Debittering of Red Bell Pepper (Capsicum annuum) Juice Retaining a High Vitamin B6 Content, Using a Styrene-divinylbenzene Adsorbent Resin

Abstract

Despite the high nutritional value of ripe red bell pepper (RBP) juice, RBP is typically considered a bitter vegetable. This excess bitterness affects the quality of RBP juice and results in rejection by consumers. We developed a manufacturing process to remove the bitter taste of RBP juice while retaining high nutrient levels. Compared to several resins, synthetic adsorbent resins had the most pronounced debittering effects, yielding the least bitter taste and green top note, while retaining a sweet taste and flavour. In particular, a styrene-divinylbenzene adsorbent resin having relatively large pores (pore radius ≥ 250 Å) removed more than 83 % of the bitterness of RBP juice, based on the quercetin-3-O-rhamnoside content. Moreover, vitamin B6 was retained at a high level, and the basic nutritional balance did not change after the treatment. The debittering effect was correlated with resin content, implying that quercetin-related flavonoids accounted for the bitter taste of RBP juice. Thus, resin adsorption is an efficient technique for the debittering and selective retention of high-value nutrition in RBP juice.

Introduction

Ripe red bell pepper (RBP; Capsicum annuum) is used widely as a vegetable and food additive, and is considered a good source of vitamins and carotenoid pigments such as capsanthin (Aizawa and Inakuma, 2009). Among the nutrients in RBP, vitamin B6 is water-soluble, and it functions as a coenzyme in many reactions involved in amino acid, carbohydrate, and lipid metabolism (Eliot and Kirsch, 2004). Pyridoxal-5′-phosphate, a vitamin B6 derivative, is the metabolically active form of vitamin B6 and is involved in approximately 100 enzymatic reactions (Toney, 2005).

In general, RBP is considered a relatively good source of vitamin B6. However, despite the high nutritional value of RBP juice, RBP is typically considered a bitter vegetable. This affects the quality of RBP juice and results in rejection by consumers. This excess bitterness is mainly assumed to be caused by the flavonoid quercetin-3-O-rhamnoside (Drewnowski and Gomez-Carneros 2000; Marín et al. 2004; Materska and Perucka 2005; Onda et al. 2013). In citrus juice production, flavonoids such as naringin and neohesperidin, which are related to a bitter flavour, are routinely adsorbed by resins. This debittering process, which has been successfully applied in the citrus industry worldwide (Puri et al., 1996; Drewnowski and Gomez-Carneros, 2000), is based on the selective adsorption of individual bitter taste-related flavonoids to suitable resins while retaining the high nutritional value of the product (Couture and Rouseff, 1992; Ribeiro et al., 2002; Lee and Kim, 2003; Singh et al., 2008; Bao et al., 2015). Intense research efforts have been made in recent decades to reduce excessive concentrations of bitter flavonoids through hydrophobic interaction and ion-exchange. Despite their wide application in the citrus industry, there are no studies examining the debittering of RBP juice by using these types of resins.

Accordingly, in this study, we performed a comparative analysis of the effects of several types of resins (synthetic adsorbent and ion-exchange resins) on RBP juice.

Materials and Methods

RBP juice    RBP, cultivar group of the species C. annuum, was grown under Mediterranean climate conditions in the Maule (VII) and Bío-Bío (VIII) regions of Chile. Fully red ripe peppers were harvested and squeezed in the same summer season of 2016 by Invertec Foods (Santiago, Chile). For sampling, the calyx and seed were removed.

Debittering adsorption analysis    Styrene-divinyl benzene (SDVB) copolymer resins were tested in this study, including Diaion HP20, Sepabeads SP70, Sepabeads SP850, Sepabeads SP207 (Mitsubishi Chemical, Tokyo, Japan), and Amberlite XAD1180N (Dow Chemical, Midland, MI, USA). Strong acid cation exchange resins were also used, including Diaion PK216, Diaion PK228 (Mitsubishi Chemical), Amberlite 200CTNa, Amberlite IR120B, and Amberlite FPX62 (Dow Chemical). Batch experiments were carried out in a glass beaker. The tested juice was prepared by diluting the RBP (Invertec Foods) juice concentrate with distilled water and adjusting to 25 % Brix. Fixed amounts of the pretreated resins were combined with the diluted RBP juice in 10 % and 20 % resins. The suspensions were stirred at 300 rpm in a water bath at 25 °C. After treatment for 60 min with 20 % resin, the juice was used for sensory evaluation. Further, after HP20 treatment for 60 min with 10 % and 20 % resins, the juice was used for sensory evaluation to determine the resin dose-dependent effect. At 0, 15, 30, 45, 60, 90, and 120 min, a 1-mL aliquot was withdrawn for high-performance liquid chromatography (HPLC) and total quercetin-related flavonoid analyses.

Sensory evaluation of debittered RBP juice    Seven trained panellists evaluated the taste, flavour, and odour characteristics of the RBP juice adjusted to 20 % Brix. Panellists preliminarily evaluated the juices treated by the 10 resins, and selected 3 resins, HP20, XAD1180N (synthetic adsorbent resins), and PK216 (cation exchange resin). Therefore, the following experiments were performed with these selected three resins. Before the evaluation sessions, the panellists collaboratively selected a set of seven attributes that aptly described the various flavour characteristics of RBP juice: “green top note”, “sweet top note”, “bitter first taste”, “bitter and acrid aftertaste”, “sweet taste”, “rich flavour”, and “flavour associated with sweet RBP”. The panellists assessed the quality of the samples using a 9-point category scale, with 0 representing weak and 9 representing strong. The scores were averaged and the differences between the control and test samples were compared using Dunnett's test. Further, in order to determine the resin dose-dependent effect, five trained panellists evaluated a set of the following three attributes: “green top note”, “bitter first taste”, and “bitter and acrid aftertaste”. The scores were averaged and differences between the test samples were compared using the Kruskal–Wallis and Steel–Dwass multiple comparison tests.

Quantitative changes in the bitter compounds and nutrients after debittering of RBP juice with different resins    The quercetin-3-O-rhamnoside content was measured as previously described (Tsushida and Suzuki, 1995; Watanabe et al., 2013). Briefly, 2 g of RBP juice (10 % Brix) was mixed with 10 mL methanol (Wako, Osaka, Japan), and the mixture was then ultrasonicated for 20 min. The obtained extracts were filtered using a PTFE filter (DISMIC-13HP pore size 0.20 µm; ADVANTEC, Tokyo, Japan), and the filtrates were analysed using an HPLC system (Hitachi Chromaster, Tokyo, Japan), which was comprised of the following components: a Chromaster 5110 quarternary pump, solvent degasser model 2003, Chromaster 5210 autosampler, Chromaster 5410 UV/VIS detector, and Chromaster 5310 column oven. The separation was performed with a C18 5-µm HPLC column (Mightysil, 250 × 4.6 mm i.d.; Kanto Kagaku, Tokyo, Japan) at 40 °C. The mobile phase consisted of water:acetonitrile:2-propanol (190:36:4, v/v/v) with 0.4 % citric acid. The detection wavelength was 360 nm at a flow rate of 1.0 mL/min. Quercetin-3-O-rhamnoside contents were determined using a calibration curve for the reference compound (quercitrin (quercetin-3-O-rhamnoside); Wako).

Effects of the resin HP20 on the total quercetin-related flavonoid content in RBP juice    Total quercetin-related flavonoid extraction was performed as previously described (Tsushida and Suzuki, 1995; Muro et al., 2015), with slight modifications. Five grams of RBP juice (10 % Brix) was mixed with 40 mL of 80 % aqueous methanol. The entire mixture was mixed thoroughly for 10 s and centrifuged at 1,600 ×g for 3 min. The supernatant was then filtered using no. 5A filter paper (Advantec, Tokyo, Japan). Aliquots of the samples were transferred into microplate cuvettes, and the absorbance was determined at 360 nm using a Corona Grating Microplate Reader SH-9000 Lab (Corona Electric, Ibaraki, Japan). Standard curves of isoquercitrin (quercetin-3-O-glucoside; Sigma, St. Louis, MO, USA) were generated for total quercetin-related flavonoid quantification. All determinations were performed in triplicate.

Nutritional composition of RBP juice following HP20 resin treatment    Compositional analysis of the RBP juice was performed at the Japan Food Research Laboratories (Tokyo, Japan) using the standard protocols recommended by the Resources Council of the Science and Technology Agency of Japan. The protein content was measured by the Kjeldahl method, using 6.25 as the conversion ratio of nitrogen to crude protein; moisture was measured as the decrease in weight after heating; fat content was measured by acid degradation; ash content was measured by the direct ashing method; dietary fibre was measured by the modified Prosky method (Prosky et al., 1988); and carbohydrate content was calculated using the following equation: 100 − (moisture + protein + lipid + ash + dietary fibre) according to the standard protocol. Vitamin B6 was bioassayed using Saccharomyces cerevisiae ATCC 9080 according to the AOAC official methods, and the capsanthin content was determined by HPLC (Aizawa and Inakuma, 2007).

Results and Discussion

Sensory evaluation of debittered RBP juice    Panellists preliminarily evaluated the juices treated by the 10 resins (data not shown). Among them, juices treated with HP20, XAD1180N (synthetic adsorbent resins), and PK216 (cation exchange resin) were debittered and had a preferable taste. Therefore, the following experiments were performed with these selected three resins. Indeed, when any of these three resins was used, the scores for “bitter first taste” and “bitter and acrid aftertaste” were significantly lower than those for the control samples (Table 1). Among the three resins, the panellists found that the HP20-treated juice also had a reduced “green top note”, whereas “sweet taste” and “flavour associated with sweet RBP” were favourably enhanced. This enhancement of preferable attributes was presumably due to the decrease of negative taste and note, indicating that a well-balanced taste had been achieved. Thus, we concluded that the debittering effects of these three resins were preferable and well balanced, and treatment with HP20 resin was the most promising for debittering of RBP juice. It was assumed that HP20 and XAD1180N adsorbed the bitter compounds through hydrophobic interactions, whereas the compounds were either ion-exchanged with PK216 or adsorbed onto the SDVB base copolymer of PK216.

Table 1. Sensory score of RBP juices debittered with different resins.

	Average scores
Attributes	Control	HP20	XAD1180N
green top note	7.71 ± 1.38	4.43 ± 2.37*	4.29 ± 1.38**
sweet top note	5.86 ± 1.68	5.29 ± 1.89	5.00 ± 2.38
bitter first taste	7.43 ± 1.40	2.71 ± 0.95**	2.29 ± 0.49**
bitter and acrid aftertaste	7.71 ± 1.60	2.43 ± 0.98**	3.29 ± 1.80**
sweet taste	5.14 ± 0.69	7.14 ± 1.21**	7.43 ± 1.27**
rich flavour	6.00 ± 1.29	5.86 ± 1.07	6.43 ± 1.40
flavour associated with sweet RBP	4.29 ± 1.25	6.86 ± 1.21**	6.14 ± 2.04

After treatment for 60 min with 20 % resin, the juice at 20 % Brix was used for sensory evaluation. Scores are expressed as an average of a scale of 0 to 9 ± standard deviations. Asterisks indicate a statistically significant difference between the control and each treated sample as analysed by Dunnett's test (*p < 0.05, **p < 0.01).

Quantitative changes in the bitter compounds and nutrients after debittering of RBP juice with different resins    HPLC analysis of the main bitter flavonoid quercetin-3-O-rhamnoside revealed significant differences in adsorption among the three types of resins. Following treatment with the synthetic adsorbent resins HP20 and XAD1180N, the juice showed a significant reduction (more than 83 %) in the bitterness-related compound (quercetin-O-rhamnoside), as shown in Table 2. In contrast, treatment with the cation exchange resin PK216 was not as effective as treatment with synthetic adsorbent resins HP20 and XAD1180N, showing only a 56 % reduction in the bitterness-related compound. HPLC analysis of the HP20-treated juice revealed significant reduction of several peaks besides that of quercetin-O-rhamnoside (Fig. 1), implying that the decrease in unknown quercetin-related flavonoids may affect the reduction of bitterness of the juice. Moreover, treatment with the synthetic adsorbent resins HP20 and XAD1180N resulted in the retention of high amounts of vitamin B6 (approximately 0.454 and 0.466 mg/100 g RBP juice, respectively; more than 80 %), whereas PK216 adsorbed more than half the amount of vitamin B6 (approximately 55 % reduction) (Table 2).

Table 2. Analysis of nutritional composition after debittering of RBP juice with different resins.

Nutritions (per 100g RBP juice)	Control	HP20	XAD1180N	PK216
Quercetin-3-O-rhamnoside (mg)	0.935	0.151	0.077	0.413
Vitamin B6 (mg)	0.531	0.454	0.466	0.257
Energy (kcal)	38	37	nt	nt
Protein (g)	1.2	1.1	nt	nt
Lipid (g)	0.3	0.2	nt	nt
Carbohydrate (g)	7.5	7.5	nt	nt
Dietary fibre (g)	0.2	0.1	nt	nt
Capsanthin (mg)	7.3	6.1	nt	nt

After treatment for 60 min with 20 % resin, the juice was used for compositional analysis. Data are expressed as the mean per 100 g RBP juice at 10 % Brix. nt indicates not tested.

Fig. 1.

HPLC chromatograms of quercetin-O-rhamnoside of non-debittered (a) and debittered RBP juice with HP20 resin (b).

After HP20 treatment for 60 min with 20 % resins, the juice was withdrawn for HPLC analysis. The arrow indicates the peak corresponding to quercetin-O-rhamnoside.

For the nonpolar SDVB-based synthetic adsorbent resins used in this study, the adsorption efficiency was mainly determined by the pore size and hydrophobicity of the target molecules. The average pore radii (Å) of HP20 and XAD1180N were relatively large (> 250 Å), indicating that this type of resin matched the molecular size of the flavonoids and adsorbed them effectively. These findings imply that the pore sizes were sufficiently large to allow the flavonoids to undergo adsorption. Further, the driving force of the adsorptive mass transfer in this process was considered to arise from the difference in polarity between the weakly polar bitter flavonoids and the polar aqueous juice matrix, resulting in the high affinity of the bitter flavonoids for the nonpolar resin. In this study, one bitter taste-related flavonoid, quercetin-3-O-rhamnoside, was also effectively adsorbed by HP20 and XAD1180N, consistent with previous reports of flavonoids in citrus juices (Couture and Rouseff, 1992; Ribeiro et al., 2002; Lee and Kim, 2003; Singh et al., 2008; Kola et al., 2010). In contrast, the cation exchange resin PK216 did not show good exchange of quercetin-3-O-rhamnoside, which was only partially adsorbed to the SDVB base copolymer, showing a 56 % reduction compared with that shown by the other ion exchange groups. However, in sensory evaluation, the juice treated with PK216 was considered debittered and well balanced, suggesting the presence of unidentified bittering substances following exchange with PK216. Additionally, following treatment with HP20 and XAD1180N, vitamin B6 was retained at a high level, whereas PK216 adsorbed more than half of the vitamin B6. This result could be explained by the positive charge of the amino group in vitamin B6, which would permit cation exchange.

Effects of the resin HP20 on the total quercetin-related flavonoid content in RBP juice    The content of total quercetin-related flavonoids, presumably related to a bitter taste, was measured following treatment with different amounts of HP20 resin. As expected, the quercetin-related flavonoid contents in the juice without adding resin were constant over a period of 120 min, as shown in Fig. 2. It was observed that with increasing resin amount, up to 20 %, the quercetin contents were reduced more extensively and rapidly. Equilibrium was reached after approximately 90 min with 20 % resin. Sensory evaluation of each treated sample after 60 min revealed that the debittering effect was correlated with resin content (Table 3). Indeed, the average scores of the three samples with different amounts of HP20 resin decreased considerably in a resin dose-dependent manner. Among these, scores of “bitter first taste” approached significance (10 % resin vs no resin, and 20 % resin vs 10 % resin, p<0.1 respectively), suggesting their correlation with resin content. These data implied that the total quercetin-related flavonoids could be responsible for the bitter taste in RBP juice and the synthetic adsorbent resin HP20 could effectively adsorb quercetin-related flavonoids other than quercetin-3-O-rhamnoside.

Fig. 2.

Total quercetin-related flavonoid content in adsorption experiments with the synthetic adsorbent resin HP20. Data are expressed as the means ± standard deviations (error bars).

After HP20 treatment with 10 % and 20 % resins at 0, 15, 30, 45, 60, 90, and 120 min, the juice was used for total quercetin-related flavonoid analyses. Measurements are mean values and based on 10 % Brix.

Table 3. Sensory score of RBP juices debittered with different amounts of HP20 resin.

	Average scores
Attributes	No resin	10% resin	20% resin
green top note	7.60 ± 0.55a	4.80 ± 1.48b	4.20 ± 2.17b
bitter first taste	7.20 ± 1.92a	3.80 ± 0.84ab	2.40 ± 0.89b
bitter and acrid aftertaste	7.60 ± 0.89a	4.40 ± 0.55b	2.80 ± 1.30b

After HP20 treatment for 60 min with 10 % and 20 % resins, the juice at 20 % Brixwas used for sensory evaluation. Scores are expressed as an average of a scale of 0 to 9 ± standard deviations. The data were analysed by Kruskal–Wallis and Steel–Dwass multiple comparison tests; scores that do not share a letter are significantly different (p < 0.05) between the test samples.

Nutritional composition of RBP juice following HP20 resin treatment    Comparison of the nutritional composition between debittered concentrates processed with 20 % HP20 resin and a control (non-debittered concentrate) showed that the nutritional balance did not change after debittering treatment, as shown in Table 2. Other nutritional components such as capsanthin were also retained after debittering, with a slight reduction of 16 % from 7.3 to 6.1 mg/100 g RBP juice. The effects of adsorbent resin treatment of citrus juices on the nutritional content have been reported by a number of researchers for different adsorbents used in debittering (Couture and Rouseff, 1992; Ribeiro et al., 2002; Lee and Kim, 2003; Kola et al., 2010). Notably, in our study, there were no significant differences in basic nutrition or capsanthin values in RBP juice.

Adsorption technology using nonpolar SDVB copolymers or slightly hydrophilic acrylic polymers has been shown to be applicable for industrial-scale processes because of the high adsorption capacities, ability to recover the adsorbed molecules, relatively low costs, and easy regeneration (Scordino et al., 2004). Adsorption processes have long been used for the debittering of citrus juices (Manlan et al., 1990), removal of browning reaction products, e.g., from apple juice (Carabasa et al., 1998; Gökmen and Serpen, 2002), or other unwanted compounds (Ferreira-Dias et al., 2002), and selective recovery of high-value compounds (Di Mauro et al., 1999, 2000). In this study, a debittering process utilising a styrene-divinylbenzene-based synthetic adsorbent resin reduced the content of bitter taste-related compounds and selectively retained the high nutritional value of RBP juice. This is the first study to apply resin treatment to RBP juice. Furthermore, this debittering effect was observed at the individual phenolic compound level for quercetin-3-O-rhamnoside. These findings may also have applications in the debittering of other vegetable juices while retaining selected nutritional compounds, which should be explored in future studies.

Conclusions

In conclusion, we found that treatment with an SDVB-based synthetic adsorbent resin having relatively large pores (pore radius ≥ 250 Å) selectively decreased the typical bitter taste of RBP juice and retained a high level of nutrients such as vitamin B at the laboratory scale. This resin could effectively adsorb quercetin-related flavonoids, including the main bitter taste-related compound quercetin-3-O-rhamnoside. This is the first report demonstrating that resin adsorption is an efficient technique for the debittering of RBP juice. Scale-up of the debittering process and development of practical applications such as for straight or mixed juices should be the main focus areas of future studies to enhance industrial implementation.

Acknowledgements    We thank members of the Material Resource Development Department, Innovation Division, KAGOME Co., Ltd. for their excellent advice and technical assistance.

References

•  Aizawa,  K. and  Inakuma,  T. (2007). Quantitation of carotenoids in commonly consumed vegetables in Japan. Food Sci. Technol. Res., 13, 247-252.
•  Aizawa,  K. and  Inakuma,  T. (2009). Dietary capsanthin, the main carotenoid in paprika (Capsicum annuum), alters plasma high-density lipoprotein-cholesterol levels and hepatic gene expression in rats. Br. J. Nutr., 102, 1760-1766.
•  Bao,  Y.,  Yuan,  F.,  Zhao,  X.,  Liu,  Q., and  Gao,  Y. (2015). Equilibrium and kinetic studies on the adsorption debittering process of ponkan (Citrus reticulata Blanco) juice using macroporous resins. Food Bioprod. Process, 94, 199-207.
•  Carabasa,  M.,  Ibarz,  A.,  Garza,  S., and  Barbosa-Cánovas,  G. (1998). Removal of dark compounds from clarified fruit juices by adsorption processes. J. Food Eng., 37, 25-41.
•  Couture,  R. and  Rouseff,  R. (1992). Debittering and deacidifying sour orange (Citrus aurantium) juice using neutral and anion exchange resins. J. Food Sci., 57, 380-384.
•  Di Mauro,  A.,  Fallico,  B.,  Passerini,  A.,  Rapisarda,  P., and  Maccarone,  E. (1999). Recovery of hesperidin from orange peel by concentration of extracts on styrene-divinylbenzene resin. J. Agric. Food Chem., 47, 4391-4397.
•  Di Mauro,  A.,  Fallico,  B.,  Passerini,  A., and  Maccarone,  E. (2000). Waste water from citrus processing as a source of hesperidin by concentration on styrene-divinylbenzene resin. J. Agric. Food Chem., 48, 2291-2295.
•  Drewnowski,  A. and  Gomez-Carneros,  C. (2000). Bitter taste, phytonutrients, and the consumer: a review. Am. J. Clin. Nutr., 72, 1424-1435.
•  Eliot,  A.C. and  Kirsch,  J.F. (2004). Pyridoxal phosphate enzymes: mechanistic, structural, and evolutionary considerations. Annu. Rev. Biochem., 73, 383-415.
•  Ferreira-Dias,  S.,  Monteiro,  J.P., and  Ribeiro,  M.H. (2002). The use of different adsorbents for selective removal of compounds from olive residue oil miscella. Eur. Food Res. Technol., 214, 400-404.
•  Gökmen,  V. and  Serpen,  A. (2002). Equilibrium and kinetic studies on the adsorption of dark colored compounds from apple juice using adsorbent resin. J. Food Eng., 53, 221-227.
•  Kola,  O.,  Kaya,  C.,  Duran,  H., and  Altan,  A. (2010). Removal of limonin bitterness by treatment of ion exchange and adsorbent resins. Food Sci. Biotechnol. 19, 411-416.
•  Lee,  H.S. and  Kim,  J.G. (2003). Effects of debittering on red grapefruit juice concentrate. Food Chem., 82, 177-180.
•  Manlan,  M.,  Matthews,  R.,  Rouseff,  R.,  Littell,  R.,  Marshall,  M.,  Moye,  H., and  Teixeira,  A. (1990). Evaluation of the properties of polystyrene divinylbenzene adsorbents for debittering grapefruit juice. J. Food Sci., 55, 440-445.
•  Marín,  A.,  Ferreres,  F.,  Tomás-Barberán,  F.A., and  Gil,  M.I. (2004). Characterization and quantitation of antioxidant constituents of sweet pepper (Capsicum annuum L.). J. Agric. Food Chem., 52, 3861-3869.
•  Materska,  M. and  Perucka,  I. (2005). Antioxidant activity of the main phenolic compounds isolated from hot pepper fruit (Capsicum annuum L.). J. Agric. Food Chem., 53, 1750-1756.
•  Muro,  T.,  Kami,  D., and  Sugiyama,  K. (2015). “Quergold”, a new cultivar of hybrid onion with high quercetin glycoside contents. Hortic. Res., 14, 305-311 (in Japanese).
•  Onda,  K.,  Nunokawa,  M.,  Musumi,  K.,  Kubota,  K., and  Morimitsu,  Y. (2013). Studies on flavor and taste of green pepper. Presented at 65th Annual Congress of The Japan Society of Home Economics, Tokyo, May 17–19, p. 91.
•  Prosky,  L.,  Asp,  N.,  Schweizer,  T.,  DeVries,  J., and  Furda,  I. (1988). Determination of insoluble, soluble, and total dietary fiber in foods and food products: interlaboratory study. J. Assoc. Off. Anal. Chem., 71, 1017-1023.
•  Puri,  M.,  Marwaha,  S.S.,  Kothari,  R.M., and  Kennedy,  J.F. (1996). Biochemical basis of bitterness in citrus fruit juices and biotech approaches for debittering. Crit. Rev. Biotechnol. 16, 145-155.
•  Ribeiro,  M.,  Silveira,  D., and  Ferreira-Dias,  S. (2002). Selective adsorption of limonin and naringin from orange juice to natural and synthetic adsorbents. Eur. Food Res. Technol., 215, 462-471.
•  Scordino,  M.,  Di Mauro,  A.,  Passerini,  A., and  Maccarone,  E. (2004). Adsorption of flavonoids on resins: cyanidin 3-glucoside. J. Agric. Food Chem., 52, 1965-1972.
•  Singh,  S.V.,  Gupta,  A.K., and  Jain,  R.K. (2008). Adsorption of naringin on nonionic (neutral) macroporus adsorbent resin from its aqueous solutions. J. Food Eng. 86, 259-271.
•  Toney,  M.D. (2005). Reaction specificity in pyridoxal phosphate enzymes. Arch. Biochem. Biophys., 433, 279-287.
•  Tsushida,  T. and  Suzuki,  M. (1995). Flavonoid in fruits and vegetables part I. Isolation of flavonoid-glycosides in onion and identification by chemical synthesis of the glycosides. Nippon Shokuhin Kagaku Kogaku Kaishi, 42, 100-108 (in Japanese).
•  Watanabe,  J.,  Muro,  T.,  Yanagida,  D.,  Yamagishi,  T., and  Takano-Ishikawa,  Y. (2013). Varietal differences in quercetin contents and antioxidant capacities of onions. Nippon Shokuhin Kagaku Kogaku Kaishi, 60, 563-566 (in Japanese).