Bioactive Constituents of F. esculentum Bee Pollen and Quantitative Analysis of Samples Collected from Seven Areas by HPLC

Bee pollen contains all the essential amino acids needed by humans. China is the largest producer of bee pollen in the world. In the present study, we identified 11 fatty acids in F. esculentum bee pollen oil by GC-MS analysis, and 16 compounds were isolated from F. esculentum bee pollen by column chromatography and identified. A high-performance liquid chromatography-diode array detector (HPLC-DAD) method was established for the quality control of F. esculentum bee pollen. A validated HPLC-DAD method was successfully applied to the simultaneous characterization and quantification of nine main constituents in seven samples collected from seven different areas in China. The results showed that all standard calibration curves exhibited good linearity (R2 > 0.999) in HPLC-DAD analysis with excellent precision, repeatability and stability. The total amount in the samples from the seven regions ranged from 23.50 to 46.05 mg/g. In addition, seven compounds were studied for their bioactivity using enzymic methods, whereby kaempferol (3) showed high α-glucosidase inhibitory activity (IC50: 80.35 μg/mL), ergosterol peroxide (8) showed high tyrosinase inhibitory activity (IC50: 202.37 μg/mL), and luteolin (1) had strong acetylcholinesterase inhibitory activity (IC50: 476.25 μg/mL). All results indicated that F. esculentum bee pollen could be a nutritious health food.


Introduction
Bee pollen (BP), one of the hive products in addition to honey, royal jelly and propolis, is gaining attention due to the presence of bioactive compounds associated with beneficial health properties [1,2]. The BP composition varies due to the plant species and is also influenced by age, nutritional condition of the plant, differences in gathering area and time as well as environmental conditions during pollen development [3][4][5][6]. Monofloral pollen pellets maintain organoleptic and biochemical properties similar to those of the original plant, whereas the multifloral pollen has variable properties of more than two original plants [7]. Generic bee pollen composition data were considered sufficient for most purposes, but now the usefulness of bee pollen-specific composition data is increasingly being acknowledged [8,9]. Recent research has also shown that bee pollen possesses therapeutic benefits for improving the cardiovascular system, stimulating body immunity [10], promoting antitumor effects, Seven batches of BPs, consisting of Fagopyrum esculentum Moench from different production areas in China, were collected in 2016. These samples were identified by the corresponding author (NB). Their voucher specimens were deposited in a refrigerator at 4 • C at the Department of Pharmaceutical Engineering, Northwest University, China. BPSX was from Shannxi, BPAH was from Anhui, BPNM was from Neimeng, BPHN was from Heinan, BPGS was from Gansu, BPHB1 was from Hebei and BPHB2 was from Hubei.

Saponification and Methylation for Fatty Acids (FAs)
The F. esculentum pollen oils were obtained by a pressing method under low temperature, and processed by saponification and methylation reactions for further GC-MS analysis. Briefly, 0.5 g of F. esculentum pollens oil prepared above was dissolved in 4 mL methanol solution containing 0.5 mol/L of potassium hydroxide, stirred for 2 min, and kept in a 60 • C water bath for 1 h in a capped test tube to induce saponification for derivatization. Then, 3 mL of methanol solution containing 10% concentrated sulphuric acid was added, the sample was again placed in a 60 • C water bath for 1 h. After reactions, the sample was cooled to room temperature. Subsequently, 3 mL of deionized water and 6 mL of n-hexane were added for extraction. After centrifugation (13,000 rpm, 10 min), the supernatant was obtained and dried using a nitrogen flow. The obtained solution was stored in a refrigerator at 4 • C and filtered through a 0.45 µm nylon-membrane filter prior to injection into the GC-MS analysis.

GC-MS Analysis
The FA profiles in the pollen oils were analyzed by GC-MS as their corresponding methyl ester. The oil sample was analyzed three times under the same conditions. The GC-MS (Agilent 6590N Network System, M 5973N) instrument was coupled with a Rtx-5MS column (30 m × 0.25 mm × 0.25 µm (5% phenylmethylsiloxane)). Helium (purity, 99.99%) was used as the carrier gas at a flow rate of 3.0 mL/min. The injection volume is 1 µL by split (1:30). Injector, source and oven temperatures were 250 • C, 200 • C and 120 • C, respectively. The initial oven temperature 120 • C held for 2.5 min and then ramped at 10 • C/min to 180 • C, ramped at 1.5 • C/min to 210 • C, ramped at 5 • C/min to 250 • C and last for 3 min.

Extraction and Isolation
The BPs (15 kg from Shanxi) was percolated with 90% EtOH at room temperature for three times, then the EtOH extracts were combined. The extract (6.32 kg) was suspended in H 2 O (6.5 L) and then extracted exhaustively with PE, EtOAc and n-BuOH (6.5 L each) 3 times. The n-BuOH solution containing 195.3 g of solid was chromatographed on a silica gel column (15 L, 8.0 cm × 120 cm) with CH 2 Cl 2 -MeOH (100:1-0:1) gradient to give seven crude fractions (Fr.1-7).

Preparation of Sample Solutions
Each BP material was ground to powder and sifted through an 80 mesh sieve. Subsequently, 5 g of pulverized samples were accurately weighed, and extracted with ultrasonication using 40 mL of 90% methanol (volume fraction) for 1.5 h. After centrifugation (13,000 rpm, 10 min), the supernatant was concentrated and transferred into a 5 mL volumetric flask. The obtained solution was stored in a refrigerator at 4 • C and filtered through a 0.45 µm filter prior to HPLC analysis.

Preparation of Standard Solutions
Nine reference standards were accurately weighed (10,000 mg) respectively, then dissolved in methanol and a small amount of dimethyl sulfoxide to final concentrations of 1.0 mg/mL. Every standard stock solution was diluted with methanol to five appropriate concentrations (500 µg/mL, 250 µg/mL, 100 mg/mL, 50 µg/mL and 10 µg/mL). All the standard solutions were stored at 4 • C in darkness for HPLC analysis.

Identification and Quantification
Identification of these analytes were carried out by comparing the HPLC retention time and UV spectra of target peaks with those of the standards. Quantification was performed on the basis of linear calibration plots of the peak areas versus the concentration. The results are presented in Figure 2. Quantification was performed on the basis of linear calibration curves. a refrigerator at 4 °C and filtered through a 0.45 μm filter prior to HPLC analysis.

Preparation of Standard Solutions
Nine reference standards were accurately weighed (10,000 mg) respectively, then dissolved in methanol and a small amount of dimethyl sulfoxide to final concentrations of 1.0 mg/mL. Every standard stock solution was diluted with methanol to five appropriate concentrations (500 μg/mL, 250 μg/mL, 100 mg/mL, 50 μg/mL and 10 μg/mL). All the standard solutions were stored at 4 °C in darkness for HPLC analysis.

Identification and Quantification
Identification of these analytes were carried out by comparing the HPLC retention time and UV spectra of target peaks with those of the standards. Quantification was performed on the basis of linear calibration plots of the peak areas versus the concentration. The results are presented in Figure  2. Quantification was performed on the basis of linear calibration curves.    (5), chlorogenic acid (6), rutin (7), 2-β-d-glucopyranosyloxy-1-hydroxytrideca-3,5,7,9,11-pentayne (10) (13), quercitrin (14), oleanolic acid (15) and tyrosol (16). a 0.2 mol/L Na 2 CO 3 solution to each well after 15 min. The absorbance at a wavelength of 405 nm was measured with a microplate reader. Acarbose was used as a positive control. The specific grouping is shown in Table 1. Table 1. Specific grouping and reaction system for α-glucosidase inhibitory activity. ; finally, 20 µL of substrate (l-tyrosine) was added to each well to trigger the reaction, and the 96-well plate was further incubated at 37 • C. After 30 min, the absorbance at a wavelength of 405 nm was measured by a microplate reader. Kojic acid was used as a positive control. The specific grouping is shown in Table 2. Table 2. Specific grouping and reaction system for tyrosinase inhibitory activity.  Table 3. Table 3. Specific grouping and reaction system for acetylcholinesterase inhibitory activity.

Optimization of Extraction Method
To determine the optimal extraction conditions and enhance the overall response of those investigated compounds in the HPLC analysis, extraction methods (refluxing and sonication), extraction solvents (100%, 90%, 75%, 50% and 25% methanol-water), number of extractions (1, 2, 3 and 4 times), and extraction times (30, 60, 90 and 120 min) were investigated individually by using a univariate approach (default values: extraction method, sonication; extraction solvent, 50% methanol; extraction number, 2 times; extraction time, 90 min) on the sample BPSX. The results suggested that the established extracted method (each sample was extracted three times with 50% methanol under ultrasonication condition for 90 min) was optimal for HPLC analysis. The results are shown in Figure 4.

HPLC Method Validation
The HPLC method was validated by determining the linearity, limit of detection (LOD), limit of quantitation (LOQ), precision (inter-day and intra-day), repeatability, stability, and accuracy.

Linearity, LODs and LOQs
Linear calibration curves were established by plotting the peak area (Y) versus the corresponding concentration (X, μg/mL) of each compound. All calibration curves exhibited good linear regressions (R 2 ≥ 0.999) within the tested concentration ranges ( Table 5). The lowest concentration of working solution for calibration use was diluted to a series of appropriate concentrations. They were then measured and checked until the signal-to-noise ratio (S/N) for each compound was approximately 3 for LOD and 10 for LOQ. The LODs and LOQs of all analytes were less than 0.23 and 0.64 μg/mL, respectively.  (6), rutin (7), catechin (13), quercitrin (14), and tyrosol (16).

HPLC Method Validation
The HPLC method was validated by determining the linearity, limit of detection (LOD), limit of quantitation (LOQ), precision (inter-day and intra-day), repeatability, stability, and accuracy.

Linearity, LODs and LOQs
Linear calibration curves were established by plotting the peak area (Y) versus the corresponding concentration (X, µg/mL) of each compound. All calibration curves exhibited good linear regressions (R 2 ≥ 0.999) within the tested concentration ranges ( Table 5). The lowest concentration of working solution for calibration use was diluted to a series of appropriate concentrations. They were then measured and checked until the signal-to-noise ratio (S/N) for each compound was approximately 3 for LOD and 10 for LOQ. The LODs and LOQs of all analytes were less than 0.23 and 0.64 µg/mL, respectively.

Precision, Repeatability, and Stability
As all nine analytes could be detected and quantified in BPSX, BPSX was selected for the tests of precision, repeatability, and stability. The intra-day and inter-day precisions were successively evaluated by a prepared sample solution under optimal conditions within one day and duplicating this analysis once a day for three consecutive days, respectively. A stability test of the sample solution was conducted at 0, 2, 4, 8, 12, 24 and 48 h. Repeatability was determined by analyzing six independently prepared solutions of sample BPSX. The relative standard deviation (RSD) of peak area for each marker compound was taken as a measure ( Table 6). The RSD values of the reference compounds were found in the ranges of 0.71-1.45%, 0.95-2.25%, 0.47-2.06%, and 1.60-3.15% for intra-day variations, inter-day variations, stability, and repeatability. Table 6. Precision, repeatability, stability, and recovery of the analytes.

Recovery
To check the accuracy of the analytical method, a recovery test was performed. In the test, a known amount of standards was added into 0.1 g of sample BPSX previously quantified. The spiked samples were extracted, processed and quantified per the methods above. Three replicates were performed for the test. The recovery values ranged from 97.25 to 102.07%, and RSDs values were less than 3.49%. These results indicate that the established method is accurate enough for quantitative analysis.

Quantitative Determination of Nine Compounds
The proposed analytical method was then successfully applied to the simultaneous quantification of seven BPSs collected from different areas in China. The results (Table 7) indicated that the contents of nine compounds varied greatly among different samples. The total content of these investigated compounds reached as high as 46.05 mg/g in sample BPNM, which was cultivated in Neimeng. However, the content was only 23.50 mg/g in BPGS, for the sample cultivated from Gansu. Flavonoids were considered the most abundant constituents in BPSs. Rutin, an important bioflavonoid, is abundantly found in various foodstuffs [17]. Rutin has been acknowledged for its protective and beneficial effects on various aspects of the biological system, rutin possesses sufficient potential for increasing immune activity by cellular and humoral mediated mechanisms [33]. Analyte 1 (luteolin) and analyte 2 (resveratrol) were detected in seven samples, the contents of analyte 1 varied from 6.23 to 10.94 mg/g and analyte 2 varied from 3.26 to 5.25 mg/g. In addition, compound 6 could not be detected in many analyzed samples, only be detected in BPSX and BPNM, and the contents were 1.45 and 1.12 mg/g, respectively. Chlorogenic acid and caffeic acid are common bioactive compounds [34,35]. Chlorogenic acid has the effect of protecting liver, inhibits cancer cell growth and has a beneficial effect on ameliorating aging-related diseases [24]. Caffeic acid has also been reported have to a wide antibacterial effect [25]. From our data, analyte 6 (chlorogenic acid) varied from 1.12 to 1.45 mg/g, and analyte 5 (caffeic acid) varied from 3.47 to 6.00 mg/g. Look at all the data, the highest contents of analyte 1 is 10.94 mg/g in sample BPNM, the lowest contents of analyte 7 (rutin) is 1.05 mg/g in sample BPHB1.

Analysis of α-Glucosidase Inhibitory Activity
Glucosidase is an important member of the sugar metabolism pathway in vivo, and α-glucosidase is directly involved in the metabolic pathways of starch and glycogen [36]. By inhibiting α-glucosidase, the chemical metabolism of sugar can be reduced, thereby achieving a hypoglycemic effect [37].

Conclusions
In this study, 16 compounds were isolated from F. esculentum bee pollen, including flavone compounds, phenolic acids, terpenoids, and 11 fatty acids were analyzed in F. esculentum bee pollen oil. Saturated fatty acids were 30.86%, unsaturated fatty acids were 69.14%, of which (9Z,2Z,15Z)-octadecatrienoic acid content was as high as 36.25%, indicating that the unsaturated fatty acid content of F. esculentum bee pollen is high, and it is a green and healthy food. A HPLC method with high stability, good repeatability and high precision was established, and the content of nine compounds in F. esculentum bee pollen from seven different habitats was quantitatively analyzed. From the results, the content of bee pollen collected in Neimeng was the highest (46.05 mg/g), and the content of bee pollen collected in Gansu was the lowest (23.50 mg/g). Kaempferol (3) showed high α-glucosidase inhibitory activity (IC 50 : 80.35 µg/mL), ergosterol peroxide (8) showed high tyrosinase inhibitory activity (IC 50 : 202.37 µg/mL), and luteolin (1) had strongly acetylcholinesterase inhibitory activity (IC 50 : 476.25 µg/mL). Furthermore, bee pollen has been used as a health food supplement for many years due to its abundant nutrient properties including proteins, polysaccharide, and lipids [42]. The average protein content of bee pollen is 24.65% (10-40% dry weight) [9]. The proportion of amino acids in bee pollen is also high and there are many kinds of amino acids. Studies have proved that the protein in bee pollen accounts for 29.18% of its dry weight [43]. Polysaccharides, namely carbohydrates, are the most abundant component in bee pollen, accounting for 18.9-57.6% of bee pollen [44]. Lipids are an important component of bee pollen, and their content accounts for about 1-20% of its dry weight [45]. Other trace elements including vitamins, minerals, enzymes, nucleic acids have also been found in pollen. Therefore, it has antioxidant, anti-inflammatory and immune-enhancing capabilities [46][47][48][49]. Our research has proved that isolated metabolites possess potential anti-aging activity. These compounds could work synergistically with nutrients in pollen and provide more evidence for the beneficial effects of pollen. In conclusion, F. esculentum bee pollen is a kind of natural nutrition and health food for anti-aging, beauty and improving human immunity.
At present, most of reports on bee pollen are aimed at the study of extracts. Therefore, the innovation of this paper is the separation and identification of extracts. The quantitative analyses of bee pollen from seven different places of origin were also carried out.

Conflicts of Interest:
The authors declare no competing interest.