kaempferol for flavone aglycones, vitexin for flavonoid C-glycosides, rutin as an external reference sample for flavonoid O-glycosides, others and unknown), and their corresponding concentrations were calculated as shown in Table 1. According to the HPLC screening data summarized in Figure 4a, flavonoids accounted for the majority of the chemical compositions and represented the dominant class of components (67.24%) in R. davurica. Due to the chemical diversities of different groups, flavonoid glycosides exhibited higher content of 40.98%, followed by the flavone aglycones of 26.26%, anthraquinones and their O-glycosides of 26.2%, respectively. In general, vitexin was the most abundant glycoside with 86.07 mg/100 g DW (dry weight); aromadendrin and question, the major components of aglycones and anthraquinones, accounted for 21.48 mg/100 g DW and 42.17 mg/100 g DW. In addition, the other components in trace quantities also showed significant variance in the constitutes shown in Table 1. Therefore, the discrepancies in the diversities of the chemical compositions and their corresponding contents could further affect the quality control and future medical exploration of R. davurica.

Antiproliferation Assays on R. davurica
The bioactive evaluations in one plant species depend on both the qualitative and quantitative knowledge on this species [39]. As we know, R. davurica has long been used as a kind of folk remedy in many Asian countries. According to the quantification above, flavonoids accounted for the majority of the chemical compositions and represented the dominant class of components in R. davurica. Flavonoids ubiquitously exist and are widely consumed as the important secondary metabolites from natural plants and have remarkable anticancer activities. However, there is very little study on their anticancer activities.
In this study, for the first time, we focused on the antiproliferative effects of R. davurica on human cancer cell lines, like HT-29 and SGC-7901 cells. As shown in Figure 4b, it is found out that R. davurica exhibited significant dose-dependent antiproliferative activities against HT-29 and SGC-7901 cells with IC 50 values of 24.96 ± 0.74 and 89.53 ± 4.11 µg/mL, respectively. Meanwhile, Figure 4c shows that the inhibitory activities against both HT-29 and SGC-7901 cells significantly increased by the treatment with R. davurica in a time-dependent manner from 24 h-96 h at a dose of 150 µg/mL, although there was a decrease on SGC-7901 cells at the time from 72 h-96 h. Considering the drug safety, the evaluation of the general toxicity of R. davurica on the normal human hepatic cells (L-O2) was also conducted ( Figure 4d). As expected, R. davurica displayed a much higher IC 50 at 229.19 ± 8.52 µg/mL on L-O2, which suggested that R. davurica showed very low or no toxicity on hepatic cell viability. In this study, for the first time, we focused on the antiproliferative effects of R. davurica on human cancer cell lines, like HT-29 and SGC-7901 cells. As shown in Figure 4b, it is found out that R. davurica exhibited significant dose-dependent antiproliferative activities against HT-29 and SGC-7901 cells with IC50 values of 24.96 ± 0.74 and 89.53 ± 4.11 μg/mL, respectively. Meanwhile, Figure 4c shows that the inhibitory activities against both HT-29 and SGC-7901 cells significantly increased by the treatment with R. davurica in a time-dependent manner from 24 h-96 h at a dose of 150 μg/mL, although there was a decrease on SGC-7901 cells at the time from 72 h-96 h. Considering the drug safety, the evaluation of the general toxicity of R. davurica on the normal human hepatic cells (L-O2) was also conducted ( Figure 4d). As expected, R. davurica displayed a much higher IC50 at 229.19 ± 8.52 μg/mL on L-O2, which suggested that R. davurica showed very low or no toxicity on hepatic cell viability. To some extent, the efficacy of the folk medicine is deemed to rely heavily on the characteristics of its complex chemical components, which could further lead to the complicated and diverse pharmacological mechanisms. Accordingly, it is important to figure out the correlations between the antiproliferative activities and the contents of the corresponding chemical compositions. Our phytochemical investigations on R. davurica have revealed that flavonoids accounted for the prominent constituents in diversity and content and could be the most promising antiproliferative components of R. davurica.

Chemicals and Reagents
The reference standards of rutin, orientin, isoorientin, vitexin, isoquercitrin, astragaloside, luteolin, apigenin, quercetin and kaempferol were purchased from Shanghai Tauto Biotech (Shanghai, China). Formic acid (FA) and acetonitrile (ACN) of HPLC grade were provided by TEDIA Company To some extent, the efficacy of the folk medicine is deemed to rely heavily on the characteristics of its complex chemical components, which could further lead to the complicated and diverse pharmacological mechanisms. Accordingly, it is important to figure out the correlations between the antiproliferative activities and the contents of the corresponding chemical compositions. Our phytochemical investigations on R. davurica have revealed that flavonoids accounted for the prominent constituents in diversity and content and could be the most promising antiproliferative components of R. davurica.

Plant Materials
The barks of R. davurica originated from the northeast of China and were provided by Jikang Pharmaceutical Co., Ltd. (Baoding, China). Those raw materials were dried at 40 • C and powdered using a high speed disintegrator, then packed in sealed polyethylene bags and stored in a refrigerator at 4 • C until use. The authentication and identification of the specimens was kindly assisted by the taxonomist (Guangwan Hu) of Key Laboratory of Plant Germplasm Enhancement and Specialty Agriculture (Wuhan Botanical Garden), Chinese Academy of Sciences. A voucher specimen (No. 0031) was deposited in the herbarium of the Key Laboratory.

Sample Preparation
First, 100 g raw powered sample of R. davurica were accurately weighed and then extracted in an ultrasonic bath (KQ-3200DE, 300 × 150 × 150 mm, SHUMEI, Kunshan, China) with 60% ethanol for 30 min at room temperature, and the residue was re-extracted twice as described above. Next, the extracts were combined and filtered. After that, the filtrates were concentrated in a rotary evaporator under reduced pressure at 40 • C to afford the crude syrup residues. Later, the crude extracts were dispersed in water (100 mL) and subjected to liquid-liquid partition with petroleum ether (PE, b.p. 60-90 • C, to remove chlorophyll) and ethyl acetate (EA), successively. Afterwards, the EA layers were concentrated and fractionated by a polyamide column (45 cm × 5.3 cm). The column was eluted with distilled water to nearly colorless to remove water soluble impurities [26] and then with 80% ethanol to give the expected fractions. Finally, the fractions were concentrated and lyophilized in a freeze dryer to dryness, and the residues were stored at 4 • C for the subsequent analysis.

Cell Culture
Human cancer cell lines of HT-29 (colon carcinoma) and SGC-7901 (gastric carcinoma) were obtained from the China Center for Type Culture Collection (CCTCC). Cells were routinely grown in DMEM medium supplemented with 10% fetal bovine serum (FBS), glutamine (2 mM) and 1% penicillin (100 U/mL)-streptomycin (100 µg/mL). Furthermore, the cell lines were sub-cultured twice a week, and incubated in a humidified atmosphere with 5% CO 2 and 90% relative humidity (RH) under 37 • C. The number of living cells was assessed using a hematocytometer and phase-contrast microscopy. Finally, the cells over 80% confluence (growth phase) were used for the following cell antiproliferation assay [40].

Sulforhodamine B Antiproliferation Assays
The antiproliferative activities against HT-29 and SGC-7901 cells were estimated by means of the protein-staining sulforhodamine B (SRB) microculture colorimetric assay with some modification [41]. The assay shows high sensitivity to total cellular protein content and linearity to cell density, which has been used for in vitro anticancer evaluation at the National Cancer Institute (Bethesda, MD, USA) [42]. Briefly, a 100-µL cell suspension of the trypsinized monolayer cells in DMEM medium was seeded into 96-well plates with a density of 1.0 × 10 4 cells per well. After incubation at 37 • C in 5% CO 2 and 90% relative humidity for 24 h to resume exponential growth and stabilization, the culture medium was removed carefully, and an aliquot of 100 µL R. davurica extracts was added into each well in the plates. All of the test samples were first dissolved in DMSO and further diluted with the medium to the final DMSO content less than 0.1%, which was innocuous on cell growth and proliferation. After incubation for another 72 h, cells were fixed with 50 µL 10% cold (4 • C) trichloroacetic acid (TCA) for 30 min at 4 • C. Afterwards, the supernatants were washed out with deionized water five times and air dried at room temperature. Subsequently, the dried plates were stained with 100 µL 4 mg/mL SRB in 1% acetic acid solution for 30 min at room temperature. Next, the plates were washed five times with 1% acetic acid to remove the unbound SRB and then air dried overnight. The protein-bound SRB was solubilized with 150 µL of 10 mM Tris base (pH 10.5), and the plates were left on a gyratory shaker for 10 min. The complete medium with less than 0.1% DMSO was used as the control. The optical density (OD) value of each well was determined with a 96-well plate reader (Tecan) at a wavelength of 540 nm. The inhibition rate (%) was calculated using the equation: inhibition = (ODC − ODT)/ ODC × 100%, where ODC and ODT were the OD values of controls and samples of R. davurica, respectively. Each test sample solution was performed in triplicate, and the results were expressed as the means ± SD (standard deviation).

HPLC Fingerprint Analysis
Experiments for phytochemical fingerprint analysis of R. davurica were carried out using a Thermo Accela 600 HPLC system (Thermo Fisher Scientific, Waltham, MA, USA). The chromatographic separation was performed on a Waters SunFire™ RP-C18 column (150 mm × 4.6 mm, 3.5 µm). The mobile phases were composed of 0.1% aqueous formic acid (A) and acetonitrile (B). An aliquot of a 10-µL sample solution was injected into the HPLC system, and the linear eluting gradient was as follows: 20% B in 0-2 min, 20%-45% B in 2-15 min, 45%-70% B in 15-31 min and 70% B in 31-35 min. The column temperature was maintained at 30 • C. The flow rate was 0.4 mL/min, and the online UV spectrum was monitored at the wavelength of 360 nm.

ESI-MS/MS Analysis
The ESI-MS/MS analysis was carried out using a TSQ Quantum Access MAX mass spectrometer (Thermo Fisher Scientific) equipped with an ESI source operating in Auto-MS n mode to obtain fragmentation. The negative ionization mode was applied, and the optimized instrument settings were set as follows: source voltage, 3.0 kV; cone voltage, 40.0 V; desolvation temperature, 350 • C; capillary temperature, 250 • C; nebulizing gas flow rate, 6.0 L/min; sheath gas (N 2 ) pressure, 40 arb; Aux gas (N 2 ) pressure, 10 arb; collision energy (CE), 10 V; collision energy grad (CE grad), 0.035 V/m. Mass spectra data were obtained with the full-scan mode for m/z in the range from 150 to 1500, and the nine most abundant ions were selected for the further MS 2 spectra. All data acquisition and analysis were performed using the Thermo Xcalibur ChemStation (Thermo Fisher Scientific).

Quantitative Analysis of Flavonoids Compounds
For the quantitative analysis, three representative standards (rutin, vitexin and kaempferol) were applied to calculate the individual component concentration present in R. davurica by HPLC. Each standard was accurately weighed, prepared in methanol, and then, the calibration curves were established by diluting the standard stock solutions into a series of concentrations measured at 360 nm. Six different concentrations (1.0-333 µg/L) of each standard were used for the calibrations and measured in triplicate. Method validations included accuracy, correlation determination (R 2 ), limit of detection (LOD), limit of quantitation (LOQ), linearity range, repeatability, precision and recovery. Finally, the quantification of each individual compound was calculated based on the calibration curves of the standards for the corresponding HPLC peak area values of R. davurica. In the present study, individual components are presented as mg per 100 g DW, and the total content was defined as the sum of each corresponding quantified component.

Conclusions
In this study, a HPLC-ESI-MS/MS method was employed to comprehensively analyze the phytochemical fingerprint profile of R. davurica. As a result, 32 peaks were detected. Structural identifications and characterizations showed that 23 of them were identified as flavonoids based on the comparisons of their ESI-MS/MS data with the corresponding standards and fragmentation pathways reported in the previous literature. Among 32 components identified, including orientin (Peak 1), isoorientin (Peak 2), diosmetin 7-O-glucoside (Peak 5), astragaloside (Peak 6) and luteolin 5-O-glucoside (Peak 7), 14 components were identified in R. davurica for the first time. For individual component quantitative analysis, flavonoid glycosides, flavone aglycones and anthraquinones exhibited higher content of 40.98%, 26.26% and 26.2%, respectively, which consisted of the major components of R. davurica. For the discrepant contents of individual components, vitexin, aromadendrin and question accounted for 86.07, 24.18 and 42.17 mg/100 g DW, respectively. Meanwhile, to better correlate the phytochemical components with their pharmacological activities, the antiproliferative activities against human cancer cells were tested in vitro and exhibited distinct inhibitory effects against HT-29 and SGC-7901 cells with IC 50 values at 24.96 ± 0.74 and 89.53 ± 4.11 µg/mL, respectively. Finally, R. davurica displayed a much higher IC 50 at 229.19 ± 8.52 µg/mL on L-O2, which suggested that R. davurica showed very low or no toxicity on hepatic cell viability. The current study revealed for the first time the correlations between the flavonoids of R. davurica with their antiproliferative activities, which indicated that the fingerprint profile of flavonoids and their anticancer activities could provide valuable information on the quality control for this herbal medicine and its derived natural remedies.