Compositional changes induced by UV-B radiation treatment of common bean and soybean seedlings monitored by capillary electrophoresis with diode array detection
Giovanni Dinelli1, Irene Aloisio1, Alessandra Bonetti1, Ilaria Marotti1,
Alejandro Cifuentes2*
1Department of Agroenvironmental Science and Technology, University of Bologna, Viale Fanin 44, 40127 Bologna, Italy.
2Institute of Industrial Fermentations (CSIC), Juan de la Cierva 3, 28006 Madrid, Spain.
Running Title: Monitoring compositional changes in beans by CE
Keywords: bean sprouts; flavonoids; UV treatments; antioxidants; food analysis.
Correspondence: Dr. Alejandro Cifuentes, Institute of Industrial Fermentations (CSIC), Juan de la Cierva 3, E-28006 Madrid, Spain. E-mail:Fax: +34-91-5644853 Tel: +34-91-5618806 Ext 315.
Abbreviations: CE, capillary electrophoresis; DAD, diode array detection; FSCE, free solution capillary electrophoresis
Summary
In this work, a new capillary electrophoresis (CE) method with diode array detection (DAD) was developed for the monitoring and quantitation of flavonoids in different beans treated and untreated with UV-B radiation. Namely, flavonoid concentration was monitored in UV-B treated and untreated sprouts of three common beans (Zolfino ecotype, cv. Verdone, cv. Lingua di Fuoco) and one soybean (cv. Pacific). After acid hydrolysis of extracts, the CE-DAD method provides reproducible quantitative determinations of daidzein, glycitein, genistein and kaempferol at ppm level in these natural matrices within a relatively short time (less than 16 min). Total flavonoid content determined by CE-DAD was 159 ± 8, 26 ± 2, 13 ± 1, 1.3 ± 0.3 µg/g fresh weight for untreated sprouts of Pacific soybean, Verdone bean, Zolfino bean and Lingua di Fuoco bean, respectively. UV-B treatment caused no significant quantitative effect on Pacific soybean sprouts, whereas it enhanced total isoflavone content 1.5, 1.8 and 3.2-fold in Verdone, Zolfino and Lingua di Fuoco beans, respectively. The proposed method shows i) the potentialities of bean sprouts as a natural source of bioactive compounds (antioxidants); ii) the technological role of UV-B treatment for sprout isoflavone enrichment and iii) the good capabilities of CE-DAD to monitor this process.
1 Introduction
Crop species belonging to the Leguminosae family, including soybean, pea, lentil, peanut, beans and other podded plants, have been cultivated for thousands of years. Recently, a new interest has grown relative to the consumption of legumes due to their recognized health benefits originating that several of these legumes can be considered as functional foods [1]. The potential health benefits of legumes have been mainly attributed to the presence of secondary metabolites such as phenolic compounds that possess antioxidant properties [2]. Among the different legumes, soybeans have attracted much attention due to their high concentration of isoflavones, phenolic compounds that in some cases have a similar structure to human estrogens. Due to these properties, isoflavones are reported to protect against many hormone-dependent diseases such as osteoporosis, cancer and cardiovascular diseases [3, 4, 5]. In soybean grains the main forms of isoflavones include various glucosidic conjugates of the aglycones daidzein, genistein and glycitein [6]. Although widely distributed in the Legumonosae family, isoflavones are present in soybeans in concentrations approximately 50-fold higher than in other legumes [7].
Sprouting of seeds is a processing that can increase the nutritive value (vitamin C, proteins) and the health qualities (isoflavones, antioxidant capacity) of legume foods [8]. Germinated bean sprouts are also served as staple vegetables and used in soups, salads, and side dishes in many Asian countries. Germination changes the nutrient composition (including functional substances) and also removes antinutrient factors of the seed making sprouts safe for the diet [9].
The health benefits of bean sprouts will mainly depend on the amount of biologically active compounds present in the bean [10]. In this regard, the stimulation of secondary metabolism following plant exposure to UV radiation or elicitation with chemical treatments seems to be an interesting field of research [11, 12]. Isoflavones are synthesized as a part of phenylpropanoid pathway. The phenylpropanoid pathway has multiple branches common to legumes, which provide numerous secondary metabolites. These compounds have an important role in normal development and can serve as protection (phytoalexins) against many environmental sources of stress such as nutrient deficiency, prolonged cold, pathogen attack, and exposure to UV light [13, 14]. Up to date, UV treatment or chemical elicitation on legume plants were investigated to improve the knowledge on the biochemical bases of plant defense against biotic and abiotic stress [13, 15, 16].
UV-irradiated or chemically treated plants have usually been monitored spectrophotometrically in a very approximate way [17] or by using more informative techniques such as high pressure liquid chromatography (HPLC) or thin layer chromatography (TLC) [18]. Interestingly, capillary electrophoresis (CE) has been shown to have a great potential in terms of efficiency and speed of analysis to investigate polyphenolic compounds [19, 20], providing in many cases complementary information to that obtained by HPLC [21]. Regarding the analysis of Leguminosae by CE, this technique has only been used, to our knowledge, to detect proteins in beans (mainly soy proteins) [22, 23], niacin in legumes [24] or daidzein and genistein [25, 26] and phospholipides [27] in soy products.
The goals of this work are i) to investigate the effect of UV-radiation on different types of Leguminosae bean sprouts and ii) to demonstrate that CE can be a suitable technique to monitor flavonoid compositional changes induced by UV-radiation in these natural matrices. For this purpose a new CE procedure based on UV-diode array detection (DAD) was developed and characterized.
2 Experimental
2.1 Chemicals and standards.
All reagents and solvents used for extractions and CE analyses were of analytical or HPLC grade. Hydrochloric acid, ammonium hydroxide, ammonium acetate, and sodium dodecyl sulfate were from Merck (Darmstadt, Germany). Methanol and acetonitrile were from Scharlau (Barcelona, Spain). Water was deionized by using a Milli-Q system (Millipore, Bedford, MA, USA). Standards of daidzein, genistein, glycitein and kaempferol were purchased by Indofine Co, Hillsborough, USA.
2.2 Plant materials.
Phaseolus vulgaris L. seeds of the Italian ecotype “Zolfino del Pratomagno” were directly sampled from local farmers (Pratomagno, Arezzo, Italy). In addition, two commercial common bean cultivars (“Verdone”, “Lingua di Fuoco”) and one soybean cultivar (“Pacific”) were employed. In our earlier studies [28], the total flavonoid content of the four accessions was determined at the seed level. In the Zolfino bean three flavonoid glycosides (kaempferol 3-O-glucoside, kaempferol 3-O-xylosylglucoside and a kaempferol monoglucoside not yet identified) were found and the total kaempferol content was 350 g/g fresh seed. The same three kaempferol derivatives, observed in the Zolfino bean, were found in the seeds of “Verdone” bean with a total kaempferol content of 410 g/g fresh seed [29]. In contrast, no flavonols were observed in the seeds of “Lingua di Fuoco” bean and no appreciable amounts of soybean isoflavones were detected in all the three investigated common bean accessions [28, 29]. Finally, in the soybean accession, after acid hydrolysis three main isoflavones were detected (daidzein, genistein and glycitein) with a total content of 1806 g/g fresh seed [29].
2.3 Seedling growth and UV-B irradiation.
Seeds of the four different legume accessions were placed in Petri dishes on moist sterile sand and incubated in the dark at 25°C in a growth chamber. After 3 days from the germination, the seedlings in the Petri dishes were placed directly under UV-B light for 30 min (305 nm, 50 mWatt cm-2, 10 cm distance from Petri dishes). After additional 12 h of growth in the dark at 25 °C the UV-treated and control seedlings were extracted.
Two levels of UV-B radiation (no UV-B and 30 min UV-B radiation) and four accessions of legume crops (three common bean and one soybean genotypes) were studied. Ten seedlings per Petri dish and four Petri dishes per legume accession were arranged.
2.4 Flavonoid extraction and hydrolysis.
Three g (fresh weight) of seedlings were dispersed in a mixture of 12 ml of acetonitrile and 3 ml of 0.1 M HCl by stirring for 3 h [28, 30, 31]. The mixtures were centrifuged at 15000 rpm for 10 min at 10° C. Clear supernatant was filtered through a 0.45 µm nylon filter (Millipore, Bedford, MA, USA) and dried by using a vacuum pump. Samples were reconstituted with 5 ml of 80% methanol in ultrapure water (v/v). One ml of these extracts was mixed with 3 ml of 1 M HCl and incubated for one night in the dark. After the incubation, the samples were heated in sealed vials at 80°C for 2 hours. The acid hydrolysis procedure was shown to completely convert glycosides in aglycones by using flavonoid standards [28]. Subsequently, flavonoids were extracted with ethylacetate, evaporated in vacuum and dissolved in an appropriate volume (between 1 and 0.25 ml) of 50% methanol in ultrapure water (v/v). These solutions were filtered through 0.45 µm nylon filters (Millipore, Bedford, MA, USA) prior to CE analysis.
2.5 CE-DAD analysis.
Experiments were performed using a P/ACE System 5500 instrument with a diode array detector (DAD) from Beckman (Fullerton, CA, USA). A GOLD software also from Beckman was used for system control and data handling. Fused-silica capillaries were from Composite Metal Services (Worcester, UK) with 75 µm of ID and 37 cm of total length and 30 cm of effective length. Before first use, all capillaries were conditioned by rinsing for 30 min with 0.1 M NaOH followed by ultra pure water for 15 min. At the beginning and at the end of the day, capillaries were rinsed for 15 min with ultrapure water. The electropherograms were obtained by applying a potential difference of 15 kV and a temperature of 25 °C. Samples were injected under constant pressure of 0.5 psi for 10 s. UV-Vis spectra were obtained in the 200-500 nm range and the electropherograms were recorded at 214 and 260 nm, simultaneously.
The separation efficacy was measured by the number of theoretical plates (N) according to the formula N = 5.54 (tm/w)2 where tm is the migration time of a compound and w is the peak width at half-peak height [32]. The resolution (R) between each pair of adjacent peaks was calculated according to the formula R = [2(tm2- tm1)/(w1+w2)] where tm1and tm2 are the migration times, and w1 and w2 are the respective baseline peak widths of two adjacent peaks [33]. The R values were determined for each pair of the four main flavonoids (peaks 1-4) and employed to calculate the mean resolution. The detection and quantification limits were calculated according to the IUPAC method [34]. If the concentration of a flavonoid analyte was below the respective limit of quantification, the sample extract was dried and properly re-dissolved (methanol 80%) in order to fall beyond the quantification limit.
3 Results and discussion
3.1 Setting up the CE-DAD method.
To carry out the optimization of the CE method, flavonoid extracts obtained as indicated in section 2.4 from treated and untreated Pacific soybean and Verdone bean were used. Initially, and in order to investigate the effect of the buffer ionic strength onto the CE separation of the four samples (i.e., untreated Pacific soybean, UV-treated Pacific soybean, untreated Verdone bean and UV-treated Verdone bean), four different buffers were prepared using ammonium acetate at different concentrations (12, 25, 50 and 75 mM) at the same pH 9. Both the migration time and the overall resolution of the peaks detected in each sample increased with increasing the concentration of ammonium acetate in all the samples although not complete resolution could be achieved for all the analytes. The resolution was slightly similar using the 50 and 75 mM ammonium acetate buffers, however, the 50 mM buffer provided the best compromise between an adequate separation time and an acceptable resolution. In order to further improve this separation, micellar electrokinetic chromatography was tested preparing buffers with different concentrations of SDS (5, 10, 20 and 40 mM) added to the 50 mM ammonium acetate at pH 9.The addition of SDS did not improve significantly the separation resolution, but caused a remarkable increasing of migration times, therefore, the use of SDS was abandoned. The effect of the buffer pH was then investigated testing pH values ranging from 8.5 to 11 in 0.5 steps using the 50 mM ammonium acetate. Additionally, the effect of methanol as organic modifier (from 0 to 40%) on the separation speed and resolution was also investigated. In general, by increasing the pH of the buffer the resolution of the separation improved without significant increasing of the analysis time for the four samples. On the other hand, an increase of the methanol percentage originated an increasing of the migration times and with that an improvement of the separation resolution. A compromise between resolution and analysis speed was found for all the samples using the 50 mM ammonium acetate at pH 10.5 buffer containing 20% of methanol and this separation medium was chosen for all subsequent analyses.
The electropherograms of the four analyzed samples are reported in Figures 1a and 1b. CE-DAD analyses of both treated and untreated samples evidenced three main peaks, hereinafter called compounds 1-3, representing more than 95% of the total area of electropherograms for Pacific soybean and Verdone bean (see Figure 1). Interestingly, the same three peaks were also observed in the electropherograms of the other treated and untreated common beans (i.e., Zolfino and Lingua di Fuoco) as can be seen in Figure 2. This was corroborated by the similar analysis time and UV spectra for peaks 1, 2 and 3 in all analyzed samples. In addition, a fourth main peak (hereinafter called compound 4) was observed in the treated and untreated Verdone (Figure 1b) and Zolfino samples (Figure 2a). Moreover, the UV spectra and migration time of peak 4 were equivalent in Verdone and Zolfino electropherograms, indicating that this compound was the same for both samples.
The four mentioned peaks were employed for the assessment of the figures of merit of this new CE-DAD method. Thus, a mean efficiency of 72000 ± 8700, 71000 ± 5700, 59000 ± 3600 and 38000 ± 4600 theoretical plates was found for compounds 1, 2, 3 and 4, respectively. In addition, also the mean resolution was satisfactory giving R values higher than 2 in all the cases. Moreover, the separation procedure exhibited an acceptable reproducibility as can be deduced from the results given in Table 1 for intra-day (6 runs in the same day) and inter-day (12 runs in three different days) experiments. Thus, relative standard deviations (RSD) for migration times were below 1.7 and 2.4% for intra-day and inter-day experiments, respectively, corroborating the usefulness of this approach. Besides, RSD values calculated for corrected peak areas (i.e., peak area divided by migration time) of compounds of Table 1 were lower than 9% for both intra-day (6 runs in the same day) and inter-day assays (12 runs in three different days). As indicated above, these reproducibility data are given only for the CE-DAD analysis, i.e., the effect of the sample pretreatment on reproducibility was not considered.
3.2 Identification and quantification of compounds separated by CE-DAD.
The identification of compounds 1, 2, 3 and 4 was based on the comparison of their migration times and UV-spectra with those of standards. As previously reported, compounds 1-3 were detected by CE-DAD in the four investigated samples including one soybean and three common beans and, moreover, these peaks show a typical isoflavonoide UV-spectra. On the other hand, peak 4 was only detected in Verdone and Zolfino beans (see Figure 1b and Figure 2a) and it showed a flavonol-like UV-spectra. Considering that daidzein, genistein and glycitein are the most abundant aglycon isoflavones in soybean seed and seedling [30, 31], the standards of these three isoflavones were considered as a good possibility. Moreover, in a previous investigation on different accessions of Zolfino bean, the most abundant flavonol observed at the seed level was the kaempferol [28], therefore, the standard of this flavonol was also included as the fourth possible compound.
Figure 3 shows the CE-DAD electropherogram obtained after injecting a solution containing the four proposed standards (i.e., daidzein, genistein, glycitein and kaempferol) under the same conditions of Figure 1 and 2. As can be seen, comparing Figure 3 with Figure 1 or Figure 2, there is a good agreement between the migration times of the four standards and the compounds 1-4 detected in the real samples, what can be considered a proof that indicates they are the same compounds.
In order to corroborate the identification of these four compounds, their UV spectra were also compared. Figure 4 shows a comparison between the UV spectra of compounds 1 to 4 from standards and from a real bean sample, in this case from Verdone. As can be seen, there is a good match between their UV spectra in all the cases with coefficients of correlations ranging from 0.994 to 0.999. Similar results were obtained for the other investigated samples (data not shown). Therefore, on the basis of comparison with migration times and UV spectra it can be concluded that compounds 1, 2, 3 and 4 found in the real samples correspond to glycitein, daidzein, genistein and kaempferol, respectively.
3.3 Flavonoid content in legume sprouts and effects of UV-B treatment.
Once the main flavonoid compounds were identified, the next step was to carry out their quantitative determination in order to investigate the effect of UV-treatment on their concentration. Thus, Table 2 shows the main parameters for the quantification of these four flavonoids. Calibration curves were based on seven different concentrations (ranging from 1 to 80 mg/L) of a mixture of standard daidzein, genistein, glycitein and kaempferol injected in triplicate. As evidenced by the determination coefficients, ranging from 0.990 to 0.999, a linear response for all flavonoids was observed in the studied concentration interval. The detection limit varied from 0.25 mg/L for glycitein to 1 mg/L for kaempferol, while the quantification limit ranged from 1.54 mg/L for glycitein to 4.11 mg/L for kaempferol. The performance of the present method in the quantification of selected flavonoids was comparable with that of other CE-DAD [26] and HPLC-DAD [3, 35] methods available in literature.The use of an internal standard (IS) would probably improve the figures of merit of the quantitative analysis, as frequently demonstrated in CE methods [36]. However, some limitations of using internal standards should also be taken into account (e.g., difficult selection of adequate IS, its comigration with analytes, etc). Noteworthy, the present CE-DAD method permitted the resolution of the three main aglycone soybean isoflavones (daidzein, glycitein and genistein) plus kampferol in a relatively short time (within 16 min) with respect to published HPLC methods based on gradient separation that can require up to 55 min although in general with better sensitivity [3, 30, 37].