Conditions of enzyme-assisted extraction to increase the recovery of flavanone aglycones from pectin waste

Abstract

The citrus pectin by-product (CPB), generated from pectin industry, is a rich-source of flavanones, but not explored until now. As most of these compounds are inside vacuoles or bound to cell wall matrix, enzymatic hydrolysis was applied on their recovery, followed by hydroalcoholic and ultrasound extraction. Different parameters were studied: enzymes (β-glucosidase, tannase, and cellulase), their concentration (5, 10, and 20 U g⁻¹ CPB), and reaction time (6, 12, and 24 h). Extracts were characterized in total phenolic content (TPC), antioxidant capacity (ORAC and DPPH assays), and polyphenolic profile (HPLC–DAD). All enzymatic treatments significantly improved CPB antioxidant capacity and TPC, compared with hydroalcoholic and ultrasound extraction. β-glucosidase (5 U) for 24 h was the most effective in polyphenol extraction and bioconversion, followed by β-glucosidase (5 U) for 12 h and tannase (5 U) for 24 h. Thus, the concentration of these enzymes was increased (10 and 20 U) to improve flavanones extraction. β-glucosidase at 20 U offered the highest amount of naringenin (77.63 mg 100 g⁻¹ of CPB) and hesperetin (766.44 mg 100 g⁻¹) obtained so far by biological processes. According to Person’s correlation analysis, TPC and antioxidant activity were highly correlated with CPB contents of hesperetin and naringenin. The aglycone flavanones are rarely found in natural sources and have higher biological potential than their glycosylated forms. Our results indicated enzyme-assisted extraction as a good choice for recovering aglycone flavanones from CPB, and increased knowledge on the biological activity of this agroindustrial waste, amplifying their application in food and pharmaceutical field.

Electronic supplementary material

The online version of this article (10.1007/s13197-020-04906-4) contains supplementary material, which is available to authorized users.

Introduction

Agroindustrial wastes have gained increasing attention because of their rich content of valuable compounds, including pigments, organic acids, flavor compounds, enzymes, and bioactive compounds (Delgado and Fleuri 2016). Orange is the world’s largest fruit crop, with 73 million tonnes harvested in 2016 (FAO 2018). Brazil produced 17 million tonnes of oranges in 2016, 70% of which was destined to juice extraction, yielding 6 million tonnes of by-products (FAO 2018). Currently, citrus juice by-products (CJB) are used for animal feed, biofuel production, single-cell protein production, and extraction of essential oils and biomolecules, such as enzymes, polyphenols, and pectin (Mamma and Christakopoulos 2014; Nadar et al. 2018). Pectin can be extracted from CJB in yields of approximately 20%, generating a significant volume of waste that has no industrial use until now (Mamma and Christakopoulos 2014). Barbosa et al. (2018) were the first investigators to chemically characterize the residue from citrus pectin by-product (CPB). The authors showed that CPB had a higher phenolic content than CJB; thus, higher yields of phenolic compounds can be obtained from CPB than from CJB.

The most important polyphenols in citrus fruits are the glycosylated flavanones hesperidin, naringin, and narirutin. They are found predominantly in citrus species but are also present in other plants in minor amounts, for instance, in tomato peels (Khan and Zill-E-Huma 2014). Citrus flavanones have been widely studied, expressing numerous activities, such as antioxidant, antiproliferative, anti-inflammatory, antimicrobial, inhibition of human platelet aggregation, and hypocholesterolemic effects (Khan and Zill-E-Huma 2014; Barreca et al. 2017). The O-glycosylation of flavanones reduces their biological potential (Ferreira et al. 2013; Madeira et al. 2014; Barreca et al. 2017). However, the aglycones hesperetin and naringenin occur less frequently than their glycoside. Interestingly, these aglycones can be obtained by hydrolyzing the sugar moieties of hesperidin, naringin, and narirutin.

Several methods have been proposed to maximize polyphenol yields from different natural sources. The most commonly used method is solvent-based extraction. Yields are negatively affected by cell wall polysaccharides, as some phenolic compounds form hydrophobic interactions and hydrogen bonds with these structures (Nadar et al. 2018). Enzyme-assisted extraction is an efficient, eco-friendly, alternative process that can disrupt the cell wall matrix under mild conditions, thereby preserving polyphenols functionality (Madeira et al. 2013). The efficiency of the method depends on enzyme type and reaction parameters, such as temperature, time, pH, and enzyme/substrate ratio (M’hiri et al. 2014). Studies have focused on the use of carbohydrases, such as pectinases, cellulases and β-glucosidases. In addition to releasing polyphenols, enzymes can hydrolyze these compounds into low molecular weight phenolic compounds, which increases availability and bioactivity (Chamorro et al. 2012; Ferreira et al. 2013). Tannase acts on polyphenol structures by hydrolyzing depsidic linkages and ester bonds (Battestin and Macedo 2007). Therefore, it can hydrolyze the glycosidic linkages of hesperidin, naringin, and narirutin (Ferreira et al. 2013).

Thus, this study aimed to apply enzymatic treatments (β-glucosidase, tannase, and cellulase) to CPB, in order to enhance polyphenols release from the cell wall matrix and increase the yields of the bioactive phenolics, as hesperetin and naringenin.

Materials and methods

Chemicals

Phenolic acids (gallic and ellagic acid), flavones (tangeretin and diosmetin), flavanones (hesperidin, hesperetin, narirutin, naringin and naringenin), tannic acid, 2,2′-azobis(2-methylpropionamidine) hydrochloride (AAPH), 2,2-diphenyl-1-picrylhydrazyl (DPPH), fluorescein, Trolox, β-glucosidase (naringinase from Penicillium decumbens), p-nitrophenyl-β-D-glucopyranoside, and p-nitrophenol were purchased from Sigma–Aldrich (St. Louis, MO, USA). Folin–Ciocalteu reagent, monobasic and dibasic sodium phosphate, sodium carbonate, ammonium sulfate, carboxymethylcellulose, glucose, and formic acid were purchased from Dinâmica Química Contemporânea (Diadema, Brazil). HPLC grade methanol was purchased from JT Baker (Center Valley, PA, USA). Cellulase (Celluclast 1.5 L) was obtained from Novozymes (Copenhagen, Denmark). All other chemicals were analytical grade.

CPB

CPB was provided by CP Kelco (Limeira, SP, Brazil). The residue was dried at 70 °C, crushed in a blender (OXY, Brazil), and sieved to obtain particle sizes lower than 1.86 mm (10 mesh sieve, Bertel Metallurgical Industries Ltd., Brazil). CPB was stored at −20 °C until analysis.

Enzymes

Three enzymes were used to treat CPB: β-glucosidase (EC 3.2.1.21), cellulase (EC 3.2.1.4), and a semi-purified extract of tannase (EC 3.1.1.20) from Paecilomyces variotii (CBMAI 1157). Tannase of P. variotii was obtained by solid-state fermentation according to Battestin and Macedo (2007), with minor modifications. Briefly, 10 g of wheat bran, 10 mL of distilled water, and 10% (w/w) tannic acid were used as fermentation medium. A suspension of 5 × 10⁷ spores mL⁻¹ of P. variotii was inoculated and incubated at 30 °C for 120 h. After fermentation, tannase was extracted with acetate buffer (20 mM at pH 5.0), and the solution was filtered and centrifuged. For partial purification of tannase, the supernatant was precipitated with ammonium sulfate (80% saturation) and dialyzed. The dialysate was freeze-dried and used as semi-purified tannase.

Enzymes activities were evaluated spectrophotometrically before their application. Tannase activity was measured using tannic acid as substrate and quantified using a standard curve of gallic acid, according to Sharma et al. (2000). β-Glucosidase activity was evaluated using p-nitrophenyl-β-D-glucopyranoside as substrate and quantified using a standard curve of p-nitrophenol, according to Matsuura et al. (1995). Cellulase activity was determined using carboxymethylcellulose as substrate followed by quantification of reducing sugars, according to the dinitrosalicylic acid (DNS) method, using a standard curve of glucose (Miller et al. 1960). Enzyme specific activities were expressed as U mg⁻¹ protein, where U is the amount of enzyme capable of generating 1 μmol of product/min of reaction.

Enzymatic treatments and polyphenol extraction

CPB was subjected to five enzymatic treatments, including: T (5 U tannase g⁻¹ CPB), β (5 U β-glucosidase g⁻¹ CPB), C (5 U cellulase g⁻¹ CPB), TC (5 U tannase g⁻¹ CPB and 5 U cellulase g⁻¹ CPB), and βC (5 U β-glucosidase g⁻¹ CPB and 5 U cellulase g⁻¹ CPB). The reaction medium consisted of 2 g of CPB in 25 mL of acetate buffer (20 mM, pH 5.0) and the respective enzyme treatment. Flasks were incubated at 40 °C for 6, 12, or 24 h at 120 rpm. A control treatment was carried out under the same conditions but without the addition of enzymes. At the end of the enzymatic reaction, hydrolysis was stopped by placing the flasks in an ice water bath (30 min). Then, polyphenols were extracted using a 1:1 (v/v) hydroalcoholic solution, as previously described by Nakajima et al. (2016). The extracted polyphenol in ethanol was placed in an ultrasonic bath for 30 min (40 kHz, 135 W, 25 °C), shaken at 200 rpm for 15 min (25 °C), and centrifuged at 5000 rpm for 15 min (25 °C). Finally, the ethanol fraction was removed by rotary evaporation at 40 °C (15 min), and the aqueous extracts were frozen at −20 °C until analysis. After the most efficient enzymatic treatment was identified, higher enzyme concentrations (10 and 20 U g⁻¹ of residue) were studied on polyphenols extraction. The yields (%) were calculated based on analyte content (mg 100 g⁻¹ DM) obtained after (A_i) and before (A₀) increasing enzymes concentration, according to Eq. (1):

$$Yield\left(\%\right)=\left(\frac{A_i}{A_0}\right)\times 100$$ 1

Analyses of polyphenols

The total phenolic content (TPC) was evaluated using the Folin–Ciocalteu method, established by Singleton and Rossi (1965), modified by Nakajima et al. (2016). Extracts (50 µL) were mixed with distilled water (800 µL) and Folin–Ciocalteu reagent (50 µL). After 3 min, Na₂CO₃ was added, mixed and incubated in the dark at room temperature for 2 h. The absorbance was measured at 725 nm wavelength, carried out on a NovoStar Microplate reader (BMG LABTECH, Germany). TPC were quantified using a standard curve of gallic acid, ranging from 15 to 300 µg mL⁻¹. Results were expressed as mg of gallic acid equivalents (GAE) per 100 g of CPB dry matter (DM).

The major polyphenol compounds of the extracts were detected and quantified using a high-performance liquid chromatography system (HPLC) (Dionex UltiMate 3000, Dreieich, Germany) equipped with a diode array detector (DAD-3000) and a C-18 Acclaim 120 column (Dionex, 3 µm, 4.6 × 150 mm) at 30 °C. Analyte separation was achieved using the mobile phases A (99.9:0.1 v/v water/formic acid) and B (99.9:0.1 v/v methanol/formic acid) in a linear gradient at a flow rate of 0.6 mL min⁻¹: 90% A (0–5 min), 20% A (5–80 min), 90% A (80–85 min), and 90% A (Caridi et al. 2007). Individual flavonoids were identified by comparison of their retention times and UV–Vis spectra. Detection was carried out at 280 nm, and quantification was performed using standard curves of naringin, narirutin, naringenin, hesperidin, hesperetin, tangeretin, gallic acid, ellagic acid, and diosmetin, ranging from 0.15 to 100 µg mL⁻¹ (Table S1). Results were expressed as mg 100 g⁻¹ DM.

Antioxidant assays

The effect of enzymatic treatment on CPB antioxidant capacity was evaluated by the DPPH and ORAC (oxygen radical absorption capacity) assays using a microplate reader (Novostar, BMG LABTECH, Germany).

DPPH radical scavenging activity was assessed as described by Bondet et al. (1997). Samples and Trolox solutions were added to DPPH reagent, and methanol was used as the reaction medium. The absorbance of decolorized DPPH was recorded during 90 min at 520 nm.

The ORAC method was performed according to Prior et al. (2003) using fluorescein as fluorescent probe and APPH as free radical generator. All chemicals and samples were prepared in 75 mM sodium phosphate buffer (pH 7.4). Fluorescence was monitored every 56 s for 75 min at 37 °C. The excitation and emission wavelength were 485 nm and 520 nm, respectively.

DPPH and ORAC results were determined against a standard curve of Trolox and were expressed as µmol Trolox equivalents (TE) g⁻¹ DM.

Statistical analysis

All measurements were performed in triplicate, and the results are presented as mean ± standard deviation (SD). The effect of the independent variables, enzyme and reaction time, on individual phenolic compounds, TPC and antioxidant capacity (ORAC and DPPH) was assessed using one-way ANOVA. When statistical difference was declared significant (p value ≤ 0.05), means were compared by Tukey HSD post hoc test. Pearson’s correlation test was used for correlation analysis between individual phenolic compounds and TPC, DPPH and ORAC. Statistical analysis was carried out using SPSS version 25 (Chicago, IL, USA), following the manufacturer’s instructions to one-way ANOVA, Pearson’s correlation test, and PCA analysis.

Results and discussion

Analyses of polyphenols

Compared with the control (hydroalcoholic with ultrasound extraction), all enzymatic treatments significantly (p < 0.05) increased the TPC of the extracts from CPB after 24 h of reaction (Fig. 1). β-glucosidase treatment for 12 and 24 h afforded the best results, they increased TPC by 2.8 fold compared with the control. However, the combined use of β-glucosidase and cellulase did not improve TPC compared with β-glucosidase treatment. After 24 h of reaction, tannase, cellulase, and “β-glucosidase + cellulase” were able to increase by approximately 1.9 times the TPC of CPB (p > 0.05). “Tannase + cellulase” was the least effective treatment, increasing TPC content by 1.66 fold.

Fig. 1.

Effect of enzymatic treatment on CPB total phenolic content. TPC: total phenolic content; DM: dry matter; GAE: gallic acid equivalents; Control: hydroalcoholic and ultrasound extraction; T: tannase; β: β-glucosidase; C: cellulase; TC: tannase + cellulase; βC: β-glucosidase + cellulase

The bioactive properties and TPC of CPB were first determined by Barbosa et al. (2018). The authors reported that CPB had a lower TPC than CJB (170 and 386 mg GAE 100 g⁻¹ DM, respectively). However, after enzymatic treatments, the extract of CPB had the same amount of TPC as CJB. Compared with other green techniques, such as ultrasound-assisted, microwave-assisted, and high-pressure extraction, enzyme-assisted extraction promotes a higher recovery of phenolic compounds (Khan et al. 2010; M’hiri et al. 2014). Wang et al. (2011) reported that the TPC of citrus peel subjected to microwave-assisted, ultrasound-assisted, and conventional solvent extraction was between 109.4 and 114.0 mg GAE 100 g⁻¹. In Lachos-Perez et al. (2018) study, ultrasound-assisted extraction increased orange peel TPC by 1.8 fold compared with conventional extraction. Thus, enzyme-assisted extraction, followed by hydroalcoholic and ultrasound extraction, is a promising method for obtaining high-TPC extracts.

Nine phenolic compounds were investigated (Table 1), including flavanones (n = 5), flavones (n = 2), and phenolic acids (n = 2). The contents of naringin, ellagic acid, and diosmetin were below than their limit of quantification or not detected. Treatments and reaction times had significant effects on the profile of polyphenol extracts (p < 0.05). In accordance with the literature, the glycosides hesperidin and narirutin, were the most abundant flavanones in the control (Khan and Zill-E-Huma 2014). On the other hand, a great content of aglycone flavanones was detected in enzymatically treated extracts. β-glucosidase treatment for 24 h was the most effective in producing hesperetin and naringenin. The content of gallic acid also increased after tannase, cellulase, “tannase + cellulase”, and “β-glucosidase + cellulase” treatments. Contrarily, tangeretin content decreased or did not change after enzymatic treatment.

Table 1.

Phenolic profile (mg 100 g⁻¹ DM) of CPB extracts after enzymatic treatment, as quantified by HPLC–DAD

Phenolic compound	Reaction time (h)	Enzymatic treatment
		Control	T	β	C	TC	βC
Gallic acid	6	3.75 ± 0.17c,A	4.66 ± 0.04b,B	3.52 ± 0.12c,A	5.83 ± 0.13a,A	3.85 ± 0.09c,C	6.03 ± 0.22a,B
	12	3.62 ± 0.12c,A	4.84 ± 0.11b.B	3.81 ± 0.21c,A	6.09 ± 0.12a,A	6.07 ± 0.06a,A	6.32 ± 0.29a,AB
	24	3.63 ± 0.09d,A	5.56 ± 0.17c,A	3.64 ± 0.08d,A	6.00 ± 0.10b,A	5.76 ± 0.13bc,B	6.75 ± 0.23a,A
Narirutin	6	29.40 ± 2.51a,A	9.25 ± 0.81c,A	8.18 ± 1.62c,A	31.40 ± 1.59a,A	23.30 ± 2.61b,A	1.84 ± 0.25d,B
	12	18.74 ± 2.54b,B	8.14 ± 0.31c,A	7.00 ± 0.36c,A	31.93 ± 0.72a,A	16.24 ± 1.64b,B	4.60 ± 0.15c,A
	24	15.68 ± 2.54b,B	8.28 ± 0.31c,A	6.19 ± 0.36c,A	29.16 ± 0.72a,A	17.61 ± 1.64b,B	2.28 ± 0.15c,B
Naringenin	6	–	11.25 ± 0.21c,B	21.15 ± 0.14a,C	–	–	16.64 ± 0.75b,B
	12	–	9.96 ± 0.26c,C	35.65 ± 1.73a,B	–	5.71 ± 0.09d,B	22.31 ± 1.32b,A
	24	–	16.86 ± 0.59c,A	41.48 ± 1.31a,A	8.24 ± 0.20d	10.61 ± 0.06d,A	24.19 ± 3.23b,A
Hesperidin	6	166.28 ± 12.10b,A	198.86 ± 4.39a,A	204.53 ± 2.61a,A	127.27 ± 18.12c,A	126.00 ± 5.23c,A	–
	12	131.85 ± 4.01c,B	141.13 ± 3.61b,C	165.81 ± 5.16a,B	115.12 ± 0.60d,A	101.41 ± 1.17e,C	–
	24	159.46 ± 2.79a,A	164.06 ± 2.24a,B	158.10 ± 2.59a,B	131.77 ± 2.86b,A	112.95 ± 1.44c,B	–
Hesperetin	6	–	69.26 ± 1.53c,B	163.50 ± 0.91a,C	–	–	97.05 ± 4.92b,A
	12	–	69.28 ± 2.38c,B	315.70 ± 15.65a,B	–	24.27 ± 0.43d,B	156.10 ± 10.01b,A
	24	–	146.45 ± 5.10b,A	407.90 ± 2.69a,A	36.05 ± 1.18d	68.69 ± 0.48c,A	169.01 ± 24.37b,B
Tangeretin	6	5.82 ± 0.44a,B	4.17 ± 0.06b,A	4.29 ± 0.07b,B	1.45 ± 0.06d	2.28 ± 0.10c,A	1.45 ± 0.04d,B
	12	7.06 ± 0.25a,A	2.41 ± 0.07c,B	5.67 ± 0.29b,A	1.78 ± 0.04d,B	1.60 ± 0.01d,C	1.73 ± 0.12d,A
	24	5.27 ± 0.09a,B	2.99 ± 0.09b,C	5.30 ± 0.10a,A	3.18 ± 0.07b,A	2.08 ± 0.06c,B	1.09 ± 0.14d,C

CPB citrus pectin by-product, DM dry matter, T tannase, β β-glucosidase, C cellulase, TC tannase + cellulase; and βC β-glucosidase + cellulase

^a,b,c,dValues (mean ± SD) within a row followed by different lowercase letters are significantly different (p ≤ 0.05) by one-way ANOVA followed by post hoc Tukey HSD

^A,B,CValues (mean ± SD) within a column followed by different uppercase letters are significantly different (p ≤ 0.05) by one-way ANOVA followed by post hoc Tukey HSD

Citrus polyphenols are usually found either inside the vacuole of vegetable cells or bound to cellular matrices (Shahidi and Yeo 2016). Considering this information, we hypothesized that cellulase improves the recovery of phenolic compounds by hydrolyzing the cellulose matrix that was previously observed by Li et al. (2006) and Madeira and Macedo (2015). After 12 h of reaction, cellulase treatment increased gallic acid and narirutin contents by 1.7 and 1.9 fold, respectively. Cellulase for 24 h also increased hesperetin and naringenin contents, indicating the capacity of cellulase to remove sugar residues from phenolic compounds (Winotapun et al. 2013). Cellulase treatment reduced hesperidin and tangeretin contents. This result might be due to the ability of phenolic compounds to complex with proteins and enzymes, end-product inhibition pathways, and by phenolics degradation when exposed to air (Li et al. 2006; Madeira and Macedo 2015).

Among tannase treatments, tannase-24 h was the most efficient in releasing hesperetin and naringenin. As previously demonstrated (Ferreira et al. 2013), tannase is able to remove sugar residues from hesperidin and naringin. Thus, along with increased hesperetin and naringenin contents, reduced concentrations of glycosylated flavanones should be expected after tannase biotransformation. This effect was observed for narirutin. However, tannase-treated extracts showed markedly higher hesperidin contents than the control, and the levels of this glycoside did not coincide with the production of the aglycone. Together, hesperidin and hesperetin contents in tannase-24 h extracts were approximately 2-fold higher than the hesperidin content in the control. This result can be explained by the presence of other enzymes in the semi-purified tannase extract, which might have enhanced the release of hesperidin from the plant matrix.

The combined enzyme treatment “tannase + cellulase” resulted in no significant (p < 0.05) increase and had no synergistic effect on phenolic extraction and conversion compared with tannase and cellulase treatments. A similar result was obtained by Chamorro et al. (2012) and Martins et al. (2016), who investigated the effect of tannase, cellulase, and pectinase on polyphenol recovery from grape pomace. In contrast, Madeira and Macedo (2015) demonstrated that the combination of cellulase and tannase was optimum for the production of hesperetin and naringenin from citrus residues. This discrepancy between results could be attributed to differences in enzyme type, the content of polyphenol in the matrix, extraction solvent, or inhibitory effects on enzyme activity (Cerda et al. 2013; Río Segade et al. 2015; Nakajima et al. 2016).

Treatment of CPB residue with β-glucosidase for 24 h produced the greatest flavanone biotransformation. As observed with tannase treatment, aglycone levels after β-glucosidase for 24 h treatment did not coincide with glycoside levels in CPB control. The sum of the contents of hesperidin and hesperetin in β-glucosidase-24 h extracts was approximately 3.55-fold higher than the hesperidin content in the control, suggesting that β-glucosidase-24 h treatment released high amounts of flavanones from cell wall structures. According to Shahidi and Yeo (2016), β-glucosidases can promote the hydrolysis of β-glycosidic linkages, e.g., cellobiose into free glucose units and release of polyphenols linked to sugar moieties. In our study, the combined use of β-glucosidase and cellulase did not improve polyphenol release. In fact, compared with β-glucosidase treatment, “β-glucosidase + cellulase” treatment afforded low amounts of all phenolic compounds, except gallic acid. Although β-glucosidase and tannase can hydrolyze similar compounds, the enzymes greatly differ in their extent of hydrolysis, as shown in Table 1.

Madeira and Macedo (2015) applied tannase and cellulase to citrus by-products with high contents of hesperidin and naringin (4438 and 490 mg 100 g⁻¹ DM, respectively). The authors obtained lower hesperetin and naringenin yields (12 and 8 mg 100 g⁻¹ DM, respectively) than those of the present study. On the other hand, our results are consistent with those of Ruviaro et al. (2018), who used cellulase, pectinase, tannase, and β-glucosidase to treat CJB.

Antioxidant capacity

In general, all enzymatic treatments had significant (p < 0.05) and positive effects on the antioxidant capacity of CPB (Fig. 2). However, differences were observed between the results of the two antioxidant methods. According to DPPH radical scavenging values, “β-glucosidase + cellulase” was the most effective treatment in increasing antioxidant activity (6.1-fold increase compared with the control), followed by cellulase-24 h (fivefold increase compared with the control) and “tannase + cellulase” (fourfold increase compared with the control). DPPH results indicate that cellulase improved the bioactivity of extracts, producing compounds that are able to donate hydrogen atoms to DPPH radicals. By the ORAC method, antioxidant capacity was consistent with TPC and the quantified aglycones. According to ORAC results, β-glucosidase was the most effective in producing compounds able to transfer hydrogen atoms to the peroxyl radical. Previous studies reported that a high level of phenolic compounds contributes to stronger antioxidant activity in citrus fruits (Sun et al. 2013). Also, according to Di Majo et al. (2005), hesperetin and naringenin have higher antioxidant activity than their glycosides. These observations are consistent with our results, which showed that β-glucosidase-12 h and β-glucosidase-24 h extracts exhibited higher antioxidant activity (in ORAC), TPC, and aglycones content.

Fig. 2.

Effect of enzymatic treatment on CPB antioxidant capacity by ORAC (a) and DPPH assays (b). TPC: total phenolic content; DM: dry matter; GAE: gallic acid equivalents; Control: hydroalcoholic and ultrasound extraction; T: tannase; β: β-glucosidase; C: cellulase; TC: tannase + cellulase; βC: β-glucosidase + cellulase

Polyphenols are usually present in higher concentrations in the fruit pomace, which explains why this material has higher antioxidant activity than the juice. According to Khan et al. (2010), 100 g of the peel of Citrus sinensis L. contains between 509 and 712 µmol TE, as measured after ethanol- and ultrasound-assisted extraction, respectively. The present study and others in the literature demonstrate that the antioxidant capacity of citrus fruits can be improved by enzymatic treatments. In a study of Madeira and Macedo (2015), the antioxidant activity of citrus residue extracts was 1.8-fold higher after cellulase and tannase treatment. Similar results were observed by Li et al. (2006) after treating orange peel with cellulase. This treatment increased the antioxidant capacity of the extract in ~ 1.33-fold and ~ 2-fold, compared to the aqueous and hydroalcoholic extract, respectively. Ferreira et al. (2013) demonstrated that naringinase and tannase were able to increase the antioxidant capacity of orange juice by approximately 1.1 and 1.6 times, respectively. Regarding a different agroindustrial residue, Martins et al. (2016) found that the antioxidant activity of grape pomace could be increased by 1.6 to 5.9 times using tannase. Overall, these findings confirm the potential of enzyme-assisted extraction for releasing antioxidant compounds, such as polyphenols, from food matrices.

Pearson’s correlation analysis

The correlation between increases in CPB phenolic compounds contents after enzymatic treatments and increases in TPC and antioxidant capacity were analyzed by Pearson’s correlation test. The aglycones, naringenin and hesperetin, showed high positive correlation (r-values between 0.84 and 0.94) with ORAC and TPC, which indicate that these compounds are strongly correlated with the increase in antioxidant capacity of CPB. As the antioxidant capacity of citrus residues is influenced by TPC, as well as the number of hydroxyl groups in phenol rings, the aglycones generated from enzymatic hydrolysis and phenolic compounds extraction from the matrix CPB are expected to increase antioxidant activity of CPB (Rice-Evans et al. 1996; Sun et al. 2013). This hypothesis also explains the glycosides (narirutin and hesperidin) coefficients with TPC and antioxidant capacity. Narirutin showed a significant negative correlation with TPC and ORAC (−0.37 and −0.53, respectively). The r-values for the correlation between hesperidin and DPPH was also significant and negative (−0.69). These results indicate that the presence of these glycosides is negatively correlated with the extraction of phenolic compounds from CPB matrix and extracts with high antioxidant capacity. The presence of glycoside molecules bonded to flavanones decrease the number of free hydroxyl groups, decreasing the antioxidant capacity of phenolic compounds (Chamorro et al. 2012). The coefficients of correlation between gallic acid and tangeritin, and TPC, and ORAC values were low, confirming the minor effect of these compounds on the studied parameters. DPPH showed positive correlation with gallic acid, indicating that this compound present strong DPPH radical scavenging activity. Conversely, the r-values for the correlation between DPPH results and tangeretin contents were significant and negative.

A heat map in Fig. 3 summarize the effects of each enzymatic treatment on CPB. A color scale ranging from red (low concentration) to green (high concentration) was used to represent the TPC, DPPH radical scavenging activity, ORAC value, and polyphenols content of CPB extracts. β-glucosidase-24 h, β-glucosidase-12 h, and tannase-24 h had fewer red squares than other treatments, which indicate that these treatments increased most of the evaluated parameters of CPB extracts and improved their bioactivity. The opposite was observed for control treatment, which was represented mostly by red squares.

Fig. 3.

Heat map summarizing the effects of each enzymatic treatment on the total phenolic content (TPC), DPPH radical scavenging activity, ORAC value, and polyphenol content of CPB extracts. The color scale ranges from red (low concentration) to green (high concentration)

Considering the results presented so far, we concluded that β-glucosidase-24 h and tannase-24 h were the most effective treatments for increasing CPB polyphenols release, polyphenols bioconversion, and antioxidant activity. Therefore, we investigated the effect of increasing the concentration of β-glucosidase and tannase on the phenolic profile of CPB extracts.

Effect of increased enzyme concentration

Higher enzyme concentrations significantly (p < 0.05) increased naringenin, hesperidin, hesperetin, and diosmetin contents (Table 2). However, no changes (p > 0.05) in gallic acid and narirutin contents were observed. The highest effects on aglycone production were observed with 20 U of β-glucosidase treatment (Fig. S1), which increased naringenin and hesperetin contents 166.6% (73.63 mg 100 g⁻¹ DM) and 205.93% (766.44 mg 100 g⁻¹ DM) higher than when 5 U g⁻¹ was used. The content of the glycoside hesperidin in CPB was greatly increased after “tannase + β-glucosidase” and 20 U of tannase (by 115.72 and 161.31%, respectively). Furthermore, “tannase + β-glucosidase” was the only treatment able to produce diosmetin in quantifiable amounts.

Table 2.

Phenolic content (mg 100 g⁻¹ DM) in CPB extracts obtained using increased enzyme concentrations

	Control	5T	10T	20T	5β	10β	20β	Tβ
Gallic acid	4.74 ± 0.73a	4.44 ± 0.49a	5.27 ± 0.10a	4.08 ± 0.90a	4.75 ± 0.51a	3.84 ± 1.10a	5.21 ± 0.40a	4.35 ± 0.13a
Yield (%)§	–	–	118.69	91.89	–	80.84	109.68	91.58–97.97†
Narirutin	27.85 ± 1.21a	5.90 ± 0.14b	7.05 ± 0.55b	5.19 ± 0.67 b	5.95 ± 0.19b	5.59 ± 0.88b	6.06 ± 0.72b	7.01 ± 0.41b
Yield (%)	–	–	119.49	87.97	–	93.95	101.85	117.89–118.81†
Naringenin	–	28.01 ± 0.78d	28.48 ± 2.02d	51.58 ± 2.38 bc	44.18 ± 0.30c	57.56 ± 1.32b	77.63 ± 6.87a	68.24 ± 0.82a
Yield (%)	–	–	101.68	184.15	–	130.29	166.66	154.46–243.63†
Hesperidin	91.72 ± 10.0c	84.82 ± 0.93c	82.80 ± 10.07c	136.32 ± 5.82 a	108.19 ± 14.64bc	95.93 ± 11.98c	86.50 ± 9.34c	125.20 ± 4.96ab
Yield (%)	–	–	97.62	161.31	–	88.67	79.95	115.72–147.61†
Hesperetin	–	219.35 ± 5.22d	241.30 ± 16.89d	470.57 ± 21.50 b	372.18 ± 2.39c	525.27 ± 11.90b	766.44 ± 54.08a	641.65 ± 7.61a
Yield (%)			110.01	185.81	–	141.13	205.93	172.40–292.52†
Diosmetin	–	–	–	–	–	–	–	3.64 ± 0.05
Yield (%)	–	–	–	–	–	–	–	–

Control (Hydroalcoholic and ultrasound extraction), 5T (5 U tannase g⁻¹); 10T (10 U tannase g⁻¹); 20T (20 U tannase g⁻¹); 5β (5 U β-glucosidase g⁻¹); 10β (10 U β-glucosidase g⁻¹); 20β (20U β-glucosidase g⁻¹); Tβ (10 U tannase g⁻¹ + 10 U β-glucosidase g⁻¹) § Analyte yield, † Compared with 5β and 5T, respectively

^a,b,c,dValues (mean ± SD) within a row followed by different letters are significantly different (p ≤ 0.05) by one-way ANOVA, followed by post hoc Tukey HSD

Several parameters can affect and limit an enzymatic reaction. Enzyme and substrate concentrations are examples of such. In general, as enzyme concentration increased, the concentration of some polyphenols also increased (Table 2). This result indicates that the reaction did not reach a plateau with enzymes at 20 U g⁻¹ and that higher enzyme concentrations, and their economic cost analysis, can be studied in the future to optimize polyphenol release. Ruviaro et al. (2018) studied the effects of tannase and β-glucosidase concentrations on the release of polyphenols from CJB. The authors showed that the reaction plateau was not achieved for β-glucosidase, whereas the plateau was reached when using 10 U tannase g⁻¹. Thus, the enzyme/substrate ratio is a key factor to maximize polyphenol yield. Substrate composition and its affinity with enzymes might also influence the reaction (Panja 2017).

As shown in Table 2, “tannase + β-glucosidase” showed a synergistic effect on polyphenol release from CPB. “β-glucosidase + cellulase” and “tannase + cellulase”, however, produced opposite effects (Table 3). According to Andrić et al. (2010), the hydrolysis rate of β-glucosidase can be significantly reduced by high glucose concentrations. Cellulase possibly promoted the release of glucose in the reaction medium, which might have inhibited β-glucosidase and tannase activities, explaining the antagonistic effect observed in “β-glucosidase + cellulase” and “tannase + cellulase”. The combined use of β-glucosidase and tannase, on the other hand, seemed to have positive effects on enzyme activity, increasing the polyphenol content of CPB extracts.

Table 3.

Pearson’s correlation coefficients (r) between phenolic compounds, TPC, DPPH radical scavenging activity, and ORAC values of CPB extracts

Phenolic compound	TPC	DPPH	ORAC
Gallic acid	0.026ns	0.438**	−0.039ns
Narirutin	−0.373**	−0.138ns	−0.535**
Naringenin	0.843**	0.076ns	0.928**
Hesperidin	−0.069ns	−0.696**	−0.079ns
Hesperetin	0.855**	−0.002ns	0.936**
Tangeritin	−0.046ns	−0.606**	0.097ns

n = 18 (Control-6 h; control-12 h; control-24 h; T-6 h; T-12 h; T-24 h; β-6 h; β-12 h; β-24 h; C-6 h; C-12 h; C-24 h; TC-6 h; TC-12 h; TC-24 h; βC-6 h; βC-12 h; βC-24 h)

^nsnot significant; *p ≤ 0.05; **p ≤ 0.0001

Naringenin and hesperetin possess higher in vitro biological activity than their glycosylated forms (Shin et al. 2013; Ferreira et al. 2013; Madeira and Macedo 2015). Because these aglycones occur less frequently in fruits than glycosides, microorganisms and enzymes have been used for their synthesis. Microbial hydrolysis results in low yield and poor productivity compared with enzymatic hydrolysis (Shin et al. 2013). For instance, P. variotii produced 5 and 6 mg of hesperetin and naringenin 100 g⁻¹ citrus residue, respectively, under optimized conditions of solid fermentation (Madeira et al. 2014). A combination of tannase and cellulase, however, produced 12 and 8 mg of hesperetin and naringenin 100 g⁻¹ residue, respectively (Madeira and Macedo 2015). Acid hydrolysis can also be used to obtain aglycones from glycosylated flavanones. Its main disadvantage is the use of toxic solvents (Timell 1964). Enzyme-assisted extraction is a viable, green process with great potential to release naringenin and hesperetin from citrus residues compared with other methods.

To the best of our knowledge, the amounts of naringenin and hesperetin obtained from CPB by 20 U of β-glucosidase treatment, followed by hydroalcoholic and ultrasound extraction, are the highest ever reported for an enzymatic process. Given that these flavanones are not commonly found in natural sources, enzyme-assisted extraction of CPB can be considered an interesting alternative to obtain naringenin and hesperetin. The results of this study add value to this agroindustrial waste and provide support for the industrial use of CPB.

Conclusion

Enzymatic treatment, followed by hydroalcoholic and ultrasound extraction, significantly improved the antioxidant capacity and phenolic content of CPB. β-glucosidase for 24 h was the most effective treatment, followed by β-glucosidase-12 h and tannase-24 h. “β-glucosidase + cellulase” and “tannase + cellulase” showed no synergistic effect or significant benefits over the use of separate enzymes on polyphenol release. Person’s correlation test showed that TPC and antioxidant activity were highly correlated with hesperetin and naringenin contents. High enzyme concentrations significantly increased naringenin, hesperidin, hesperetin, and diosmetin contents in CPB. The treatment with 20 U of β-glucosidase produced the highest amounts of aglycone flavanones ever obtained by an enzymatic process. Enzymatically treated CPB extract is an important source of high-quality antioxidant compounds, which can be incorporated into functional foods to meet the growing demands of health and food markets for these products. The results of this study add value to citrus pectin industry waste, a material that currently has no industrial application. Further research is needed to develop food products and nutraceuticals containing polyphenolic-rich extracts of CPB.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Compliance with ethical standards

Conflict of interest

None.