Introduction

Flavonoids consist of large family of more than 4000 ubiquitous secondary plant metabolites and can further be divided into five subclasses: flavonols, flavones, flavanones, anthocyanins, and catechins (Wang and Huang 2004). Most of the edible vegetables and fruits contain flavonoids, but their type varies according to the dietary sources they are obtained from (Awad et al. 2000), e.g., flavanones and flavones are usually found in the same plant, yet flavones and flavonols are generally not present together, nor are flavanones and anthocyanins (Merken and Beecher 2000). The most common flavonol in the diet is quercetin (3,3′,4′,5,7-pentahydroxy flavone) (Ignat et al. 2011). It is present in plants in glycosidic form with quercetin-3-rutinoside, and also called rutin (Erlund 2004). Over the past years, both quercetin and rutin have gained tremendous interest, up to the present, rutin has been used clinically as the therapeutical medicine, and over 130 preparations containing quercetin or rutin are registered as drugs worldwide. Flavanones occur almost exclusively in citrus fruits. The highest concentrations are found in the solid tissues, but concentrations of several hundred milligrams per liter are present in the juice as well (Awad et al. 2000). Naringin (4′,5,7-trihydroxyflavanone-7-rhamnoglucoside) is a flavanone present in many citrus fruits while naringenin is the aglycone of naringin (Fang et al. 2006).

Beneficial effects of regular consumption of fruits and vegetables are associated with high content of flavonoids present in them (Lotito and Frei 2004). Their dietary significance exhibits a wide range of biological effects, including anti-oxidative, anti-microbial, anti-bacterial, anti-ulcer, anti-mutagenic, anti-carcinogenic, anti-inflammatory, anti-aging, anti-allergic, anti-thrombic, anti-viral, anti-neopasic, anti-tumor, vasodilatory, and cardioprotective actions (Biesaga 2011; Gil and Couto 2013 ; Awad et al. 2000; Volikakis and Efstathiou 2005; Gattuso et al. 2007).

Ziziphus mauritiana L. belongs to the Rhamnaceae family, fruit of which is commonly known as “ber” in India and Pakistan (Memon et al. 2012). The three cultivated varieties of Z. mauritiana L. in Pakistan (particularly in Sindh Province) are locally known as Gola ber (var. Gola Lemai, Gola Achri, Gola Lootri, Gola Dehli), Soofi ber, and Kheerol (Chambeli) ber (Memon et al. 2013). Ber fruit is consumed fresh or dried and is used to prepare jams, candied fruit, beverages, and other food products (Memon et al. 2012). Ber fruit is known to contain a wide array of phytochemicals and minerals such as amino acids, carbohydrates, ascorbic acid, flavonoids, phenolic acids, vitamins A and C, phosphorus, and calcium (Memon et al. 2013).

Several analytical methods have been described for the separation and determination of flavonoids using various techniques, such as TLC (Males and Medic 2001; Vogel et al. 2005), HPLC (Wang and Huang 2004; Ishii et al. 2003; Pascual et al. 1998), HPLC-MS (Lin et al. 2000; Wu et al. 2004; Sanchez et al. 2003), GC-MS (Fiameogos et al. 2004; Deng and Zito 2003) and Capillary Electrophoresis (CE) (Akiyama et al. 2000; Bo et al. 2002). Compared with conventional chromatographic methods, capillary electrophoresis (CE) has many valuable advantages, including excellent separation efficiency, high resolution, short analysis time, easy automation, and low solvent and sample consumption.

Flavonoids have been analyzed by different modes of CE. A micellar electrokinetic capillary chromatographic (MEKC) method was developed and validated for the simultaneous determination of flavonoids (Ganzera et al. 2008). A selective and sensitive microemulsion EKC (MEEKC) method with electrochemical detection has been developed for separation and determination of flavonoids (Yu et al. 2008). Among different modes in CE separation, capillary zone electrophoresis (CZE), using an uncoated capillary column requiring less maintenance is the most frequent and simplest separation mode, which makes the optimization of conditions easy (Liu et al. 2006). A number of papers are reported every year using CZE for the separation and determination of flavonoids from various matrices. For example, Liu et al. (Liu et al. 2006) developed a CZE method coupled with diode-array detection for simultaneous determination of four flavonoids including icariin, epimedin A, epimedin B, and epimedin C in Epimedium, in which 50 mM borate buffer containing 22 % acetonitrile was used at pH 10 at 15 kV. Wu et al. (Wu et al. 2007) determined five flavonoids (hesperidin, naringin, hesperedin, naringenin, and rutin) and ascorbic acid in grapefruit peel and juice by CE-ED, using 60 mM borate buffer of pH 9 and analytes were well separated within 25 min. Zhang et al. (Zhang et al. 2008) simultaneously determined six flavonoids including apigenin, luteolin, kaempferol, quercetin, (+)-catechin, (−)-epicatechin in chrysanthemum by CZE with amperometric detection (CZE-AD), where β-cyclodextrin and the mixture of methanol and ethanol were used as running buffer modifiers and the baseline separation of six analytes was achieved in 20 min by using 36 mM borate-phosphate buffer of 8.8 pH at 18 kV. Ren et al. (Ren et al. 2009) developed a CZE method for the simultaneous determination of nine flavonoids in Anaphalis margaritacea, in which 25 mM Na2B4O7 and 10 mM NaH2PO4 buffer solution was used at pH 9.6 at 20 kV. Zhang et al. (Zhang et al. 2010) simultaneously determined flavonoids and anthraquinones (emodin, kaempferol, apigenin, luteolin, and rhein) in chrysanthemum by CE-AD, where all five analytes were well separated within 17 min, using borate-phosphate buffer of pH 9 at 19 kV.

The purpose of this work is to develop a simple, fast, effective, and reliable CZE method for simultaneous determination of four flavonoids including two flavonols (quercetin and rutin) and two flavanones (naringin and naringenin) in different fruits such as apples (Malus domestica), oranges (Citrus sinenis), grapefruits (Citrus paradisi), and ber (Z. mauritiana L.), using photodiode array (PDA) detector. To the best of our knowledge, no one method is reported to describe the simultaneous determination of these four flavonoids in different fruit samples.

Experimental

Standard and Reagents

Naringin, naringenin, ethyl alcohol, and hydrochloric acid were purchased from Sigma-Aldrich (China and UK). Quercetin dihydrate and rutin were purchased from Fluka (Switzerland). Boric acid and sodium hydroxide were purchased from E. Merck (Germany). Sodium borate decahydrate was obtained from Daejung Chemicals (Korea). Methanol was purchased from Fischer Chemicals (UK). All reagents were HPLC grade and used without further purification. Double-deionized water was used throughout the experiments.

Stock solutions of flavonoids: naringin, naringenin, rutin, and quercetin were prepared separately by dissolving 0.1 g of each flavonoid in 100 mL of methanol. The background electrolyte (BGE) for separation was 10 mM borate buffer (pH 8.5) which was prepared by dissolving the appropriate amounts of sodium borate decahydrate and boric acid in double-deionized water. The pH of the background buffer was adjusted with either 0.1 M NaOH or 0.1 M HCl. All solutions were filtered through 0.45-μm cellulose acetate filter paper prior to use for CE experiments.

Samples Collection

Fresh fruits of apple, orange, grapefruit, and four species of ber (Soofi ber, Kheerol ber, Gola Lemai ber, and Lootri ber) were purchased from local market of Hyderabad and commercial juices of apple and orange were purchased from superstores of Hyderabad.

Fresh fruits of apples, oranges, and grapefruit were treated in a commercial home juicer and juice was collected, whereas the ber fruit was washed with tap water and pulp was separated from its seed. The edible pulp part of the fruit was cut into small pieces and freeze dried. After drying, the samples were ground into a fine powder and stored in a freezer until analyzed.

Instrumentation and Electrophoretic Conditions

All the experiments were conducted with a Beckman P/ACE MDQ CE instrument (Beckman Coulter Inc., USA) equipped with a photodiode array detector with 32 Karat Software. Uncoated fused-silica capillaries (Polymicro Technologies, Phoenix) with dimensions of 57 cm total length, 50 cm effective length, 75 μm inner diameter, and 375 μm outer diameter were used. Capillary was conditioned and regenerated as the previously reported method (Memon et al. 2015).

Samples were injected by means of the autosampler using hydrodynamic method (0.5 psi). An injection time of 10 s was used for all analyses. The separation voltage was 25 kV while the temperature of capillary was maintained at 25 °C. The pH of solutions was measured using an inoLab pH meter.

Sample Preparation

Samples were prepared by a reported method (Wu et al. 2007) with some modifications. Ber flavonoids were extracted by keeping mixture of 1 g of freeze-dried ber samples and 10 mL of ethyl alcohol in an ultrasonic bath for 2 h. The extract was then filtered through a filter paper (Whatman # 42) and the extraction procedure was repeated thrice. After filtration through 0.45-μm cellulose acetate filter paper, 80 μL of sample was diluted with 10 mM borate buffer to 1 mL and the clear solution was injected in CE to record electropherogram.

Juice samples of apple, orange, and grapefruit were separately centrifuged at 5000 rpm for 10 min; then the juice was filtered through 0.45-μm cellulose acetate filter paper; then 80 μL samples was diluted with 10 mM borate buffer to 1 mL; and the clear solution was run for recording electropherogram.

Results and Discussion

Optimization of the Method

The mixture of four flavonoids was analyzed by an optimized CZE method. All four flavonoids are soluble in methanol and show appreciable absorbance at 214 nm. Separation was optimized with respect to the pH and concentration of the background buffer. The separation voltage and injection time were also optimized to resolve the standard mixture of all four flavonoids in shortest time.

Buffer Selection

Sodium hydrogen phosphate (pH 7.5–9.0), sodium acetate (pH 6–7) and sodium tetraborate (pH 8–10) buffers were tried to obtain the optimum CE conditions. Sodium tetraborate was found to have better efficiency in terms of peak shape, peak height, and shorter migration time. Therefore, it was selected and further optimized.

Effect of Buffer Concentration

The concentration of buffer is an important parameter in CE. The effect of buffer concentration in the range of 10–50 mM in intervals of 10 mM was studied. Figure 1 shows the effect of buffer concentration on migration time, peak area, and resolution. Migration time of the four analytes was increased by increasing buffer concentration due to the strong interaction of borate and flavonoids at higher buffer concentration. Peak area was also increased by increasing the concentration of the buffer since the interaction of borate ions and flavonoids considerably enhances UV absorbance of the analytes (Luo et al. 2007). The increase of migration time is because of decreasing electrosmotic flow (EOF) since the buffer concentration acts directly on the magnitude of the EOF, the higher the concentration lowers the EOF and vice versa, while low EOF produces a long migration time, attributed to the better resolution (Ren et al. 2009). Considering migration time, resolution, and peak area, the best separation was achieved at 10 mM borate buffer. Therefore, 10 mM was selected for further studies.

Fig. 1
figure 1

Effect of buffer concentration on the migration behavior of analytes. CE conditions: 25 °C, 25 kV, and detection at 214 nm. Peak identification: 1 naringin, 2 rutin, 3 naringenin, 4 quercetin

Full size image

Effect of Buffer pH

In CE the migration time and separation of analytes are greatly affected by pH of buffer solution as it affects the zeta potential, the EOF, as well as the overall charge of the analytes (Xu et al. 2005a, 2005b). Therefore, the effect of buffer pH was studied from 8–10 in increments of 0.5, while keeping the concentration of buffer at 10 mM. Figure 2 indicates the increase in migration time of analytes with increasing pH; the phenomenon may be attributed to the dissociation of flavonoid hydroxyl groups with rising pH value (Xu et al. 2005a, 2005b). Moreover, at pH 8, the signal peaks of four analytes were very weak. At pH 9, quercetin was not detected and at pH 9.5 and 10, there was a large increase in migration time; therefore, borate buffer of pH 8.5 was selected because of reasonable migration time, better separation, greater absorbance, and fairly good peak shapes.

Fig. 2
figure 2

Effect of buffer pH on the migration behavior at 10 mM concentration of borate. Other conditions and peak identification are shown in Fig. 1

Full size image

Effect of Applied Voltage and Injection Time

The effect of applied voltage was examined in the range of 10–30 kV with an interval of 5 kV, under the selected conditions. As the applied voltage has a negative relationship with migration time of the analytes, with increasing applied voltage a decrease in migration time of analytes was recorded but baseline noise increased and results in poor detection limits. As the voltage is increased, there is more joule heat generated in the capillary, which can be unfavorable to the separation because it led to broader peaks, which possibly results sample decomposition, or the formation of the bubbles inside the capillary. The better the heat is dissipated by the equipment, the higher the voltage that can be used (Pancorbo et al. 2004). In the present study, 25 kV was selected as the optimum applied voltage in terms of shorter analysis time, current intensity, and high separation efficiency.

Samples were injected into the capillary by hydrodynamic method and the effect of injection time was investigated by changing it from 4–12 s, at intervals of 2 s, at a constant pressure of 0.5 psi, and voltage of 25 kV. It was observed that by increasing injection time, peak current was increased, and peaks were broadened as injection time affects both the peak current and the peak shape (Xu et al. 2005a). When the injection time exceeded 10 s, the peak current increased and peak broadening became more severe. Therefore, considering resolution and sensitivity, 10 s was selected as the optimum injection time.

Figure 3 shows the typical electropherogram for standard mixture solution at the final optimized conditions as running buffer 10 mM borate buffer (pH 8.5), applied voltage 25 kV, hydrodynamic injection (10 s at 0.5 psi), and capillary temperature 25 °C. Baseline separation for all four flavonoids has been achieved within 10 min.

Fig. 3
figure 3

Electropherogram of standard mixture solution containing naringin, rutin, naringenin, and quercetin at optimized conditions: 10 mM borate buffer, pH 8.5, 214 nm, applied voltage 25 kV. Peak identification is given in Fig. 1

Full size image

Validation of the Method

Linearity

The calibration curves were plotted (concentration vs. peak area) in the concentration range from 3.125 to 200 μg/mL for naringin, naringenin, and rutin, whereas 6.25 to 200 μg/mL for quercetin. The response was linear with coefficient of determination (R 2) = 0.987 for naringin, 0.987 for naringenin, 0.990 for rutin, and 0.982 for quercetin. Regression equations, linear ranges, and values of coefficient of determination are shown in Table 1.

Table 1 Analytical data for the determination of analytes
Full size table

LOD and LOQ and Selectivity

The limit of detection (LOD) and limit of quantification (LOQ) measured as signal to noise ratio of 3:1 and 10:1 were obtained within 0.3–0.6 and 1.0–1.9 μg/mL, respectively. Table 1 shows the LOD and LOQ data of each analyte. Standard errors of intercept and slope were also calculated by Microsoft Excel data toolpak for all four analytes and values are listed in Table 1. Major possible interfering constituent of natural samples can be sugars, vitamin C, and some organic acids such as citric acid and oxalic acid. Method’s selectivity in presence of all these were checked and found to have no significant interference with quantitative or qualitative determination of all four analytes.

Repeatability

The repeatability of the method was studied by performing inter- and intra-day assay by the analysis of the replicate injections of standard solution mixture. The intra-assay precision (n = 5) was performed under the same conditions, by the same analyst on the same day with the interval of 2 h, whereas the inter-assay precision (n = 3) was performed under the same conditions, by the same analyst on three consecutive days. The reproducibility of the separation in terms of migration time and peak height for all four analytes was examined and %RSD values of each are shown in Table 1.

Application of Developed CZE Method to Different Fruits

The optimized CZE method was applied for the determination of naringin, rutin, naringenin, and quercetin in different fruit samples. The detected flavonoids in samples were identified by comparing their migration time and UV spectra with those of the standards using PDA detector. Quantitation was done by external calibration plot and the results are summarized in Table 2.

Table 2 Assay results of the analytes in real samples
Full size table

All four flavonoids of interest were present in fresh grapefruit juice, which were successfully identified and quantified, whereas in Kheerol ber, only three flavonoids were detected, i.e., rutin, naringenin, and quercetin. High amounts of naringin and naringenin were detected in grapefruit juice, as naringin is the main flavonoid of grapefruit. Fresh apple juice contained high amounts of quercetin. As shown in Table 2, naringin was only determined in Lootri ber, rutin was detected in Kheerol ber, naringenin was detected in Kheerol ber, Soofi ber (small), and Gola Lemai ber species, whereas quercetin was detected in Soofi ber (large) and Kheerol ber species. In orange juice, only two flavonoids, i.e., naringenin and quercetin, were detected. These findings were in general agreement with the literature data in terms of flavonoid distribution. The developed sensitive and rapid method resolved the flavonoids within 10 min.

Quantitative data of all real samples was cross checked by already reported HPLC method (Memon et al. 2013). The results obtained with HPLC method were in good agreement with developed CE method with the standard deviation of ±1.2–2.0.

Comparison of Developed Method with Reported Methods

The developed method has been compared with already reported methods of these analytes on CE. As shown in Table 3, only one report is available by PDA and our method is superior to that method both in terms of analysis time and LOD. Method developed in this study has shorter analysis time, simple background system, and comparable LOD values with most of the reported methods.

Table 3 Comparison of developed method with reported methods
Full size table

Conclusion

This paper describes the successful development of a rapid and easy CZE method for the simultaneous determination of four flavonoids, i.e., naringin, naringenin, rutin, and quercetin in different fruits like apple, orange, grapefruit, and ber using PDA detector. Under optimized conditions, all four analytes could be well determined within 10 min in a 10 mM borate buffer (pH 8.5) at an applied voltage of 25 kV. The developed CZE method demonstrates high separation efficiency and sensitivity, short analysis time, satisfactory repeatability with low operating cost, and minimal sample preparation and volume requirement.