**Biochemical Characterization of Traditional Varieties of Sweet Pepper (***Capsicum annuum* **L.) of the Campania Region, Southern Italy**

**Florinda Fratianni 1,\*, Antonio d'Acierno 1,\*, Autilia Cozzolino 2, Patrizia Spigno 3, Riccardo Riccardi 3, Francesco Raimo 4, Catello Pane 4, Massimo Zaccardelli 4, Valentina Tranchida Lombardo 5, Marina Tucci 5, Stefania Grillo 5, Ra**ff**aele Coppola <sup>2</sup> and Filomena Nazzaro <sup>1</sup>**


Received: 18 May 2020; Accepted: 24 June 2020; Published: 26 June 2020

**Abstract:** Bioactive compounds of different Campania native sweet pepper varieties were evaluated. Polyphenols ranged between 1.37 mmol g−<sup>1</sup> and 3.42 mmol g−1, β-carotene was abundant in the red variety "Cazzone" (7.05 μg g−1). Yellow and red varieties showed a content of ascorbic acid not inferior to 0.82 mg g−1, while in some green varieties the presence of ascorbic acid was almost inconsistent. Interrelationships between the parameters analyzed and the varieties showed that ascorbic acid could represent the factor mostly influencing the antioxidant activity. Polyphenol profile was different among the varieties, with a general prevalence of acidic phenols in yellow varieties and of flavonoids in red varieties. Principal Component Analysis, applied to ascorbic acid, total polyphenols and β-carotene, revealed that two of the green varieties ("Friariello napoletano" and "Friariello Sigaretta") were well clustered and that the yellow variety "Corno di capra" showed similarity with the green varieties, in particular with "Friariello Nocerese". This was confirmed by the interrelationships applied to polyphenol composition, which let us to light on a clustering of several red and yellow varieties, and that mainly the yellow "Corno di capra" was closer to the green varieties of "Friariello".

**Keywords:** biodiversity; *Capsicum annuum* L.; antioxidant activity; β-carotene; ascorbic acid; polyphenols; statistical analysis

#### **1. Introduction**

Sweet pepper (*Capsicum annuum* L.), belonging to the Solanaceae family, is one of five cultivated species of the genus *Capsicum*, including sweet and hot peppers. It is a component present in the diet of many populations in the world; as such, therefore it represents an important source of income for farmers and operators in the agro-industrial sector. *C. annuum* L., originated from Central and South America and arrived in Europe in the sixteenth century with the Spanish and Portuguese expeditions to the lands of the New World. Once introduced in cultivation, pepper soon became a very common

vegetable in the kitchens of all the countries of Europe. Today, *C. annuum* L. is cultivated in a large number of varieties all over the world, reaching a cultivated area that exceeds 1.5 million hectares. Italy is one of the most important countries in the world for this crop. The commercial produce of sweet pepper is the fruit, having different forms and size, as well as several colors, which range from yellow to red, from intense purple to dark green to black, depending on cultivar, maturity, growing conditions, and postharvest manipulation. Sweet pepper represents an excellent source of several antioxidant molecules, in particular carotenoids [1,2] ascorbic acid, and polyphenols, such as quercetin [3], which received huge interest for to their antioxidant properties [4]. A considerable weight of evidence suggests that consumption of fruit and vegetables is beneficial for human health and may help in preventing several chronic diseases [5]. Since last decades, secondary metabolites of vegetables have come into light for their presumed role in fighting cancer and cardiovascular diseases [6,7] as well as in slowing down aging and atherosclerosis [8].

Campania is among the most prominent regions of Europe in terms of production and number of varieties of fruit and vegetables [9]. Over the millennia, indigenous plant varieties, in a sort of ideal symbiosis with the territory, have managed to express their quality and nutritional value at the best. Several of these varieties are still widespread, while others are cultivated as niche products, or have been abandoned, supplanted by commercial varieties, which are more productive but often have lower quality. Pepper is one of the most important crops for this area and several varieties have become traditional of the Campania region, giving very worthy inputs to its rich cultivated biodiversity. Plant biodiversity includes the enormous amount of vegetal germplasm differentiated in the course of the long history of biological evolution of species. In agriculture, biodiversity is also the work of human selection from a "wild" gene pool to obtain breeds and varieties adapted to various ecological, economic and social conditions. Phytochemicals of sweet peppers undoubtedly arouse great interest mainly act as antioxidant agents. They represent also important element for human health and can prevent the occurrence of disease linked to oxidative stress, including cardiovascular and neurodegenerative diseases, and cancer [10]. Certain green, red, orange, and yellow pepper showed interesting capacity to inhibit some key enzymes linked to Alzheimer disease [acetylcholinesterase (AChE), butyrylcholinesterase (BChE), and β-secretase (BACE1)] [11]. Due to the broad biochemical and nutritional variations existing among the different varieties within each plant species, the identification of the best genotypes assumes a particular importance for both breeders and consumers, which thereby can select and consume products with high nutritional quality, respectively. Keeping this in view, the present study focused on the following objectives: (a) to analyze the contents of bioactive compounds of native pepper varieties from the Campania region, Southern Italy, including β-carotene, total phenolics, phenolic profiles, ascorbic acid; (b) to determine their in vitro antioxidant activities through the DPPH radical-scavenging activity; (c) to correlate the antioxidant activity of the varieties to total phenolics, ascorbic acid and carotenoids, with the aim of identifying the main factors influencing the antioxidant properties of the product. Interrelationships between the parameters analyzed and the different varieties were investigated by principal component analysis (PCA).

#### **2. Materials and Methods**

#### *2.1. Chemicals*

Caffeic, ferulic, p-coumaric, gallic, chlorogenic acids, epicatechin, rutin, quercetin, 2,2-diphenyl-1-picrylhydrazyl (DPPH), β-carotene, ascorbic acid, HPLC-grade methanol, sulphuric, metaphosphoric, acetic and formic acids, acetonitrile, petroleum ether, ethanol and acetone were purchased from Sigma-Aldrich (Milano, Italy). Apigenin was purchased from Extrasynthese (Genay, France). The Folin–Ciocalteu reagent was purchased from BIO-RAD (Milano, Italy). Water was distilled and filtered through a Milli-Q apparatus (Millipore, Milano, Italy) before use.

#### *2.2. Plant Material*

Plant material (Figure 1) included seven types of yellow and red sweet pepper, (*Capsicum annuum* L.) ("Papaccella napoletana", "Papaccella liscia", "Corno Marconi", "Corno di capra", "Cornetto di Acerra", "Cazzone", "Sassaniello"), corresponding to fourteen varieties, and three varieties of green sweet pepper ("Friariello napoletano", "Friariello nocerese", "Friariello a sigaretta") listed by the Official Bulletin of the Campania Region (B.U.R.C. n◦42, 145, 2009), grown and collected in the farm of the "Cooperativa ARCA 2010" sited in Acerra (NA), Italy. Acerra (40.9441◦ N, 14.3714◦ E) is characterized by a Mediterranean climate with an average of air temperature (*T*), humidity (*U*) and rainy days (*R*) *T* = 22.7 ◦C; *U* = 63.8%; *R* = 6.6 during the growing season [12]. The seeds of the varieties of pepper are stored and preserved by the gene bank "banca del germoplasma" at the "Consorzio Arca 2010" on behalf of the Campania region. Seedlings were transplanted in three rows with 60 plants ("Cazzone" and "Sassaniello"); 100 plants ("Friariello"); 83 plants ("Papaccella napoletana", "Papaccella napoletana liscia", "Corno di capra"). Cultivation techniques included stakes (1.2 m) as support and twine threads to tie the plants. Microirrigation was used as technique of irrigation. Replicates were 3. Each replicate included 10 plants collected for the determination of marketable production. Four harvests were generally performed from the beginning of August to the beginning of October. Before analysis, fruits were gently cleaned. Peduncles and seeds of pepper fruits were removed; the comestible portion was cut and immediately stored at −26 ◦C.

**Figure 1.** Traditional sweet pepper varieties of the Campania region analyzed in the present work. Legend: (**a**) Papaccella liscia; (**b**) Sassaniello; (**c**) Cornetto di Acerra; (**d**) Corno di capra; (**e**) Corno Marconi (**f**) Papaccella Napoletana; (**g**) Friariello a sigaretta; (**h**) Friariello Napoletano; (**i**) Friariello Nocerese; (**l**) Cazzone. The photos were made by the authors.

#### *2.3. Dosage of Ascorbic Acid*

Dosage of ascorbic acid was performed following the method of Nazzaro et al. [13]. All samples were cut, squeezed and incubated in three volumes of metaphosphoric acid (4%) and maintained for 1 h at 4 ◦C, avoiding the exposition to the light. Extracts were subjected to centrifugation (11,600× *g* for 10 min at 4 ◦C, Biofuge, Beckman Italia, Cassina de' Pecchi, Milano, Italy), and filtration (0.45 μm mesh, Millipore, Milano, Italy) to recovery the supernatant. A Gold System chromatograph equipped with an UV detector (Beckman Italia, Cascina dè Pecchi MI, Italy) and a Khromasil KR 100-5 C18 column (25 cm × 4.6 mm) was used to assess, through RP-HPLC, the amount of ascorbic acid present in the samples (run condition = mobile phase: sulphuric acid 0.001 M in HPLC-grade water; injection volume: 20 μL; flow rate: 1.0 mL min−1; detection wavelength: 245 nm; Temperature: room temperature). Ascorbic acid, previously dissolved in the mobile phase, was used as standard to generate the standard curve.

#### *2.4. Carotene Content*

Extraction was carried out according to the method described by Nazzaro et al. [13], modified as follows: fresh sample was cut and squeezed in ethanol (1:1 *w*/*v*), and then petroleum ether was added (1.5:1 *v*/*v*). The mixture was vigorously shaken; the supernatant was recovered by centrifugation (11,600× *g*, 15 min; Biofuge, Beckman Italia, Milano, Italy). The steps were repeated until the complete disappearance of the color, then the supernatants were put together. The amount of carotenoids was evaluated at 450 nm with petroleum ether as blank, and using the extinction coefficient ε = 2592, using the spectrophotometer Cary 50 Uv/Vis (Varian-Agilent Italia, Cernusco sul Naviglio, Italy).

#### *2.5. Total Polyphenols and Antioxidant Activity*

Samples were cut and squeezed (1:3 *w*/*v*) in methanol (containing acetic acid 1%) overnight at 4 ◦C. After centrifugation (11,600× *g*, 15 min; Biofuge, Beckman Italia), supernatants were recovered and the polyphenols amount and profile as well as the antioxidant activity was evaluated. The content of total polyphenols was spectrophotometrically evaluated (Cary 50 Varian-Agilent Italia) at λ = 760 nm, following the method described by Singleton and Rossi [14] with Folin Ciocalteau reagent. Gallic acid was used as standard. Results were indicated as mMol gallic acid equivalent g−<sup>1</sup> of fresh sample. Radical-scavenging activity was assayed through the use of the stable radical 2,2-diphenyl-1-picrylhydrazyl (DPPH assay) following the method of Ombra et al. [15]. The analysis was performed in microplates by adding 15 μL of extract to 300 μL of a methanol- DPPH solution (6 <sup>×</sup> 10–5 M). The absorbance was measured at <sup>λ</sup> = 517 nm (Cary 50 MPR Varian-Agilent Italia). The scavenging activity was expressed in terms of EC50, indicating the amount of sample amount (mg) necessary to inhibit DPPH radical activity by 50% for the duration of 60 min of incubation.

The experiments were performed in triplicate. Results were expressed as the mean values ± standard deviation.

#### *2.6. Chromatographic Analysis*

Ultra-high-performance liquid chromatography (UPLC) analysis was made using the ACQUITY Ultra Performance LCTM system (Waters, Milford, MA, USA) connected to a PDA 2996 photodiode array detector (Waters), characterized by a low dispersion with enlargement of the band lower than 10 μL; automatic control of the temperature; technology Smart-Start for the gradient, to perform a controlled mixing of the solvents until pressure = 15,000 psi; controlled degassing and automatized firmness of the solvents; direct setting of gradients in terms of pH, molarity and/or organic composition. Detector PDA: linear answer in the interval of wavelength ranging between 190 nm and 500 nm and for values of absorbance until 2.0 AU: Deviation of 1.3% at 2.0 AU; Deviation at 5.0% at 2.8 AU. The acquisition and processing of the relative data, as well as the control of the instruments was performed through the Empower software. The analysis was performed following the method described by Fratianni et al. [16]

and Pane et al. [17]. All the extracts and standards were dissolved in methanol, and filtered through Whatman 0.45 μm (Waters, Milford, MA, USA). The analyses were performed at 30 ◦C. Running conditions = Injection volume: 5 μL. Mobile phase: solvent A (7.5 mMol acetic acid) and solvent B (acetonitrile); flow rate: 250 <sup>μ</sup>L min<sup>−</sup>1; column: reversed phase column (BEH C18, 1.7 <sup>μ</sup>m, 2.1 <sup>×</sup> 100 mm Waters, based on ethyl bridge silanes, stable in a range of pH between 2 and 12 and temperatures up to 90 ◦C). The analysis was performed with a gradient elution (5% B for 0.8 min; 5–20% B over 5.2 min; 20% B for 0.5 min; 20–30% B for 1 min; 30% B for 0.2 min, 30–50% B over 2.3 min, 50–100% B over 1 min, 100% B for 1 min; 100–5% B over 0.5 min). Finally, the column was restored to the initial conditions for 2.5 min. Conditions of the apparatus= Pressure ranging from 6000–8000 psi; Scanning range of LC detector: 210–400 nm, resolution: 1.2 nm).

#### *2.7. Statistical Analysis*

Data were expressed as mean ± standard deviation of triplicate measurements. The PC software "Excel Statistics" was used for the calculations. Interrelationships between the parameters analyzed and the different varieties were investigated by principal component analysis (PCA), following the method of Fratianni et al. [18] and using the software package MATLAB.

#### **3. Results and Discussion**

#### *3.1. Ascorbic Acid Content*

Ascorbic acid, β-carotene, total polyphenols contents and antioxidant activity are shown in Table 1.

**Table 1.** Content of ascorbic acid, β-carotene, total polyphenols and antioxidant activity exhibited by the yellow, red and green traditional varieties of sweet pepper of the Campania region. Data represent the average (± SD) of three independent experiments. For details, see the Section Materials and Methods.


Among the yellow varieties, the amount of ascorbic acid ranged between 0.80 mg g−<sup>1</sup> ("Sassaniello") and 1.72 mg g−<sup>1</sup> ("Corno Marconi") of fresh product. We observed a less wide situation in red varieties,

where the concentration of ascorbic acid ranged between 0.82 mg g−<sup>1</sup> ("Papaccella liscia) to 1.20 mg g−<sup>1</sup> ("Corno Marconi"). Both in yellow and red varieties, "Corno Marconi" showed the highest amount of ascorbic acid. Among the red varieties is to remark also the high amount of ascorbic acid in the "Cornetto di Acerra" (1.19 mg g−<sup>1</sup> of fresh product). The three green varieties showed the smallest content of ascorbic acid. "Friariello nocerese" contained a slightly higher content of ascorbic acid than "Friariello Napoletano" and "Friariello sigaretta". For the most of the analyzed varieties, the amount of ascorbic acid was also higher than that reported by Howard et al. [19], and in some cases it was similar (in the case of yellow "Cazzone" or slightly inferior (red "Cazzone" and yellow "Corno di capra") than reported by Mennella et al. [20]. Considering a portion of 100 g of raw pepper, all varieties of sweet pepper analyzed, with the exception of the two green varieties "Friariello Napoletano" and "Friariello a sigaretta", reached a concentration of vitamin C at least two-fold the recommended daily dosage suggested by different international committees (WHO, the USA National Academy of Sciences, the USA Ministry of Health), and the *Codex alimentarius*, which indicate a minimum of 30 mg/die of ascorbic acid in human diet [21].

#### *3.2.* β*-Carotene*

β-Carotene or pro-vitamin-A is an abundant carotenoid present, as stable all-trans isomer, in many vegetal tissues including those of pepper. Due to its nutritional and coloring properties, it is used for several purposes, such as food coloring and preserving agent, in pharmaceuticals, as drug or drug ingredient that help to counteract vitamin A deficiency (VAD), thus preventing anomalies in growth, development, vision and immune function and cosmetics, for example as protective skin agent [22]. Pepper represents one of the primary dietary sources of pro-vitamin A [23] and generally, β-carotene is more abundant in fresh hot pepper than in sweet cultivars [23,24]. β-Carotene concentration is directly correlated to the total carotenoid content, being the precursor for the predominant orange and red carotenoids of pepper [25]. The data about the content of β-carotene present in the varieties of sweet pepper analyzed are shown in Table 1. In the yellow varieties, it ranged between 2.55 μg g−<sup>1</sup> "(Corno Marconi") and 5.79 μg g−<sup>1</sup> of fresh product ("Cazzone giallo"); on the other hand, the red varieties exhibited always a content of β-carotene higher than the analogous yellow varieties, with values ranging from 3.96 μg g−<sup>1</sup> (Sassaniello that therefore was the only variety to show a content of β-carotene inferior to the corresponding yellow "Sassaniello") to 7.05 μg g−<sup>1</sup> of fresh product (in the variety "Cazzone rosso"). Such values could be considered superior, for instance, respect to those reported by Thuphairo et al. [11], who indicated high values of carotenoids but considering the dry weight of the product. Respect to the shape of the varieties of peppers analyzed, we could say that the round red varieties of sweet pepper ("Papaccella napoletana" and "Papaccella napoletana liscia") almost always had a higher β-carotene content than the analogous yellow varieties and that among the elongated varieties, "Cazzone" had a slightly greater β-carotene content respect to the others. The green varieties "Friariello a sigaretta" and "Friariello Napoletano" showed a lower content of β-carotene. Also in this case, "Friariello nocerese" exhibited a different behavior, with 2.24 μg of β-carotene g<sup>−</sup>1.

#### *3.3. Total Polyphenol Content and Antioxidant Activity*

Total polyphenol (TP) content ranged between 1.37 mmol g−<sup>1</sup> (in the variety "Friariello sigaretta") and 3.42 mmol g−<sup>1</sup> (observed in the yellow "Papaccella napoletana") of fresh product (Table 1). Such values were lower respect to those reported by Moktar et al. [26] and Hallmann and Rembiałkowska [27], but in some cases similar to those observed by Nazzaro et al. in elongated yellow and red varieties of sweet pepper cultivated in the Sicilia region [13] or to those cultivated in Romania, in different conditions of growth [28]. On the other hand, they resulted in some cases higher than those reported by Shotorbani et al. [29]. This indicates that such an important biochemical parameter might be related not only to the variety but also to the different geo-climatic conditions and to methods of treatment applied for their extraction. Among the round varieties of "Papaccella Napoletana", the yellow type showed

slightly higher content of TPs than the red type (3.42 and 3.13 mmol g−<sup>1</sup> of fresh product, respectively); on the other hand, the difference of total polyphenols was more marked between the two "Papaccella liscia", whose yellow type exhibited a TPs content 21.08% higher than the correspondent red type (2.80 mmol g−<sup>1</sup> and 2.21 mmol g−<sup>1</sup> of fresh product, respectively). Among the elongated varieties, the trend was different. Thus, while the red type of "Cazzone" contained less total polyphenols than the corresponding yellow type (2.66 mmol g−<sup>1</sup> and 3.14 mmol g−1, respectively), the red "Corno Marconi" showed higher total polyphenols than the yellow type (3.36 mmol g−<sup>1</sup> and 2.79 mmol g−<sup>1</sup> of fresh product, respectively). A much more marked difference was observed within the variety "Corno di capra", whose red type contained 1 mmol g−<sup>1</sup> of total polyphenols more than the corresponding yellow type. Instead, the two varieties "Cornetto di Acerra" and "Sassaniello" showed almost the same amount of total polyphenols in both red and yellow types, with a very slight predominance in the yellow ones. The unusual difference between red and yellow "Corno di capra" was also observed in a genetic survey of the same set of traditional Campania peppers analyzed in the present study. The survey showed that the two varieties belong to two different and unrelated genetic groups according to Simple Sequence Repeats (SSR) DNA molecular markers (Tranchida-Lombardo et al., in prep.). Indeed, the "Corno di capra" yellow type is more related to the three green peppers, while the "Corno di capra" red type is more associated to "Cornetto di Acerra" (Tranchida-Lombardo et al., in prep.). The three green varieties of sweet pepper showed a much lower content of total polyphenols than those red and yellow. "Friariello nocerese" represented an exception, with total polyphenol similar to the yellow "Corno di capra" However, this did not correspond to a similar antioxidant activity, which resulted less effective (EC50: 7.27 versus 2.70, respectively) in "Friariello nocerese" than the "Corno di capra" that therefore exhibited higher amounts of β-carotene and ascorbic acid. Its content of TPs was also higher than that of the red "Papaccella liscia". However, also in this case, to a higher amount of total polyphenols did not correspond a similar antioxidant activity; on the contrary, the antioxidant activity of red "Papaccella liscia" was two-time stronger respect to that of "Friariello nocerese" (EC50: 7.27 versus 3.64, respectively). The other two green peppers, "Friariello a sigaretta" and "Friariello Napoletano", which showed the lowest values of all three parameters (ascorbic acid, β-carotene and total polyphenols) had the lowest effective antioxidant activity (EC50 = 9.57 mg and EC50 = 13.18 mg, respectively). Interestingly, the biochemical relationships observed among green peppers were also confirmed by SSR markers analysis, which showed that "Friariello nocerese" is less related to the two other green peppers ("Friariello sigaretta" and "Friariello napoletano") and more related to the "Corno Marconi" variety (red and yellow types) (Tranchida-Lombardo et al. in prep.).

Statistical analysis, performed taking into account these biochemical parameters, allowed us to potentially identify the contribution of ascorbic acid, β-carotene and total polyphenols on the antioxidant activity. Thus, as shown in Figure 2, although all three parameters concurred to affect the antioxidant activity of the varieties of sweet pepper, the content of total polyphenols and ascorbic acid seemed better correlated with the antioxidant activity (corr = −0.75 and corr = −0.81, respectively) than the β-carotene (corr = −0.51).

**Figure 2.** Correlation analysis. Red and yellow varieties are indicated with red and yellow squares, respectively. Green varieties are represented with green squares, except samples of "Friariello Nocerese", represented by green circles, to highlight that at least two (of three samples) resulted very similar to red and yellow samples.

In any case, for all analyses of correlation, we always observed a group formed by the red and yellow varieties and another distinct group formed by the green varieties. For this last group, the exception was represented by the "Friariello Nocerese", which seemed most moved towards the red and yellow varieties. Therefore, the sweet pepper 'Friariello', widely cultivated in the Campania region, is also one of the most common Italian varieties [30], for which distinct group types were ascertained by Parisi et al. [31], on the basis of the genetic, morphological traits and some qualitative traits.

Principal Component Analysis, obtained using MATLAB and applied to ascorbic acid, total polyphenols and β-carotene, revealed that the first principal component accounted for 66% of the total variance, while the first two components accounted for about 89% of the total variance of data (Figure 3). Interestingly, two of the green varieties ("Friariello napoletano" and "Friariello Sigaretta", indicated as green squares) are well clustered and completely located near the PCA1 axis. On the other hand, we observed a significant overlapping between red and yellow varieties. The green variety "Friariello Nocerese" -indicated in the Figure 3 with green circled- showed, once again, greater similarity with the complex group of yellow and red varieties, in particular with the yellow "Corno di capra" than with the other two green varieties of sweet pepper.

**Figure 3.** PCA obtained considering β-carotene, ascorbic and total polyphenols. Yellow and red-colored symbols indicate the yellow and red varieties, respectively.

#### *3.4. Polyphenol Profile*

Polyphenol profile, from qualitative and quantitative point of view, may vary in relation to genetic variation but also to different growth and geographic conditions. Obtaining as much information as possible on the polyphenolic composition becomes crucial for identifying the best varieties, from a qualitative and health-nutritional point of view, existing in a given territory. As far as we know, there are no studies that have characterized such a large number of traditional Campania varieties from a biochemical point of view. In particular, there are no scientific studies reporting the analysis of the qualitative and quantitative profile of the polyphenols present in many of the traditional sweet pepper varieties of the Campania region. This study was carried out using UPLC, an analytical approach that proved to be a powerful method for the analysis of individual polyphenols in a few minutes (in our case in an analysis time not exceeding 15') and capable of providing a detailed analysis of the phenolic molecules and, above all, of their concentration, contained in the varieties. All data, reported as μg g−<sup>1</sup> respect to the metabolites identified through UPLC are shown in Table 2.

**Table 2.** Polyphenol composition of the yellow, red and green traditional varieties of sweet pepper of the Campania region, obtained by UPLC analysis, expressed as <sup>μ</sup>g g−<sup>1</sup> <sup>±</sup> Standard Deviation.



**Table 2.** *Cont.*

Legend: molecules = GAL: gallic acid; CHL: chlorogenic acid; CAF: caffeic acid; CUM: *p*-coumaric acid; FER: ferulic acid; CAT: catechin; EPI: epicatechin; QUE: quercetin; RUT: rutin; NAR: naringenin; API: apigenin; Varieties = CG: Cazzone, yellow; PNG: Papaccella napoletana, yellow; PLG: Papaccella liscia, yellow; CAG: Cornetto di Acerra, yellow; SG: Sassaniello, yellow; CCG: Corno di Capra, yellow; CMG: Corno Marconi, yellow; CR: Cazzone, red; PNR: Papaccella napoletana, red; PLR: Papaccella liscia, red; CAR: Cornetto di Acerra, red; SR: Sassaniello, red; CCR: Corno di Capra, red; CMR: Corno Marconi, red; FNAP: Friariello Napoletano, green; FNOC: Friariello Nocerese, green; FSIG: Friariello Sigaretta, green.

Gallic acid and chlorogenic acid were the most abundant phenolic acids in all varieties. Among the yellow and red varieties, gallic acid ranged between 44.80 μg g−<sup>1</sup> (in the red "Papaccella liscia") and 185.70 μg g−<sup>1</sup> (found in the yellow type of "Corno Marconi"). "Friariello Nocerese" showed a content of gallic acid more similar, in terms of concentration, to some of yellow and red varieties (such as to the red variety of "Papaccella napoletana" (117.34 μg g−<sup>1</sup> and 112.27 μg g<sup>−</sup>1, respectively), than to the other two green varieties of Friariello, which content did not exceed 69.26 μg g−<sup>1</sup> ("Friariello napoletano"). In general, almost all the traditional varieties analyzed exhibited a content of gallic acid undoubtedly superior if compared to that present in some commercial varieties such as Roberta or Berceo [27,32], but similar to such varieties for the high content of chlorogenic acid, the low amount of caffeic acid

and for the negligible quantity of apigenin, therefore present only in very few varieties. Chlorogenic acid showed a sharped distribution along the varieties. It is to highlight that the green variety "Friariello Nocerese" contained the highest amount of this metabolite (273 μg g−1); on the contrary, we observed that the yellow type of "Cornetto di Acerra" showed a ninth of chlorogenic acid respect to "Friariello Nocerese". Probably in the "Cornetto di Acerra" other metabolites, not recognized with the available standards, were also present, justifying the amount of polyphenols that, in the yellow "Cornetto di Acerra", was 1mmol higher than "Friariello Nocerese". Caffeic acid was much less abundant both in yellow and red varieties, not exceeding 77.11 μg g−<sup>1</sup> (in the yellow "Corno di Capra"), while it was enough abundant in the three varieties of green pepper, reaching also 97.10 μg g−<sup>1</sup> in the "Friariello napoletano". *p*-Coumaric acid was absent in all varieties, except two green varieties ("Friariello Napoletano and "Friariello Nocerese"), which content of this metabolite did not exceeded 8.06 μg g−1. The trend of caffeic acid and chlorogenic acid was in agreement with Blanco-Rios et al. [33], which, analyzing different varieties of Mexican sweet pepper, found a higher amount of caffeic acid and chlorogenic acid in the green varieties, with a decreasing quantity starting from the red varieties, and following towards the orange and yellow ones. The almost complete absence of p-coumaric acid, observed only in the green varieties of "Friariello" is in agreement with Dimitriu et al. [34], which led to an increase of this metabolite at to 0.0487 mg GAE in 100 g FW only under microorganism fertilization in cultivars of pepper where it was not detected. Among flavonoids, catechin was detected in all varieties (except in the yellow "Corno Marconi" and in the red "Papaccella liscia") reaching also 166.68 μg g−<sup>1</sup> (in the red "Sassaniello"). Its isomer epicatechin was slightly less abundant than catechin (121.93 μg g−<sup>1</sup> as the highest concentration, in the red "Sassaniello"). The content of catechin observed by us was certainly higher than that reported by Ghasemnezhad et al. [35] in some varieties of sweet pepper, which contained a concentration of this metabolite ranging between 3.2 μg g−<sup>1</sup> FW and 15.54 μg g−<sup>1</sup> FW. The presence of such an important amount of catechin and epicatechin in some varieties of sweet pepper is very important. Catechins can reduce the risk of occurrence and development of some diseases. These metabolites can concur to regulate the glucose/insulin and lipid metabolism [36]; they have also neuroprotective [37], hearth protective [38], and anti-inflammatory effects [39]. Catechin can also protect from eye macula [40]. Catechin has also antibacterial activity. Nazzaro et al. found a strict correlation between the presence of catechin in extra virgin olive oil (EVO) and antibacterial activity exhibited by the EVO extract against *Staphylococcus aureus* [41]. Recently, catechin extracted from green tea-based polyphenol was also associated to rare earth (RE) metal ions to prepare catechin–RE complexes that showed significant anti-biofilm properties against *Pseudomonas aeruginosa*, *Staphylococcus sciuri* and *Aspergillus niger*, acting through the damage of microbial cell membrane [42]. Epicatechin and its metabolites can enhance muscle performance, improve symptoms of cardiovascular and cerebrovascular diseases, and support human health preventing diabetes and protecting the nervous system [43]. The presence of a so high amount of catechins and epicatechin in the varieties of sweet pepper resulted still more interesting, taking into account that, in some cases, flavonoid contents of pepper extracts can be enhanced with thermal process, such as by increasing temperature to 65 ◦C [29].

Quercetin was detected in the varieties "Papaccella napoletana" (both yellow and red), "Corno Marconi" (both yellow and red, 17.23 μg g−<sup>1</sup> and 124.86 μg g<sup>−</sup>1, respectively), in the red "Cornetto di Acerra" (104.35 μg g<sup>−</sup>1) and in the red "Corno di Capra" (that showed 7.36 μg g<sup>−</sup>1). Also in this case, our results, at least for the green varieties, were in agreement with Blanco-Rois et al. [33]; however, unlike these last, we found a certain amount of quercetin also in some varieties of red and yellow sweet peppers. Quercetin is one of the metabolites with the most known healthy effects. Recently Pingili et al. [44] ascertained an important role of quercetin as liver protective agent against different drugs and toxic agents. Therefore, due to its healthy properties [45], quercetin has been also used in the formulation of novel functional foods, such as quercetin-fortified bread [46]. Yellow "Papaccella napoletana" exhibited a content of quercetin also superior respect to that observed by Ghasemnezhad et al. in the variety Arona [35]. Rutin was detected in some varieties, such as "Cazzone" (both yellow

and red), "Papaccella napoletana" (both yellow and red), and in the red types of "Cornetto di Acerra" and "Sassaniello". It was also identified and quantified in the two green varieties of sweet pepper "Friariello napoletano" and "Friariello sigaretta" that, on the other hand did not contain naringenin, present only in the third variety of green pepper "Friariello nocerese" and detected also in other two varieties, the yellow "Corno di capra" and the red "Sassaniello".

From a global analysis of polyphenolic profile, we could observe that the red "Sassaniello" and the yellow "Papaccella napoletana" showed almost all metabolites, except *p*-coumaric acid and quercetin in the "Sassaniello" and *p*-coumaric acid and naringenin in the "Papaccella napoletana". Therefore, the red type "Papaccella liscia" missed even seven polyphenols, in particular caffeic acid among the phenolic acids, and almost all flavonoids except epicatechin. This clearly shows in general a different metabolic pathway between the varieties of sweet pepper, which affects the presence and amount of these secondary metabolites.

PCA applied to the polyphenol composition of all varieties of sweet pepper analyzed, revealed that the first principal component accounted for 34.31% of the total variance, while the first two components accounted for about 52% of the total variance (Figure 4).

**Figure 4.** PCA obtained considering polyphenol composition. Yellow and red-colored symbols indicate the yellow and red varieties, respectively.

Although PCA covered 52% of the total variance, it allowed us to observe the clustering of several yellow and red varieties and that the yellow varieties "Corno di capra" and "Sassaniello" were close to the green varieties of "Friariello".

Calculating the percentages of the individual polyphenols with respect to the sum of the polyphenols recognized by the chromatographic analysis, and assembling phenolic acids and flavonoids in two groups, we could observe the general distribution of acidic phenols and flavonoids among the varieties of sweet pepper. Results are shown in Figure 5.

**Figure 5.** Distribution of phenolic acids and flavonoids and phenolic acids, respect to the molecules recognized by UPLC, in yellow, red and green traditional varieties of sweet pepper of the Campania Region. Varieties = CG: Cazzone, yellow; PNG: Papaccella napoletana, yellow; PLG: Papaccella liscia, yellow; CAG: Cornetto di Acerra, yellow; SG: Sassaniello, yellow; CCG: Corno di Capra, yellow; CMG: Corno Marconi, yellow; CR: Cazzone, red; PNR: Papaccella napoletana, red; PLR: Papaccella liscia, red; CAR: Cornetto di Acerra, red; SR: Sassaniello, red; CCR: Corno di Capra, red; CMR: Corno Marconi, red; FNAP: Friariello Napoletano, green; FNOC: Friariello Nocerese, green; FSIG: Friariello Sigaretta, green.

By the whole, the composition exhibited by the yellow and red varieties was quite different. Thus, we found that the majority of yellow varieties contained much more phenolic acids than flavonoids, in particular "Corno Marconi" (73.34% phenolic acids and 26.66% flavonoids), "Cornetto di Acerra" (72.5% phenolic acids and 27.5% flavonoids), "Sassaniello" (71.9% phenolic acids and 28.1% flavonoids) and "Corno di capra" (64.1% phenolic acids and 35.9% flavonoids). In the varieties "Cazzone" and" Papaccella liscia" the percentage of phenolic acids and flavonoids was almost equal, with a slightly preponderance of flavonoids. The trend was completely opposite in the red varieties, which contained generally much more flavonoids than phenolic acids. This was the behavior of "Corno Marconi (68.9% flavonoids and 31.31% phenolic acids), "Sassaniello" (containing 63.91% flavonoids and 36.09% phenolic acids) and "Cazzone" (62.68% flavonoids and 37.72% phenolic acids). Two red varieties, "Papaccella napoletana" and "Corno di capra" showed a similar content of phenolic acids and flavonoids. Only "Papaccella liscia" contained much more phenolic acids (73.81%) than flavonoids (26.19%). The three varieties of "Friariello" behaved completely different from each other. Thus, while the "Friariello Nocerese" showed a trend more similar to that of the yellow varieties, in particular "Cornetto di Acerra" and "Sassaniello", with a content of acid phenols (71.22%) much higher than that of the flavonoids (28.78%). The "Friariello napoletano" and the "Friariello a sigaretta" instead practically the same percentage of acid phenols (49.63% and 49.33%, respectively) and flavonoids (50.36% and 50.67% respectively). The behavior of "Friariello Nocerese", which in general showed highest values of phenolic acids and flavonoids than the other two green varieties, was in agreement with the results of Mennella et al. [20], which found similarities between 'Friariello Nocerese' and some yellow varieties of sweet pepper, in spite of the clear differences that the two varieties show as regard their morphological traits and color at maturation.

#### **4. Conclusions**

Breeding for high-value vegetables is an increasingly preferred ambition and the enhancement of crop bioactive compounds might be strategic to encounter the actual ideas of the market and consumers indeed. As far as we know, this is the first time that a similar study was performed on different yellow, red and green traditional varieties of sweet pepper of the Campania region, grown at the same conditions and collected in the same period, so to avoid as more as possible all those variables that could affect the metabolic pathway of these products [47]. Our results suggest the great potential of these landraces in terms of phytochemicals of health interest. The interrelationships between the parameters analyzed and the different varieties let us to light on a clustering of several red and yellow varieties, and that mainly the yellow "Corno di capra" was closer to the green varieties of "Friariello". This could contribute in the exploitation and improvement of both cultivation and breeding programs at regional, national and international level, with the aims to safeguard the crop biodiversity, taking into account the health of consumers and without forgetting that an increase in the cultivation of these traditional varieties could rise the farmers' income.

**Author Contributions:** Conceptualization, F.F. and F.N.; Funding acquisition, P.S., R.R. and M.Z.; Investigation, F.F., A.d.A., A.C., P.S., R.R., F.R., M.Z., V.T.L., M.T., S.G. and F.N.; Methodology, F.F., A.d.A., C.P., R.C. and F.N.; Software, A.d.A.; Supervision, F.N.; Writing—original draft, F.N., M.T. and A.d.A.; Writing—review & editing, F.N., M.T., V.T.L., S.G. and A.d.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** The work was partially supported by the Campania regional Projects "SALVE" and "AGRIGENET" PSR 2007-2013, az. f2. The authors are really grateful to Mario Parisi, which support contributed to enrich the discussion of the work.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **Voltammetric Behavior, Flavanol and Anthocyanin Contents, and Antioxidant Capacity of Grape Skins and Seeds during Ripening (***Vitis vinifera var. Merlot***,** *Tannat***, and** *Syrah***)**

### **Nawel Benbouguerra 1, Tristan Richard 2, Cédric Saucier <sup>1</sup> and François Garcia 1,\***


Received: 29 July 2020; Accepted: 25 August 2020; Published: 27 August 2020

**Abstract:** Skin and seed grape extracts of three red varieties (Merlot, Tannat, and Syrah) at different stages of ripening were studied for their total phenolic content (TPC) by using the Folin-Ciocalteu assay and for their total antioxidant capacity (TAC) by using spectrophotometric and electrochemical assays. Flavanol and anthocyanin compositions were also investigated using Ultra Performance Liquid Chromatography coupled with Mass Spectrometry (UPLC-MS). Results showed that seeds had the highest phenolic content and the highest antioxidant potential compared to skins at all stages of ripening. The highest TPC and TAC values were measured in seeds at close to veraison and veraison ripening stages. In skins, the highest values were found at the green stage, it was in accordance with the flavanols content. The voltammetric measurements were carried out using disposable single walled carbon nanotubes modified screen-printed carbon electrodes (SWCNT-SPCE). Three peaks on voltammograms were obtained at different oxidation potentials. The first anodic peak that oxidized at a low potential describes the oxidation of ortho-dihydroxy phenols and gallate groups, the second peak corresponds to the malvidin anthocyanins oxidation and the second oxidation of flavonoids. The third voltammetric peak could be due to phenolic acids such as *p*-coumaric acid and ferulic acid or the second oxidation of malvidin anthocyanins. The high linear correlation was observed between antioxidant tests and flavanols in skins (0.86 ≤ *r* ≤ 0.94), while in seeds, '*r*' was higher between electrochemical parameters and flavanols (0.64 ≤ *r* ≤ 0.8).

**Keywords:** skins; seeds; *Vitis vinifera*; antioxidant activity; cyclic voltammetry; phenolic compounds

#### **1. Introduction**

*Vitis vinifera* is the most economically important species of grape vine in the world with 78 million tons of grapes production in 2018 (see http://www.oiv.int/en/oiv-life/oiv-2019-report-on-the-worldvitivinicultural-situation). Grapes consumed as fresh fruits, juices, and other processed products, contain many phenolic compounds which are mostly located in seeds and skins [1]. These compounds are synthesized in response to various biotic and abiotic stress such as fungal invasion, UV irradiations, ozone, and heavy metal ions [2]. Their content changes depending on the grape variety, soil, climatic conditions, and the ripening stages [3].

Polyphenols are commonly present in the plant kingdom and they bring more and more interest [4]. Phenolic compounds can be divided in two groups, flavonoids and non-flavonoids, according to their carbon skeleton [4]. The flavonoids (C6-C3-C6) are located in both skins and seeds and the anthocyanins and flavanols are the most abundant compounds [5]. The non-flavonoids such as stilbenes and phenolic

acids are found in the skins [6]. The synergy between the various classes of polyphenols increases sample efficiency and activity [7]. Polyphenols protect plants against biotic and abiotic stresses and they are involved in organoleptic and qualitative properties of food and beverages derived from these plants [8]. Many studies have reported their biological activities. They have potent antioxidant capacity [7,9–18]. They may prevent diabetes [19,20], obesity [21–23], cardiovascular [24,25], and neurodegenerative diseases [25,26].

Radical scavenging capacity (DPPH and ABTS) and ferric reducing capacity, which are spectrophotometric assays, are usually used in order to determine the antioxidant capacity of foods and beverages [13]. In the last years, electrochemical techniques have been more widely used as alternative methods due to their sensitivity, rapidity, ease of use, and due to their minimal environmental effects [27]. Among these electrochemical techniques, cyclic voltammetry (CV) was the first and the most commonly used to characterize and determine the total polyphenols and the total antioxidant capacity [27]. The main CV (anodic curve) parameters are:


Electrode made of glassy carbon electrode is widely used but recently, carbon nanotubes electrode have become one of the most promising material [29]. This electrode is classified into two categories depending on the number of layers on multi-walled carbon nanotubes (MWCNTs) and single-walled nanotubes (SWCNTs) [29,30]. Actually, disposable screen-printed carbon electrodes modified with carbon nanotubes attract the attention of many researchers because of their numerous advantages including disposability [31], reproducibility, practicality, high sensitivity, the ability to be miniaturized to minimize the consumption of samples, and the low detection limits [32,33].

The aims of this work were:


#### **2. Materials and Methods**

#### *2.1. Chemicals and Reagents*

Folin-Ciocalteu reagent, sodium carbonate, 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS), persulfate de potassium, 1,1-diphenyl-2-picrylhydrazyl free radical (DPPH), sodium acetate trihydrate, ferric chlorure, 2,4,6-tri(2-pyridyl)-s-triazine (TPTZ), iron(II) sulfate heptahydrate, phloroglucinol, ascorbic acid, sodium acetate, tartaric acid, sodium hydroxide, gallic acid, trolox, catechin, caffeic acid, *trans*- resveratrol, hydrochloric acid, and glacial acetic acid were purchased from Sigma-Aldrich (Saint-Quentin Fallavier, France). Oenin chloride was obtained from Extrasynthese (Genay, France). Acetonitrile, methanol, and water UPLC-MS were purchased from Biosolve Chimie (Dieuze, France) and trifluoroacetic acid from Carlo Erba Reagents (Peypin, France).

#### *2.2. Samples*

Three *V. vinifera* varieties (Merlot, Tannat, and Syrah) were harvested on 2017 at different stages of ripening: Green stage (GS), close to veraison (CV), veraison (V), and maturity (M) (Supplementary Data, Table S1) from INRAe experimental vineyard (Montpellier, France) (coordinates: 43◦37 02.7" N 3◦51 22.3" E, average annual temperature: 15.85 ◦C, average annual precipitation: 629 mm (the weather was quite dry), and soil: Gravels and river sand). The whole grapes were stored at −80 ◦C in plastic bags until polyphenols extraction.

#### *2.3. Samples Preparation*

Seeds and skins of thirty Merlot, Tannat, and Syrah berries were manually removed from the pulp. The polyphenols were extracted with 100 mL of acetone/water (70/30 *v*/*v*) deoxygenated with nitrogen for 5 min. The solutions were filtered through a 0.45 μm filter paper after stirring during 18 h in the dark, and they evaporated in a rotavapor under low pressure at 37 ◦C. The resulting products were freeze-dried and stored at −20 ◦C until their use in antioxidant and other analytical assays [34]. Three biological replicates were done. After extraction, skin and seed extracts were weighted (dry weight: DW) and they were stored at −20 ◦C between 5 and 12 months before being used in the experiments.

#### *2.4. Determination of Phenolic Composition*

#### 2.4.1. Flavanols

The assay on flavanols was performed as described by [35]. Briefly, a solution of 0.1 N HCl in MeOH, containing 50 g/L phloroglucinol and 10 g/L ascorbic acid was prepared. Seed and skin grape extracts were dissolved in methanol and reacted for 20 min at 50 ◦C in this solution, and then combined with five volumes of 40 mM aqueous sodium acetate to stop the reaction.

The UPLC system was a Waters Acquity (Saint-Quentin-en-Yvelines, France), with a photodiode array detector (PDA), LC pump, and an auto sampler. The column used was a reversed phase UPLC with an Acquity UPLC BEH C18 column (2.1 × 50 mm, 1.7 μm particle size) (Saint-Quentin-en Yvelines, France). The method used a binary gradient with mobile phases: Mobile phase A containing 1% *v*/*v* aqueous trifluoroacetic acid and mobile phase B containing acetonitrile. The 20 min elution method at flow 0.45 mL/min was 0 min 2% B, 8 min 6% B, 14 min 20% B, 14.1 min 99% B, 16 min 99% B, 16.1 min 2% B, and 20 min 2% B. The column temperature was 40 ◦C. Eluting peaks were monitored at 280 nm. The catechin calibration curve was used. Results were expressed as mg/g of DW.

#### 2.4.2. Anthocyanins

Skin grape extracts were solubilized in MeOH/water (80/20 *v*/*v*) at an appropriate concentration then injected directly after filtration as described previously with some modifications [36].

The conditions of the chromatographic apparatus are the same as those mentioned in experimental Section 2.4.1. The column temperature was set at 50 ◦C. The 40 min elution method at flow 0.25 mL/min was 0 min 1% B, 5 min 8.8% B, 30 min 20.6% B, 30.5 min 96% B, 34 min 96% B, 34.1 min 1% B, and 40 min 1% B. The detection was monitored at 520 nm. The malvidin-3-O-glucoside calibration curve was used. Results were expressed as mg malvidin-3-O-glucoside equivalent (M3GE)/g of DW.

#### *2.5. Determination of Total Phenolic Content*

Skin and seed grape extracts (dry weight) were solubilized in methanol at a concentration of 5 g/L. The same solution was used to determine the total phenolic content (TPC) and total antioxidant capacity (TAC) assays. To measure the absorbance, an Agilent Cary 60 UV-Vis spectrophotometer (Santa Clara, CA, USA) connected to the Cary win UV software was used.

The Folin-Ciocalteu method was used to determine the total phenolic content (TPC) [3,13,37]. Twenty μL of the diluted extract (see Section 2.5) and 100 μL of Folin-Ciocalteu reagent were added to 1.58 mL of water. After 30 s, 300 μL of sodium carbonate solution (20%) were added; the reaction mixture was thoroughly shaken and left for 120 min in the dark at room temperature (20 ◦C). Then, the absorbance was measured at 700 nm against the blank (sample was replaced by the methanol). The gallic acid calibration curve was used to determine the concentration of phenolic compounds in samples. The results were expressed as mg gallic acid equivalent (GAE)/g DW.

#### *2.6. Determination of Antioxidant Capacities*

#### 2.6.1. Radical Scavenging Activity: DPPH• Assay

DPPH antioxidant capacity was determined according to a published protocol [38]. Fifty μL of diluted extract (see Section 2.5) was added to 1.95 mL of a DPPH (6 <sup>×</sup> 10−<sup>5</sup> M) methanolic solution. After 30 min of incubation in the dark at room temperature (20 ◦C), the absorbance was measured at 515 nm. The trolox calibration curve was used. The results were expressed as μmol TE/g DW.

#### 2.6.2. Radical Scavenging Activity: ABTS Assay

ABTS antioxidant capacity was determined according to [39]. To generate ABTS• radical, 20 mL of ABTS solution (7 mM) was added to 20 mL of a potassium persulfate solution (2.45 mM). The mixture was incubated at room temperature in the dark all night. The stock solution was diluted with water/ethanol (50/50 *v*/*v*) to an absorbance of 0.7 ± 0.02 at 734 nm. One hundred μL of diluted extract (see Section 2.5) was mixed with 1 mL of ABTS• solution. After 10 min, the absorbance was measured at 735 nm. The trolox calibration curve was used. Results were expressed as μmol TE/g DW.

#### 2.6.3. Ferric-Reducing Antioxidant Power: FRAP Assay

FRAP antioxidant capacity was determined according to reference [40]. Fifty μL of diluted extract (see Section 2.5) and 150 μL of distilled water were added to 1.5 mL freshly prepared FRAP reagent (mixture of 10 volumes of a 300 mmol/L acetate buffer pH 3.6 with 1 volume of 10 mmol/L TPTZ in 40 mmol/L hydrochloric acid and 1 volume of 20 mmol/L ferric chloride). The solution was incubated at 37 ◦C for 4 min. Absorbance was measured at 593 nm. The FeSO4·7H2O calibration curve was used. Results were expressed as mmol Fe<sup>+</sup>2E/g DW.

#### 2.6.4. Electrochemical Apparatus and Measurements

Electrochemical measurements were carried out with potentiostat/galvanostat, Autolab PGSTAT 302N controlled by the Nova 2.1.3 software (Metrohm, Switzerland) in the personal computer (Supplementary Data, Figure S1). Tartaric acid buffer (3.3 mM tartaric acid adjusted with 1 M NaOH to obtain a pH 3.6) was used to prepare standard phenolic compounds solutions as well as diluted extracts (see Section 2.5) at appropriate concentrations (100 mg/L for skins and 20 mg/L for seeds). The scan rate was 100 mV/s.

#### **Disk Electrode**

Voltammetric measurements were carried out in a standard three-electrode electrochemical cell using an Ag/AgCl (KCl, 3 M) reference electrode, platinum counter electrode, and a glassy carbon electrode (GCE) of 3 mm diameter (Metrohm, Switzerland) as working electrode. Before each test, the working electrode surface was carefully polished with 3 μm alumina powder, then washed with purified water and cleaned for 5 min in an ultrasonic bath.

#### **Disposable Single-Walled Carbon Nanotubes Electrodes**

Single-walled carbon nanotubes electrodes (4 mm diameter, Dropsens, Spain) were also used in a three-electrode configuration comprising single-walled carbon nanotubes (SWCNTs-SPCE) with a silver reference electrode and carbon counter electrode. An aliquot of 200 μL of a solution of standard polyphenols or samples was cast onto the surface of the electrode, and the electrochemical measurements were performed immediately.

#### *2.7. Statistical Analysis*

The ANOVA and correlation tests were calculated by using the XLSTAT software (Addinsoft version 19.02, Paris, France). A Tukey test was carried out and where *p*-values < 0.05 was considered as significant. Pearson's correlation coefficient was carried out for the determination of correlations between the different antioxidant assays (spectrophotometric and electrochemistry) and between the antioxidant assays and phenolic composition (anthocyanins and flavanols).

#### **3. Results and Discussion**

#### *3.1. Flavanol and Anthocyanin Content of Skin and Seed Grape Extracts during Ripening*

The results of the evolution of total flavanols and anthocyanins content in skin and seed grape extracts are presented in Table 1.

**Table 1.** Phenolic composition of skin and seed grape extracts of the studied varieties at different stages of ripening.


Values represent means of triplicate determination ± SD. Different letters indicate the significant differences between stages according to Tukey's test, *p* < 0.05. DW: Dry Weight; MG3E: Malvidin-3-O-Glucoside Equivalent.

#### 3.1.1. Flavanols

#### **Skins**

For the three varieties, the highest flavanol content was determined at the green stage then it decreased significantly until maturity. It declined from 224 mg/g DW at the green stage to 19 mg/g DW at maturity in Tannat grape extracts. A similar evolution was shown in the literature [41,42]. On the opposite, an increase of flavanols content during ripening was also observed in other study [43].

#### **Seeds**

The highest content of flavanols was reached at close to veraison compared to the green stage and the maturity for all varieties. It increased from 424 mg/g DW at the green stage to 530 mg/g DW at close to veraison, then it declined significantly to 382 mg/g DW at maturity in seed Tannat grape extracts. This evolution was in accordance with a previous study [41]. The decline of flavanols content was explained by the oxidation of these compounds after veraison [44].

Flavanols are present in both skins and seeds at all stages of ripening with an abundance in seeds [27,45]. It has been shown that in Syrah skins at maturity the content was about 250 mg/g DW and about 455 mg/g DW in seeds [46]. There is an important variability in the literature concerning the phenolic composition content due to the extraction solutions, methods, and unit used to express results.

#### 3.1.2. Anthocyanins

The anthocyanin synthesis started at close to veraison and they accumulated until maturity in Merlot and Syrah skins, in Tannat skins, the anthocyanin synthesis started at veraison. The content increased from 2 mg M3GE/g DW to 22 mg M3GE/g DW at maturity in skin Merlot extracts. A similar evolution was reported in the literature [41–43,47–50]. In the case of Syrah, the anthocyanins content decreased at maturity from 28 to 14 mg M3GE/g DW, this decline may be due to the degradation of anthocyanins by the peroxidases and glycosidases present in skins [47].

Anthocyanins, the pigmented compounds, are present only in skin red grapes. As flavanols, the anthocyanins content differs considerably in the literature. It increases from 1.80 to 3.81 mg/g DW in Tannat skins [51] and it is about 86.68 mg/g DW at maturity in another study [16]. As mentioned previously, the anthocyanins content is also greatly affected by weather, climatic conditions, soil conditions, cultivars, irrigation [49], temperature, and light [52].

#### *3.2. Electrochemical Behavior of Polyphenol Standards and Skin and Seed Extracts for Various Cultivars at Di*ff*erent Stages of Ripening*

#### 3.2.1. Electrochemical Behavior of Standard Polyphenols

Cyclic voltammograms of polyphenol standards in tartaric acid buffer (pH 3.6) at glassy carbon electrode (GCE) in a potential range from 0 to 1100 mV (vs. Ag/AgCl-KCl 3M) and at single-walled nanotubes (SWCNT) in a potential range from 0 to 800 mV (vs. Ag) are illustrated in Figure 1 and peak potentials are given in Table 2. For caffeic acid, only one anodic peak was present. This peak corresponds to the oxidation of the *ortho*-diphenols to form the corresponding o-quinone. The potential values for the concentration 0.1 mM are 445 mV (vs. Ag/AgCl-KCl 3M) for GCE and 139 mV (vs. Ag) for SWCNT. Two peaks were observed for catechin and gallic acid at 0.1 mM. With GCE (vs. Ag/AgCl-KCl 3M), the voltage values were 483/826 mV for gallic acid and 472/766 mV for catechin. With SWCNT (vs. Ag), the voltage values were 132/468 mV and 122/465 mV for catechin and gallic acid, respectively. For both catechin and gallic acid, the first anodic peak correspond to the oxidation of the hydroxyl groups on the B-ring to quinone [53]. This oxidation was reversible generating cathodic peak in the negative scan for caffeic acid and catechin. The second peak corresponds to the oxidation of the hydroxyl group on the C-ring of catechin and can also correspond to the oxidation of the third phenol group adjacent to the ortho-diphenol group in gallic acid which is in agreement with previous results [54]. Other phenolic standards characterized corresponding to the anthocyanins and the flavonols are present mostly in skin grapes. Oenin chloride and rutin at 0.1 mM presented two anodic peaks at 377/669 mV and at 201/460 mV with SWCNT (vs. Ag), respectively, and at 652/987 mV and at 260/898 mV with GCE (vs. Ag/AgCl-KCl 3M), respectively (Figure 1).

The classification obtained considering only the first peak potential for the studied standards at the same concentration (0.1 mM) by increasing potential was: Gallic acid 122 (mV) < catechin (132 mV) < caffeic acid (139 mV) < rutin (201 mV) < oenin chloride (377 mV) was found [55] since catechin, caffeic, and gallic acid oxidized at lower potential.

**Figure 1.** Cyclic voltammograms of catechin with SWCNT (**A**) and GCE (**B**), caffeic acid with SWCNT (**C**) and GCE (**D**), gallic acid with SWCNT (**E**) and GCE (**F**), oenin chloride with SWCNT(**G**) and GCE (**H**), rutin with SWCNT (**I**) and with GCE (**J**) at a concentration of 0.1 mM (blank subtracted). GCE: Glassy Carbon Electrode; SWCNT-SPCE: Single Walled Carbon Nanotubes modified Screen Printed Carbon Electrodes.


**Table 2.** Voltammetric behavior of the studied standard polyphenols in tartaric acid buffer (pH 3.6) with SWCNT-SPCE and GCE for a concentration of 0.1 mM.

GCE: Glassy Carbon Electrode; SWCNT-SPCE: Single Walled Carbon Nanotubes modified Screen Printed Carbon Electrodes.

#### 3.2.2. Electrochemical Characterization of Skins and Seeds

Voltammetric measurements were performed on the extracts of each variety. For all varieties, cyclic voltammograms had three anodic peaks at different potentials depending on grape part (skins or seeds) (Figure 2) and the ripening stage. Syrah grape seed extracts were studied with both types of electrodes, as with SWCNT, three anodic peaks were also obtained with GCE (Figure 2).

**Figure 2.** Cyclic voltammograms of skin (**A**–**C**) and seed grape extracts (**D**–**F**) with SWCNT and those of Syrah seed grape extract (**G**) with GCE at different stages of ripening (blank subtracted). GS: Green Stage; CV: Close to Veraison; V: Veraison; M: Maturity.

For skin Merlot grape extracts, the first anodic peak was measured at 137, 134, 159, and 157 mV at the green stage, close to veraison, veraison, and maturity, respectively. This peak corresponds to the more oxidizable compounds that oxidized at a low potential as catechin-type flavonoids, including larger oligomeric and polymeric molecules, gallic acid, caffeic acid, and flavonols. The second anodic peak appeared at 391, 383, 363, and 370 mV at the green stage, close to veraison, veraison, and maturity, respectively. This peak may result from the oxidation of malvidin anthocyanins and stilbene derivatives overlapped with the second oxidation of the catechin flavonoids [56]. The third peak close to 600 mV corresponds to the phenolic acids such as vanillic and *para*-coumaric acid or the second oxidation of malvidin anthocyanins [54]. The same behavior was observed for the two other varieties.

In grape seed extracts, the first anodic peak was obtained at the same potential in all stages of ripening, it was around 130 mV. For Syrah grape extracts for example, the first peak appeared at 136 mV, 127, 129, and 126 mV at the green stage, close to veraison, veraison, and maturity, respectively. The second peak followed the same trend of the first potential with the following potential values for Syrah at 396, 438, 409, and 391 mV for the different stages of ripening. This peak could be attributed to the oxidation of the hydroxyl group on the C-ring of catechin derivatives. The third anodic peak corresponds to the higher oxidation potential compound which produces a peak at around 600 mV [57].

*3.3. Total Phenolic Content and Total Antioxidant Capacity by Spectrophotometric and Electrochemical Assays*

#### 3.3.1. Total Phenolic Content and Total Antioxidant Capacity by Spectrophotometric Assays

The total phenolic content and the antioxidant capacity of skin and seed grape extracts during ripening were determined using different spectrophotometric methods: Folin-Ciocalteu, DPPH, ABTS as well as FRAP assays, respectively. The results were summarized in Table 3.

#### **Skins**

The highest total phenolic content was detected at the green stage of ripening then it decreased significantly at maturity in the three varieties. For example, in Syrah grape extracts, the total phenolic content (TPC) was 212 mg GAE/g DW at the green stage then declined to 63 mg GAE/g DW at maturity.

The antioxidant capacities were measured using a single electron transfer (DPPH, ABTS, and FRAP). The highest total antioxidant capacity (TAC) was found in the green stage compared with maturity. The same evolution was obtained with the three assays on the three varieties. For example, in the skin Syrah grape extract, DPPH values decreased significantly from 853 at the green stage to 557 μmol TE/g DW at maturity, ABTS values from 843 to 357 μmol TE/g DW, and FRAP values from 2159 to 780 mmol Fe<sup>+</sup>2E/g DW.

#### **Seeds**

The TPC in seed grape extracts increased before veraison then decreased after veraison with the highest content at close to veraison for both seed Tannat and Syrah grape extracts, whereas for seed Merlot grape extract, the content decreased significantly from 867 at the green stage to 571 mg GAE/g DW at maturity.

The antioxidant capacity of seed grape extracts followed the same trend with the three antioxidant assays. The antioxidant capacity at close to veraison was higher than that found at the green stage and maturity. Among the samples tested, for seed Syrah grape extracts, the DPPH values increased from 2677 to 2915 μmol TE/g DW then decreased to 1991 μmol TE/g DW. ABTS values raised from 1171 to 1325 then declined to 590 μmol TE/g DW. FRAP values increased from 3979 to 5386 mmol Fe<sup>+</sup>2E/g DW then decreased to 3460 mmol Fe<sup>+</sup>2E/g DW.

Several methods were used to determine total phenolic content and antioxidant capacity of samples to take into account not only the composition of the extracts but also the mode of action and the specificity of the antioxidant [58,59]. Due to its ease of use, the Folin-Ciocalteu assay is the common used method to determine the TPC. The principle is the transfer of electrons from phenolic compounds to phosphomolybdic/phosphotungstic complexes [60]. The weakness of this method is the overestimation of the phenolic content due to the lack of specificity [55,60,61] which can react with other compounds particularly aromatic amines, ascorbic acids, and sugars [61]. In addition, the phenolic compounds react with the Folin-Ciocalteu reagent only under the basic conditions [61]. The three colorimetric methods used to determine the antioxidant capacity DPPH, ABTS, and FRAP

are considered as assays based on the electron transfer [58,61]. DPPH assay is an easy method widely used to determine the antioxidant capacity of natural extracts. The drawback of this method is the variation of reaction time with different phenolic compounds. Caffeic acid, for example, reacts quickly with DPPH whereas the catechin reacts slowly. The results obtained with this method differ depending on the time of readings (from 16 min to some hours) [55]. The FRAP assay is a simple, fast, and robust method used in the determination of the concentration of the most easily oxidized compounds [61]. It is based on the ability to reduce Fe3<sup>+</sup> to Fe2<sup>+</sup> quantified at 593 nm. Fe(III)/TPTZ reagent is more stable than DPPH• and gives results in shorter times [55]. The ABTS assay is based on the reduction by an antioxidant of the generated blue/green ABTS<sup>+</sup> [62]. DPPH and ABTS assays are the easiest to implement and yield the most reproducible results [58]. FRAP and DPPH methods are still used as they are the easy and accurate methods to measure the antioxidant activity [60].

The results of this work confirm that the total phenolic content in skins were lower than in seeds [13]. In skins, the highest antioxidant capacity was found at the green stage but a previous study [48] found the highest TAC at maturity. This difference may depend on the extraction method used but also on the protocol of the test. The total polyphenolic content increased when the berry weight decreased in accordance with previous studies [51].


**Table 3.** TPC and antioxidant capacities of skin and seed grape extracts of the three studied varieties at different stages of ripening by spectrophotometric methods.

Values represent means of triplicate determination ± SD. Different letters indicate the significant differences between stages according to Tukey's test, *p* < 0.05. TPC: Total Phenolic Content; DPPH: 1,1-diphenyl-2-picrylhydrazyl free radical; ABTS: 2,2 -azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)diammonium salt; FRAP: Ferric Reducing Antioxidant Potential; DW: Dry Weight; GAE: Gallic Acid Equivalent; TE: Trolox Equivalent; Fe<sup>+</sup>2E: Fe<sup>+</sup><sup>2</sup> Equivalent.

#### 3.3.2. Antioxidant Capacity by Electrochemical Method of Skin and Seed Grape Extracts

Different parameters shown in Table 4 allowed the estimation of the antioxidant capacity of extracts by cyclic voltammetry. The total charge Q800mV corresponds to all oxidizable phenolic compounds that will contribute to the total antioxidant capacity of the extract. Q240mv represents the electrochemically of the easily oxidizable polyphenols that have consequently the highest antioxidant capacity. Q520mv estimates the most antioxidant compounds which oxidize until 520 mV (until the second peak of the voltammogram). Q520mv-Q240mv corresponds to the compounds that have the lesser antioxidant capacity that oxidize until 520 mV. Finally, Q240mv/Q800mv ratio indicates the contribution of the most antioxidant compounds to the total antioxidant capacity of extract.

#### **Skins**

Q800mV, Q240mV, and Q520mV values presented the same evolution for all grape skin extracts. They declined from the green stage to maturity. For example, in Merlot, Q800mV values decreased from 262 to 118 μC/g DW, Q240mV from 44 to 22 μC/g DW, and the antioxidant capacity until 520 mV diminished from 153 to 75 μC/g DW. The contribution of the most antioxidant compounds to the total antioxidant capacity was also determined. It followed the same evolution of the other parameters except for Merlot grape extracts where the percentage increased from 17 to 22% then decreased to 19%.

#### **Seeds**

Electrochemical parameters of seed grape extracts have the same evolution in the three varieties. They raised from the green stage to close to veraison and veraison then declined until maturity. In Merlot, Q800mV values increased from 1232 to 1471 μQ/g DW then decreased to 1036 μC/g DW at maturity. The antioxidant capacity at 240 mV was about 358 μC/g DW at the green stage, 379 μC/g DW of extract at veraison, and 252 μC/g DW at maturity. Antioxidant capacity of seed extracts until 520 mV has the same trend than the other parameters, it stated from 905 μC/g DW at the green stage then increased to 944 μC/g DW at veraison, and declined to 639 μC/g DW at maturity. The most antioxidant compounds almost contribute with the same percent at all stages of ripening except for Merlot where the percent of Q240mV/Q800mV decreased from 30% to 24% at maturity.

Electrochemical parameters in both skins and seeds have the same trend than TPC, TAC values, and flavanol content. The higher TPC and TAC were found in seed grape extracts compared with skin grape extracts, in agreement with the literature [3,45,63]. The percent of Q240mV/Q800mV in seed grape extracts was more important than in skin grape extracts. It follows the same evolution of the other parameters in skins except for Merlot, in seeds there is no among differences between stages. At the charge Q240mV corresponding to the oxidation of flavanols, this result suggests the abundance of these compounds in seeds compared with skins.

#### *3.4. Correlation between TPC, Antioxidant Capacity, and Phenolic Composition*

Table 5 shows the Pearson correlation coefficients between TPC, electrochemical parameters, antioxidant assays, and phenolic composition for which: *r* < 0.39 weak correlation, 0.4 < *r* < 0.69 moderate correlation 0.7 < *r* < 0.89 strong correlation, and 0.9 < *r* < 1 very strong correlation [64].



*Antioxidants* **2020**, *9*, 800



In skins, a strong correlation was found between TPC and electrochemical parameters (*r* = 0.88 vs. Q240mV, *r* = 0.90 vs. Q520mV, and *r* = 0.84 vs. Q800mV). It was shown in the literature that TPC is significatively correlated with electrochemical responses [65], especially with cumulative response up to relatively high potentials [32]. In this study, TPC was better correlated with Q240mV than Q800mV. Colorimetric antioxidant assays (DPPH, ABTS, and FRAP) were strongly correlated with all electrochemical parameters. The best correlation was found with Q240mV (*r* = 0.81 vs. DPPH, *r* = 0.89 vs. ABTS, and *r* = 0.86 vs. FRAP) than with Q800mV (*r* = 0.69 vs. DPPH, *r* = 0.69 vs. ABTS, and *r* = 0.75 vs. FRAP). The methods used are well correlated because they are all based on electron transfer from antioxidant to oxidized compounds [32]. Flavanols content were well correlated with colorimetric assays (*r* = 0.93 vs. Folin-Ciocalteu, *r* = 0.86 vs. DPPH, *r* = 0.86 vs. ABTS, *r* = 0.94 vs. FRAP) as well as electrochemical parameters (*r* = 0.87 vs. Q240mV, *r* = 0.72 vs. Q800mV). The strong correlation between flavanols and Q240mV compared with Q800mV indicates that these compounds are the easiest antioxidant compounds that oxidized at a low potential (240 mV). A negative correlation between anthocyanins and the antioxidant tests have been shown, this result is an agreement with a previous study [59].

In seed grape extracts, the best correlation was found between flavanols and electrochemical parameters (*r* = 0.80 vs. Q240mV, *r* = 0.80 vs. Q520mV, and *r* = 0.64 vs. Q800mV) than with spectrophotometric methods (*r* = 0.67 vs. Folin, *r* = 0.66 vs. DPPH, *r* = 0.71 vs. ABTS, and *r* = 0.58 vs. FRAP). A strong correlation was observed between Folin-Ciocalteu, DPPH, and ABTS (*r* = 0.78 vs. DPPH and *r* = 0.77 vs. ABTS) whereas a moderate correlation was found between Folin-Ciocalteu and FRAP (*r* = 0.67). Contrary to skin grape extracts, in seed grape extracts, FRAP have the lowest correlation with all assays compared with the other colorimetric methods. This result illustrates the specificity of each assay and the variability of phenolic composition between skin and seed grape extracts.

The antioxidant capacity was mainly related to the TPC of extracts in accordance with previous results [9,12–14,58,62,66,67] and especially to the flavanols content [14]. The antioxidant capacity of polyphenols is mainly linked to their structures, compounds that have more than one aromatic ring, more than one hydroxyl groups in different positions are able to have a highest antioxidant capacity. This may explain the variability of Pearson correlation between the different methods and between skins and seeds.

#### **Principal Components Analysis (PCA)**

Figure 3 shows the Biplot graphic that represents the association of the phenolic composition with the antioxidant assays on skin and seed grapes extracts during ripening. The first two principal components explained 94.2% of the total variability. The first axis accounted for 88.6% and the second axis only for 5.6%. From the Biplot, skin grape extracts are separated in the left side from seed grape extracts in the right side.

For skin grape extracts, the stages of maturity were well separated depending mainly on the content of anthocyanins, flavanols, as well as antioxidant capacity, down the stages before veraison (have the highest flavanols content and antioxidant capacity) and up the stages from veraison to maturity (beginning of synthesis and accumulation of anthocyanins, low antioxidant capacity, and flavanols content). For seed grape extracts, it is more difficult to separate the different stages of maturity because the variables are very close.

Flavanols were compounds with the highest positive contribution to the antioxidant capacity, while the anthocyanins were the highest negative contribution in the three varieties studied. As it can be seen in Figure 3, the content of flavanols and the antioxidant capacity were higher in seed than in skin grape extracts.

**Figure 3.** Biplot of the first PCs obtained from PCA for seeds (SD) and skins (SK) of three grape varieties Merlot (M), Tannat (T), and Syrah (S) at different stages of ripening (green stage (GS), close to veraison (CV), veraison (V), and maturity (M)).

#### **4. Conclusions**

Total phenolic content, antioxidant capacity, flavanol, and anthocyanin contents of grape skin and seed extracts of three red grape varieties were studied at different stages of ripening. At all stages of ripening, the total phenolic content was higher in seed than in skin grape extracts. The green stage had the highest total phenolic content in grape skin extracts, whereas in grape seed extracts, they were the close to veraison and the veraison that had the highest content.

To measure the antioxidant capacity of extracts, different colorimetric methods were used (DPPH, ABTS, and FRAP) in addition to cyclic voltammetry. In skin grape extracts, the total antioxidant capacity was higher at the green stage than at maturity, in seed grape extracts, they were the close to veraison and the veraison that had the highest content with all assays. Generally, stages that had the highest phenolic content presented also the highest antioxidant capacity.

The correlation between electrochemical results using disposable electrodes and the colorimetric assays indicates that electrochemical assays can be considered as an alternative to these routine tests in the determination and the characterization of the antioxidant capacity in a short time.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-3921/9/9/800/s1. Table S1: Dates corresponding to the different stages of ripening for the three varieties Merlot, Tannat, and Syrah; Figure S1: The experimental electrochemical set up using GCE (A) and SWCNT (B) electrodes.

**Author Contributions:** Conceptualization, N.B. and F.G.; methodology, N.B. and F.G.; formal analysis, N.B.; investigation, N.B. and F.G.; resources, N.B. and C.S.; data curation, N.B.; writing—original draft preparation, N.B. and F.G.; writing—review and editing, N.B., F.G., C.S., and T.R.; visualization, F.G. and C.S.; supervision, F.G., C.S., and T.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported in part by a PhD grant (N.B.) from the government of Algeria (Ministère algérien de l'Enseignement Supérieur et de la Recherche Scientifique).

**Acknowledgments:** The authors would like to thank the Algerian government for financing this thesis. Marie Zerbib is also thanked for her experimental work assistance.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Comparative Studies on Di**ff**erent** *Citrus* **Cultivars: A Revaluation of Waste Mandarin Components**

**Giulia Costanzo 1, Maria Rosaria Iesce 2, Daniele Naviglio 2, Martina Ciaravolo 2, Ermenegilda Vitale <sup>1</sup> and Carmen Arena 1,\***


Received: 17 May 2020; Accepted: 10 June 2020; Published: 12 June 2020

**Abstract:** Peel, pulp and seed extracts of three mandarin varieties, namely Phlegraean mandarin (*Citrus reticulata*), Kumquat (*Citrus japonica*), and Clementine (*Citrus clementina*) were compared and characterised in terms of photosynthetic pigment content, total polyphenols amount, antioxidant activity and vitamin C to assess the amount of functional compounds for each cultivar. The highest polyphenols content was found in the Phlegraean mandarin, especially in peel and seeds, whereas Kumquat exhibited the highest polyphenols amount in the pulp. The antioxidant activity was higher in the peel of Phlegraean mandarin and clementine compared to Kumquat, which showed the highest value in the pulp. The antioxidant activity peaked in the seeds of Phlegraean mandarin. The vitamin C in the Phlegraean mandarin was the highest in all parts of the fruit, especially in the seeds. Total chlorophyll content was comparable in the peel of different cultivars, in the pulp the highest amount was found in clementine, whereas kumquat seeds showed the greatest values. As regards total carotenoids, peel and pulp of clementine exhibited higher values than the other two cultivars, whereas the kumquat seeds were the richest in carotenoids. Among the analysed cultivars Phlegraean mandarin may be considered the most promising as a source of polyphenols and antioxidants, compared to the clementine and Kumquat, especially for the functional molecules found in the seeds. Moreover, regardless of cultivars this study also highlights important properties in the parts of the fruit generally considered wastes.

**Keywords:** ascorbic acid; antioxidant activity; chlorophyll and carotenoid content; phenolic compounds

#### **1. Introduction**

In recent years, clinical trials and epidemiological studies have established an inverse correlation between the intake of fruits and vegetables and the occurrence of chronic diseases, the most prevalent causes of death in the world [1,2]. This protective effect has been ascribed to the antioxidant properties of different compounds, which coordinate and balance the body system to protect tissues and fluids from damage by reactive oxygen species (ROS) or free radicals [1,3,4].

Besides health and nutritional benefits, antioxidants have an important role for the food industry. These compounds prevent the propagation reaction of free radicals during the oxidative process preserving food quality and shelf life during handling and storage [5–7].

In general, citrus fruits are considered as one of the natural sources of antioxidants. In fact, they contain an appreciable amount of ascorbic acid, flavonoids, and phenolic compounds [8–11] and even some essential minerals important for human nutrition [12–14].

Mandarin is a product with many desirable characteristics for consumers who are health aware [15]. Continuous improvements in transportation logistics have allowed its widespread distribution to consumers throughout the world. These reasons have increased the world demand for mandarin cultivation so that its products are in continuous growth [16].

In food manufacturing, citrus is mainly used for producing fresh juice or citrus-based drinks, so a large amount of citrus wastes such as peels and seeds must be discarded. The global volume of citrus processed every year is about 31.2 million tons [17], 50%–60% of which represents waste called "pastazzo" [18,19]. The management of a such amount of wastes represents a critical issue for the citrus industry due to the high costs involved for its disposal [18,20].

This encourages the implementation of recycling policies to promote potential new and innovative uses of citrus by-products. Currently, several technological innovations have been developed to valorise citrus wastes in order to convert possible environmental risks into a valuable resource, thus reducing the environmental impact [19]. Citrus by-products find utilisation in biogas production [21], ruminant feeding [22], and essential oil extraction [23].

Moreover, a very interesting perspective would be also to utilise citrus by-products as a source of bioactive compounds for human diet [24,25]. In fact, recent studies suggest that citrus waste could be used as natural sources of antioxidants [26,27]. However, it is well known that the chemical composition of fruits may be subjected to variations according to climate, cultivation practices, soil type, cultivar, fruit maturity, and even between parts of the same fruit [1]. In addition to the expected changes of fruit quality, variations in antioxidant properties during ripening have also been described [28,29]. For a revaluation of citrus by-products, it would be appropriate to focus the attention on varieties with very different provenance and traits, in order to have an indication about the potential associated to the wastes of diverse cultivars.

Starting from these considerations, the aim of this study was to determine the amount of specific functional compounds such as chlorophyll and carotenoid, total polyphenols, vitamin C as well as the antioxidant capacity in the pulp and more specifically in peel and seeds of three different cultivars of mandarin, namely Phlegraean mandarin (*Citrus reticulata* Blanco), Kumquat (*Citrus japonica*), and Clementine (*Citrus* × *clementina*). These cultivars have been selected for specific characteristics. In detail, clementine was chosen for its large demand and commercial consumption worldwide due mainly to ease of peeling and seedlessness, kumquat is very appreciated in food preparation but has a niche consumption, while the Phlegraean mandarin is specifically diffused only in a peculiar volcanic area of Southern Italy (Campi Flegrei, Naples, Italy). The mandarin cultivar from the Phlegraean fields, planted for the first time in Naples in the 19th Century, is considered a traditional product from Campania region. The mandarin and the liqueur derived from it, namely "mandarinetto", have been included by the Italian Ministry of Agriculture, Food and Forestry in the list of typical, traditional products of the Campania region (GU 168/2015). This citrus fruit variety has shown its maximum expression in the Phlegraean fields, a peculiar area surrounding the supervolcano, Vesuvius; the fertile soil, typical of a volcanic area as well as the proximity of the sea and the mild climate make the Phlegrean land unique and particularly favourable to the agriculture.

The outcomes of this study will be useful for a valorisation of mandarin waste products, encouraging their use and thus favouring the recycling practices and the bioeconomy strategies.

#### **2. Materials and Methods**

#### *2.1. Plant Material and Sample Preparation*

Three cultivars of mandarins (Phlegraean mandarin, *Citrus reticulata*; Kumquat, *Citrus japonica*; and Clementina, *Citrus clementina*) collected in the seasons 2019–2020 were used in this study. Fruits, used for the experiments, were homogeneously collected from five selected mandarin trees, for each cultivars. Phlegrean mandarin fruits were collected in the area of Bacoli, Phlegraean fields, (Naples, Southern Italy); mandarins from Clementine and Kumquat cultivars were sampled in private plantations

at Sorrento peninsula (Naples, Southern Italy). Sampling was carried out during the harvest period typical for each cultivar, generally from November to January. The samples were placed in plastic bags and stored in dry ice, then they were transferred immediately to the laboratory and stored at −80 ◦C for subsequent analysis. Analyses were performed on each component of the fruit: peel, pulp and seeds for all the collected samples.

The fruits were washed with tap water, separated into the three components and homogenised, using mortar and pestle, by preventively powdering in liquid nitrogen. Powdered citrus tissues, obtained for each component of the fruit, were placed in test tubes and stored at −20 ◦C until analysis. At least three different trees for each mandarin cultivar were chosen to collect samples. For each mandarin cultivar a total of ten samples were analysed. Mandarins of two collections (seasons 2019 and 2020) for all cultivars were analysed in this study.

#### *2.2. Photosynthetic Pigments Content Determination*

Total chlorophylls and carotenoids contained in peel, pulp and seeds were determined according to Lichtenthaler (1987) [30]. Briefly, pigments were extracted from 0.25 g of powered sample in ice-cold 100% acetone and centrifuged (Labofuge GL, Heraeus Sepatech, Hanau, Germany) at 5000 rpm for 5 min. The absorbance was measured by spectrophotometer (Cary 100 UV-VIS, Agilent Technologies, Santa Clara, CA, USA) at wavelenghts of 470, 645 and 662 nm and pigment concentration expressed as mg g−<sup>1</sup> fresh weight (FW).

#### *2.3. Total Polyphenol Content*

Total polyphenol content was measured according to the reported procedure [31]. Briefly, 0.25 g of powered sample was extracted with aqueous 80% methanol, at 4 ◦C (for 30 min) and then centrifuged at 11,000 rpm for 5 min. Extracts were combined with 1:1 (*v*/*v*) 10% Folin–Ciocalteu phenol reagent and water. After 3 min, 700 mM Na2CO3 solution was added to the resulting mixture in 5:1 (*v*/*v*). Samples were incubated for 2 h in darkness. Then, the absorbance at 765 nm was measured with a spectrophotometer (UV-VIS Cary 100, Agilent Technologies, Palo Alto, CA, USA). Gallic acid was used as a standard. Calibration curve was constructed analysing standard solutions in the interval of concentration 5–500 ppm. The total polyphenols concentration was calculated and expressed as gallic acid equivalents (mg GAE g−<sup>1</sup> FW) from the calibration curve (*R*<sup>2</sup> = 0.996) using gallic acid.

#### *2.4. Determination of Antioxidant Capacity and Ascorbic Acid*

The antioxidant activity of the cultivars was evaluated by the Ferric Reducing Antioxidant Power (FRAP) assay, according to the reported method [32]. Briefly, 0.25 g of powered sample was mixed with 60:40 (*v*/*v*) methanol/water solution and centrifuged at 14,000 rpm for 15 min (4 ◦C). FRAP reagents (300 mM Acetate Buffer pH 3.6; 10 mM tripyridyltriazine (TPTZ), 40 mM HCl and 12 mM FeCl3) were added to the extracts of each sample in 16.6:1.6:1.6 (*v*/*v*), respectively. After 1 h in darkness, the absorbance at 593 nm was measured with a spectrophotometer (UV-VIS Cary 100, Agilent Technologies, Palo Alto, CA, USA). Trolox (6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid) was used as the standard and total antioxidant capacity was quantified and expressed as mmol Trolox equivalents (μmol Trolox Eq. mg−<sup>1</sup> FW).

The ascorbic acid (AsA) content was determined using the Ascorbic Acid Assay Kit (MAK074, Sigma-Aldrich, St. Louis, MO, USA), following the reported procedure [33]. Briefly, 10 mg of sample was homogenized in 4 volumes of cold AsA buffer, and then centrifuged at 13,000 rpm for 10 min at 4 ◦C to remove insoluble material. The liquid fraction was mixed with AsA assay buffer to a final volume of 120 μL. The assay reaction was performed by adding the kit reagents to the samples.

In this assay, the AsA concentration was determined by a coupled enzyme reaction, which develops a colorimetric (570 nm) product, proportionate to the amount of ascorbic acid contained in the sample. The concentration of ascorbic acid in the samples was referred to a standard curve and expressed in mg L<sup>−</sup>1.

#### *2.5. Statistical Analysis*

Statistical analysis was performed using Sigma Plot 12.0 (Jandel Scientific, San Rafael, CA, USA). Statistically significant differences among varieties were checked by one-way ANOVA followed by Holm Sidak test for multiple comparison tests, based on a significance level of *p* < 0.05. The normal distribution of data was verified by Shapiro–Wilk and Kolmogorov–Smirnov tests. Spearman correlation coefficient was used to test associations between variables. All data were expressed as means ± standard error (SE) (*n* = 6).

#### **3. Results**

#### *3.1. Total Polyphenol Content*

Total phenolic content of the peel, pulp and seeds extracts strongly varied among *Citrus* varieties and components of the fruit (Figure 1). Highest amounts (*p* < 0.05) of total polyphenols were found in *C. reticulata* peel (2.21 <sup>±</sup> 0.19 mg GAE g−<sup>1</sup> FW), followed by *C. japonica* (1.24 <sup>±</sup> 0.013 mg GAE g−<sup>1</sup> FW) and *C. clementina* (0.24 <sup>±</sup> 0.011 mg GAE g−<sup>1</sup> FW). No statistically significant difference was detected in the total polyphenol content for the pulp of all the examined Citrus varieties.

**Figure 1.** Total polyphenols in peel, pulp and seeds of the three different mandarin cultivars: *C. reticulata* (Phlegrean mandarin), *C. japonica* (Kumquat) and *C. clementina* (Clementine). Each bar represents the mean ± SE (*n* = 6). Different letters indicate statistically significant differences among mandarin varieties (*p* < 0.05). Results were analysed by one-way ANOVA followed by Holm–Sidak post hoc test for multiple comparisons. GAE: gallic acid equivalents; FW: fresh weight.

Among the fruit components, seeds of the *C. reticulata* cultivar exhibited the highest (*p* < 0.001) total polyphenol content (5.43 <sup>±</sup> 0.04 mg GAE g−<sup>1</sup> FW); this concentration was eight-fold higher than that measured in *C. japonica* seed extracts (0.65 <sup>±</sup> 0.011 mg GAE g−<sup>1</sup> FW).

#### *3.2. Total Soluble Antioxidant Capacity*

The total antioxidant capacity of the different cultivars is shown in Figure 2 for different parts of mandarin fruit: peel, seeds and pulp. The peel extracts of *C. reticulata* cultivar exhibited the highest (*<sup>p</sup>* <sup>&</sup>lt; 0.01) total antioxidant capacity (14.69 <sup>±</sup> 1.80 mmol eq Trolox mg−<sup>1</sup> FW) compared to *C. clementina* (7.5 <sup>±</sup> 1.0 mmol Trolox eq mg−<sup>1</sup> FW) and *C. japonica*, which showed the lowest antioxidant capacity (2.1 <sup>±</sup> 0.2 mmol Trolox eq mg−<sup>1</sup> FW).

**Figure 2.** Total antioxidant capacity of peel, pulp and seeds in the three different mandarin cultivars: *C. reticulata* (Phlegrean mandarin), *C. japonica* (Kumquat) and *C. clementina* (Clementine). Each bar represents the mean ± SE (*n* = 6). Different letters indicate statistically significant differences among mandarin varieties (*p* < 0.05). Results were analysed by one-way ANOVA followed by Holm–Sidak post hoc test for multiple comparisons.

In the *C. clementina* pulp extracts, the total antioxidant capacity was found to be three-fold higher (7.1 <sup>±</sup> 1.27 mmol Trolox eq mg−<sup>1</sup> FW) (*<sup>p</sup>* <sup>&</sup>lt; 0.01) than that measured in *C. reticulata* (2.6 <sup>±</sup> 0.5 mmol Trolox eq mg−<sup>1</sup> FW) and almost six-fold higher compared to *C. japonica* (1.2 <sup>±</sup> 0.06 mmol Trolox eq mg−<sup>1</sup> FW).

As observed for total polyphenols, among the different fruit parts, the seeds of *C. reticulata* showed the highest (*p* < 0.001) antioxidant capacity (55.6 <sup>±</sup> 2.6 mmol Trolox eq mg−<sup>1</sup> FW). This value was seventeen-fold more abundant compared to *C. japonica* (3.2 <sup>±</sup> 0.15 mmol Trolox eq mg−<sup>1</sup> FW).

#### *3.3. Total Chlorophyll and Carotenoid Content*

The concentration of pigments varied significantly among both cultivars and fruit components. Regarding peel, no difference in total chlorophyll content was detected among cultivars (Figure 3). Conversely, the total carotenoid content increased significantly (*p* < 0.01) in *C. clementina* peel (128.16 <sup>±</sup> 3.03 mg g−<sup>1</sup> FW) compared *C. reticulata* (70.99 <sup>±</sup> 3.01 mg g−<sup>1</sup> FW) and *C. japonica* (42.97 <sup>±</sup> 1.72 mg g−<sup>1</sup> FW) (Figure 3).

In the pulp of *C. clementina* cultivar the total chlorophyll content was very high (*p* < 0.01) (37.13 <sup>±</sup> 0.67 mg g−<sup>1</sup> FW) compared to values observed for *C. japonica* (2.93 <sup>±</sup> 0.23 mg g−<sup>1</sup> FW) and *C. reticulata* (0.73 <sup>±</sup> 0.07 mg g−<sup>1</sup> FW). The same trend was observed for carotenoid content in the pulp (Figure 3).

The total amount of chlorophyll increases significantly (*p* < 0.001) in *C. japonica* seeds (94.01 <sup>±</sup> 7.92 mg g−<sup>1</sup> FW), compared to *C. reticulata* (25.05 <sup>±</sup> 1.78 mg g−<sup>1</sup> FW). The highest concentration (*p* < 0.01) of carotenoids was detected in *C. japonica* seeds (Figure 3).

**Figure 3.** Total chlorophyll (a + b) and total carotenoid (x + c) content of peel, pulp and seeds in the three different mandarin cultivars: *C. reticulata* (Phlegrean mandarin), *C. japonica* (Kumquat) and *C. clementina* (Clementine). Each bar represents the mean ± SE (*n* = 6). Different letters indicate statistically significant differences among mandarin varieties (*p* < 0.05). Results were analysed by one-way ANOVA followed by Holm–Sidak post hoc test for multiple comparisons.

#### *3.4. Ascorbic Acid*

Ascorbic acid (AsA) content was very different among mandarin cultivars and fruit components. The highest values (*p* < 0.01) were found in peel, pulp and seed extracts of *C. reticulata* (Figure 4).

**Figure 4.** Total soluble antioxidant capacity expressed as ascorbic acid equivalents in peel, pulp and seeds in the three different tangerine cultivars: *C. reticulata* (Phlegrean mandarin), *C. japonica* (Kumquat) and *C. clementina* (Clementine). Each bar represents the mean ± SE (*n* = 6). Different letters indicate statistically significant differences among mandarin varieties (*p* < 0.05). Results were analysed by one-way ANOVA followed by Holm–Sidak post hoc test for multiple comparisons.

In particular, in the peel of *C. reticulata* the ascorbic acid concentration was higher (*p* < 0.01) (13.32 <sup>±</sup> 0.96 mg L<sup>−</sup>1) compared to that measured in *C. clementina* (2.75 <sup>±</sup> 0.10 mg L<sup>−</sup>1) and *C. japonica* (2.13 <sup>±</sup> 0.16 mg L−1). In pulp the trend was the same with the highest (*p* < 0.01) AsA concentration in extracts of *C. reticulata* (7.72 <sup>±</sup> 0.97 mg L−1) followed by *C. clementina* (4.37 <sup>±</sup> 0.12 mg L−1) and *C. japonica* (1.39 <sup>±</sup> 0.07 mg L<sup>−</sup>1) cultivars.

Finally, in the seeds of *C. reticulata*, the AsA amount was almost twice higher (*p* < 0.01) than that found in *C. japonica* (20.61 <sup>±</sup> 0.57 mg L−<sup>1</sup> and 12.78 <sup>±</sup> 0.45 mg L<sup>−</sup>1, respectively).

#### *3.5. Relationship among Bioactive Compounds in the Seeds*

The correlations among bioactive compounds in seeds are reported in Table 1. *C. reticulata* showed a significant positive correlation between antioxidant capacity and total chlorophyll, antioxidant capacity and carotenoids as well as between total chlorophylls and carotenoids. *C. japonica* exhibited a significant negative correlation between antioxidant capacity and ascorbic acid and a significant positive correlation between chlorophylls and carotenoids.

**Table 1.** Pearson correlations among bioactive compounds in the seeds (\* *p* < 0.05; \*\* *p* < 0.01); total polyphenols—TP; antioxidant capacity—AC; total chlorophyll—Chl (a + b); total carotenoids— Car (x + c); ascorbic acid—AsA.


#### **4. Discussion**

In this study, a comparison between different Citrus varieties was carried out to quantify the content of some bioactive compounds in the peel, pulp and seeds. Data collected allowed us to screen the most promising Citrus cultivar in terms of nutraceutical compounds and to valorise the parts of the fruit, such as peel and seeds, generally considered useless by-products of the production chain. Among cultivars, compelling experimental evidence emerges mainly from *C. reticulata* cultivar. The Phlegraean mandarin is a typical product of the volcanic Phlegraean area (Southern Italy, Naples), characterized by mild climate conditions and very fertile soils. This cultivar is not widespread outside the Campania region boundaries, obscured by the most famous Clementine and Kumquat varieties, with a broader national market.

From our data, it is evident that all fruit parts of the investigated citrus species showed strong antioxidant properties, especially peels and seeds, considered waste products, with limited or even without market value for the food industry [18,19].

Consistent with findings of other authors [34], the total pigment content was very high in the peel compared to the pulp in all tested varieties of mandarin, confirming the valuable antioxidant properties of this fruit part. Among cultivars, *C. clementina* showed the highest pigment content in both peel and pulp.

It is interesting to note that in seeds, *C. japonica* presented the highest content of carotenoids and chlorophylls compared to *C. reticulata*. This difference may be due to intrinsic characteristics of the species. It has been demonstrated that *Citrus* fruits are complex sources of pigments, especially carotenoids, with a broad diversity among the different species and cultivars in terms of types and

amounts [35–37]. Therefore, most *Citrus* species show the same carotenoid profile, although some of them have higher concentrations [38].

Generally, the carotenoids are absent in seeds compared to peel and pulp, conversely to the concentration of chlorophylls. The presence of chlorophyll in the seeds of both*C. reticulata* and*C. japonica* cultivars is due to the chlorophyllous embryo. Especially in Kumquat, the cotyledon primordia were particularly evident under the outer seed integument (personal observation). The amount of chlorophyll in seeds is a valuable attribute because the chlorophyll presence ensures the activation of light capture mechanisms, as soon as germination begins [39]. Moreover, there are also some references about fruit photosynthesis: this process may be affected by light able to penetrate in fruit, temperature and ontogeny [40]. It is interesting to consider that the occurrence of a slight level of photosynthesis in mandarin fruit could determine the enrichment of the fruit biochemical profile.

Differently from chlorophylls, the presence of carotenoids in seeds and non-green tissues is common in many species such as maize, pumpkins, sunflowers [41].

The occurrence of carotenoids in plant tissues is associated with a protective function. They develop more abundantly in seedlings at a later time, to enlarge the sunlight harvesting and to defend the photosynthetic apparatus from damages due to excess of light. The carotenoids contribute, as antioxidants, to contrast the deterioration of the membranes induced by free radicals and ageing [42,43].

The function of carotenoids in the seed is less clear than in other plant tissues; however, previous researches have demonstrated that carotenoid presence in the grain is essential for the production of abscisic acid (ABA) and the induction of seed dormancy [44]. The dormancy is one of the mechanisms by which plants can delay germination when the environmental conditions are unfavourable to sprouting [45]. Generally, the number of plants with dormant seeds increases with increasing distance from the equator, in response to seasonality and habitat diversity [46]. It may be hypothesised that the higher content of carotenoids in the *C. japonica* seeds serves for higher production of ABA, in response to the plant's need to maintain seeds quiescent for a longer time compared to *C. reticulata*. On the other hand, *C. reticulata* does not require a quiescent strategy for seeds because it is particularly diffused in Mediterranean ecosystems where the mild climate favours germination conditions.

Furthermore, carotenoids in seeds also contribute to the antioxidant defence of embryonal tissues limiting the membrane deterioration due to free radical and ROS production during seed ageing [42,43]. The presence of carotenoids ensures healthy and long-lived mandarin seeds and together with chlorophylls contributes to improve the fruit value, as chlorophyll and carotenoids are important antioxidant compounds in the human diet [47].

The most relevant result of our study regards the antioxidant power found in mandarin by-products. Our data demonstrate that considerable amounts of polyphenols, water-soluble antioxidants and ascorbic acid were found in the peel and extremely high concentrations of these compounds were measured in seeds, especially in *C. reticulata* cultivar. The peels, and even more so the seeds, are considered wastes of the citrus industry and currently to our knowledge, they are not used for the human diet, but only for ruminant nutrition [22].

The peel is generally the tissue richer in polyphenols and antioxidant activity than other fruit parts [26], even if the antioxidant properties are strictly related to the species. Our results evidence an interesting exception for the *C. reticulata* cultivar, where the highest concentration of total polyphenols and the most elevated antioxidant capacity was found in the seeds. According to other authors [48], we assume the elevated content of compounds with antioxidant action, located in particular in peel and seeds, represents a defence mechanism of the fruit and the embryo against external agents. Indeed, it is well known that these compounds are involved in the protection of the fruit and the embryo against herbivores and fungal pathogen attacks [49].

For the highest cumulative capacity to scavenge free radicals compared to peel and pulp, the seeds of *C. reticulata* are worthy of particular attention. The concurrent increase in total polyphenol and ascorbic acid contents suggests that these molecules may, in part, contribute to the seed antioxidant power in this species. The *C. japonica* seeds exhibited different nutraceutical traits compared to *C. reticulata*, showing a lower antioxidant capacity and polyphenol and ascorbic acid content, but a higher amount of chlorophylls and carotenoids. *C. reticulata* seeds also showed a positive relationship among chlorophylls, carotenoids and antioxidant capacity, while *C. japonica* revealed a positive relationship between chlorophylls and carotenoids but a negative relationship between ascorbic acid and antioxidant capacity, indicating that ascorbic acid does not contribute so much to the antioxidant power.

The absence of relationships among ascorbic acid and antioxidant capacity in the peel and pulp of mandarin cultivars suggests that other compounds should be responsible for the antioxidant power of these fruit components [27].

Several studies report that various phenolic compounds, compared to ascorbic acid, mainly influence antioxidant activity [50,51]. However, according to the literature [6], it may be argued that the phenolic compounds in citrus fruits contribute less than vitamin C to the antioxidant activity. This evidence seems to be confirmed for the seeds of the *C. reticulata* cultivar, where we have found a very high content of vitamin C. However, it cannot be excluded that other antioxidant compounds, such as flavonoids may be present in seeds enhancing the antioxidant power. In addition, other factors may contribute to this distinctive trait in Phlegrean mandarin, such as the plant age or the stage of fruit ripeness [52]. It is noteworthy that antioxidant properties found in the different parts of mandarin fruit are strictly related to intrinsic species characteristics. In fact, conversely to Phlegrean mandarin, in other citrus fruits such as orange, lemon and grapefruit, the higher amount of phenolic compounds, flavonoids, vitamin C, and antioxidant activity, were found in the peels compared to the inner wasted parts (pulp and seeds) [26]. Starting from this evidence, the high antioxidant charge of Phlegrean mandarin seeds assumes a great commercial value for diverse purposes. Lyophilised or fresh seeds extracts might be used as food additives or included in pharmaceutical formulations as promising alternatives to the common preparations [53]. The current challenge is to improve the extraction techniques of bioactive compounds from vegetable wastes to preserve the properties of these molecules over time and obtain a full by-product valorisation.

#### **5. Conclusions**

The comparison among diverse mandarin cultivars and fruit components revealed some important insights about the nutraceutical value associated with species and fruit parts. For all tested cultivars the peel is more abundant in antioxidants compared to pulp and may be potentially used as a dietary supplement. However, the most significant result concerns the seeds of the Phlegraean mandarin, which have proved to be highly rich in polyphenols, ascorbic acid, and antioxidant activity, compared to the other parts of the fruit and other citrus varieties.

Seeds could be inexpensive and readily available resources of bioactive compounds (such as natural antioxidants) for use in the food and pharmaceutical industries. The seed consumption would also reduce the problem of large amounts of wastes derived annually from the agri-food industry.

A sustainable re-utilisation of seeds and peels for industrial and pharmacological applications could represent a strong boost toward circular economy initiatives in Southern Italy.

Further analyses are needed to improve this initial research, allowing a deepen characterisation of bioactive molecules responsible for the high antioxidant power of the seeds.

From the economic point of view, the evidence that the Phlegraean mandarin is richer in bioactive compounds than the most commercialised varieties (i.e., kumquat and clementine) promotes the valorisation of a potentially unexploited resource typical of the Campania region (Southern Italy), whose economy is mainly based on the tourism and agriculture.

**Author Contributions:** Conceptualization, C.A., M.R.I.; methodology, C.A., G.C.; investigation, C.A., G.C., E.V., M.C.; resources, M.R.I., D.N.; data curation, E.V.; writing—original draft preparation, C.A., G.C.; writing—review and editing, C.A., M.R.I, D.N.; supervision, C.A., M.R.I. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


of chemicals, materials and fuels. Current situation and global perspective. *Energy Environ. Sci.* **2013**, *6*, 426–464. [CrossRef]


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

### **Electroactive Phenolic Contributors and Antioxidant Capacity of Flesh and Peel of 11 Apple Cultivars Measured by Cyclic Voltammetry and HPLC–DAD–MS**/**MS**

#### **Danuta Zieli ´nska \* and Marcin Turemko**

Department of Chemistry, University of Warmia and Mazury in Olsztyn, Plac Lodzki 4, 10-727 Olsztyn, Poland; marcin.turemko@uwm.edu.pl

**\*** Correspondence: danuta.zielinska@uwm.edu.pl; Tel.: +48-89-523-39-35

Received: 29 August 2020; Accepted: 26 October 2020; Published: 28 October 2020

**Abstract:** In this study, 11 apple cultivars were characterized by their total phenolic content (TPC) and total flavonoid content (TFC) and antioxidant, reducing, and chelating capacity by 2,2-diphenyl-1-picrylhydrazyl (DPPH) test, cyclic voltammetry (CV), and ferric reducing antioxidant power (FRAP) assays; and ferrous ion chelating capacity. The phenolic compounds in flesh and peel were determined by liquid chromatography coupled to mass spectrometry and diode array detector (HPLC–DAD–MS/MS) and their electroactivity by CV. The results showed higher TPC, TFC, and antioxidant capacity by DPPH test in the peels of all apple cultivars as compared to the respective flesh. The peel extracts also showed two-fold higher FRAP values as compared to the flesh extracts. The reducing capacity of the peel and flesh determined by CV measurements confirmed the results achieved by spectrophotometric methods of evaluating antioxidant capacity. There was no significant difference in chelating capacity in the peel and flesh. The HPLC–DAD–MS/MS analysis showed the presence of 11 phenolic compounds in the peel and flesh which varied in antioxidant, reducing, and chelating activity. The order of the phenolic compound content in flesh and peel in Quinte cultivar, which showed the highest antioxidant capacity, was as follows: epicatechin > chlorogenic acid > quercetin 3-arabinoside > quercetin 3-glucoside > cyanidin 3-galactoside > quercetin 3-rhamnoside > catechin > phloridzin > rutin > phloretin = quercetin. CV results were highly correlated with those obtained by spectrophotometry and HPLC–DAD–MS/MS, providing evidence to support the use of cyclic voltammetry as a rapid method to determine the phenolic profile and reducing the power of apple flesh and peel. The association between antioxidant assays and phenolic compound content showed that the highest contribution to the antioxidant capacity of apple peel and flesh was provided by catechin, epicatechin, and cyadinin-3-galactoside, while phloretin, phloridzin, and chlorogenic acid were the main contributors to chelating activity. Results from this study clearly indicate that removing the peel from apples may induce a significant loss of antioxidants.

**Keywords:** apples; phenolic compounds; antioxidant; reducing and chelating capacity; cyclic voltammetry; HPLC–DAD–MS/MS

#### **1. Introduction**

Apple is a popularly consumed fruit, mostly because of the pleasant taste and the fact that it is cultivated worldwide. Apples are a significant part of the human diet and are ranked in the top five consumed fruits in the world [1]. The beneficial health effects of apples have been ascribed to the polyphenolic compounds, a group of secondary plant metabolites, of which several thousand structurally different compounds have been identified [2,3]. Phenolic compounds are generally recognized as the main determinants of the biological activities of apples, such as the prevention of cardiovascular diseases, asthma and other lung dysfunctions, diabetes, obesity, and cancer [4–7] as well as age-related neurodegeneration [8,9]. The potential health benefits of polyphenols have been reviewed by Scalbert et al. [10]. Moreover, the content and composition of polyphenols present in apples are important because of their contribution to the sensory quality of fresh fruit and processed apple products [11].

Apples contain a variety of phenolic compounds that can be classified into five major sub-classes, with procyanidins being the most abundant class (between 40% and 89%), followed by hydroxycinnamic acids, dihydrochalcones, flavonols, anthocyanins, and flavan-3-ols [12]. Anthocyanins that contribute to the red color of apple fruits are exclusively found in the peel and represent less than 8% of total phenolics [13–15].

The flavan-3-ols can be found in the form of monomers, oligomers, and polymers (procyanidins), and flavonols are often associated with sugar moieties. The main sugars involved in glycosylation are galactose, glucose, rhamnose, arabinose, and xylose, and rutinose, a disaccharide, has also been found in apple. Dihydrochalcones are mainly associated with glucose and xyloglucose [16]. Moreover, the distribution and profile of phenolic compounds vary considerably among different cultivars, genotypes, ripening stages, and environmental conditions, and also within different tissues [13,15,17–21].

The antioxidant activity of polyphenolics has been studied extensively [22–24]. These compounds usually have a high redox potential, which allows them to act as reducing agents, hydrogen donors, and singlet oxygen quenchers [22]. Therefore, several methods to measure the antioxidant activity of polyphenols have been proposed and were recently reviewed [23,24]. Among other methods, scavenging stable radicals such as 2,2-diphenyl-1-picrylhydrazyl (DPPH) and 2,2'-azinobis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS), oxygen radical absorbance capacity (ORAC), total radical trapping antioxidant parameter (TRAP), ferric reducing antioxidant power (FRAP), and cupric ion (Cu2<sup>+</sup>) reducing power (CUPRAC) were employed in foods [23]. The determination of antioxidant activity by electrochemical methods is of increasing interest. Electrochemical methods used to determine antioxidant activity are still being developed. Among electrochemical techniques, the most widely used for this purpose is cyclic voltammetry (CV). In contrast to the aforementioned methods, electrochemical assays are low-cost and usually do not require time-consuming sample preparation. CV is based on an analysis of the anodic current (AC) waveform, which is a function of the reactive potential of a given compound in the sample. The CV tracing indicates the ability of a compound to donate electrons at the potential of the anodic wave. Therefore, in the past couple of years, CV has proven to be highly practical and efficient in determining the phenolic composition and reducing the power of complex matrices, including fruit extracts, honey, wine, tea, coffee, and kiwifruit [25–29]. Methods involving cyclic voltammetry (CV) have also been suggested as an instrument in evaluating the reducing activity of a wide spectrum of bioactive compounds and food extracts [25,30,31].

In many works, the content and antioxidant properties of polyphenols present in all parts of the apple fruit (skin, pulp, and seeds) were determined for various cultivars [32–35]. However, these You jumped the numbers in between.studies mainly focused on the relationship between antioxidant activity and total phenolic content, while a limited amount of data were available on phenolic profiles and their contribution to the antioxidant activity of apple extracts. Additionally, the correlation between the different antioxidant activity evaluation assays, chelating activity, and contents of individual apple polyphenols has not yet been fully investigated. However, the feasibility of electrochemical methods in determining the antioxidant activity in the phenolic compounds of apple and extracts from the peel and flesh samples is yet to be studied.

In this study, the antioxidant properties and major phenolic contributors present in the flesh and peel extracts of 11 apple cultivars from Poland were addressed. We attempted a novel approach by investigating the feasibility of applying cyclic voltammetry (CV) to determine the reducing activity of major phenolic compounds and predicting the antioxidant capacity of apple extract from peel and flesh. The aims of this study were as follows: (1) to determine the antioxidant capacity of apple flesh and peel

by peel by cyclic voltammetry and spectrophotometric assays, (2) to determine the profiles of phenolic compounds in the flesh and peel of popular apple cultivars by sensitive liquid chromatography (HPLC) coupled to mass spectrometry (MS) using the electrospray ionization (ESI) and diode array detector (DAD) methodology, (3) to determine the antioxidant activity of the identified phenolic compounds by cyclic voltammetry (CV) and spectrophotometric assays, and (4) to show the relationship between the content of individual phenolic compounds and the antioxidant capacity of apple flesh and peel.

#### **2. Materials and Methods**

#### *2.1. Chemicals and Reagents*

Chlorogenic acid, gallic acid, rutin, quercetin, quercetin-3-*O*-glucoside, quercetin-3-*O*-rhamnoside, (-)-epicatechin, and cyanidin-3-*O*-galactoside were supplied by Extrasynthese (Genay, France). Quercetin-3-*O*-arabinoside, (+)-catechin, phloretin, phloridzin, and other compounds were obtained from Sigma-Aldrich (Munich, Germany); 2,2-diphenyl-1-picrylhydrazyl (DPPH), 2,4,6-tri(2-pyridyl)-s-triazine (TPTZ) and 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox) were purchased from Sigma (Sigma Chemical Co., St. Louis, MO, USA). Folin–Ciocalteu's reagent and others of reagent-grade quality were from POCh (Gliwice, Poland). Ultrapure water was purified with a Millipore Direct-Q UV 3 System (Bedford, MA, USA). Flavonoids and solvents were HPLC-grade quality, and other reagents were at least reagent-grade quality.

#### *2.2. Sample and Standard Preparations*

The studied material consisted of 11 first-quality grade apple cultivars at their ripe period of growth. Early varieties of apples such as Antonówka, Delikates, Early Geneva, Papierówka, Paulared, Sunrise, and Quinte were harvested in August, while Gloster, Jonagored, Ligol, and Rubinola cultivars, which are late varieties, were harvested in September, both during the 2019 season. All fruits were purchased from the Experimental and Production Institute "Pozorty" Sp. z o.o. in Olsztyn (Poland). Fruit samples (10 apple fruits randomly selected) were washed with distilled water to remove foreign substances and manually peeled using a hand peeler (1–2 mm thickness), cored, and cut into small pieces. The weighed apple flesh and peels were pooled separately. Before freeze drying (FD), the apple flesh and peels were frozen overnight at −25 ◦C and dried in the FreeZone 2.5 freeze dryer (Labconco, CA, USA). During FD, pressure was reduced to 16 Pa. The temperature in the drying chamber was −55 ◦C, and the shelf temperature was 26 ◦C. Apple flesh and peels were kept in the drying chamber for 24 h. The lyophilized samples were ground in a laboratory mill and stored at −20 ◦C up to further analysis. The moisture content of the peel from all apple cultivars ranged from 93.44 to 94.83%, and that of flesh was in the range from 91.63 to 93.01%. About 100 mg (for spectrophotometric methods) and 250 mg (for the electrochemical method) of lyophilized flesh and peel were extracted with 1 mL of 80% methanol by 30 s sonication. Then, the mixture was vortexed for 30 s, again sonicated and vortexed, and centrifuged for 5 min (13,200 rpm). This step was repeated five times and supernatant was collected in a 5 mL flask. Finally, all extracts were kept in dark-glass vials at −20 ◦C for further analysis.

#### *2.3. Spectrophotometric Determination of Total Phenolic and Flavonoid Content*

Total phenolic content (TPC) was determined according to the Folin–Ciocalteu (FC) assay as described by Shahidi and Naczk [36] with a slight modification. A volume of 90 μL of sample extract (10 mg/ml), 90 μL of FC reagent (diluted 1:10, *v*/*v*), 180 μL of saturated solution of Na2CO3, and 1440 μL of H2O were mixed and allowed to react for 25 min in a thermomixer (Comfort, Eppendorf) at room temperature. The absorbance was measured at 725 nm in a UV-1800 spectrophotometer (Shimadzu, Japan) with a temperature controller (TCC-Controller, Shimadzu, CA, USA). Catechin was used as a standard and the results were based on the standard curve equation of catechin (0.0625–1.0 mM) and expressed as a catechin equivalent (CAE) in μg/g of fresh weight. The measurements were made in triplicate.

Total flavonoid content (TFC) was determined based on the method by Jia et al. [37] with some modifications. A volume of 1230 μL of extract (10 mg/mL) was mixed with 62 μL of 5% NaNO2 solution (m/v). After incubation in the thermomixer (Comfort, Eppendorf) at room temperature for 6 min, 123 μL of 10% AlCl3·6H2O was added. Again, the mixture was incubated under the same conditions for 6 min, then 410 μL of 1M NaOH was added and the mixture was centrifuged for 5 min at 2000 rpm (Centrifuge 5424, Eppendorf). The absorbance of the reaction mixture was measured against the reagent blank at 510 nm with the UV-1800 spectrophotometer with a temperature controller (TCC-Controller, Shimadzu, city, State, country). Catechin was used as a standard and the results were based on standard curve equation of catechin (0.0625–0.5 mM) and expressed as catechin equivalent (CAE) in μg/g of fresh weight. The measurements were done in triplicate.

#### *2.4. Analysis of Phenolic Compounds by HPLC–DAD–MS*/*MS*

The identification of phenolic compounds was done by means of liquid chromatography (HPLC) coupled to mass spectrometry (MS) using the electrospray ionization interface (ESI). Quantification of phenolic compounds was carried out by using HPLC with diode array detector (DAD). The analysis of phenolic compounds was performed on a micro-HPLC system (LC200, Eksigent, Dublin, CA, USA) with pump, autosampler, column oven, and system controller. The micro-HPLC was connected in series to a QTRAP 5500 mass spectrometer (AB Sciex, Canada) equipped with a triple quadrupole, ion trap, and an ion source of electrospray ionization (ESI). The analytical column was a Halo C18 column (50 mm × 0.5 mm, 2.7 μm i.d.; Eksigent, Dublin, CA, USA). Eluent A was water/formic acid, 99.05/0.95 (*v*/*v*); eluent B was acetonitrile/formic acid, 99.05/0.95 (*v*/*v*). A gradient elution program was used: 5–5–90–90–5–5% B in 0–0.1–2–2.5–2.7–3 min. Before chromatographic analysis, apple extract was centrifuged (20 min, 13,000 *g*). Aliquot (2 μL) of sample solution was injected, with flow rate of 15 μL/min, at a column temperature of 45 ◦C. Phenolic compounds detected in the apple extracts were identified according to their MS/MS fragmentation spectrum (*m*/*z* values). Mass spectrometry data were obtained in positive- and negative-ion mode. The optimal identification of phenolic compounds was achieved under the following conditions: curtain gas: 25 L/min, collision gas: 9 L/min, ions pray voltage: 5400 V (for positive-ion mode) and −4500 V (for negative-ion mode), temperature: 350 ◦C, 1 ion source gas: 35 L/min and 2 ion source gas: 30 L/min; and entrance potential: 10 V (for positive-ion mode) and −10 V (for negative-ion mode). Phenolic compounds were quantified from the determined multiple reaction monitoring pairs (MRM) as shown in Table 1 and the calibration curves of external standards (range of 10–1000 nM).


**Table 1.** MS data of phenolic compounds from apple.

#### *2.5. Ferric Reducing Antioxidant Power (FRAP) Assay*

The FRAP assay was carried out with some modifications according to the method of Benzie and Strain [38]. Briefly, the FRAP-2,4,6-tri(2-pyridyl)-s-triazine (TPTZ) reagent was prepared from the sodium acetate buffer (300 mM, pH 3.6), 10 mM 2,4,6-tri(2-pyridyl)-s-triazine (TPTZ) solution (40 mM HCl as solvent), and 20 mM FeCl3·6H2O in a volume ratio of 10:1:1. The FRAP reagent was freshly prepared on the day of the measurements. An aliquot of 75 μL of the extract was mixed with 225 μL of ultrapure water and 2250 μL of FRAP reagent (pre-incubated for 5 min at 25 ◦C). The absorbance of the reagent mixture was measured at 593 nm after 30 min incubation at 25 ◦C. Samples were measured in 3 replicates. The standard curve was prepared using Trolox solution (0.034–0.612 mM), and the results were expressed as μmol of Trolox equivalent per gram of apple fresh weight (μmol TE/g FW).

#### *2.6. DPPH Assay*

DPPH assay was based on the method of Brand-Williams et al. [39]. The samples were diluted to a proper concentration to make sure that the test results were readable between the absorbance values of 0.2–0.8. A volume of 100 μL of sample was mixed with 2 mL of methanol and then reacted with 250 μL of DPPH solution (10 mg DPPH in 25 mL of methanol) The reaction mixture was incubated in the dark at room temperature for 20 min, after which the absorbance at 517 nm was recorded. The test was performed in triplicate. The Trolox calibration curve (0.1–1.0 mM) was plotted as a function of the decrease in absorbance. The percentage of inhibition of DPPH radical by tested samples was calculated using the following equation, expressed as μmol TE/g FW:

Scavenging activity % = 100 − [(Abs. of sample − Abs. of blank) × 100/Abs. of control]

#### *2.7. Cyclic Voltammetry (CV) Assay*

Cyclic voltammograms were recorded using a Gamry G 750 potentiostat (Warminster, PA, USA). The working electrode was a 3 mm diameter glassy carbon disk electrode (BAS MF-2012). An Ag/AgCl reference electrode was used in conjunction with a platinum wire as a counter electrode. Given the effect of polyphenol adsorption on the voltammetric response, which was observed in our previous work [40], it was considered important to run the voltammograms in as consistent a manner as possible. The following electrode cleaning procedure was undertaken between each run. The electrode was thoroughly hand-polished with 0.05 μm alumina powder (BAS CF-1050) on the polishing cloth (BAS MF-1040) and rinsed thoroughly with Milli-Q grade water. Before taking the cyclic voltammogram of the test solution, a background cyclic voltammogram was run in the buffer solution in the potential range from <sup>−</sup>0.1 to 1.3 V at a scan rate of 0.1 V s−1, and the electrode was rinsed with Milli-Q grade water and methanol and dried. Apple extracts were diluted with Britton-Robinson (B-R) buffer (0.1 M, pH 7.4) at a 1:1 ratio, and the final extract concentration was 25 mg/mL. Cyclic voltammograms were taken in the potential range from <sup>−</sup>0.1 to 1.3 V at scan rate of 0.1 V s−1, with only the first scan being recorded. Background cyclic voltammograms were subtracted from those obtained for apple extracts to allow the oxidation and reduction processes to be more clearly revealed. For the purpose of testing, the total anodic peak wave area of the voltammograms was calculated within the range of 0 to 1.1 V. This method was based on the correlations between the total anodic peak wave area of a cyclic voltammogram and the antioxidant capacity of the sample and reference substance. For the reference, a solution of Trolox within the concentration range of 0.15–1.00 mM was used, and the results were expressed as μmol TE/g FW.

#### *2.8. Chelating Activity on Ferrous Ions*

The chelating activity of ferrous irons was measured by the inhibition of the formation of an Fe2<sup>+</sup>-ferrozine complex after treating the apple extracts with Fe2<sup>+</sup> according to Mladénka et al. [41]. Briefly, 0.4 mL of apple extract (0.5 mg/mL) and 0.2 mL of HEPES (pH 7.5, 0.12 mM) were added

to a solution of 0.4 mM FeSO4·7H2O (0.2 mL) and mixed for 2 min. Then, a volume of 800 μL of ferrozine solution (0.5 mM) was added and the absorbance of the ferrous ion–ferrozine complex was immediately measured at 562 nm. Ferrous ion solution was prepared daily and purged by argon 5.5 grade quality (Linde, Germany). For the comparison of ferrous chelating activity, deferoxamine (DEF) was used as a standard iron chelator. The amount of remaining ferrous ion was calculated from the difference of absorbance between the apple extract sample (with ferrozine) and its control blank (without ferrozine) was divided by the difference of the control sample (known amount of ferrous ion without apple extract) and its corresponding blank. A standard curve of Fe2<sup>+</sup> ions was prepared within the range of concentration of 5–60 μM. The ferrous chelation efficiency of tested apple extract was expressed in %. Measurements were done at least in triplicate.

#### *2.9. Analysis of Antioxidant Activity of Phenolic Compounds Identified by HPLC–DAD–MS*/*MS*

The antioxidant activity of the phenolic compounds identified by HPLC–DAD–MS/MS in apple peel and flesh was provided by cyclic voltammetry and by spectrophotometric assays (DPPH, FRAP, and chelating activity). Stock solutions of each standard compound were also dissolved in 80% methanol (*v*/*v*, pH 6.0) and stored at −80 ◦C. Results were expressed as mM of Trolox of 9 independent experiments (*n* = 9). The ferrous chelation efficiency of apple phenolics was expressed in % of 9 replicates (*n* = 9).

#### *2.10. Statistical Analyses*

The analyses were performed in triplicate, and the results were displayed as the mean ± standard deviation (SD). The differences in identified phenolic contents in the peel and flesh of 11 apple cultivars were determined by one-way analysis of variance (ANOVA) with Fisher's least significant difference test (*p* < 0.05). Correlations between the antioxidant capacity assays and polyphenol compounds were analyzed using the Pearson correlation coefficient test. All analyses were performed using Statistica software (v. 12; StatSoft, Tulsa, OK, USA). Statistical significance thresholds for correlations were set at *p*-value< 0.05 (\*), *p* < 0.01 (\*\*) and *p* < 0.001 (\*\*\*).

#### **3. Results and Discussion**

#### *3.1. Total Phenolic Content (TPC) and Total Flavonoid Content (TFC)*

According to several authors, the phenolic and flavonoid contents vary among different cultivars of fruits and vegetables, and within different tissues [42,43]. With respect to this condition, for this study, the cultivars were selected in order to eliminate the impact of soil, fertilizing method, and climatic conditions on apples, because all fruits were grown exclusively in one orchard (Pozorty, Olsztyn). It can be supposed that the antioxidant activity of apples depends, to a large extent, on the cultivar. The 11 apple cultivars selected for this study were characterized by an over-color of the peel that ranged from green-yellow (Papierówka and Antonówka) to red (Paulared, Quinte, Gloster, and Rubinola). The distribution of polyphenol compounds between the peel and flesh of analyzed cultivars for total phenolic content (TPC) for total flavonoid content (TFC) is shown in Table 2.

For all the studied cultivars, both the TPC and TFC were higher in the apple peel extract than in the flesh extract. Furthermore, significant differences were found between the cultivars (*p* < 0.05) in TPC and TFC. TPC ranged between 1821.3 and 3278.6 μg CAE/g fresh peel (Table 2). Quinte had the highest TPC, followed closely by Early Geneva and Jonagored (3278.6, 3147, and 3123.1 μg CAE/g fresh peel, respectively), whereas Ligol and Antonówka had the lowest TPC (1821.3 and 2051.6 μg CAE/g fresh peel, respectively). The remaining cultivars were intermediate, with a TPC varying between 2194.7 and 2916.1 μg CAE/g fresh peel. However, the total phenolics were lower in flesh than in peel, ranging from 535.5 to 1740.3 μg CAE/g fresh flesh (Table 2). Ligol and Gloster presented low contents with less than 600 μg CAE/g fresh flesh, whereas Quinte showed a concentration level of 1740.3 μg CAE/g fresh flesh. TPC values for the flesh and peel extracts of the studied apple cultivars were comparable

with those previously reported by Tsao et al. [13] and Carbone et al. [18]. Tsao et al. [13] reported, for the eight most popular apple cultivars grown in Ontario, that the TPC ranged from 1016.5 to 2350.4 μg GAE/g of FW in the peel and 177.4 to 933.6 μg GAE/g of FW in the flesh. These values were found to be in agreement with our TPC results obtained for the 11 tested apple cultivars (1821.3 to 3278.6 μg CAE/g FW and 535.5 to 1740.3 μg CAE/g FW for peel and flesh, respectively). In the present study, the flavonoid content (TFC) ranged from 25% to 44.7% of the TPC in the peel and from 28.9% to 61.0% in the flesh, and these results are in agreement with those reported by Carbone et al. [18].

**Table 2.** Total phenolic content (TPC), total flavonoid content (TFC), Fe2<sup>+</sup> chelation activity, and antioxidant activity by ferric-reducing/antioxidant power assay (FRAP), 2,2-diphenyl-1-picrylhydrazyl (DPPH), and cyclic voltammetry (CV) of the peel and flesh of different apple cultivars.


Values represent the mean (*n* = 3) ± SD. Different letters a–j in the same column for peel or flesh indicate significant differences by ANOVA test (*p* < 0.05). Results for ferric reducing antioxidant power (FRAP), DPPH, and CV are expressed in Trolox equivalent per gram of apple fresh weight (μmol TE/g FW), and for chelating activity in % of chelating of Fe(II). TPC and TFC results are expressed in μg catechin equivalent (CAE)/g FW.

Recent studies also have demonstrated the influence of the apple cultivar on the fruit's phytochemical content [13,18] as well as a possible relationship between the color of different cultivars and their nutritional values [44]. In West Himalayan apple varieties, it was confirmed that there is a significant difference in phenolic content among cultivars and locations [45]. It was also confirmed that the variety and maturity of apples have a significant impact on chemical composition, the concentration of polyphenols, and level of antioxidant activity [35]. In a study of 120 apple varieties, a large diversity in polyphenolic compound content was observed depending on the variety [35]. According to this research, the highest phenolic content was found in Ozark Gold (~2116.03 mg/kg), and the lowest concentration was for Ligol (~814.17 mg/kg). The quality and quantity of polyphenols in apples directly influences their antioxidant activity [46].

While a TPC assay can adequately differentiate between apple cultivars that are high and low in polyphenols, it was less useful as a forecaster of potential health benefits. This is because the TPC measurements include nonabsorbable polymeric polyphenols as well as smaller, potentially absorbable polyphenolic compounds, which are thought to be mainly responsible for the observed physiological effects. Although some degradation products of polymeric polyphenols are absorbed in

the colon, it is still not fully explained whether they have beneficial physiological effects. Whether the magnitude of polyphenol content has any relevance to the health properties of apples must then be tested by measuring the individual small molecular weight polyphenols.

#### *3.2. Antioxidant and Chelating Capacity of Apple Flesh and Peel Determined by Spectrophotometric Assays*

The antioxidant capacity of food should be evaluated with a variety of methods that can address the different mechanisms [47,48]. In the present study, spectrophotometric methods such as DPPH radical-scavenging activity and ferric reducing antioxidant power (FRAP) were used to determine the antioxidant capacity of apple extracts. The DPPH assay is based on a mixed mechanism of free radical DPPH• stabilization: hydrogen atom transfer and electron transfer. This assay presents some critical analytical points [49], but it has the great advantage of being easy to use. The FRAP assay is based on the ability of antioxidants to reduce ferric(III) ions to ferrous(II) ions by the electron transfer mechanism. The antioxidant capacity of apple peel and flesh extracts, evaluated by DPPH and FRAP assays and expressed as Trolox equivalent (μmol TE/g FW), is shown in Table 2.

The results obtained in the DPPH assay showed that, in general, the flesh and peel have intermediate radical-scavenging activity, with peels being better scavengers than flesh. Apple peel extracts from Jonagored (8.65 μmol TE/g FW), Gloster (8.19 μmol TE/g FW), and Rubinola (7.91 μmol TE/g FW) had the highest and extracts from Papierówka (5.62 μmol TE/g FW) and Antonówka (5.27 μmol TE/g FW) showed the lowest free radical scavenging capacity among the tested apple cultivars. In the case of apple flesh extracts, the DPPH values ranged between 2.23 and 4.65 μmol TE/g FW for Gloster and Quinte cultivars, respectively. These results were consistent with those reported for different apple varieties by Carbone et al. [18]. A highly significant relationship was found between the DPPH antiradical activity of apple flesh extracts and the concentration of TPC (r = 0.96; *p* < 0.001) and TFC (r = 0.82; *p* < 0.01). Panzela et al. [50] also found a good correlation between the percentage of reduced DPPH and the concentration of total polyphenols (r = 0.79) and total flavan-3-ols (r = 0.77).

For all the tested apple cultivars, peel extracts had a much greater ferric-reducing antioxidant power than flesh extracts, with a two-fold difference (Table 2). Quinte and Jonagored peel extracts had the highest FRAP values (21.31 μmol TE/g FW and 20.89 μmol TE/g FW, respectively), and Antonówka and Ligol peel extracts had the lowest (12.73 μmol TE/g FW and 12.40 μmol TE/g FW, respectively). These results were consistent with the total polyphenol and total flavonoid concentration in Early Geneva, Quinte, and Jonagored and in Ligol. The FRAP activity of the extracts from different apple cultivars was positively correlated with both total phenolic and total flavonoid content (r = 0.9972 and 0.8229, respectively). These results were in agreement with the findings of Tsao et al. [13], who studied the total phenolic compounds of eight apple cultivars and also obtained a good correlation between the TPC and FRAP activity (r = 0.95).

The phenolic compounds of apple may act as reducing agents, hydrogen donors, free radical scavengers, and singlet oxygen quenchers, and may exhibit antioxidant activity via the chelation of metal ions [23]. In this study, ferrous ion chelating activity was measured by the inhibition of the formation of a Fe(II)–ferrozine complex after the treatment of peel and flesh extracts with Fe(II). The chelating capacity of apple peel ranged from 51.80 (Quinte cultivar) to 19.30% (Ligol cultivar), while in flesh it was within the range of 49.46 (Papierówka cultivar) to 18.31% (Ligol cultivar). The chelating activities of peels of individual apple cultivars were significantly different and the same observation was noted for apple flesh (*p* < 0.05).

#### *3.3. Reducing Capacity of Apple Flesh and Peel Determined by Cyclic Voltammetry*

In this study, we conducted a critical evaluation of the cyclic voltammetry method for the determination and rapid screening of the reducing capacity of peel and flesh of 11 apple cultivars compared with DPPH and FRAP assays. The representative cyclic voltammograms of peel and flesh extracts (25 mg/mL) were recorded from −0.1 to 1.3 mV at a scan rate of 100 mV/s (Figures 1 and 2).

**Figure 1.** Cyclic voltammograms of peel and flesh extracts of selected apple cultivars: (**A**) Jonagored, (**B**) Antonówka. CV of electrolyte solution shown as dotted gray line. Operative conditions: extract concentration 25 mg/mL; pH 6.0; scan rate 0.1 V/s.

**Figure 2.** Cyclic voltammograms of selected apple cultivars: (**A**) extracts from peel; (**B**) extracts from flesh. CV of electrolyte solution shown as dotted gray line. Operative conditions: extract concentration 25 mg/mL; pH 6.0; scan rate 0.1 V/s.

The obtained voltammograms show that the apple extracts exhibited well defined oxidation and reduction voltammetric peaks. The area of each voltammetric peak was related to the concentration of antioxidants. A broad anodic peak between 0.4 and 1.0 V was observed. This peak was due to the response of several antioxidants with different oxidation potentials, mainly flavonoids and phenolic acids. The results show that the samples contained multiple reducing agents in the respective extracts. Therefore, the area under the anodic current waveform (area under the curve, AUC) was taken to reflect the reducing capacity of the samples compared to a set of Trolox solutions, as suggested by Chevion et al. [51], Martinez et al. [52], and Zieli ´nska and Zieli ´nski [40]. This provides a marked advantage in some cases, particularly when the AUC wave represents more than a single component. Higher AUC indicates a higher reducing capacity of the investigated extract.

The reducing capacity of peels as shown by CV ranged from 6.80 to 4.35 μmol TE/g FW (Table 2). The highest reducing capacity was noted for Paulared, Jonagored, and Quinte cultivars and the lowest for Rubinola, Antonówka, and Ligol. In contrast, the reducing capacity of apple flesh was noted to be

at least twice as low (Table 2), ranging from 3.95 (Quinte) to 1.44 (Ligol) μmol TE/g FW. The reducing capacity of peels as shown by CV was positively correlated with both the total phenolic content (r = 0.867; *p* < 0.01) and total flavonoid content (r = 0.752, *p* < 0.01). The same trend in correlation efficient values was noted for apple flesh. This result was in agreement with other studies, in which CV was shown to be an efficient instrumental tool for evaluating the reducing capacity of plant, food, and biological samples. An advantage of electrochemical measurement compared to DPPH and FRAP is that it is fast. CV measurement is carried out in <10 min, so it is less tedious. Moreover, it is not necessary to use expensive reagents that include free radicals, thus lowering the cost. In addition, the use of a small amount of organic solvent reduces the amount of organic waste produced.

#### *3.4. Profile and Content of Phenolic Compounds in Apple Flesh and Peel Measured by HPLC–DAD–ESI-MS*/*MS*

Detailed knowledge of the polyphenol profile and content in apple cultivars is necessary in order to evaluate their antioxidant activity and potential beneficial health effects. A comprehensive qualitative analysis of the phenolic compounds of the studied apple peel and flesh samples was achieved by HPLC–DAD–MS/MS. The composition and concentrations of identified compounds in the peel and flesh are presented in Table 3. Eleven polyphenolic compounds belonging to five major groups were identified: chlorogenic acid (hydroxycinnamic acid), phloretin and phloridzin (dihydrochalcones), catechin and epicatechin (flavonols), and quercetin, and four derivatives (flavonols) and cyanidin 3-galactoside (anthocyanins). It was found that phenolic acids and flavonols were the two main groups of polyphenols identified in the studied apple cultivars, as previously reported [53]. Concentrations of individual phenolic compounds in peel and flesh identified by HPLC–DAD–MS/MS are shown in Table 3.

Among flavan-3-ols, epicatechin was the major compound of this group in apple peel and flesh. The highest concentration of epicatechin in peel was found for Quinte (297.77 μg/g FW) and Early Geneva (278.11 μg/g FW), while Delikates had the lowest content (94.79 μg/g FW). In the flesh, the Quinte cultivar was also found to possess the highest concentration of epicatechin (325.04 μg/g FW), followed by Early Geneva and Paulared (270.76 and 221.56 μg/g FW, respectively). These results are in accordance with those reported by Tsao et al. [13] for eight apple cultivars, in which epicatechin ranged from 17.9 to 591.6 μg/g FW in the peel and 16.0 to 142.3 μg/g FW in the flesh. Catechin was present in smaller amounts in peel extracts (3.83 to 92.16 μg/g FW) as well as flesh extracts (1.64 to 75.90 μg/g FW). These data are also in agreement with those obtained for eight traditional apple cultivars of Southern Italy [50], for which the concentration of catechin ranged from 0 to 76.7 μg/g FW).

Chlorogenic acid, which is the major compound of hydroxycinnamic acids, was mostly located in the apple flesh, except for Gloster, Jonagored, and Ligol cultivars, where the content of chlorogenic acid in peel was higher than in flesh. In the flesh, chlorogenic acid levels ranged between 18.21 and 451.53 μg/g FW, with the highest amounts being recorded in Papierówka, followed by Quinte, Antonówka, and Rubinola (Table 3). The lowest amounts were found in Delikates and Gloster (20.01 and 18.21 μg/g FW, respectively). The concentrations of chlorogenic acid obtained for some cultivars tested in this work, particularly Papierówka, were twice those found in the cultivars studied by Khanizadeh et al. [53] and Panzella et al. [50]. These observations indicate that the range of differences between the polyphenol profiles of apples are highly cultivar dependent.


#### *Antioxidants* **2020**, *9*, 1054

Differences were also observed in the content of dihydrochalcones. Phloridzin (phloretin 2 -glucoside) was the predominant dihydrochalcone found and identified in all tested apple peel and flesh extracts. Phloretin and phloretin derivatives have occasionally been found in apple in trace amounts [13]. Phloridzin concentration was higher in apple peel, with a mean value of 33.28 μg/g FW compared to 16.38 μg/g FW in flesh (Table 3). Khanizadeh et al. [53] reported an average concentration of phloridzin in the peel and flesh of 10.4 and 55.4 μg/g FW, respectively. Among the tested apple cultivars, Papierówka in particular was characterized by the highest level of phloridzin (84.10 μg/g FW in peel), whose anti-diabetic properties have recently been reported by Masumoto et al. [54]. Even though dihydrochalcones exist in relatively low amounts due to the uniqueness of the apple and their different profiles among different cultivars, they have been used to distinguish apple from a number of other fruits and to identify apple cultivars [13].

Flavonols represent the second largest group in terms of concentration in apple peel. These polyphenols were constituted mainly by quercetin 3-arabinoside, followed by 3-rhamnoside and 3-glucoside, and slightly by quercetin and rutin. Depending on the cultivar, the total flavonols in the peel varied from 193.09 to 808.25 μg/g FW, with Sunrise showing the highest concentration (Table 3). These data are consistent with those reported by Tsao et al. [13]. On the other hand, in the flesh extracts of studied apple cultivars, only small amounts of quercetin 3-arabinoside, 3-rhamnoside, and 3-glucoside were detected (Table 3).

The major anthocyanins in apple are cyanidin glycosides, among which 3-galactoside is the predominant individual compound. Anthocyanins were found only in apples characterized by red and partially red skin (Quinte, Paulared, Gloster, and Rubinola), and only cyanidin 3-galactoside was identified in our study (Table 3). Cyanidin 3-galactoside content ranged from 1.15 to 103.69 μg/g FW and was the highest for Rubinola. This observation was consistent with that reported by Khanizadeh et al. [53].

According to Kschonsek et al. [43], the most abundant flavonoids that occur in apples (raw, with skin) are (-)-epicatechin, (+)-catechin, and cyanidin. The same flavonoids were found in apples without skin, but also a high amount of (-)-epigallocatechin. The main polyphenols that can be found in apples are quercetin, (-)-epicatechin, (+)-catechin, procyanidines, and anthocyanidines; dihydrochalcones; phloretin and phloridzin derivatives; and other phenolic compounds, such as chlorogenic acid. In addition, it was important that apples were shown to have the highest portion of free phenolics when compared to other fruits [4]. Apple's bound phenolics have lower bioavailability as compared to free phenolics since they need to be released from the food matrix after digestion [55].

#### *3.5. Antioxidant, Reducing, and Chelating Activities of Phenolic Compounds in Apple Flesh and Peel*

The antioxidant, reducing, and chelating activities of phenolic compounds identified in apple peel and flesh by HPLC–DAD–MS/MS are shown in Table 4. The antioxidant activity of phenolic compounds from apple flesh and peel, determined as free radical-scavenging activity against stable, nonbiological relevant DPPH radicals, is expressed as Trolox equivalent. As it was defined, antioxidant activity is equal to the millimolar concentration of a Trolox solution that has antioxidant capacity equivalent to a 1.0 mM solution of the substance under investigation.

Quercetin, cyanidin 3-galactoside, rutin, catechin, and chlorogenic acid (2.09–1.45 mM Trolox) showed the highest ability to scavenge DPPH radicals, followed by quercetin 3-glucoside, epicatechin, quercetin 3-rhamnoside, and quercetin 3-arabinoside (1.42–0.95 mM Trolox), while the ability to scavenge DPPH radicals by phloretin (0.19 mM Trolox) and phloridzin (0.06 mM Trolox) was the lowest.

The order of reducing activity as shown by the FRAP assay was as follows: cyanidin 3-galactoside > quercetin > chlorogenic acid > quercetin 3-glucoside > catechin > epicatechin > quercetin 3-rhamnoside > rutin > quercetin 3-arabinoside > phloretin > phloridzin (5.69–0.18 mM Trolox). This order was supported by a study on the structure–activity relationship (SAR) of flavonoids [56,57].


**Table 4.** Antioxidant, reducing, and chelating activities of phenolic compounds identified in apple peel and flesh by HPLC–DAD–MS/MS.

Results were provided by DPPH radical scavenging activity assay. FRAP: ferric-reducing/antioxidant power assay; CV: cyclic voltammetry assay; FZ: ferrozine assay. Data expressed as the mean ± standard deviation (*n* = 9).

In this study, cyclic voltammograms of the phenolic compounds identified in apple flesh and peel were recorded in the range of –100 to +1300 mV at a scanning rate of 100 mV s<sup>−</sup>1. Cyclic voltammograms of 0.25 mM solutions of examined compounds in 0.1 M Britton–Robinson (B–R) buffer (pH 6.0) in 80% methanol are shown in Figure 3.

**Figure 3.** Cyclic voltammograms of 0.25 mM of standard solutions (final concentration) of phenolic compounds in apple cultivars identified by HPLC–DAD–MS/MS analysis in Britton–Robinson (B–R) buffer (0.1 M; pH 7.4) recorded from –100 to +1300 mV; scan rate 100 mV s<sup>−</sup>1.

The cyclic voltammograms showed that all compounds had well defined reversible waves with the first oxidation peak potential. The first anodic peak potential (Epa) of the investigated compounds varied according to the following gradation: phloretin (0.815 V) > phloridzin (0.759 V) > cyanidin 3-galactoside (0.618 V) > catechin (0.571 V) > quercetin 3-rhamnoside (0.515 V) > quercetin 3-arabinoside (0.512 V) ≥ quercetin 3-glucoside (0.511 V) > chlorogenic acid (0.391 V) ≥ rutin (0.390 V) > epicatechin (0.339 V) ≥ quercetin (0.334 V) as compared to Trolox (0.346 V). Higher Epa values were associated with lower reducing activity of the tested compound. Therefore, taking into account the values of the first oxidation potential of the studied compounds, almost all phenolic compounds identified in apple flesh and peel can be described as having high (intermediate) antioxidant power (Ep < 0.8 V), while phloretin had low antioxidant power (0.8 V < Ep < 1.3 V). This conclusion was drawn based on the work by Blasco et al. [58], in which the differentiation of the antioxidant power of phenolic compounds was based on their oxidation potential. When the calculation of antioxidant activity was based on the area under the anodic current waveform within the range of 0 to 1100 mV for each compound and Trolox, then the order of antioxidant activity was as follows: quercetin > epicatechin > cyanidin 3-galactoside > rutin ≥ phloretin > catechin > chlorogenic acid > phloridzin ≥ quercetin 3-rhamnoside ≥ quercetin 3-glucoside > quercetin 3-arabinoside (0.90–0.19 mM Trolox). The gradation of samples for reducing activity as determined by CV mirrored that obtained with the FRAP assay.

The highest ferrous ion chelating activity was shown by chlorogenic acid (88.47%), followed by rutin, quercetin, catechin, epicatechin, quercetin 3-glucoside, quercetin 3-arabinoside, and quercetin 3-rhamnoside (68.13%); activity twice as low was noted for cyanidin 3-galactoside, and the lowest was for phloridzin and phloridzin (5.68 and 1.15%, respectively).

#### *3.6. Phenolic Contribution to Antioxidant Activity*

A correlation analysis was performed to assess the contribution of polyphenolic compounds to antioxidant capacity. Interestingly, the association between antioxidant assays and the content of bioactive compounds differed between apple peel and flesh (Table 5).

In apple peel, catechin and epicatechin content was positively correlated to FRAP assay, while cyadinin-3-galactoside was highly associated with DPPH. Moreover, a strong correlation was found between the total phenolic and flavonoid content and FRAP and CV (Table 5).


**Table 5.** Correlation coefficient between phenolic compounds and antioxidant capacity tests and linear correlation coefficient between different methods for antioxidant capacity assessment in apple peel and flesh.


**Table 5.** *Cont.*

\*, \*\*, \*\*\* Significant correlation at *p* < 0.05, *p* < 0.01, *p* < 0.001, respectively.

In the apple flesh, FRAP showed a very strong correlation with epicatechin, chlorogenic acid, and cyaniding-3-galactoside. Similarly, DPPH and CV were correlated with the same compounds, and CV was additionally correlated with catechin. The chelating activity assay showed the strongest correlation with phloretin, phloridzin, and chlorogenic acid (Table 5). Among the antioxidant capacity tests, the strongest correlation was observed between CV and FRAP in apple peel, and between all assays in apple flesh.

#### **4. Conclusions**

It needs to be noted that the ripe apples are a good source of phenolic compounds, being present in both peel and flesh. According to the available literature, the geographical origin and variety of apples influence the content of phenolic compounds and are highly related to their antioxidant activity fluctuation. In this study, the apple peel and flesh of early varieties such as Antonówka, Delikates, Early Geneva, Papierówka, Paulared, Sunrise, Quinte showed higher total phenolic and flavonoids content as compared to the late varieties such as Gloster, Jonagored, Ligol and Rubinola. The HPLC–DAD–MS/MS analysis showed that the dominant compounds were catechin, epicatechin, chlorogenic acid, quercetin 3-glucoside, quercetin 3-arabinoside, quercetin 3-rhamnoside, cyanidin 3-galactoside and phloridzin whereas phloretin, quercetin and rutin were present in low concentration. These phenolic compounds varied considerably among apple cultivars and their content was higher in peels than in flesh. Information about cultivar–typical apple polyphenol content and profile is important for bioactivity studies and, consequently, essential for the development of consumer-relevant products with particular nutritional functionalities. Therefore, it can be concluded that whole ripe apples should be used as a relevant source of phenolics in our diet since the removal of peel from apple may induce a significant loss of antioxidants. On the other hand, apple peels can be a good component to formulate functional foods after currently proposed "cold-pressing technology" as an effective method for peeling and deseeding apple fruits, with a positive effect on phenolic compounds retention in pomace. This study showed that the application of HPLC–DAD–MS/MS analysis of phenolic compounds with the spectrophotometric methods for the determination of their antioxidant, reducing and chelating capacity and their electroactivity provided by cyclic voltammetry was essential to show their contribution to the antioxidant capacity of apple peel and flesh. This study also provides

evidence to support the application of cyclic voltammetry as a rapid method in determining the phenolic profile and reducing power of apple flesh and peel.

**Author Contributions:** Conceptualization, D.Z.; methodology, D.Z., M.T.; formal analysis, D.Z., M.T.; investigation, writing—original draft preparation, D.Z., M.T.; writing—review and editing, D.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** The project was financially supported by the Ministry of Science and Higher Education within the Regional Initiative of Excellence program for 2019–2022, project no. 010/RID/2018/19, amount of funding PLN 12,000,000.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

### **Identification of a New Variety of Avocados (***Persea americana* **Mill. CV. Bacon) with High Vitamin E and Impact of Cold Storage on Tocochromanols Composition**

### **Celia Vincent 1,2, Tania Mesa 1,2 and Sergi Munne-Bosch 1,2,\***


Received: 21 April 2020; Accepted: 6 May 2020; Published: 9 May 2020

**Abstract:** (1) Background: Tocochromanols are a group of fat-soluble compounds including vitamin E (tocopherols and tocotrienols) and plastochromanol-8, and just one avocado can contain up to 20% of the required vitamin E daily intake. (2) Methods: HPLC and LC-MS/MS analyses were performed in avocados of various varieties and origin for the identification and quantification of tocopherols, tocotrienols and plastochromanol-8. After selection of the variety with the highest vitamin E content, we evaluated to what extent short- (4 h) and long-term (10 d) cold storage influences the accumulation of tocochromanols. (3) Results: Analyses revealed that "Bacon" avocados (*Persea americana* Mill. cv. Bacon) were the richest in vitamin E compared to other avocado varieties (including the highly commercialized Hass variety), and they not only accumulated tocopherols (with 110 μg of α-tocopherol per g dry matter), but also tocotrienols (mostly in the form of γ-tocotrienol, with 3 μg per g dry matter) and plastochromanol-8 (4.5 μg per g dry matter). While short-term cold shock did not negatively influence α-tocopherol contents, it increased those of γ-tocopherol, γ-tocotrienol, and plastochromanol-8 and decreased those of δ-tocotrienol. Furthermore, storage of Bacon avocados for 10 d led to a 20% decrease in the contents of α-tocopherol, whereas the contents of other tocopherols, tocotrienols and plastochromanol-8 were not affected. (4) Conclusions: It is concluded that Bacon avocados (i) are very rich in α-tocopherol, (ii) not only contain tocopherols, but also tocotrienols and plastochromanol-8, and (iii) their nutritional vitamin E value is negatively influenced by long-term cold storage.

**Keywords:** avocados (*Persea americana* Mill.); low temperatures; plastochromanol-8; tocotrienols; tocopherols; tocochromanols

#### **1. Introduction**

Tocochromanols are a group of amphiphilic molecules that includes tocopherols, tocotrienols and plastochromanol-8 [1,2]. These are all composed by a polar chromanol head and a highly apolar polyprenyl side chain that provide them with the capacity to exert an antioxidant function in membranes, from cyanobacteria and plants where they are synthetized until a variety of tissues in animals and humans, which incorporate tocochromanols regularly from their daily dietary intake [1,2]. While tocopherols have a saturated phytyl-derived side chain, tocotrienols and plastochromanol-8 tails are more unsaturated since they derive from geranylgeranyl-diphospate and solanesyl-diphospate, respectively [3]. Both tocopherols and tocotrienols include various homologues according to the

position and methylation degree in the chromanol head, thus identifying four different molecules for each group (α-, β-, γ- and δ-tocopherols and -tocotrienols). All tocochromanols exert an efficient antioxidant activity by inhibiting the propagation of lipid peroxidation through scavenging lipid peroxyl radicals and by preventing it through the (physical) quenching and (chemical) scavenging of singlet oxygen, not only in plant tissues where they are synthetized, but also in humans, although such a role is mainly generally attributed to α-tocopherol in the human body, since this is the major form transported by a specific protein [4].

While α-tocopherol is a universal molecule found in plants, animals and humans, tocotrienols have not been described in all photosynthetic tissues; rather, they accumulate in seeds and fruits of some plant species only [5,6]. Similarly, plastochromanol-8 is not present universally, despite being found in many plant tissues such as seeds, leaves, buds, flowers and fruits of several species [2,7,8]. Interestingly, in vitro studies in hydrophobic solvents show a higher antioxidant activity against singlet oxygen for tocotrienols and plastochromanol-8 than for tocopherols, which has been attributed to their more apolar structure of the side chain [2]. New nutraceutical functions have been recently attributed to tocotrienols. For instance, antiangiogenic properties against osteoporosis, atherosclerosis, inflammatory processes and many types of cancer (like colorectal, prostate, lung and pancreas cancer) have been reported, mainly for δ- and γ-tocotrienol forms [9]. Otherwise, despite not being considered a molecule belonging to the vitamin E family, the antioxidant activity of plastochromanol-8 is of great relevance in plants and it may also probably display beneficial properties in humans, although to our knowledge this has not been studied thus far in detail.

Cold storage of fruits is an effective means to prevent food deterioration, particularly in climacteric fruits such as avocados (*Persea americana* Mill), where low temperatures reduce ethylene production and therefore inhibit fruit over-ripening [10,11]. Indeed, introducing avocado fruits in cold chambers is a common technique implemented for preventing their early deterioration, thus increasing fruit marketability. However, the cellular redox balance in fruits may be threatened by the extent of low temperature exposure in storage chambers before fruits reach the consumers. Indeed, low temperature shocks or long-term cold exposure can cause a loss of cellular antioxidant defenses in fruits [12–14]. As a result, oxidative reactions may occur in an uncontrolled manner resulting in sustained oxidative stress and tissue damage [13]. To fight against oxidation reactions at low temperatures, tocopherols, mainly α-tocopherol, have been suggested to be essential for attaining acclimation in plants [3]. Additionally, some studies have reported that tocotrienols may be as effective as tocopherols in protecting leaves from photooxidation processes under low temperatures [1]. Although several studies have been performed in leaves of *Arabidopsis* and other model plant species to link tocopherol accumulation and low temperature acclimation [15–18], very few studies have investigated thus far the influence of cold temperature storage on the accumulation of tocochromanols in fruits; except for an increase in tocopherols upon 3 d of storage at cold temperatures in sweet cherries, which is a fruit with low concentrations of fatty acids and vitamin E [19] and a maintenance of constant α-tocopherol contents in oil of Fuerte avocados upon exposure to 5 ◦C for three weeks [20]. To our knowledge, the effects of low temperatures on the accumulation of tocochromanols in other avocado varieties, and of tocotrienols and plastochromanol-8 in fruits in general have not been investigated thus far.

Avocados (*Persea americana* Mill) are highly valuable fruits with increasing interest for consumers thanks to their nutraceutical properties due to antioxidant contents, such as ascorbate (vitamin C, 8.8 mg/100 g) and B-type vitamins (such as vitamin B6, 0.29 mg/100 g), fiber (6.8 g/100 g), phytosterols (83.1 mg/100 g), monounsaturated fatty acids (9.8 g/100 g) and vitamin E (2.36 mg/100 g) [21]. These are originally from Mexico where tropical environmental conditions permitted hybridization techniques that have led to the wide number of varieties currently found worldwide [22]. Nevertheless, the main producing countries are still those with warm climates which export a huge percentage of their production to other continents [23]. Hass is the most widely produced and commercialized avocado variety worldwide, while Bacon, a hybrid variety originally cultivated in 1954 by James Bacon in California, occupies the third position (after Hass and Fuerte) in terms of agricultural production in

Spain, the main avocado producer in Europe [24]. Here, avocado fruits (*Persea americana* Mill) were investigated aiming to determine (i) the amounts and composition of tocochromanols in the edible part of various avocado varieties and (ii) how cold storage in the short and long term influences tocochromanol contents in Bacon avocados. This study shows to what extent cold storage implemented along the supply chain can negatively influence the nutritional quality of avocados in terms of vitamin E accumulation.

#### **2. Material and Methods**

#### *2.1. Plant Material and Samplings*

Avocados (*Persea americana* Mill.), either collected at commercial harvest maturity in the field or purchased from supermarkets in a non-ripe stage (depending on the experiments, see details below), were immediately brought to the laboratory at the University of Barcelona (Barcelona, NE Spain) and used for assays. In all cases, fruits were selected for homogeneity according to their size and lack of pathogen symptoms. Three independent experiments were performed.

For the identification of tocochromanols in various avocado varieties and origins (experiment 1), fruits were purchased in local supermarkets or markets, as follows. Non-ripe Hass avocados originated from Brazil, Perú and Spain, and identified as such in their label, were obtained from a supermarket in Barcelona (NE Spain) and immediately transported at room temperature by car to the laboratory. Non-ripe Hass and Fuerte avocados from Chile and Govín avocados from Cuba were obtained from local markets and transported by plane and car to the laboratory. Finally, Bacon avocados were obtained from a commercial orchard in Málaga (south Spain) at a mature stage and brought to the laboratory after 12 h of transportation at 8–10 ◦C. All fruits were then exposed to room temperature in the laboratory at the University of Barcelona and when firmness attained levels of 3N for all fruits from all varieties and origins, then the mesocarp tissue of four fruits per variety and origin was sampled and immediately frozen in liquid nitrogen and stored at −80 ◦C until analyses. With these samples, tocochromanols were quantified by high-performance liquid chromatography (HPLC) and identification of compounds confirmed by liquid chromatography coupled to electrospray ionization mass spectrometry in tandem (LC-ESI-MS/MS).

The influence of short-term, cold shock exposure on tocochromanol accumulation in Bacon avocados (experiment 2) was examined by performing samplings just before and after 4 h of cold shock of fruits at 4 ◦C in a cold storage chamber (Frimatic, S.A., Barcelona, Spain). Samples from 18 randomly selected fruits at each time point including 0 h and 4 h were immediately frozen in liquid nitrogen and stored at −80 ◦C until analyses. With these samples, tocochromanols were quantified by HPLC while the extent of lipid peroxidation and changes in photosynthetic pigments were estimated spectrophotometrically, as described below.

The influence of long-term exposure to low temperatures on tocochromanol accumulation in Bacon avocados (experiment 3) was examined by performing samplings just before and during exposure for a period of 10 d of cold storage of fruits at 4 ◦C using the same cold chamber (Frimatic, S.A.). Mesocarp samples from 18 randomly selected fruits at times including 0 d, 2 d, 5 d, 7 d and 10 d of low temperature exposure were immediately frozen in liquid nitrogen and stored at −80 ◦C until analyses. With these samples, tocochromanols were quantified by HPLC while the extent of lipid peroxidation and changes in photosynthetic pigments were estimated spectrophotometrically, as described below.

#### *2.2. Tocochromanol Analyses*

The quantification of the different tocochromanol forms, including tocopherols, tocotrienols and plastochromanol-8, was performed as described previously [25] with some modifications. One-hundred mg of avocado (mesocarp) sample was extracted with 1 mL of methanol containing 0.01% (*w*/*v*) butyl-hydroxytoluene (BHT) and 5 ppm (*w*/*v*) of tocol as an internal standard. Extraction was performed using 30 min of ultrasonication (Bransonic ultrasonic bath 2800, Emerson Industrial, Danbury, CT, USA) just after vortexing for 20 s. Then, samples were centrifuged at 600 *g* during 10 min at 4 ◦C to subsequently recover supernatants with a hydrophobic PTFE filter 0.22 μm (Phenomenex, Torrance, CA, USA). Tocochromanols were separated by HPLC at room temperature using an Inertsil 100A column (5 μm, 30 × 250 mm, GL Sciences Inc., Tokyo, Japan). Quantification was performed using a Jasco fluorescence detector (FP-1520, Tokyo, Japan) and a calibration curve established with each of the tocochromanols analyzed and corrected with the tocol recovery, which was always above 97%.

The identification of tocochromanols was confirmed by using high-performance liquid chromatography coupled to electrospray ionization mass spectrometry in tandem (LC-ESI-MS/MS) as described previously [26]. Methanolic extracts were obtained as described before for the HPLC analyses and used here for the identification of tocochromanols by LC-ESI-MS/MS. Tocochromanol separation was performed with an Inertsil 100A column (5 μm, 30 × 250 mm, GL Sciences Inc. (Tokyo, Japan), and an isocratic flow of hexane:dioxane (95.5:4.5 *v*/*v*) mobile phase. The MS acquisition was performed using negative ionization between *m*/*z* 100 and 650, with the Turbo Ionspray source. In addition, quadrupole time-of-flight (QqToF) mass spectrometry was used to obtain product ion information. The MS parameters were: ion spray voltage, −4200; declustering potential (DP), −40; focusing potential (FP), −150; declustering potential two (DP2), −10; ion release delay (IRD), 6 V; ion release width (IRW), 5 ms; nebulizer gas, 50 (arbitrary units); curtain gas, 60 (arbitrary units), and auxiliary gas N2, 6000 cm<sup>3</sup> min−<sup>1</sup> heated at 500 ◦C.

#### *2.3. Lipid Peroxidation Assays*

To determine the extent of lipid peroxidation, primary (lipid hydroperoxide) and secondary (malondialdehyde, MDA) lipid peroxidation products were analyzed, as follows. For lipid hydroperoxides analyses, frozen samples (100 mg) were repeatedly (three times) extracted with 1 mL methanol + 0.01% BHT (*w*/*v*) at 4 ◦C using 30 min of ultrasonication (Bransonic ultrasonic bath 2800). After centrifugation, supernatants were collected, combined and used for analyses using the Fox-2 reagent (consisting in a solution of 90% methanol (*v*/*v*) containing 25 mM sulfuric acid, 4 mM butylhydroxyltoulene (BHT), 250 μM iron sulfate ammonium (II) and 10 μM xylenol orange) as described in Bou et al. [27]. Absorbances were measured at 560 nm and 800 nm. A calibration curve using hydrogen peroxide 37% (*v*/*v*) was used for quantification.

For estimation of the MDA content, the thiobarbituric acid-reactive substances (TBARS) assay, which considers the possible influence of interfering compounds, was used [28]. In short, 100 mg of sample was extracted with 3 mL of ethanol 80% (*v*/*v*) containing 0.01% (*w*/*v*) BHT, vortexed for 20 s and exposed to ultrasonication for 15 min (Bransonic ultrasonic bath 2800). After centrifuging at room temperature for 13 min at 600 *g*, the supernatant was recovered, and the pellet re-extracted twice using the same procedure. Then, two tubes were used: (a) − TBA, with 1 mL extract + 1 mL 20% trichloroacetic acid (*w*/*v*) with 0.01% BHT (*w*/*v*) and (b) + TBA, with 1 mL extract + 1 mL 20% trichloroacetic acid (*w*/*v*), 0.01% BHT (*w*/*v*) and 0.65% thiobarbutiric acid (*w*/*v*). Tubes were incubated for 25 min at 95 ◦C and then the reaction was stopped by maintaining them at 4 ◦C for 10 min. After centrifugation at 600 *g* at room temperature for 5 min, MDA content in samples were analyzed by spectrophotometry at 440, 532 and 600 nm and quantified using the equations developed by Hodges et al. [28].

#### *2.4. Chlorophyll Content*

To determine total chlorophyll content, samples (100 mg) were extracted in 1 mL of methanol + 0.01% BHT as explained before, using vortex and ultrasonication for 30 min at 4 ◦C. Supernatants were collected after centrifugation for 10 min at 600 *g* and 4 ◦C. Chlorophylls were measured spectrophotometrically reading absorbances at 653, 666 and 750 nm and measuring chlorophyll content as described [29].

#### *2.5. Statistical Analyses*

Statistical analyses were performed by one-way ANOVA and Tukey posthoc tests were used for multiple comparisons among time (IBS SPSS Statistics 19; SPSS Inc., Chicago, IL., USA). Differences were considered significant when *p* values were under the significance level α = 0.05.

#### **3. Results**

#### *3.1. Identification of Tocochromanols in Various Avocado Varieties*

In order to determine the presence and amount of tocochromanols in the mesocarp (edible part of the fruit) of different avocado varieties, HPLC and LC-ESI MS/MS analyses were performed. Among the various varieties and origins tested, Bacon avocados from Spain showed the largest amount of vitamin E (Table 1A). Bacon avocados contained 2.4 mg α-tocopherol per 100 g of edible fruit, which coincided with a very low quantity of its precursor γ-tocopherol in the tissue (Table 1A). Bacon was the variety with the highest tocochromanol content among all studied varieties. By contrast, Govín from Cuba, Hass from Brazil and Hass from Perú were the avocados showing the lowest amounts of total tocochromanols.



(**A**) Per 100 g fresh weight and (**B**) per g dry weight (DW). Data, which were obtained using the mesocarp of fruits in their optimum stage of ripening, show the mean of *n* = 4 fruits. Lower case letters (a–d) indicate differences between avocado varieties when *p* < 0.05. Trace amounts of δ-tocopherol and α-tocotrienol could not be properly quantified and are not shown here. T, tocopherol; TT, tocotrienol; PC-8, plastochromanol-8; ND, not detected.

A comparison of four origins of Hass avocados (Chile, Spain, Perú and Brazil) revealed that the origin had a very strong effect on tocochromanol contents, including α-tocopherol (Table 1). All avocado varieties behaved similarly to Bacon avocados from Spain in terms of accumulating most of the tocochromanols in the form of α-tocopherol but both Bacon from Spain and Govín from Cuba presented a larger amount of α-tocopherol than the highly commercialized Hass variety irrespective of its origin. Furthermore, although plastochromanol-8 was present in all avocado varieties, its contents were higher in Hass varieties (irrespective of the origin) than in Bacon and Govín (from Spain and Cuba, respectively). Notably, δ-tocotrienol seemed to be exclusively present in Bacon (Table 1A). Results differed slightly when comparing the vitamin E amounts per unit of dry weight in different varieties; Bacon occupied the second position in terms of vitamin E accumulation, just after Govín, as the contents of α-tocopherol were higher in these two varieties than in Hass or Fuerte (Table 1B). Moreover, total tocochromanol contents were also higher in Bacon when compared to the other highly commercialized varieties, except for Govín which occupied the first position just before Bacon variety (Table 1B).

The major tocochromanol present in the mesocarp (edible tissue) of Bacon avocados was α-tocopherol (with an 87.8%), as clearly observed in the HPLC chromatogram (Figure 1A), followed by plastrochromanol-8, β- and γ-tocopherols, and δ- and γ-tocotrienols (Table 1A). HPLC identification by retention time was confirmed by LC-ESI MS/MS using the corresponding authentic standards, which showed exactly the same fragmentation patterns as the corresponding peaks in the samples (Figure 1). This tocochromanol profile in Bacon avocados, enriched in the α-tocopherol form and with the presence of δ-tocotrienol, is different from that found for the Hass variety (Table 1, see also [30,31]).

**Figure 1.** (**A**): Separation (by HPLC, *left*) and identification (by liquid chromatography coupled to electrospray ionization mass spectrometry in tandem [LC-ESI-MS/MS], *center* and *right*) of tocochromanols in Bacon avocados. (**B**): Chemical formula of tocochromanols identified in Bacon avocados.

#### *3.2. Cold-induced Changes in Tocochromanol Composition in Bacon Avocados*

After a low temperature shock for 4 h, the contents of the major tocochromanol present in the mesocarp of Bacon avocados, α-tocopherol, were not altered (Figure 2). The same was observed for β-tocopherol, but not for the other tocochromanols. While the contents of plastochromanol-8, γ-tocopherol and γ-tocotrienol increased, those of δ-tocotrienol decreased, with the latter showing a reduction by 60% under cold treatment (Figure 2). This cold-induced shift in the tocochromanol composition was accompanied by an increase in the extent of lipid peroxidation, as indicated by 60% increases in lipid hydroperoxides and malondialdehyde contents, while chlorophyll levels and the chlorophyll a/b ratio remained unaltered (Figure 3).

**Figure 2.** Influence of short-term (4 h) exposure of Bacon avocados to cold temperatures (4 ◦C) in the contents of tocochromanols. Data represent the mean ± SE of *n* = 18 fruits. Differences were considered significant when *p* < 0.05. DW, dry weight.

**Figure 3.** Influence of short-term (4 h) exposure of Bacon avocados to cold temperatures (4 ◦C) in the contents of chlorophylls, chlorophyll a/b ratio and the extent of lipid peroxidation (estimated as the contents of lipid hydroperoxides and malondialdehyde, as indicators of primary and secondary lipid peroxidation, respectively). Data represent the mean ± SE of *n* = 18 fruits. Differences were considered significant when *p* < 0.05. DW, dry weight.

Total tocochromanol contents showed a decrease by 16% after 10 d of storage at low temperatures, which was mostly due to a significant decrease in tocopherols but not tocotrienols (Figure 4). Part of this loss was related to the decrease in the major tocochromanol form in Bacon avocados, α-tocopherol, which decreased by 20% after 10 d of cold storage (Figure 4). When α-tocopherol contents were expressed on a fresh weight basis (either per 100 g FW, per fruit, half fruit or serving), a decrease in its contents was also observed, thus offering a lower amount (by 15%) of α-tocopherol per amount of fruit consumed (Figure 4). This reduction in vitamin E contents occurred progressively over time, as revealed by the time-course evolution of α-tocopherol contents (Figure S1), but most particularly between 5 d and 10 d of cold storage. In contrast to short-term exposure to cold temperature, the other tocochromanol forms were not clearly affected by long-term cold storage, although γ-tocopherol and δ-tocotrienol showed slight variations over time (Figure S1). Reductions of α-tocopherol during long-term storage at low temperatures was coincident with a 3.4-fold increase in malondialdehyde contents after 10 d of cold storage (Figure S2).

**Figure 4.** Influence of long-term (10 d) storage of Bacon avocados at cold temperatures (4 ◦C) in the contents of total tocochromanols, tocopherols, tocotrienols and α-tocoperol. Contents of α-tocopherol are also shown in mg per 100 g of mesocarp in fresh weight (FW), per one fruit (136 g FW), per a half (68 g FW) and per serving according to the Nutrition Labelling and Education Act (NLEA; corresponding to 30 g FW). Data represent the mean ± standard error of *n* = 18 fruits. Differences were considered significant when *p* < 0.05. DW, dry weight.

#### **4. Discussion**

#### *4.1. Bacon is a Variety of Avocados with High Tocochromanol Contents*

Among the avocado varieties examined in our study, Bacon from Spain was the one showing the highest vitamin E content. All varieties showed a similar tocochromanol composition, so that the major tocochromanol was α-tocopherol, except Bacon, which also accumulated some amounts of δ-tocotrienol. Hence, tocochromanols composition was in general enriched in α-tocopherol, with a diminished accumulation of other tocopherols and tocotrienols, with the overall most notable exception of γ-tocochromanols in all varieties and additionally of δ-tocotrienol in Bacon. Varietal differences might be associated not only with the geographical origin of the fruit, as shown in our results, but also with the highly heterozygous genetic origin of avocado races. *Persea americana* includes *P. americana var. drymifolia* (commonly known as Mexican race)*, var. guatemalensis* (known as Guatemalan race) and *var. americana* (or West Indian race). While Bacon is obtained from the hybridization of Mexican x Guatemalan races, Hass variety is generally reported to have a pure Guatemalan origin [32]. However, breeding strategies, cross- and self-pollination techniques, different strategies of cultivation and the posterior selection according to farmer preferences, like high yield, fruit quality and long shelf life, usually give rise to quite heterogenic crops in the same variety, which might lead to the observed

differences in the Hass avocados from different origins studied here that showed notable differences in the accumulation of tocochromanols.

Plastochromanol-8 accumulation in fruits may be of particular relevance since this is also a powerful antioxidant, even showing higher antioxidant activity than α-tocopherol in hydrophobic environments due to its more highly unsaturated prenyl chain [2]. Plastochromanol-8 was found in mesocarp tissue of avocado fruit at relatively low amounts compared to α-tocopherol, but some differences between cultivars were observed. In this case, Hass was the variety with the greatest amount of this compound compared to the other studied varieties, including Bacon. Furthermore, we reported here on the accumulation of tocotrienols in avocados, which contrasts with a recent report [8] showing the accumulation of plastochromanol-8 but not of tocotrienols in Hass avocado. Our study shows that tocotrienols may accumulate in avocado fruits, in particular in some varieties such as Bacon. Notably, δ-tocotrienol was only found in Bacon among all studied varieties. Beneficial properties for humans have recently been attributed to this compound, in particular to help in the prevention of the development of various cancers, including breast, colorectal, lung and many other types of cancer, apart from providing anticholesterolemic and antidiabetic benefits [33–35]. Although the contents of δ-and γ-tocotrienols were relatively low compared to that of α-tocopherol in Bacon avocados, these compounds might, to some extent, exert an additional beneficial response in the human body, an aspect that deserves further investigations.

#### *4.2. E*ff*ects of Short and Long-Term Storage on Tocochromanol Contents*

While a low temperature shock for 4 h did not alter α-tocopherol contents in the mesocarp of Bacon avocados, long-term storage for 10 d led to significant decreases in vitamin E contents. In contrast to unaltered contents of tocopherols after 4 h of low temperature exposure, the contents of plastochromanol-8, γ-tocopherol and γ-tocotrienol increased, and those of δ-tocotrienol decreased, the latter showing a reduction by 60% under cold treatment. Therefore, cold shock led to significant reductions in the levels of δ-tocotrienol and the cold-induced shift in the tocochromanol composition was accompanied by an increase in the extent of lipid peroxidation, as indicated by 60% increases in lipid hydroperoxides and malondialdehyde contents. Interestingly, chlorophyll contents were unaltered during the same period and Bacon avocados stored for 5d did not show alterations in the extent of lipid peroxidation, as indicated by the same measurements. This suggests that the cold-induced shift in tocochromanol composition was mainly due to metabolic alterations that resulted in transient lipid peroxidation, but this was not accompanied with a quality loss. In contrast, α-tocopherol contents decreased during long-term storage of Bacon avocado fruits at low temperatures, a decrease that was accompanied by an increase in the extent of lipid peroxidation, which was reflected by an increase in malondialdehyde. This result contrasts with a previous study [20] showing that α-tocopherol in oil obtained from Fuerte avocados keeps stable after 3 weeks of storage at 5 ◦C. This difference may be due to different reasons, including not only the study of a different variety (Fuerte in [20] and Bacon in our study), but also to a higher stability of α-tocopherol in oil at 5 ◦C [20] than in entire fruits at 4 ◦C in our study. Unfortunately, little research has been performed thus far to evaluate how cold storage temperature influences tocochromanol composition in vitamin E-rich fruits and further studies are required to better understand the causes of vitamin E instability in avocado fruits and oils.

#### *4.3. Balance between Storage and Nutritional Value*

According to the Nutritional Labelling and Education Act (NLEA) and the National Health and Nutrition Examination Survey (NHANES), the serving of avocado is recommended to be of 30 g or half of an avocado, respectively, which corresponds to an intake of 0.59 mg and 1.34 mg of α-tocopherol, respectively. Interestingly, when the nutritional value in terms of vitamin E was measured over time of cold storage, a decrease in tocochromanols occurred, which was mostly attributed to a progressive drop in α-tocopherol content, so that the daily vitamin E intake is significantly lower if avocados are stored for 10 d at 4 ◦C. In contrast, other tocochromanols were not affected. According to the results

presented in our study, avocados stored for 10 d at low temperatures start to suffer oxidation processes, which might be related to the stress situation experienced by the mesocarp due to cold temperatures, which can lead to oxidative damage. Indeed, α-tocopherol levels dropped up to 20% after 10 d of cold storage, and this loss of α-tocopherol contents may slightly contribute to a lesser intake and absorption of vitamin E in the human diet under the levels of reference [36]. Furthermore, the loss in detoxifying oxygen radicals function by a loss of antioxidants such as vitamin E due to cold storage of fruits for long periods may contribute to a higher risk of suffering from cardiovascular diseases like atherosclerosis, cancer and cataracts, among other diseases related to degenerative processes [37–40].

#### **5. Conclusions**

In conclusion, Bacon has been shown to be the variety with very high tocochromanol contents relative to other studied varieties, presenting values greater than those of the highly commercialized Hass variety. Furthermore, Bacon variety marketing should be fostered not only because of the high amounts of vitamin E but also because it was the only variety showing δ-tocotrienol, a compound that might have additional beneficial effects. Moreover, according to procedures implemented along the supply chain which consist of introducing fruits into cold chambers, our study showed that 10 d might be the threshold where cold stress is starting to induce losses in vitamin E, hence decreasing nutritional value and fruit quality.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-3921/9/5/403/s1, Figure S1: Variations in the contents of tocochromanols during long-term (10 d) storage of "Bacon" avocados at cold temperatures (4 ◦C), Figure S2: Variations in the contents of chlorophylls, chlorophyll a/b ratio and the extent of lípid peroxidation (estimated as the contents of lípid hydroperoxides and malondialdehyde, as indicators of primary and secondary lípid peroxidation, respectively) during long-term (10 d) storage of "Bacon" avocados at cold temperatures (4 ◦C).

**Author Contributions:** Conceptualization, C.V. and S.M.-B.; methodology, C.V. and T.M.; software, C.V.; validation, C.V., T.M. and S.M.-B.; formal analysis, C.V.; investigation, C.V. and T.M.; resources, S.M.-B.; data curation, C.V.; writing—original draft preparation, C.V. and S.M.-B.; writing—review and editing, C.V., T.M. and S.M.-B.; visualization, C.V.; supervision, S.M.-B.; project administration, S.M.-B.; funding acquisition, S.M.-B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Generalitat de Catalunya, grant number 2017 SGR 980.

**Acknowledgments:** We are very grateful to Paula Muñoz and Maren Müller for their help in the quantification and identification of tocochromanols, respectively, to Camila Ribalta for providing the avocados from Chile, and to Marina Pérez and Andrea Casadesús for their help in samplings.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Chemical Profile and Antioxidant Activity of the Kombucha Beverage Derived from White, Green, Black and Red Tea**

#### **Karolina Jakubczyk, Justyna Kałdu ´nska, Joanna Kochman and Katarzyna Janda \***

Department of Human Nutrition and Metabolomics, Pomeranian Medical University in Szczecin, 24 Broniewskiego Street, 71-460 Szczecin, Poland; jakubczyk.kar@gmail.com (K.J.); justynakaldunska@wp.pl (J.K.); kochmaan@gmail.com (J.K.)

**\*** Correspondence: Katarzyna.Janda@pum.edu.pl; Tel.: +48-091-441-4818

Received: 27 April 2020; Accepted: 21 May 2020; Published: 22 May 2020

**Abstract:** Kombucha is a fermented tea beverage prepared as a result of the symbiotic nature of bacterial cultures and yeast, the so-called SCOBY (*Symbiotic Cultures of Bacteria and Yeasts*). Kombucha is characterised by rich chemical content and healthy properties. It includes organic acids, minerals and vitamins originating mainly from tea, amino acids, and biologically active compounds—polyphenols in particular. Kombucha is prepared mainly in the form of black tea, but other tea types are increasingly often used as well, which can significantly impact its content and health benefits. This work shows that the type of tea has a significant influence on the parameters associated with the antioxidant potential, pH, as well as the content of acetic acid, alcohol or sugar. Red tea and green tea on the 1st and 14th day of fermentation are a particularly prominent source of antioxidants, especially polyphenols, including flavonoids. Therefore, the choice of other tea types than the traditionally used black tea and the subjection of these tea types to fermentation seems to be beneficial in terms of the healthy properties of kombucha.

**Keywords:** kombucha; tea; fermentation; antioxidant; flavonoids; polyphenols

#### **1. Introduction**

Unhealthy lifestyle, intense physical exercising, stress, and environmental pollution are factors that influence the excessive synthesis of reactive oxygen species. The disturbance in homeostasis caused by free radicals leads to the formation of oxidative stress and damage to the structures of the human organism [1–4]. Illnesses that can be caused by free radical disorders include atherosclerosis, neurodegenerative diseases, such as Parkinson's or Alzheimer's disease, or even obesity. In order to maintain the balance between the production and removal of reactive oxygen species, it is important to search for easily accessible sources of antioxidants [1]. The main and the most widespread antioxidants are vitamins E, A and C, as well as polyphenolic compounds [1–3]. Phenolic compounds are an essential part of the human diet and are of considerable interest due to their health-promoting properties, including antioxidant effects. They are capable of capturing peroxide anions, lipid radicals, hydroxyl radicals and reactive oxygen species. Plant-derived polyphenols have a beneficial impact on slowing down the ageing process and reducing the risk of age-related neurodegenerative conditions, such as Alzheimer's disease, Parkinson's disease or ischaemic brain injury [5,6]. Antioxidant sources are mainly searched for in natural plant resources. Antioxidants are present in many easily available sources, such as tea, coffee, fruits, vegetables, spices and herbs. They complement everyday diet, contributing to good health.

Kombucha is a fermented tea drink created with the use of symbiotic cultures of bacteria and yeast, the so-called SCOBY (*Symbiotic Cultures of Bacteria and Yeasts*). Kombucha is prepared by combining tea

with sugar (10%), sourdough from previous fermentation (10%) and SCOBY. SCOBY, when added to sugared tea, initiates fermentation, which results in the formation of various new bioactive compounds. The fermentation is conducted at room temperature for a period of 7–14 days. Various tea types can be used to produce kombucha, including green tea, as well as fermented, e.g., red, black or yellow tea. However, black tea and white sugar (saccharose) are considered the traditional and best ingredients that condition the proper content of the drink as well as its healthy properties. The taste of the drink is described as sour, slightly fruity and delicately sparkling, but after a few days of storage it becomes similar to the taste of wine vinegar [7].

Studies of kombucha proved its anti-bacterial, antioxidant, anti-diabetic properties, as well as its ability to reduce the concentration of cholesterol, to support the immune system and to stimulate the detoxification of the liver [8,9]. Kombucha drinks also feature minerals originating mainly from tea (potassium, manganese, fluoride ions), vitamins (E, K, B), amino acids (especially theanine, a derivative of glutamine), as well as other compounds that are formed as the result of numerous reactions occurring during the fermentation of the tea. During the oxidation of polyphenolic compounds, catechins, flavonoids and other compounds with health benefits for the organism are formed [8,10,11].

Various parameters influence the properties and content of kombucha, including the type of tea, fermentation time, the content of SCOBY colonies, and temperature. Despite the increase in the popularity of this drink's consumption, information regarding the influence of the many parameters or tea types on the properties and content is still not fully available. Hence, the aim of this research was to analyse the antioxidant properties and the content of the drink prepared from black, green, white and red teas at different time points of fermentation [10].

#### **2. Material and Methods**

#### *2.1. Plant Material*

The material consisted of four types of leaf tea (*Camellia sinensis*): black Ceylon originating in India, green Gunpowder, white tea and red tea (Pu-ERH) originating in China or India.

#### *2.2. Preparation of Kombucha*

The kombucha starter cultures, also known as SCOBY (which generally consists of *Acetobacter xylinum*, *Gluconobacter*, *S. cerevisiae*), were obtained from a commercial source from Poland. The starter culture used in the present article was stored in a refrigerator (4 ◦C) and consisted of sour broth and cellulosic layer (SCOBY floating on the liquid surface). One hundred grams of sugar (100.0 g/L, 10.0%), eight grams of tea (8.0 g/L, 0.8%) and 1 litre of hot, distilled water (90 ◦C) were mixed. The solution was infused for 10 min in a sterile conical flask. After cooling (30 ◦C), the tea decoction was filtered through nylon filters (0.45 μm, diam. 25 mm, Sigma-Aldrich, Pozna ´n, Poland) into clean glass bottles.

#### *2.3. Fermentation of Kombucha*

Kombucha cultures were kept under aseptic conditions. Fermentation was carried out by incubating the kombucha culture at 28 ± 1 ◦C for 1, 7 and 14 days. Replicates were prepared so that each replicate was completely collected after its stipulated period of fermentation. The kombucha obtained was filtered and analysed.

#### *2.4. Antioxidant Activity by the DPPH Methods*

The antioxidant activity of samples was measured with the spectrophotometric method using synthetic radical DPPH (2.2-Diphenyl-1-Picrylhydrazyl, Sigma, Pozna ´n, Poland) according to Brand-Williams et al. and Pekkarinen et al. [12,13]. The spectral absorbance was immediately measured at 518 nm (Agilent 8453UV). All assays were performed in triplicate. The results are shown in % of DPPH radical inhibition.

Antioxidant potential (antioxidant activity, inhibition) of tested solutions has been expressed by the percent of DPPH inhibition, using the following formula:

$$\% \text{ inhibition} = \frac{A0 - A\text{s}}{A0} \times 100$$

where:

A0—absorbance of DPPH solution at 518 nm without tested sample As—absorbance of DPPH solution at 518 nm with tested sample

#### *2.5. The Determination of the Ferric Ion Reducing Antioxidant Power (FRAP) Method*

The FRAP method, used to determine the total reduction potential, which also means the antioxidant properties of tested ingredient, is based on the ability of the test sample to reduce Fe3<sup>+</sup> ions to Fe2<sup>+</sup> ions. The FRAP unit determines the ability to reduce 1 micromole Fe3<sup>+</sup> to Fe2<sup>+</sup> according to Benzie and Strain [14,15]. Absorbance at 593 nm was measured (8453UV, AGILENT TECHNOLOGIES, Santa Clara, CA, USA). All assays were performed in triplicate. The ferric ion reducing antioxidant power was determined from the calibration curve using Fe(II)/L as the reference standard (0–5000 μM Fe(II)/L).

#### *2.6. The Determination of the Total Polyphenols Content (TPC)*

Determination of polyphenols was performed according to ISO 14502-1; Singleton and Rossi method using the Folin-Ciocalteu reagent [16]. Absorbance at 765 nm was measured (8453UV, AGILENT TECHNOLOGIES, Santa Clara, USA). All assays were performed in triplicate. The content of polyphenols was determined from the calibration curve using gallic acid as the reference standard (0–200 mg/L of gallic acid).

#### *2.7. The Determination of the Total Flavonoids Content (TFC)*

Determination of total flavonoids content was performed according to the P ˛ekal and Pyrzynsk and Hu methods [17,18]. Different concentrations of flavonoids were used in the plotting of the standard calibration curve. The content of flavonoids was determined from the calibration curve using rutin equivalent as the reference standard (0–120 mg/L of rutin equivalent). Absorbance at 510 nm was measured (8453UV, AGILENT TECHNOLOGIES, Santa Clara, CA, USA). All assays were performed in triplicate.

#### *2.8. The Determination of pH*

The pH of both the fermented beverage and the unfermented control was determined by a pH meter (SCHOTT Instruments; SI Analytics Mainz, Mainz, Germany).

#### *2.9. The Determination of Acetic Acid*

Samples of tea and kombucha at 1, 7 and 14 days of fermentation were filtered through nylon filters (0.45 μm, diam. 25 mm, Sigma-Aldrich, Pozna ´n, Poland). Acetic acid (AA) was analysed by high performance liquid chromatography (HPLC) using a 1200 series HPLC connected to a 1100 series RI detector (Agilent Technologies, Santa Clara, CA, USA) with a Rezex ROA-Organic Acid H<sup>+</sup> (8%) column (Phenomenex, Torrance, CA, USA). The column was eluted with a degassed mobile phase containing 5 mM H2SO4, pH 2.25 at 60 ◦C with a flow rate of 0.5 mL/min for 30 min per sample [19,20]. The results are shown in mg acetic acid/L.

#### *2.10. The Determination of Alcohol*

The alcohol content was measured using an alcoholometer (Browin, Łód´z, Poland). The alcoholometer was immersed in the liquid and the result was read from the scale.

#### *2.11. The Determination of Sugar Content*

The total sugar content was measured with a laboratory refractometer RL3 (Polish Optical Works, Warsaw, Poland) from Brix scale.

#### *2.12. Statistical Analysis*

In all the experiments, three samples were analysed, and all the assays were carried out at least in triplicate. The statistical analysis was performed using Stat Soft Statistica 13.0 and Microsoft Excel 2017 (StatSoft Polska, Poland. The results are expressed as mean values and standard deviation (SD). To assess the differences between the examined parameters, the Tukey post hoc test was used. Differences were considered significant at *p* ≤ 0.05. To control type I errors, the false discovery rate (FDR) approach was used. The calculations were performed using the p. adjust function of the stats package in R (R Foundation for Statistical Computing, Vienna, Austria).

#### **3. Results**

#### *3.1. The Analysis of the Antioxidant Properties of Kombucha*

The analysis of the antioxidant potential of the studied samples revealed that the content of antioxidant compounds was in the range between 70.62% and 94.61% DPPH inhibition (Table 1). The time of fermentation and the type of tea had an influence on the anti-radical properties of kombucha. In terms of the type of tea, kombucha prepared from green tea was characterised by the highest antioxidant potential, achieving the highest value on the first day of fermentation. In the case of each of the analysed kombucha drinks, the ability to deactivate free radicals decreased with the increase in the time of fermentation.

The highest content of reductive compounds labelled by the FRAP method was observed in all tea types before the fermentation process (5374.1–4486.7 μM Fe(II)/L). The addition of sourdough caused a rapid decrease in the reductive properties of kombucha (3626.3–2274.0 μM Fe(II)/L), but after 7 days of fermentation, the potential increased (4801.1–2725.9 μM Fe(II)/L), then it became lower on the 14th day of fermentation (3172.9–1573.9 μM Fe(II)/L). When analysing the type of the selected tea, kombucha made from green tea was characterised by the highest reductive potential (Table 1).

The analysis of the total content of polyphenols in kombucha, as well as the tea types used for its preparation, revealed that the content of compounds belonging to this group fluctuated in the range from 183.12 mg/L in black tea before the addition of sourdough and SCOBY to 320.12 mg/L in kombucha prepared from green tea on the 14th day of fermentation. In the case of kombucha from green, red and white teas, the highest polyphenol content was observed on the 14th day of fermentation. In kombucha made from green and white tea, the concentration of polyphenolic compounds increased proportionally with the increase in the duration of fermentation. The content of flavonoids, a compound from the group of polyphenols, was the highest for all tea types before starting the fermentation process (395.93 mg/L in red tea). The addition of sourdough significantly reduced flavonoid content in the analysed samples. The decrease in flavonoid content was progressing, achieving the lowest values on the 7th day of fermentation. During the next labelling (14th day of fermentation), there was another increase in the content of this compound (Table 1). In the case of most of the studied parameters, statistically significant differences were observed between tea types as well as the time of fermentation (Table 1).

Table 2 presents the statistically significant correlations between polyphenol content, flavonoids, antioxidant potential (DPPH, FRAP) and the duration of fermentation. Statistical analysis of the results showed significant correlations between the parameters characterising the kombucha antioxidant potential. It was shown that the correlations between the tested parameters are very different, depending on the type of tea (Table 2).



\* FDR *p* ≤ 0.05 between type of Kombucha (0, 1, 7, 14 days of fermentation), *p* ≤ 0.05 between particular subgroup: 1—GK tea, 2—GK 1, 3—GK 7,4—GK 14,5—BK tea,6—BK 1, 7—BK 8—BK 14, 9—WK tea, 10—WK 1, 11—WK 7, 12—WK 14, 13—RK tea, 14—RK 1, 15—RK 7, 16—RK.



#### *3.2. The Analysis of pH, Content of Acetic Acid, Sugar and Alcohol in Kombucha*

During the analysis of pH values, it was observed that the pH of all of the studied samples decreased with the increase in the duration of fermentation and the increase in the content of acetic acid. The rapid decrease in this parameter (1.8 unit in the case of kombucha prepared from black tea, up to 2.97 in the case of white kombucha) was caused by the addition of sourdough and SCOBY culture (1st day of fermentation). Further fermentation did not have a significant influence on the change in pH values. No significant differences were observed in terms of pH between drinks prepared from different tea types (Table 3).

With time, the acetic acid content of the fermentation increased, regardless of the type of tea used to prepare kombucha. On the 14th day of fermentation, acetic acid concentration was the highest for all tested beverages (9071.02–9147.40 mg/L) (Table 3).

The refractometric analysis of sugar content showed that all of the tea types were characterised by the highest concentration of saccharose before the beginning of the fermentation process. When it comes to kombucha prepared with the use of black or white tea, with progressing fermentation the content of saccharose decreased, achieving the lowest value on the 14th day of fermentation (7.5 and 9.5 ◦Bx, respectively). However, in the case of kombucha made from red and green tea types, the content of saccharose decreased directly after the addition of sourdough, increasing and approaching initial values at the moment of measurement on the 7th day of fermentation. The continuation of the process caused a slow decrease in the content of saccharose in these samples (Table 3).

The concentration of alcohol increased with time, achieving maximum value on the 7th day of fermentation—from 3.0% to 3.5% depending on the tea type. Subsequently, a decrease in alcohol content was observed in all types of kombucha drink (14th day of fermentation).

In the case of most of the studied parameters, statistically significant differences were observed between the time of fermentation. The smallest statistically significant differences were observed between kombucha drinks prepared from various tea types using the same fermentation time (Table 3). Statistical analysis of the results showed significant correlations between the parameters characterising the basic chemical composition of kombucha (Table 4).


**Table 3.** The content of alcohol, sugar, pH and acidity in Kombucha tea. \* FDR *p* ≤ 0.05 between type of Kombucha (0, 1, 7, 14 days of fermentation), *p* ≤ 0.05betweenparticularsubgroups:1—GK0,2—GK1,3—GK7,4—GK14,5—BK0,6—BK1,7—BK7,8—BK14,9—WK0,10—WK1,11—WK7,12—WK14,13—RK



#### **4. Discussion**

The popularity of fermented drinks is increasing as consumers perceive fermentation as a mild method for the preservation of food and value the products themselves for their health benefits. Kombucha, as a fermented tea drink, is consumed not only in Asia, where it originally comes from, but also increasingly often in Europe. It is mainly formed from black tea, but other forms of kombucha made from different tea variants, such as green, white or red tea, are becoming increasingly available on the market. Despite the fact that kombucha has been researched in detail in terms of its microbiological content and antibacterial properties, there are not enough studies regarding the various tea types and their health benefits. This is why our study includes different, most frequently consumed tea types (black, green, white and red) and this is why we analysed the content, antioxidant potential depending on the time of fermentation and the type of tea selected for the preparation of kombucha.

This study has demonstrated that the health benefits as well as the chemical content depend both on the type of tea as well as fermentation time. Kombucha is characterised by high antioxidant potential. Green tea was characterised by the most significant antioxidant properties, slightly lower potential was observed for red and white tea types, whereas black tea featured the lowest values. The same tendency was observed for kombucha prepared from a given tea type. In the case of DPPH, the fermentation process had an influence on the increase in antioxidant properties in reference to tea, and with subsequent days of fermentation the potential decreased regardless of the tea type. A reverse situation was observed in the case of the reductive potential (FRAP). Fermentation had an influence on the decrease in reductive properties with reference to tea. The highest reductive potential was observed for kombucha on the 7th day. Therefore, a strong positive correlation was observed between the time of fermentation and the reductive potential (FRAP) as well as polyphenol content. On the other hand, a negative correlation was observed between the time and the antioxidant potential (DPPH). The differences in the antioxidant potential measured by FRAP and DPPH methods are due to different mechanisms of both methods. In the latter method, the DPPH radical uses the free electron transfer reaction, and the FRAP method utilizes metal ions for oxidation. Additionally, the DPPH method does not allow for the determination of hydrophilic antioxidant activity. FRAP was primarily used to determine the absolute reduction in body fluid. Recently, it has also been adapted for plant-based antioxidant research. In our study, both methods showed high reproducibility. However, the DPPH method has been shown to be more stable [21].

These results are similar to those achieved by Gaggia et al. in a study where the highest antioxidant potential (DPPH) was observed in relation to green tea, slightly lower for white tea, and the lowest for red tea. However, in this case, the 7th day of fermentation had the most positive influence on this parameter. It should be highlighted that the authors did not study kombucha on day one. In all cases, the fermentation process increase the antioxidant properties of the drink [19]. An increase in the antioxidant potential of kombucha in comparison to tea was also observed in the study by Chakravorty et al. [22]. The DPPH and ABTS (2,2'-Azino-bis-3-ethylbenzthiazoline-6-sulphonic acid) radicals' scavenging activity increased by 39.7% and 38.36%, respectively, after 21 days [22]. It was also observed that the microbiological content is the most diverse on the 7th day of fermentation. This might indicate that the increase in the diversity of microorganisms plays a significant role in the increase in the antioxidant properties of kombucha tea. Moreover, the change of the domination of yeast to lactic acid bacteria on the 7th day is also responsible for the increased antioxidant activity [22].

Tea, which is also the main ingredient of the drink, is rich in catechins-theaflavin and tearubigin. Polyphenols present in tea are responsible for the antioxidant activity of kombucha. A positive correlation was observed between the content of polyphenols and reductive potential. This study confirms the observations carried out by Chakravorty et al., in which an increase in polyphenols was observed during fermentation [22]. During fermentation, there is an increase in polyphenols, including flavonoids, whereas tearubigin is transformed into theaflavin, resulting in the change in kombucha's colour from dark to light with the progressing time of fermentation [22]. On the basis of our studies, it can be concluded that the general content of polyphenols depended on the type of

tea. The highest concentration was observed for green tea, slightly lower for red and white tea types, the lowest for black tea. Fermentation time had an influence on the increase in the content of these compounds. Furthermore, an increase in the content of polyphenols in kombucha in comparison to tea alone has also been observed. For kombucha prepared from green and black tea types, the content of polyphenolic compounds increased with the time of fermentation, achieving the highest concentration on the 14th day. Our studies confirm those of other authors. The highest antioxidant potential was also observed in green tea, but on the 7th day of fermentation (100.33 mg/g). Kombucha prepared from red tea included the smallest amount of polyphenols, but they were stable and their concentration did not change during the fermentation process. This kombucha included a lot of flavonoids [19]. The increase in the content of polyphenolic compounds can be associated with numerous reactions occurring during the fermentation of tea, e.g., the oxidation of polyphenolic compounds by some enzymes leads to the formation of catechins, flavonoids and other compounds with healthy properties, including antioxidant properties, which is the result of a microbial hydrolysis reaction [10]. Moreover, microorganisms such as *Candida tropicalis* are able to degrade various polyphenols [23]. Catechins included in the tea can be broken down through the activity of bacteria and yeast into simpler particles, increasing antioxidant strength [10,24]. In addition, fermentation induces the structural breakdown of plant cell walls, leading to the liberation or synthesis of various antioxidant compounds. These antioxidant compounds can act as free radical terminators, metal chelators, singlet oxygen quenchers or hydrogen donors. The production of protease, α-amylase and some other enzymes might be influenced by fermentation possessing metal ion chelation activity [25].

Our study provides an extensive body of evidence that red tea and kombucha are good sources of polyphenols, among them flavonoids with undisputed antioxidant effects Additionally, they helps seal blood vessels, have anti-inflammatory properties and support immune system function [26]. Flavonoids, present in large quantities in red tea, can significantly contribute to its antioxidant properties. A good source of flavonoids also seems to be green tea and kombucha prepared from this variant. However, fermentation contributes to the degradation of this compound. Its highest concentration in red tea subjected to fermentation was observed on day 1 and day 14: 292.54 and 242.5 mg/L, respectively. The value for tea alone was 395.9 mg/L. In comparison, buckwheat—considered as one of the best sources of flavonoids—contains 62.30 mg/100 g of fresh weight of the resource. The tea drink available on the market included only 1.968 mg/L of the resource. Out of 14 of the studied infusions from various tea types, green tea included the highest content of flavonoids—37.13 mg/L [27].

Lactic fermentation is responsible for the breakdown of glucose, which results from the activity of bacteria of lactic fermentation. Another fermentation type is alcoholic fermentation. Yeast, a constituent of the drink's microflora, is responsible for the breakdown of glucose into ethyl alcohol with the appearance of carbon dioxide. Yeast consists of *Schizosaccharomyces pombe*, as well as *Candida krusei* and *Issatchenkia orientalis* [8]. In our study, on day 7, the highest concentration of alcohol was achieved with as much as 3.5% for kombucha prepared from white and red tea types, 3.25% for green tea and 3.0% for black tea. On the 14th day the content of alcohol slightly decreased in the case of all of the studied variants to the level of 2–3%. In a study by Gaggia et al., the content of alcohol on day 14 was higher at the level of 5.83% for white tea, 4.18% for green tea and only 1.14% for black tea, but this depends on the fermentation conditions, such as temperature or microbiological composition [19]. In the next phase, *Acetobacter* bacteria [8] use ethyl alcohol as a substrate to create acetic acid. The dominating bacteria included in kombucha are the bacteria of acetic acid AAB: *Acetobacter xylinoides, Acetobacter aceti, Acetobacter pasteurianus, Bacterium gluconicum and Gluconobacter oxydans*. This is why on the 14th day of the fermentation process, the content of alcohol decreased, and there was an increase in acidity as well as the production of organic acids, including acetic acid. Acetic acid, which is the dominating acid present in a fermented solution, contributes to the decrease in pH from 5 to as low as 3 [10,28].

An important parameter that undergoes change during fermentation is pH and acidity, and thus the content of organic acids. The microorganisms present in SCOBY process the substances included in tea and sugar, producing various metabolites. This is why these parameters change with fermentation time. In this study, the pH of teas was from 5.34 to 6.53. In the case of kombucha, there was a significant decrease in this parameter: from 2.31 to 2.53 on the 14th day of fermentation. There was also a small decrease in pH between the 7th and 14th day of fermentation, which indicates that the reactions responsible for the decrease in this parameter were inhibited. Our results are similar to the findings of other authors [29–31]. Chakravorty et al. observed that the initial pH before fermentation was about 5.03 and decreased abruptly to 2.28 after 7 days of fermentation [22]. It has to be remembered that consuming drinks with a very low pH may negatively influence the digestive system [32]. This is why the fermentation time of kombucha is important, as well as the amount of the consumed drink.

The organic acids present in kombucha include acetic, glucuronic, gluconic, tartaric, malic, citric, lactic, succinic and malonic acids [8,10]. The biochemical content of the drink may slightly differ due to the change of parameters, such as: the amount of sugar, the type and quantity of tea, temperature, pH and the time of fermentation. In this study, for all kombucha types, there was a significant increase in the content of acids during fermentation. The sudden production of organic acids occurred after the 7th day of fermentation. The content of acetic acid on the 14th day of fermentation was the highest for green tea (9147.40 mg/L) and white tea (9132.20 mg/L), the lowest for red tea (9071.02 mg/L) and black tea (9083.03 mg/L). These results correspond to those present in other studies. The research showed differences in metabolite content between the drinks prepared from black tea, green tea and rooibos on different days of fermentation [19]. The content of acetic acid on the 7th day of fermentation present in a study by Gaggìa et al. was the highest in white tea (9.18 mg/mL) and green tea (7.65 mg/mL), while the lowest in rooibos (4.89 mg/mL) [19]. Shahbazi et al. determined that acetic acid was the main acid present in kombucha, and its content significantly decreased during fermentation [29]. Chen and Liu (2000) observed that the concentration of acetic acid increased to 8000 mg/L at the end of the storage period Jayabalan et al. (2007) studied the changes in organic acids of kombucha tea during fermentation [10,33]. They observed that green tea was characterised by the highest content of acetic acid (9500 mg/L) on the 15th day of fermentation [10]. The concentration of lactic acid significantly increased during fermentation. Its concentration was at the level of 145.71 mg/L on the 16th day of fermentation [29]. Malbaša, Lonˇcar and Djuri´c (2008) used molasses as a source of sugar for the fermentation of kombucha. The content of lactic acid was from 0.16 to 0.4 g/L [34]. It is also worth highlighting that the pH of the solution and the presence of some organic acids determines the growth of microorganisms, and so also the chemical content of the drink [19]. Low pH and high acidity enable the growth of only those microbes that are able to colonise such a niche, so those that can provide a certain kind of protection against unwanted microorganisms [35].

Sugar content in kombucha also changes in time and depends on fermentation. The initial increase in reducing sugar content can be attributed to the hydrolysis of saccharose into glucose and fructose by yeast. With the progressing fermentation, yeast uses sugar in an oxygen-free way to produce ethanol [10]. In our study, the content of sugar decreased with the time of fermentation. The highest decrease (32%) was observed for black tea vs. kombucha on the 14th day of fermentation. Gaggìa et al. checked the content of glucose, fructose and saccharose in kombucha prepared from black, green and red tea types on the 7th and 14th day of fermentation. The content of complex carbohydrates, i.e., saccharose, decreased during fermentation, while the content of simple carbohydrates–glucose–increased. The concentration of fructose increased during fermentation. The highest content of sugars on the 14th day of fermentation was observed in kombucha prepared from red tea [19].

In this study, a strong positive correlation was observed between time, acetic acid and pH, whereas a negative correlation was observed between acetic acid and the content of alcohol and sugar. The observed correlations confirm the changes occurring in kombucha during the process of fermentation. The increase in acidity and pH with the time of fermentation, as well as the decrease in alcohol and sugar content are associated with the production of organic acids and the use of substrates for their production.

Kombucha has many health-promoting properties, including antioxidant ones. Therefore, to support one's antioxidative response, a regular diet should include kombucha, especially in cases of increased exposure to mental and physical stress. Considering the antioxidant properties of kombucha, the most valuable one is derived from red and green tea. However, longer fermentation leads to a decrease in the pH of the drink, which is why consumption of kombucha should be avoided by people suffering from ulcers or gastrointestinal reflux. Of note, kombucha may contain lead from an inadequate vessel, which may be another health hazard [36,37].

#### **5. Conclusions**

Kombucha, the fermented tea, has strong antioxidant properties associated with high polyphenol content, particularly flavonoids. Therefore, it should be consumed by people particularly exposed to oxidative stress. The antioxidant activity of kombucha is diverse and depends on the type and composition of the tea infusion before fermentation and on the content of SCOBY, which determines the character of the forming metabolites and conditions the type of the forming products of polyphenol compound transformation. A particularly rich source of antioxidants, especially flavonoids, are red and green tea types on the 1st and 14th day of fermentation. Therefore, the selection of tea other than black tea, and the subsequent subjection to fermentation, are beneficial to human health.

**Author Contributions:** Conceptualization, K.J. (Karolina Jakubczyk) and K.J. (Katarzyna Janda); Data curation, J.K. (Joanna Kochman); Funding acquisition, K.J. (Karolina Jakubczyk) and K.J. (Katarzyna Janda); Investigation, K.J. (Karolina Jakubczyk) and J.K. (Justyna Kałdu ´nska); Methodology, K.J. (Karolina Jakubczyk) and J.K. (Justyna Kałdu ´nska); Project administration, K.J. (Karolina Jakubczyk) and K.J. (Katarzyna Janda); Resources, J.K. (Joanna Kochman); Supervision, K.J. (Karolina Jakubczyk) and K.J. (Katarzyna Janda); Writing—original draft, K.J. (Karolina Jakubczyk); Writing—review & editing, K.J. (Karolina Jakubczyk) and J.K. (Joanna Kochman). All authors have read and agreed to the published version of the manuscript.

**Funding:** The project is financed from the program of the Minister of Science and Higher Education under the name "Regional Initiative of Excellence" in 2019-2022 project number 002/RID/2018/19 amount of financing 12 000 000 PLN.

**Acknowledgments:** The authors are thankful to company Naturalnie naturalni (https://naturalnienaturalni.com/) for providing the materials (Kombucha SCOBY) for this research.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **Antioxidant Metabolism and Chlorophyll Fluorescence during the Acclimatisation to Ex Vitro Conditions of Micropropagated** *Stevia rebaudiana* **Bertoni Plants**

**José Ramón Acosta-Motos 1,2,**†**, Laura Noguera-Vera 1,**†**, Gregorio Barba-Espín 1, Abel Piqueras <sup>1</sup> and José A. Hernández 1,\***


Received: 19 November 2019; Accepted: 30 November 2019; Published: 3 December 2019

**Abstract:** In this study, the functioning of antioxidant metabolism and photosynthesis efficiency during the acclimatisation of *Stevia rebaudiana* plants to ex vitro conditions was determined. A high percentage of acclimatised plants (93.3%) was obtained after four weeks. According to the extent of lipid peroxidation, an oxidative stress occurred during the first hours of acclimatisation. A lower activity of monodehydroascorbate reductase (MDHAR) than dehydroascorbate reductase (DHAR) was observed after 2 days of acclimatisation. However, after 7 days of acclimatisation, stevia plants activated the MDHAR route to recycle ascorbate, which is much more efficient energetically than the DHAR route. Superoxide dismutase and catalase activities showed a peak of activity after 7 days of acclimatisation, suggesting a protection against reactive oxygen species. Peroxidase activity increased about 2-fold after 2 days of acclimatisation and remained high until day 14, probably linked to the cell wall stiffening and the lignification processes. In addition, a progressive increase in the photochemical quenching parameters and the electronic transport rate was observed, coupled with a decrease in the non-photochemical quenching parameters, which indicate a progressive photosynthetic efficiency during this process. Taken together, antioxidant enzymes, lipid peroxidation, and chlorophyll fluorescence are proven as suitable tools for the physiological state evaluation of micropropagated plants during acclimatisation to ex vitro conditions.

**Keywords:** acclimatisation; antioxidant defences; chlorophyll fluorescence; in vitro culture; peroxidase; stevia plants

#### **1. Introduction**

The application of in vitro culture techniques is a powerful vegetative proliferation tool for many plant species [1]. However, this process can be limited due to significant losses during acclimatisation to ex vitro conditions. For this raison, a better knowledge of the physiology and biochemistry of in vitro cultured plants that subsequently will be adapted to ex vitro conditions are of major interest. Light availability, and therefore the process of photosynthesis, is a key factor for ex vitro acclimatisation. Improvement of the photosynthetic activity is a critical step to reach a high survival rate during acclimatisation of in vitro plantlets [2]. In other words, proper photosynthesis activation is the key point to change the way to acquire carbon from heterotrophic or mixotrophic (in vitro conditions) to autotrophic (ex vitro conditions) sources. In grapevine, net photosynthesis and biomass are dependent

on the increase in light intensity [2]. However, a distinct response was found in chestnut under the same conditions, where symptoms of photoinhibition were found during the acclimatisation process [2]. The transition from in vitro to ex vitro conditions has not been studied extensively, and an appropriate research model is still missing. Some of them have used the chlorophyll fluorescence technique as a non-destructive indicator to follow the acclimatisation process [1–3].

Micropropagated plants are very susceptible to environmental challenges after transferring to ex vitro conditions. For example, ex vitro plants are normally subjected to higher photosynthetic photon flux density (PPFD) than plants grown under in vitro conditions. In addition, the relative humidity (RH) is also lower under ex vitro conditions, thus plants are prone to suffer desiccation. Both phenomena, which contribute to photoinhibition damage and water stress, can induce the overproduction of reactive oxygen species (ROS). However, plants are provided with an efficient antioxidant defence mechanism to defend against the harmful effects of ROS. These defences include the ascorbate-glutathione (ASC-GSH) cycle enzymes (ascorbate peroxidase (APX), monodehydroascorbate reductase (MDHAR), dehydroascorbate reductase (DHAR) and glutathione reductase (GR)) and ROS-scavenging enzymes (superoxide dismutases (SODs), peroxidases (POX) and catalase (CAT)). The knowledge about the behaviour of the antioxidant machinery during ex vitro acclimatisation is very scarce, and only a few researchers have studied the changes on the enzymatic and non-enzymatic antioxidants during this process [1,3–5].

Stevia (*Stevia rebaudiana* Bertoni) is a perennial shrub belonging to the Asteraceae family. The leaves of *S. rebaudiana* contain a high concentration of steviol glycosides, stevioside and rebaudioside A being the prevalent forms, and used as natural sweeteners as a substitute for saccharose [6]. However, stevia seeds have little viability and the plant requires specific humidity, light, and nutrient conditions. The accumulation of steviol glycosides within *S. rebaudiana* is very variable due to significant genetic variability. The total steviol glycoside content changed not only between plants of the same cultivar, but also among similar plants in the same developmental stage [7]. In addition, a high antioxidant capacity of *S. rebaudiana* leaf extracts, related to their function as ROS-scavengers, has been reported [7–9]. These positive roles have been primarily associated with the presence of phenolic compounds [7]. Moreover, health-related effects of stevioside against type-II diabetes, hypertension, metabolic syndrome, and atherosclerosis have been reported [7]. Therefore, the production of in vitro clonal plants with a similar stevioside profile can be of commercial interest.

Accordingly, this work has focused on the acclimatisation to ex vitro conditions of stevia clones, originated from the micropropagation of plants previously characterised as high accumulators of steviol glycosides [10]. During the process of acclimatisation, the evolution of different parameters, including antioxidant metabolism, lipid peroxidation as an oxidative stress parameter, and chlorophyll fluorescence, were monitored to determine the oxidative stress that stevia plants might be suffering during the aforementioned process.

#### **2. Material and Methods**

#### *2.1. Plant Material and Experimental Design*

The plants were obtained from micropropagated stevia shoot cultures [10] (solid Murashige and Skoog (MS) medium supplemented with 60 mg L−<sup>1</sup> phloroglucinol, 30 mg L−<sup>1</sup> sequestrene, 0.8 mg L−<sup>1</sup> meta-topolin, 6 mg L−<sup>1</sup> adenine sulphate, 0.040 mg L−<sup>1</sup> indole butyric acid, 3% sucrose, and a pH of 5.8). For elongation and rooting, shoots with three internodes were transferred to 1/2 MS medium without growth regulators, containing 40 mg L−<sup>1</sup> sequestrene, 80 mg L−<sup>1</sup> phloroglucinol, 250 mg L−<sup>1</sup> MES buffer, 0.7% Agar, and a pH of 5.8. Under these conditions, the shoots elongated and rooted in 6 weeks. All cultures were maintained at 25 ± 2 ◦C in a growth chamber with a 16 h photoperiod (80 μmol m−<sup>2</sup> s−<sup>1</sup> photosynthetically active radiation, PAR). When the plantlets reached ca. 8–9 cm shoot length, the acclimatisation stage was initiated. For this, the rooted shoots were washed with distilled water to remove the agar and grown in an acclimatisation chamber (UBBINK propagator, (Northampton, UK)), consisting of a plastic tray and a transparent plastic cover, containing 2 vents for the control of the RH\* in a mixture of perlite and peat (1:2, *v:v*). The substrate was moistened with distilled water and a systemic fungicide–bactericide (Beltanol-L, Probelte, Murcia, Spain) at 0.1% (*v*/*v*), which was also applied to the plantlets. These plantlets were kept in a culture chamber with a 16 h photoperiod and 25 ◦C, firstly at 150 μmoles cm−<sup>2</sup> s−<sup>1</sup> PAR for a period of ten days, followed by 18 days at 350 μmoles cm−<sup>2</sup> s−<sup>1</sup> PAR to complete the acclimatisation to ex vitro conditions. During this period, the respirators were gradually open to decrease the humidity progressively. For the different analysis, samples were taken from in vitro plantlets, and from plants at 2, 7, 14, 21, and 28 days of acclimatisation.

#### *2.2. Measurement of Chlorophyll Fluorescence*

Chlorophyll fluorescence was measured with a chlorophyll fluorimeter (IMAGIM-PAM M-series, Heinz Walz, Effeltrich, Germany) during the acclimatisation period of stevia plants to ex vitro conditions, at 2, 7, 14, 21, and 28 days of the initiation of the acclimatisation period. After a dark incubation period (20 min), the minimum and the maximal fluorescence yields of the stevia leaves were monitored. Kinetic analyses were carried out with actinic light (81 μmol quanta m−<sup>2</sup> s−<sup>1</sup> PAR) and repeated pulses of saturating light at 2700 μmol quanta m−<sup>2</sup> s−<sup>1</sup> PAR for 0.8 s, and at intervals of 20 s [11,12]. The following parameters were also analysed: effective PSII quantum yield (Y(II)); the quantum yield of regulated energy dissipation (Y(NPQ)); the non-photochemical quenching (NPQ); the maximal PSII quantum yield (Fv/ Fm); the coefficients of non-photochemical quenching (qN); the photochemical quenching (qP); and quantum yields of non-regulated energy dissipation Y(NO) [13].

#### *2.3. Lipid Peroxidation*

Stevia leaves were snap-frozen in liquid nitrogen and stored at −80 ◦C until use. The extent of lipid peroxidation was estimated by determining the concentration of thiobarbituric acid-reactive substances (TBARS) using a UV/Vis V-630 Bio spectrophotometer (Jasco, Tokyo, Japan). Leaf samples (0.2 g) were ground in liquid nitrogen into a fine powder and extracted in 1 M perchloric acid solution (1/10, w/v). Homogenates were centrifuged at 12,000× *g* for 10 min and 0.5 mL of the supernatant obtained was added to 1.5 mL 0.5% TBA in 1 M perchloric acid. The mixture was incubated at 90 ◦C in a shaking water bath for 20 min, and the reaction was stopped by placing the reaction tubes in an ice water bath. Then, the samples were centrifuged at 10,000× *g* for 5 min, and the absorbance of the supernatant was read at 532 nm. The value for non-specific absorption at 600 nm was subtracted. The amount of TBARS (red pigment) was calculated from the extinction coefficient 155 mM−<sup>1</sup> cm−<sup>1</sup> [10]. The lipid peroxidation was measured in plantlets under in vitro conditions as well as during the acclimatisation period of stevia plants to ex vitro conditions at 2, 7, 14, 21, and 28 days after the initiation of the acclimatisation process.

#### *2.4. Enzyme Extraction and Analysis*

Leaf samples were homogenized in liquid nitrogen and an extraction medium (1/5, *w*/*v*) containing 50 mM Tris-acetate buffer (pH 6.0); 0.1 mM EDTA; 2 mm cysteine; and 0.2 % (*v*/*v*) Triton X-100. For the ascorbate peroxidase (APX) activity, 20 mM sodium ascorbate was added to the extraction buffer. The extracts were centrifuged at 10,000× *g* for 20 min. The supernatant fraction was filtered on Sephadex NAP-10 columns (GE Healthcare, Chicago, IL, USA) equilibrated with the same buffer used for homogenisation and used for the enzymatic determinations. For the APX activity, 2 mM of sodium ascorbate was added to the equilibration buffer.

The enzymatic analyses were measured in plantlets under in vitro conditions as well as during the acclimatisation period of stevia plants to ex vitro conditions at 2, 7, 14, 21, and 28 days of the initiation of the acclimatisation process. The antioxidant enzyme determinations were carried according to protocols set up in our laboratory using a UV/Vis V-630 Bio spectrophotometer (Jasco) [11,14,15]. Specifically, APX (EC 1.11.1.11) was determined following the decrease at 290 nm due to the ascorbate

oxidation by H2O2. MDHAR (EC 1.6.5.4) was determined following the decrease at 340 nm due to the NADH oxidation. DHAR (EC1.8.5.1) was determined by following the increase at 265 nm due to ascorbate formation. The reaction rate was corrected for the nonenzymatic reduction of DHA by reduced glutathione (GSH). GR (EC 1.6.4.2) was assayed by the decrease at 340 nm due to the NADPH oxidation. The reaction rate was corrected for the non-enzymatic oxidation of NADPH by oxidized glutathione (GSSH). SOD (EC 1.15.1.1) was assayed by the ferricytochrome c method using xanthine/xanthine oxidase as the source of superoxide radicals. CAT (EC 1.11.1.6) was measured following the decrease at 240 nm due to H2O2 consumption [14,15]. POX activity (EC. 1.11.1.7) was analysed following the oxidation of 4-methoxy-a-naphtol at 593 nm [15].

Protein contents were analysed according to [16] using a plate reader (Epoch2, BioTek, Winooski, VT, USA) and bovine serum albumin as standard.

#### *2.5. Statistical Analysis*

The data were analysed by one-way ANOVA followed by Tukey's Multiple Range Test (*p* ≤ 0.05) to separate treatment means, using the SPSS 20.0 software (SPSS Inc., 2002, Chicago, IL, USA). Multivariate analysis using the StatGraphics Centurion XV software (StatPoint Technologies, Warrenton, VA, USA) were conducted by Principal Component Analysis (PCA), followed by a partial least squares discriminant analysis to assign the principal components displaying eigenvalues greater than or equal to 1.0.

#### **3. Results**

In the present work, the acclimatisation of in vitro *Stevia rebaudiana* Bertoni plants to ex vitro conditions was achieved with a high success rate, since 93.3% of the plants survived.

#### *3.1. Chlorophyll Fluorescence Measurement*

During the acclimatisation process, the evolution of the fluorescence parameters was monitored. At Day 2, the plants displayed higher values of the non-photochemical quenching parameters (Y(NPQ), Y(NO), NPQ and qN) and low values of the photochemical quenching parameters (Y(II), qP) (Table 1, Figure 1), as well as of the electron transport rate (ETR) (Figure 2). During the acclimatisation process, a progressive decrease in the non-photochemical quenching parameters and a constant increase in the photochemical-quenching parameters were observed. In that regard, Y(NPQ) continuously decreased, reducing its values by 40% and 50% after 21 and 28 days of the acclimatisation, respectively (Table 1, Figure 1). Concomitantly, Y(NO) declined during the acclimatisation assay, reaching a decrease near 40% and 30% after 21 and 28 days of acclimatisation, respectively (Table 1, Figure 1). NPQ displayed increases and decreases during the acclimatisation process. At first, after 7 days of acclimatisation this parameter increased by 22%. Then, the NPQ value increased by 63% after 14 days of acclimatization in relation to the precedent value (Day 7). One week later (day 21), again the NPQ value raised by 29% in comparison to the value observed in the second week (14 days). Finally, after 28 days of acclimatisation, a 30% decrease in the NPQ parameter was observed in relation to the value observed after 21 days (Table 1, Figure 1). However, although the qN values decreased during the acclimatisation process, the changes produced were not statistically significant (Table 1, Figure 1). Regarding the photochemical quenching parameters (Y(II) and qP), a progressive increase occurred during the acclimatisation. In both cases, the values increased near 3-fold after 7 and 14 days of acclimatisation, and about 5-fold after 21 and 28 days of the process (Table 1, Figure 1). Fv/Fm showed the lowest values after 2 days of acclimatisation. This parameter increased after 7 and 14 days, and then slightly decreased after 21 and 28 days of acclimatisation, but their values remaining statistically higher than the initial values (Table 1). The changes observed in Y(II) and qP correlated with the evolution of the ETR values, reaching an increase of near 6-fold at the end (28 days) of the acclimatisation process (Figure 2).

**Table 1.** Evolution of photochemical (Y(II), qP, Fv/Fm) Fv/Fm and non-photochemical quenching parameters (Y(NPQ), Y(NO), NPQ, qN) during the process of acclimatization to ex vitro conditions of *Stevia rebaudiana* Bertoni plants. <sup>a</sup> F: values from one-way ANOVA for the different chlorophyll fluorescence parameters at 99.9% level of significance (\*).


Different letters in the same column indicate significant differences according to Tukey's Multiple Range Test (*p* ≤ 0.05).


**Figure 1.** Evolution of the chlorophyll fluorescence parameters during the acclimatisation process of *S. rebaudiana* Bertoni plants to ex vitro conditions.

**Figure 2.** Evolution of the electron transport rate (ETR) during the acclimatization process of *S. rebaudiana* Bertoni plants to ex vitro conditions. Different letters in the same column indicate significant differences according to Tukey's Multiple Range Test (*p* ≤ 0.05).

Complementarily, PCA was utilized as a mathematical tool to determine associations among the different chlorophyll fluorescence parameters, ETR, and the acclimatisation evolution (Figure 3). The first component (PC1), which explains 63% of the variability of the experiment (Table S1, Table S2 and Figure S1), indicated the greater relevance of the photosynthetic efficiency mechanisms to adapt to ex vitro conditions, as reflected by the positive values that show all the measured parameters related to the photochemical quenching parameters, such as Y(II), qP and Fv/Fm, and ETR. On the contrary, the contribution of the non-photochemical quenching parameters, related to heat dissipation in a regulated or unregulated way, was negative in the PC1. This was the case for qN, the NPQ, and Y (NPQ) in the first case (regulated dissipation), and Y (NO) in the second case (unregulated dissipation). Regarding the second component (PC2), which explained 24% of the variability of the experiment (Table S1, Table S2 and Figure S1), it indicated that in a second level of importance for the acclimatisation to ex vitro conditions the excess of light energy was also relevant, being not destined to perform photosynthesis, but dissipated safely as heat by regulated mechanisms as observed in the high positive values that show all the related measured parameters (qN, NPQ and Y(NPQ)). Again, as in PC1, any unregulated heat dissipation mechanism played a harmful role in the plant as reflected in the negative values of Y(NO).

**Figure 3.** A principal component analysis applied to the fluorescence chlorophyll and electron transport rate (ETR) parameters. Two principal components (PC1 and PC2) resulted in a model that explained 87% of the total variance. The arrows denote eigen vectors characterised by the direction and the strength of the variable relative to PC1 and PC2.

#### *3.2. Antioxidant Metabolism*

#### 3.2.1. Lipid Peroxidation Assay

During the first hours of the acclimatisation process, stevia plants seemed to experience stress due to the modification on the culture conditions, as observed by the increase in the lipid peroxidation levels, measured as TBARS. In that regard, a peak after 2 days was detected, increasing by 86% with respect to values in the plantlet (Figure 4). Thereafter, and as the acclimatisation process to ex-vitro conditions progressed, the lipid peroxidation values progressively decreased reaching initial values (Figure 4).

**Figure 4.** Levels of lipid peroxidation (LP) during the acclimatization process of *S. rebaudiana* Bertoni plants to ex vitro conditions. Different letters (a,b,c) in the same column indicate significant differences according to Tukey's Multiple Range Test (*p* ≤ 0.05).

#### 3.2.2. Antioxidant Enzymes

During the acclimatisation process of stevia plants to ex vitro conditions, the levels of some antioxidant enzymes, including the ASC-GSH cycle enzymes, superoxide dismutase (SOD), peroxidase (POX), and catalase (CAT), were analysed.

Regarding the ASC-GSH cycle enzymes, ascorbate peroxidase (APX) activity showed its highest activity in in vitro plants after 2 days of the acclimatisation process. The lowest APX activities were observed after 14 and 28 days, representing a decrease of ca. 35% with respect to the other time points (Figure 5). The monodehydroascorbate reductase (MDHAR) activity displayed its lower values at the beginning of the process of acclimatisation, i.e., under in vitro conditions and after 2 days of the acclimatisation process. Then, MDHAR activity progressively increased, especially after 21 and 28 days of acclimatisation, showing a 2.7- and 3.6-fold increase, respectively, with respect to values after 2 days (Figure 5). Regarding dehydroascorbate reductase (DHAR) activity, it behaved contrary to MDHAR activity. In that sense, DHAR showed its higher values after 2 days of the acclimatisation process, with a 40% increase in relation to the values observed in in vitro plants (Figure 5). Subsequently, DHAR activity progressively declined, with decreases ranging from 34% to 64% depending on the acclimatisation period, and a minimum DHAR activity observed after 14 days of acclimatisation (Figure 5). The GR activity behaved similarly to MDHAR activity. In that regard, GR increased as the plant acclimatisation process progressed, reaching their maximum values after 21 and 28 days of acclimatisation (4.2- and 3.2-fold increases, respectively) (Figure 5).

**Figure 5.** Evolution of the activity of the ASC-GSH cycle enzymes during the process of acclimatisation to ex vitro conditions of *S. rebaudiana* Bertoni plants. Different letters (a,b,c,d) in the same column indicate significant differences according to Tukey's Multiple Range Test (*p* ≤ 0.05).

The activity SOD remained statistically invariable at all times except after 7 days of acclimatisation, where a 3.5-fold increase in relation to the initial values was observed (Figure 6). A similar response was observed for the CAT activity: a 1.7-fold increase occurred after 7 days of acclimatisation, whereas in the rest of the period the CAT activity displayed no significant differences in relation to the initial values (Figure 5). POX activity increased remarkably after 2 days, maintaining the value after 7 days

(2.4-fold increase with respect to in vitro plants). Thereafter, POX activity progressively declined until reaching a 4.2-fold decrease with respect to the values after 7 days (Figure 6).

**Figure 6.** Evolution of the activity of SOD, CAT, and POX during the process of acclimatisation to ex vitro conditions of *S. rebaudiana* Bertoni plants. Different letters (a,b,c) in the same column indicate significant differences according to Tukey's Multiple Range Test (*p* ≤ 0.05).

Complementarily, a PCA study was carried out to analyse the associations between the different antioxidant enzymes monitored, as well as lipid peroxidation during the evolution of acclimatisation to ex vitro conditions (Figure 7). The resulting model with two components explained 63% of the total variance. The first component (PC1), which explained 37% of the variability within the dataset (Table S3, Table S4 and Figure S2), indicates the importance of the defence mechanisms to protect from the cellular damage associated with lipid peroxidation (LP), allowing a faster acclimatisation to ex vitro conditions. The efficient antioxidant mechanisms are reflected by the positive values observed in the ROS-eliminating enzymes (SOD, POX and CAT), as well as MDHAR and GR, which guarantees redox homeostasis of the plant. However, the contribution of DHAR and APX was less important, as denoted by their negative values. With respect to the second component (PC2), explaining 26% of the variability of the experiment (Tables S3 and S4), the loadings for LP, DHAR, CAT, and POX

appeared as the dominant variables, and were opposite to the positive contribution of APX. Moreover, the loadings for DHAR and LP clustered together, which may relate to the parallel evolution of both variables during the acclimatisation period.

**Figure 7.** A principal component analysis applied to the antioxidant enzymes and lipid peroxidation variables. Two principal components (PC1 and PC2) resulted in a model that explained 63% of the total variance. The arrows denote eigen vectors characterised by the direction and the strength of the variable relative to PC1 and PC2.

#### **4. Discussion**

The acclimatisation to ex vitro conditions is a critical step for the survival of micropropagated in vitro plants. Light availability is an important factor for a successful process of acclimatisation, involving a proper activation of the photosynthesis process [2]. The transition from in vitro to ex vitro conditions has been achieved in numerous plant species, both herbaceous and woody plants [2,4,8,10,17–20].

The micropropagation and acclimatisation of stevia plants can be an excellent tool to ensure the production of clonal, uniform, and true-to-type plant material. This may be used as a by-pass to the problems associated to the low germination rate and the great variation in the profile of steviol glycosides of *S. rebaudiana*. As a consequence, the production of uniform *S. rebaudiana* plants with the same steviol glycosides profile is of important commercial interest. This is particularly interesting for rebaudioside A, which provides a superior flavour to food products compared to other steviosides [6]. In a previous work, carried out in our laboratory, we reported that the main steviol glycoside detected in stevia plants after 10 week of acclimatisation was stevioside, whereas after 12 weeks, the main steviol glycoside was rebaudioside A, whose levels were not affected by the presence of NaCl [10]

#### *4.1. Chlorophyll Fluorescence*

The use of the chlorophyll fluorescence technique has been used for different researchers to evaluate the acclimatisation process [1–3,21,22]. These authors found that acclimatised plants to ex vitro conditions increased the photosynthesis rate, which correlated with higher chlorophyll contents as well as Fv/Fm and Y(II) fluorescence parameters. In other studies, the authors used the photosynthesis rate or the chlorophyll a fluorescence technique, as well as the measurement of the antioxidant defences as indicators of the evolution of the acclimatisation process [1–4,17,19]. Accordingly, the monitoring of some physiological, biochemical, and molecular parameters can be useful tools to determine the success of ex vitro acclimatisation of in vitro plantlets.

Light intensity is another relevant abiotic factor influencing the acclimatization process. This was the case of grapevine and *Die*ff*enbachia* plantlets. In both cases, the increase in biomass or net photosynthesis was depending on the light intensity [2,5]. In grapevine, the photochemical quenching parameters, Y(II) and qP, showed higher values in the presence of 300 μmol m−<sup>2</sup> s−<sup>1</sup> of photons than in lower light intensities (100 μmol m−<sup>2</sup> s−1), indicating a better photosynthetic efficiency at higher intensities. However, during the acclimatisation of *Die*ff*enbachia* plantlets, the Fv/Fm decreased with the increase in light intensity, demonstrating that photoinhibition occurred after transplanting micropropagated plants [5]. In grapevine, immediately after transferring to ex vitro conditions, a reversible photoinhibition occurred, as denoted by an initial decrease in Fv/Fm and F'q/F'm parameters followed by a progressive increased until Day 7 [3]. The mentioned responses could be due to the presence of poorly developed chloroplasts in the in vitro established leaves, resulting in a low resistance against photoinhibition [23].

In this work, a decrease in the non-photochemical quenching parameters and an increase in the photochemical-quenching parameters were observed during acclimatisation to ex vitro conditions. This response was correlated with a progressive increase in ETR values, indicating a gain of chloroplast efficiency during the process. In this sense, Y(II) represents the proportion of the light absorbed by chlorophyll associated with PSII that is used for photochemistry, whereas qP gives an indication of the proportion of the PSII reaction centres that are open [13].

In contrast, the high values observed in the non-photochemical quenching parameters in the first steps of the acclimatisation process would indicate that part of the light captured by the chloroplasts may dissipate as heat, protecting the chloroplast against the excess of radiation and avoiding the formation of ROS, especially singlet oxygen. Different authors have previously described the increase in non-photochemical parameters as an efficient mechanism for dissipating excess light energy and thus minimizing ROS generation [13]. In addition, it has been demonstrated that NPQ and Y(NPQ) are parameters very sensitive in the early detection of stress conditions by using fluorescence images [24]. Both NPQ and Y(NPQ) are related to the energy dissipated as heat by regulated mechanism (i.e., the xanthophyll cycle) [25]. In contrast, Y(NO) reflects the fraction of energy passively dissipated as heat and fluorescence, mainly due to closed PSII reaction centres. Therefore, high values of Y(NO) are related to the inability of plants to protects itself from excess light. In that regard, after 2 days of acclimatisation, stevia plants showed the highest values of Y(NO), which progressively decreased during the acclimatisation process, reflecting a better regulation [26]. On the other hand, high values of the non-photochemical parameters indicated that plants are suffering a stress. However, as the plant adapts to the new ex vitro conditions, these parameters decreased.

#### *4.2. Antioxidant Response*

Several authors used biochemical parameters, including antioxidant enzymes and lipid peroxidation, to follow the acclimation of in vitro plantlets. In the ornamental African violet, SOD, CAT, and glutathione-peroxidase (GSH-POX) increased after 28 days of acclimatisation, concomitantly with the increase in the light irradiance, suggesting that plants were suffering an abiotic stress [4]. However, the lipid peroxidation extent, measured as malondialdehyde contents, was 2-fold higher under low irradiance (35 μmol m−<sup>2</sup> s−1) than in the presence of higher light intensities (70–100 μmol m−<sup>2</sup> s<sup>−</sup>1), indicating the presence of an oxidative stress. This response was associated with a lower activity of the antioxidants enzymes as well as a low reduced glutathione (GSH) content. In fact, GSH not only acts as an antioxidant substrate in the ASC-GSH cycle but also as an antioxidant to minimize oxidative stress [27]. In a recent work, [5] analysed the activity of some antioxidant enzymes (SOD, CAT, and GSH-POX) during the acclimatisation of different *Die*ff*enbachia* cultivars grown at different light intensities. SOD and GSH-POX activity were much higher in the presence of the higher photosynthetic photon flux density (PPFD). However, the effect of PPFD on CAT activity was less evident, although

the activity was higher in the presence of 100 PPFD (μmol m−<sup>2</sup> s<sup>−</sup>1) than with lower light intensities (35 and 75 PPDF) [5]. According to these authors, this response of the antioxidant defences could be due to the higher ROS production induced by light stress.

In the present work, during the first days of ex vitro acclimatisation, an oxidative stress also occurred, as indicated by the lipid peroxidation levels. This variable is considered as an oxidative stress marker [12]; consequently, a damage to membrane lipids occurred, probably due to the change on the culture conditions. However, once this change on the culture conditions took place, the plants became progressively adapted to the new conditions, as reflected by the decrease in lipid peroxidation. A similar response has been described by [3] during the acclimatisation to ex vitro conditions of grapevine plantlets. These authors monitored the H2O2 levels, another oxidative stress marker, observing a 50% increase in H2O2 contents after 24 h of the acclimatisation process. Thereafter, H2O2 levels declined to the initial values and even lower [3]. In the same work, the increase in H2O2 after transfer to ex vitro conditions was followed by an increase in the expression of some antioxidant enzymes, including *APX1*, *GR1*, *SOD1*, and *SOD2* [3]. This response was partially similar to the observed in acclimatized stevia plants. In the present work, both MDHAR and GR activities increased during the acclimatisation process to ex vitro conditions, and other antioxidant enzymes as SOD, CAT, and POX peaked after 7 days of acclimatisation. In accordance with this, under in vitro conditions and after 2 days of acclimatisation, stevia plantlets displayed the highest DHAR activity, with values superior to MDHAR activity—the other ASC-recycling enzyme of the ASC-GSH cycle—at the same time. In that regard, after 2 days of acclimatisation the ratio DHAR/MDHAR was nearly 2. This suggests that, at that stage, DHAR activity is predominant in recycling ascorbate in stevia plants, using GSH as the electron donor. Subsequently, DHAR decreased and MDHAR progressively increased, reaching a DHAR/MDHAR ratio of 0.22 after 28 days of acclimatisation, where MDHAR activity was near 5-times higher than DHAR activity. Therefore, after 2 days of acclimatisation, stevia plants used the MDHAR way, spending NADH as a reducing power. It is necessary to clarify that the utilization of NADH to recycle the ASC is more efficient energetically than the use of GSH. Thus, different possibilities can be speculated to explain the higher DHAR activity under in vitro conditions and after 2 days of acclimatisation. The first is that under in vitro conditions, the culture media contained sucrose, so the plants had enough carbon source to generate energy through the glycolysis and respiration pathways, and thus they can afford the use GSH to recycle ASC (the inefficient way). The second possibility is that after 2 days of the acclimatisation process, the plants suffered an oxidative stress, as monitored by the lipid peroxidation data. Given that overexpression of *DHAR* has been associated with environmental stress tolerance [28], the higher DHAR activity observed at this stage could have a role coping with the stress resulting of the acclimatisation conditions. The third explanation is linked to the role of DHAR in plant growth and development [29]. Probably, after 2 days of acclimatisation the increased DHAR could have a function in plant growth and development processes. We also observed that MDHAR activity was enhanced after 7 days of acclimatisation. At this stage, photosynthesis seemed to work properly, as observed by the chlorophyll fluorescence and ETR values. Therefore, from that moment on, the plants produced their own sugars and energy to support plant growth. Probably, for this reason, plants changed the manner to recycle the ascorbate to an efficient way—via NADH.

On the other hand, it is known that DHAR and GR work together in the ASC-GSH cycle. We observed that under in vitro conditions and after 2 days of acclimatisation the ratio DHAR/GR was ca. 1, supporting the function of DHAR activity in the ascorbate recycling by using GSH as reducing power. However, from that moment on, the GR activity progressively increased, whereas DHAR significantly declined. This response supports the idea that most of the GSH produced in the GR reaction is used for other purposes, including the maintenance of redox homeostasis in stevia plants.

The peroxidase (POX) activity increased more than 2-fold after 2 days of acclimatisation and remained high until day 14, probably linked to the cell wall stiffening and the lignification processes that lead to hardening of the cell wall, as part of the plant differentiation process [30–32]. The overexpression of horseradish *POX* stimulated the growth of tobacco and hybrid aspen plants [33]. In pea seedlings and plants, a correlation between POX activity increase and growth was observed [34,35]. Thus, the observed increases in POX activity during the acclimatisation period can be linked to the plant growth and differentiation processes. However, a role of the POX eliminating H2O2 during the acclimatisation process cannot be ruled out. In fact, SOD, POX, and CAT displayed their maximum activities after 7 days of acclimatisation, indicating that these enzymes could work sequentially to eliminate O2•<sup>−</sup> and H2O2 at that time, since SOD generates H2O2, which is eliminated by the action of the H2O2-scavenging enzymes, such as POX, CAT, or even APX, which maintained its activity after 7 days of acclimatisation.

The results obtained by PCA confirmed the importance of the antioxidant mechanisms and the photosynthesis in the acclimatization of stevia plants. In general, greater efficiency in acclimatization to ex vitro conditions was more evident after 7 days of acclimatization and in the last two weeks (24 and 28 days), with increases in the antioxidant enzymes CAT, GR MDHAR, POX, and SOD, and also accompanied by decreases in lipid peroxidation (LP) in these same time periods.

#### **5. Conclusions**

Taken together, the data suggested that antioxidant enzymes, lipid peroxidation, and chlorophyll fluorescence parameters can be suitable tools for the evaluation of the physiological state of micropropagated plants during the acclimatisation to ex vitro conditions of stevia plants, providing very useful information to monitor the stress state of the plants during the process of acclimatisation. This work has practical implications, since clonal plants of stevia with a known and stable profile of steviol glycosides are a suitable source of edulcorates and natural antioxidants to a diet.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-3921/8/12/615/s1, Figure S1: Sedimentation graph where two PCA with eigenvalues greater than or equal to 1.0 were obtained to determine associations among the different chlorophyll fluorescence parameters, ETR and the acclimatisation evolution, Figure S2: Sedimentation graph where two PCA with eigenvalues greater than or equal to 1.0 were obtained to analyse the associations between the different antioxidant enzymes monitored, as well as lipid peroxidation during the evolution of acclimatisation to ex vitro conditions, Table S1: Weight of the Components for the fluorescence chlorophyll parameters and ETR, Table S2: Eigenvalues and percentage of variance of components for the fluorescence chlorophyll parameters and ETR, Table S3: Weight of the Components for the antioxidant enzymes and lipid peroxidation data (LP), Table S4: Eigenvalues and percentage of variance of components for the antioxidant enzymes and lipid peroxidation data.

**Author Contributions:** Conceptualization, J.A.H and A.P.; methodology, J.A.H, L.N.-V. and A.P.; formal analysis, J.A.H. and J.R.A.-M.; investigation J.A.H, L.N.-V. and A.P.; data curation J.A.H., J.R.A.-M and G.B.-E.; writing—original draft preparation, J.A.H.; writing—review and editing, J.A.H, J.R.A.-M. and G.B.-E.

**Funding:** G.B.E. thanks the "Fundación Séneca" of the Agency of Science and Technology of the Region of Murcia for his research contract.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Domestic Sautéing with EVOO: Change in the Phenolic Profile**

### **Julián Lozano-Castellón 1,2, Anna Vallverdú-Queralt 1,2, José Fernando Rinaldi de Alvarenga 3, Montserrat Illán 1, Xavier Torrado-Prat <sup>1</sup> and Rosa Maria Lamuela-Raventós 1,2,\***


Received: 11 December 2019; Accepted: 13 January 2020; Published: 16 January 2020

**Abstract:** (1) Background: The health benefits of extra-virgin olive oil (EVOO), a key component of the Mediterranean diet, are attributed to its polyphenol profile. EVOO is often consumed cooked, and this process may degrade and transform polyphenols. (2) Methods: In this work, we determined how temperature, time, and the interaction between them affects the EVOO polyphenolic profile during a domestic pan-frying process, simulating the cooking conditions of a home kitchen, without the control of light or oxygen. Applying a 2<sup>2</sup> full factorial design experiment, "Hojiblanca" EVOO was processed at two temperatures (120 ◦C and 170 ◦C) either for a short time or a long time, mimicking a domestic process, and polyphenol content was analyzed by UPLC-ESI-QqQ-MS/MS. (3) Results: Temperature degraded the polyphenols of EVOO during the sauté cooking process, whereas time had an effect on some individual phenols, such as hydroxytyrosol, but not on the total phenol content. The polyphenol content decreased by 40% at 120 ◦C and 75% at 170 ◦C compared to raw EVOO. (4) Conclusions: Cooked EVOO still meets the parameters of the EU's health claim.

**Keywords:** home-cooking; extra virgin olive oil; UPLC-ESI-QqQ-MS/MS; healthy cooking; Mediterranean diet

#### **1. Introduction**

Extra virgin olive oil (EVOO), the main source of fat in a Mediterranean diet, displays a singular fatty acid composition with a higher content of phenolic compounds and other antioxidants than other edible oils. Its health benefits are mainly attributed to these minor components, above all to simple phenols and polyphenols (both referred to henceforth as polyphenols) [1]. Its consumption has shown to play a protective role against a wide range of diseases [1,2], such as cancer [3], cardiovascular diseases [4], neurodegeneration [5], and diabetes [6]. EVOO phenolic concentration can be improved by changing agronomic and technical factors, such as the simple minimization of bruising by a selection of the variety [7].

The problem is that the Mediterranean consumption of EVOO is not only carried out by using it as a final seasoning; EVOO is also used in Mediterranean cuisine for roasting, sautéing (pan-frying), stir-frying, and deep-frying. All of these culinary techniques are thermal processes that could diminish the minor components of EVOO, such as polyphenols, by substances leaching (especially of more polar compounds) into the medium or by the degradation and transformation of its polyphenol

content [8,9]. In addition to the loss of antioxidants, pro-oxidants formation can occur, especially when cooking at high temperatures, notably as a consequence of the lipid oxidation [10,11]. Nevertheless, EVOO polyphenols have been shown to reduce the heat-induced formation of undesired compounds, such as the cancerogenic heterocyclic amines [12], and the formation of acrolein and hexanal [13]. Finally, the polyphenols can act as lipid-derived carbonyl scavengers [14].

Most of the studies on cooking-induced changes in the polyphenol composition of EVOO have been carried out in laboratory conditions [15,16], applying non-conventional Mediterranean cooking techniques, like microwaving [17], or exploring the addition of a phenolic extract rather than EVOO [18]. Their results may not match those produced in a domestic setting because of the differences in oxygen and light availability or because polyphenol degradation in EVOO may be influenced by its content of other minor compounds [19].

On the other hand, previous studies carried out under more true-to-life conditions have focused on comparing the polyphenol content between raw and cooked foods [20] and between foods prepared with different cooking techniques [9]. However, they were not focused on evaluating EVOO polyphenol degradation or how this is affected by cooking factors, like temperature or time. When cooking factors, such as time, were explored, oil was heated for a longer time than the real cooking time, i.e., for 25 or even 36 h [21,22]. Furthermore, some works explored the degradation of total polyphenols measured by the Folin-Ciocalteau method, which is not selective and measures all antioxidant compounds [23]. Consequently, more research is required to determine the extent to which the loss of polyphenols during cooking is counteracted by the beneficial effects of EVOO, or how the phenolic profiles are altered during domestic cooking.

In this context, the aim of the present study was to determine changes in the EVOO polyphenolic profile during a domestic sautéing process commonly used in the Mediterranean diet [24], using a 2<sup>2</sup> full factorial design to assess the effect of time, temperature, and the interactions between these two factors, mimicking real conditions (without oxygen or light control). The polyphenolic profile was measured using ultra-high performance liquid chromatography coupled to a tandem mass spectrometer detector (UPLC-ESI-QqQ-MS/MS), providing information on how the polyphenolic profile changed and how individual polyphenols degraded at different rates.

#### **2. Materials and Methods**

#### *2.1. Chemicals and Standards*

Acetonitrile, methanol, formic acid, and acetic acid were purchased from AppliChem, Panreac Quimica SA (Barcelona, Spain). Hexane, *p*-coumaric acid, ferulic acid, luteolin, oleuropein, oleocanthal, and pinoresinol were purchased from Sigma-Aldrich (St. Louis, MO, USA). Hydroxytyrosol was acquired from Extrasynthese (Genay, France) and apigenin from Fluka (St. Louis, MO, USA). Ultrapure water was obtained using a Milli-Q purification system (Millipore, Bedford, MA, USA).

#### *2.2. Samples*

Polyphenol degradation was assessed in the common Spanish "Hojiblanca" variety of EVOO, which has a medium concentration of polyphenols [25]. It was provided by the Fundación Patrimonio Comunal Olivarero and was produced from olives milled in December 2016 in Spain.

#### *2.3. Domestic Sauté Process*

To simulate the home-cooking process of sauté, EVOO was heated in a pan (20 cm diameter, 0.8 mm thickness, stainless steel 18/10, Excalibur, Pujadas, Girona, Spain), and the influence of the cooking process on polyphenol degradation was monitored at two different temperatures: moderate (120 ◦C) and high (170 ◦C). In order to assess the influence of time, short and long cooking times were determined for each temperature, corresponding to the time needed to obtain "al dente" and well-cooked textures, respectively. For determining these times, 200 g of potatoes and 100 g of chicken

(an average portion) were pan-fried at both temperatures, and the selected times for 120 ◦C were 30 and 60 min and the times for 170 ◦C were 15 and 30 min, time being a qualitative factor. A full-factorial design was performed (22) with three replicates per point to assess the effect of the temperature and time of cooking and the possible interaction between these two factors. The levels and the processing conditions are shown in Table 1.


**Table 1.** Levels and conditions of the full factorial design.

The domestic sautéing was performed at the Food Torribera Campus, University of Barcelona (Santa Coloma de Gramenet, Spain). The pan was heated on an electrical cooking plate (180 mm diameter, 1500 W, model Encimera EM/30 2P, Teka®, Madrid, Spain) until the required temperature was reached. The temperature was monitored with a laser thermometer (error: ±1 ◦C, ScanTemp 410, TFA Dostmann GmbH & Co. KG, Wertheim, Germany) and maintained by turning the heat up or down as necessary. When the target temperature was achieved, 20 g of EVOO were added to the pan and heated for the chosen time. The pan was then removed from the heat and after a short cooling period, the oil was stored in a vacuum bag at −20 ◦C until extraction. The oxygen or light were not controlled to mimic the process carried out in a normal kitchen.

#### *2.4. Polyphenol Extraction and Analysis*

#### 2.4.1. Polyphenol Extraction

The liquid-liquid extraction of phenolic compounds was performed following the method proposed by Kalogeropoulos et al. (2007) with minor modifications [26]. All of the extraction process was carried out over an ice bed. Briefly, 0.5 g of EVOO was suspended with 5 mL of methanol in a 10 mL centrifuge tube and stirred for 30 s. It was centrifuged for 3 min at 3000 rpm and 4 ◦C. The methanolic fraction was then transferred into a flask and the extraction was repeated. Both methanolic fractions were combined and evaporated under a reduced pressure. The residue was reconstituted with 2 mL of acetonitrile and washed twice with 2 mL of hexane. The acetonitrile was evaporated under a reduced pressure and the residue was reconstituted with 800 μL of MeOH:H2O (4:1 *v*/*v*), filtered with Polytetrafluoroethylene syringe filters (0.2 μm), and was transferred to an amber glass vial and stored at −80 ◦C until analysis.

#### 2.4.2. Polyphenol Analysis by UPLC-ESI-QqQ-MS/MS

The identification and quantification of phenolic compounds, except oleocanthal, oleacein and oleuropein and ligstroside aglycones, was performed following the method proposed by Suárez et al. (2008) with minor modifications [27], using an AcquityTM UPLC (Waters; Milford, MA, USA) coupled to an API 3000 triple-quadruple mass spectrometer (PE Sciex, Framingham, MA, USA) with a turbo ion spray source. The separation of compounds was achieved using an Acquity UPLC® BEH C18 Column (2.1 <sup>×</sup> 50 mm, i.d., 1.7 <sup>μ</sup>m particle size) (Waters Corporation®, Wexford, Ireland) and an Acquity UPLC® BEH C18 Pre-Column (2.1 <sup>×</sup> 5 mm, i.d., 1.7 <sup>μ</sup>m particle size) (Waters Corporation®, Wexford, Ireland). The exact chromatographic conditions were as detailed elsewhere [28].

The quantification of oleocanthal, oleacein, oleuropein aglycone, and ligstroside aglycone was performed using a methodology proposed by Sánchez de Medina et al. (2017) with some modifications [29]. Separation was achieved using an Acquity UPLC® BEH C18 Column (2.1 <sup>×</sup> 50 mm, i.d., 1.7 μm particle size) (Waters Corporation®, Wexford, Ireland) and Acquity UPLC® BEH C18

Pre-Column (2.1 <sup>×</sup> 5 mm, i.d., 1.7 <sup>μ</sup>m particle size) (Waters Corporation®, Wexford, Ireland). The exact chromatographic conditions were as detailed elsewhere [28].

Ionization was performed using an electrospray (ESI) interface operating in the negative mode [M − H], and all of the compounds were monitored in the multiple reaction monitoring mode (MRM). The exact ionization and spectrometric conditions are detailed in the previous study [28], and the energies and retention times for each analyzed compound are shown in Tables S1 and S2. The system was controlled by Analyst version 1.4.2 software supplied by Applied Biosystems (Waltham, MA, USA).

Quantification was performed by an external standard calibration method, standards showed linearity in the concentration range 1–20 mg/L. Quantification was performed using oleuropein for hydroxydecarboxymethyl oleuropein aglycone (HDCM-OA), hydroxyoleuropein aglycone (HOA), elenolic acid, and hydroxyelenolic acid; hydroxytyrosol for hydroxytyrosol and hydroxytirsol acetate; the respective standards for ferulic acid, *p*-coumaric acid, pinoresinol, apigenin and luteolin were used; and oleocanthal was used for oleocanthal, ligstroside aglycone, oleacein, and oleuropein aglycone.

#### *2.5. Statistical Analysis*

The statistical differences between samples of EVOO taken in different cooking conditions were analyzed by Statistica version 10.0.228.8 (StatSoft Inc., Tulsa, OK, USA) using the factorial ANOVA test. The assumption of normalization was graphically checked. To assess the importance of the contributing factors, multiple linear regressions were calculated. The form of the regression is as follows:

$$\text{Concentration} = \beta\_0 + \beta\_1 \cdot \text{T} + \beta\_2 \cdot \text{t} + \beta\_3 \cdot \text{Tt} \tag{1}$$

where T stands for temperature, t stands for time, and each β is the contribution of these factors. If its *p*-value is lower than 0.05, then β is significantly different from 0. The statistic *R*<sup>2</sup> is adjusted to the size of the model and can decrease if insignificant factors are added [30]. This parameter measures the proportion of the total variability explained by the model [31,32]. Even if the factors were not statistically significant, they were added to the model as confusing variables, and it was assessed if the model was more accurate with or without them. The model with the largest adjusted *R*<sup>2</sup> was selected. In order to build the model, the low temperature was the point −1 and the high temperature point was +1, and the same was applied for the short (−1) and long (+1) cooking time. Then, the value of β multiplied per 2 is the difference between the two levels of a factor.

#### **3. Results and Discussion**

#### *3.1. Total Polyphenols*

The concentrations of different polyphenols and of the groups found in raw and cooked EVOO samples are presented in Table 2.

When EVOO was heated in a pan, the sumatory of polyphenolic content decreased by around 40% at the low temperature (120 ◦C) and 75% at the high temperature (170 ◦C). Casal et al. (2010) reported a decrease of 50% in the total phenolic content, measured by the Folin-Ciocalteu method, after heating olive oil in a domestic deep-fat fryer at 170 ◦C for 3 h. [23]. However, in this study, the oil was deep fried so the samples were less exposed to oxygen and light, which may explain why the results are different to those presented here. Moreover, Folin-Ciocalteau methods are not selective, so this variation is not measuring only the phenol content, but also other reducing compounds. For this reason, it is also difficult to compare the results with those showed by one recent study, in which the degradation of the total phenolic content during a sautéing process was evaluated. The authors showed a decrease of approximately 50% of the antioxidant capacity measured by the Folin-Ciocalteau method after sautéing typical Mediterranean vegetables (potato, eggplant, tomato, and pumkin) for 10 min at 100 ◦C [33].


**Table 2.** Polyphenolic concentration of raw and processed extra-virgin olive oil (EVOO) expressed in mg/kg of EVOO.

HDCM-OA: Hydroxydecarboxymethyloleuropein Aglycone; HOA: Hydroxyoleuropein Aglycone; T: temperature; t: time; ↑ high level of the factor; ↓ low level of the factor.

For the ANOVA and multiple regression models, the normality of residuals was verified. To check this assumption, normal probability plots of the residuals were plotted for each compound. The graph for the sum of polyphenols is shown in Figure 1. The results of the ANOVA test and the linear regression models are shown in Table 3. The temperature was mainly responsible for the polyphenols depletion and there were no significant effects from time or the interaction. These results are in accordance with those reported by Goulas et al. (2015), who showed that heating the oil at 180 ◦C for 1 h or for 5 h made no difference in polyphenol content decrease [8].

**Figure 1.** Normal probability plot of the sum of phenols.

*Antioxidants* **2020**, *9*, 77


**Table 3.** Statistical results of the ANOVA and the lineal models.

The model for the sum of polyphenols was great fitted, with a *R*<sup>2</sup> of 0.97, in which 97% of the variance is explained by the model. The slope for temperature was significantly different from 0, suggesting that a longer cooking period does not change the polyphenolic fraction when the EVOO is processed only once. As the low level is −1 and the high level is +1, and the β of the temperature is −131, then cooking using a high temperature decreased the polyphenol content 232 mg/kg more than applying a moderate temperature, which represents 27% of the raw EVOO concentration.

#### *3.2. Secoiridoids*

Secoiridoids are the largest group of EVOO polyphenols. Secoiridoids include oleuropein, ligstrosides, and their derivatives. Some of them have been reported to have important benefits to health, such as oleocanthal or oleacein [34]. Oleocanthal has demonstrated anti-inflammatory effects [35] and a protective role against some diseases, such as Alzheimer disease [34], and oleacein has proven to protect against cardiovascular diseases, reducing hypertension [36] and inhibiting neutrophils adhesion [37].

During the cooking process, secoiridoids decreased 45% at the low temperature and 70% at the high temperature. Among this group, a different behavior was observed in hydroxyelenolic acid, which is not a polyphenol but a related compound produced by the ester breakdown of ligstroside, oleuropein, and their aglycones [38]. Thus, the formation of hydroxyelenolic acid was enhanced by processing at a moderate temperature. However, a longer cooking period and a higher temperature promoted its degradation, giving a lower concentration.

According to ANOVA analysis, the factor responsible for the depletion of secoiridoids was temperature. However, different results were found for each secoiridoid, in which oleacein and oleuropein aglycone were also affected by the interaction of time and temperature, and hydroxydecarboxymethyl oleuropein aglycone and hydroxyoleuropein aglycon were affected by all of the evaluated factors. These results are in accordance with those reported by Attya et al. (2010), in which heating at 90 ◦C was shown to cause almost no degradation of oleocanthal and oleacein in EVOO, but at 170 ◦C the concentration of both compounds was reduced by half, reflecting the major role played by temperature in polyphenol degradation [16].

The models were properly fitted for most of the secoiridoids analyzed, however, oleocanthal showed a *R*<sup>2</sup> of 0.7 because of its high reactivity, which also prompted the development of a new method for its specific analysis [29]. Oleocanthal presents keto-enolic tautomerism, which impeded its proper analysis, as it reacts with the solvent during the chromatographic separation. Oleocanthal may have reacted more in some samples than others, but even with only a 70% model fit, the result indicates that the cooking time did not change the oleocanthal concentration. In contrast, it decreased by 100 mg/kg of oil after cooking at a high temperature compared to the moderate temperature.

In the case of the sum of secoiridoids, the ANOVA results showed that the interaction factor was not significant, although in the multiple regression model the β for the interaction was different to 0, indicating that there was an effect. This difference occurred because the test used for ANOVA and the test used for the regression model were different: ANOVA applies a *F*-test, and, for the regression models, a t-test is used. Despite the ANOVA result giving a *p*-value of over 0.05, it showed a trend that this factor had an effect on secoiridoid degradation (*p*-value = 0.052).

According to the slopes of the models, the temperature was mainly responsible for the depletion of seicoiridoids during a domestic sautéing process. Slopes were analyzed as a percentage of the initial concentration (in raw EVOO), as the initial concentration for each polyphenol differed substantially, making it difficult to compare the models between the different polyphenols. The results are shown in Table 3. The slopes represent the values ranging from 12% (ligstroside aglycone) to 20% (oleuropein aglycone) of their original concentrations. The most different one was oleocanthal, which showed just a 6.5% depletion and withstands better the temperature than oleuropein aglycone. The compounds with a *o*-diphenol group were the most reactive, with a slope representing between 18% (hydroxydecarboxymethyl oleuropein aglycone) an 20% (oleacein) of their initial concentration. On the

other hand, oleocanthal and ligstroside aglycone (compounds with just one hydroxyl group) showed less reactivity. *ortho*-Diphenols are the most reactive, as they can be converted easily to *ortho*-quinones through a radical reaction [39–41]. Also, the intermediates of the reaction are radicals too, so they stabilized by the hydroxyl in the ortho position [40]. This rapid conversion may be responsible for the higher degradation compared to single phenols. This difference in the reactivity is also reflected in the activation energy, in which oleocanthal presents lower than oleacein because a higher temperature change is needed to degrade oleocanthal at the same rate as oleacein [16].

As mentioned above, hydroxyelenolic acid was an exception, its concentration was affected by the cooking time and the interaction between time and temperature, but not the temperature alone. For this compound and for elenolic acid, the time factor had a positive effect as the slopes were positive, indicating that frying with EVOO for longer periods may increase their concentration. Although the hydroxyelenolic acid model was not well fitted, the elenolic acid model showed an 82% fitness result. Like hydroxyelenolic acid, elenolic acid is not a phenol, but a derivative of oleuropein and ligstroside aglycones. Thus, despite a long cooking process degrading some of the polyphenols, it can enhance some related and new compounds.

#### *3.3. Phenolic Alcohols and Others*

Phenolic alcohols, mainly hydroxytyrosol and hydroxytyrosol acetate, are derivates from oleuropein, like the secoiridoids, but as their chemical behaviors are different, they are classified in a different group [27,28].

When frying at a low temperature for a short time, only 9% of phenolic alcohols decreased, although a depletion of 85% (90% in the case of hydroxytyrosol), when applying a high temperature and a long cooking time, was observed. At a low or moderate temperature, the hydroxytyrosol degradation, formed by the ester breakdown of oleuropein and its aglycones (Figure 2), may be counteracted by the rate of its generation. However, at a high temperature, its degradation is more likely to occur, resulting in a substantial reduction. Similar results are observed by Ramírez-Anaya et al. (2019), who showed that after sautéing typical Mediterranean vegetables at 100 ◦C, the hydroxytyrosol content only decreased between 25% and 50% [33]. Furthermore, a similar behavior was found by Krichene et al. (2015), who showed an increase in hytroxytyrosol concentration during the first months of storage due to the transformation of oleuropein and derivates in the compound, but after some months there was a high decrease in its concentration [42].

The concentration of phenolic alcohols was affected by temperature and time and hydroxytyrosol was also affected by their interaction, however, the β value for the temperature factor was higher, indicating that it is mainly responsible for their degradation.

Hydroxytyrosol was the most degraded compound by temperature, its slope represents 30% of its initial concentration. Thus, the amount of hydroxytyrosol diminished greatly—about 60% of the hydroxytyrosol concentration in the raw EVOO cooked at a high temperature compared to the EVOO cooked at a low temperature. So, cooking at low temperature should be recommended due to the fact that, according to the European Food and Safe Authority, it protects low-density lipoproteins (LDL) from oxidative damage [43] that have proven health effects.

The minor groups of polyphenols in EVOO are phenolic acids, lignans, and flavones. There was no possibility to build a model for those groups because of their low concentration. The only model properly fitted (>80%) for flavones was luteolin that was mainly affected by temperature. In the case of lignans, the only compound present in quantifiable amounts was pinoresinol, which increased during cooking probably because of the transformation of 1-acetoxypinoresinol and because of its high temperature stability [44].

**Figure 2.** Normal probability plot of the sum of polyphenols.

#### **4. Conclusions**

In this work, we determined changes in the EVOO polyphenolic profile during a domestic sautéing process commonly used in the Mediterranean diet, simulating the cooking conditions of a home kitchen, without the control of light or oxygen. The cooking temperature was the most important factor in the degradation of EVOO polyphenols. In the case of time, it was only a significant factor for some polyphenolic compounds and was not a significant factor for the sum of the polyphenols. Sautéing at a low temperature changes the polyphenolic profile of EVOO by increasing the concentration of hydroxyelenolic acid and depleting other compounds. Besides, this oil would still have the amount of polyphenols, with values higher than 250 mg/kg of hydroxytyrosol, tyrosol, or the derivates necessary to inhibit LDL oxidation [43]. Furthermore, research is needed to determine if there are differences in EVOO polyphenol degradation when proteins or complex sugars are present and whether the presence of these phenolics for their antioxidant properties could avoid the formation of secondary undesirable compounds that originated from the cooking and food processing.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-3921/9/1/77/s1, Table S1: Ionization conditions for compounds analyzed with the method 1, Table S2: Ionization conditions for compounds analyzed with the method 2.

**Author Contributions:** Conceptualization, J.F.R.d.A. and R.M.L.-R.; Formal analysis, J.L.-C. and J.F.R.d.A.; Investigation, J.L.-C., A.V.-Q., M.I. and X.T.-P.; Methodology, J.L.-C. and J.F.R.d.A.; Supervision A.V.-Q. and R.M.L.-R.; Writing—original draft preparation, J.L.-C.; Writing—review and editing, A.V.-Q., J.F.R.d.A. and R.M.L.-R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by CICYT [AGL2016- 75329-R]. Julián Lozano-Castellón thanks the Ministry of Science Innovation and Universities for the FPI contract [BES-2017-080017]. Anna Vallverdú-Queralt thanks the Ministry of Science Innovation and Universities for the Ramon y Cajal contract (RYC-2016-19355). José Fernando Rinaldi de Alvarenga thanks the Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP) for the post-doc grant [2019/11324-8].

**Acknowledgments:** The authors wish to thank the CCiT-UB for the mass spectrometry equipment. The authors also wish to thank the Fundación Patrimonio Comunal Olivarero for the EVOO.

**Conflicts of Interest:** Lamuela-Raventós reports receiving lecture fees from Cerveceros de España and receiving lecture fees and travel support from Adventia. The other authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Phenolic Compounds and Bioaccessibility Thereof in Functional Pasta**

#### **Valentina Melini \*, Francesca Melini and Rita Acquistucci**

CREA Research Centre for Food and Nutrition, Via Ardeatina 546, I-00178 Roma, Italy; francesca.melini@crea.gov.it (F.M.); rita.acquistucci@crea.gov.it (R.A.)

**\*** Correspondence: valentina.melini@crea.gov.it

Received: 6 April 2020; Accepted: 20 April 2020; Published: 22 April 2020

**Abstract:** Consumption of food products rich in phenolic compounds has been associated to reduced risk of chronic disease onset. Daily consumed cereal-based products, such as bread and pasta, are not carriers of phenolic compounds, since they are produced with refined flour or semolina. Novel formulations of pasta have been thus proposed, in order to obtain functional products contributing to the increase in phenolic compound dietary intake. This paper aims to review the strategies used so far to formulate functional pasta, both gluten-containing and gluten-free, and compare their effect on phenolic compound content, and bioaccessibility and bioavailability thereof. It emerged that whole grain, legume and composite flours are the main substituents of durum wheat semolina in the formulation of functional pasta. Plant by-products from industrial food wastes have been also used as functional ingredients. In addition, pre-processing technologies on raw materials such as sprouting, or the modulation of extrusion/extrusion-cooking conditions, are valuable approaches to increase phenolic content in pasta. Few studies on phenolic compound bioaccessibility and bioavailability in pasta have been performed so far; however, they contribute to evaluating the usefulness of strategies used in the formulation of functional pasta.

**Keywords:** phenolic compounds; bioactive compounds; functional pasta; gluten-free pasta; bioaccessibility; bioavailability; whole grain; composite flour; legumes; food by-products

#### **1. Introduction**

Phenolic compounds are secondary plant metabolites with strong antioxidant activity [1]. The consumption of food products rich in phenolic compounds has been associated with a reduced risk of chronic disease onset and ageing [2,3]. Currently, Phenol-Explorer, the first comprehensive database on polyphenol content in foods, reports the content for 500 phenolic compounds in 400 foods, for a total of 35,000 values. Fruit and vegetables are the main source of these secondary plant-metabolites.

Cereal grains contain significant amounts of phenolic compounds, as well [4,5]. Nevertheless, daily consumed cereal-based products, such as bread and pasta, are not a carrier of phenolic compounds, since they are produced with refined flour or semolina. Most bioactive compounds are concentrated in the outer layers of cereal grains which are discarded as bran, while flour and semolina are obtained from the starchy endosperm layer [6]. Hence, phenolic compounds are commonly lost during milling.

Pasta is one of the staple foods of the Mediterranean diet. It composes the base of the food pyramid and a daily consumption is recommended [7]. Pasta is a good source of carbohydrates and energy. One serving of 100 g of pasta (cooked, unenriched, without added salt) contains about 31 g of carbohydrates, 26.01 g starch, 1.8 g total dietary fibre, 5.8 g protein, and 0.93 g lipid (fat), and provides about 158 kcal [8]. When pasta is cooked al dente, it also has a low glycemic index, ranging around 32–40, depending on the pasta type [9]. Pasta glycemic index is far lower than that of bread. Additionally, pasta can possibly slow digestion rates and may contribute to longer

satiety [10–13]. Pasta has also additional unquestionable advantages, such as ease of preparation, long shelf-life, low price and global consumption. It is consumed by people of all ages and from all walks of life. Hence, it may be an optimal carrier of phenolic compounds.

Currently, the focus of nutritional science has shifted toward the concept of optimal nutrition, which aims at optimizing the daily diet in terms of nutrients and non-nutrients. Hence, the demand for functional food products with a well-balanced nutritional composition and contributing to maintaining wellbeing and health, has grown.

In this framework, novel formulations of functional pasta have been proposed and innovation in pasta-making has been prompted. The aim of this paper is to identify which formulations of functional pasta contribute to a higher intake of phenolic compounds, and greater bioaccessibility and bioavailability thereof. The consumption of food products with a high number of bioactive compounds does not necessarily imply beneficial effects on human health. Bioaccessibility studies are, thus, mandatory, to evaluate the bioactivity of a functional product. To this aim, the strategies used so far in formulation of functional pasta rich in phenolic compounds, both gluten-containing and gluten-free, will be reviewed. In addition, studies on phenolic compound bioaccessibility and bioavailability in pasta will be discussed, in order to evaluate the usefulness of these strategies and provide a basis for further investigations.

#### **2. Dietary Phenolic Compounds**

#### *2.1. Structure*

Phenolic compounds are a heterogeneous group of bioactive compounds produced in plants, via either the shikimate or the acetate pathway [14]. They include a variety of chemical structures having one or more phenolic groups as a common structural feature.

Based on the number of phenol rings and the structural elements that bind rings one to another, they can be classified into: (i) simple phenols; (ii) phenolic acids; (iii) flavonoids; (iv) xanthones; (v) stilbenes; and (vi) lignans [15], while a broader classification divides phenolic compounds into flavonoids and non-flavonoids [16]. Flavonoids show a distinctive benzo-γ-pyrone skeleton and occur as aglycones, glycosides and methylated derivatives. They comprise flavonols, flavan-3-ols, flavones, isoflavones, flavanones, anthocyanidins and dihydrochalcones. Non-flavonoids include diverse classes of polyphenols, such as phenolic acids and stilbenes [16]. Among non-flavonoids of dietary significance, phenolic acids play a pivotal role and are a major class in grains. They include hydroxybenzoic acids (C6–C1), such as gallic, *p*-hydroxybenzoic, vanillic, syringic, protocatechuic and ellagic acids, as well as hydroxycinnamic acids (C6–C3), namely *p*-coumaric, caffeic, ferulic, sinapic and chlorogenic acids (Table 1).

Phenolic compounds may occur in free, soluble conjugated, and bound form, depending on whether they are bound to other constituents, or otherwise. Hence, they can be classified as free phenolic compounds (FPCs), soluble conjugated phenolic compounds (EPCs) and insoluble bound phenolic compounds (BPCs) [17]. EPCs are esterified to other molecules such as fatty acids, while BPCs are covalently bound to cell wall constituents, such as pectin, cellulose, arabinoxylans and structural proteins. BPCs are the main fraction of phenolic compounds in wheat grains [18,19].


**Table 1.** Major classes of dietary phenolic compounds, skeleton structure thereof and common representatives.

#### *2.2. Bioaccessibility, Biotransformation and Bioavailability*

The concept of bioavailability in nutrition has been borrowed from pharmacology. In this discipline, the term "bioavailability" refers to the fraction of the administered dose of drug that enters systemic circulation, so as to access the site of action [20]. In nutrition, bioavailability refers to the amount of a nutrient or bioactive compound which becomes available for normal physiological functions or storage, after absorption by the gut [21].

The first step, necessary for a food component to become bioavailable, is the release from the food matrix. The extent at which a nutrient or bioactive molecule is released from the food matrix into the gastrointestinal tract and is in the right form to be absorbed, is referred to as bioaccessibility [22].

The bioaccessibility and bioavailability of phenolic compounds are affected by factors related to phenolics, food matrix and host (Figure 1).

**Figure 1.** Factors affecting bioaccessibility and bioavailability of phenolic compounds.

As regards the relationship between phenolic characteristics and bioavailability, it has been observed that chemical structure (degree of polymerization and molecular size), glycosylation and conjugation with other phenolics, and solubility are critical factors [23]. For example, phenolic acids, isoflavones, catechins and quercetin glucosides are easily absorbed, while large polyphenols are poorly absorbed.

Generally, phenolic compounds in liquid foods are more bioaccessible than those in solid foods. However, differences in phenolic bioavailability among liquid matrices have been observed. The occurrence of alcohol, dietary fibre or other nutrients, such as carbohydrates, lipids and proteins, may in fact influence phenolic compound bioavailability because of the interactions between phenolics and matrix constituents. Food processing may positively or negatively affect phenolic compound bioaccesibility and bioavailability, as well. Lafarga et al. observed that cooking increased the bioaccessibility of phenolic compounds in pulses [24]. Zeng et al. found that the content of bioaccessible phenolics in brown rice and oat significantly decreased (by 31.09% and 30.95%, respectively) after improved extrusion-cooking treatment, while in wheat they were almost unchanged, possibly because of differences in the cereal matrix [25]. It should be also considered that processing can cause a loss of phenolic compounds while promoting their bioaccessibility. Hence, bioavailability is a compromise between the compounds lost during processing and those absorbed into the organism [26].

Host-related factors—such as physiological conditions, disorders or pathologies, gastric emptying, enzyme activity, intestinal transit time and colonic microflora—may influence bioaccessibility and bioavailability of phenolic compounds, as well.

The bioavailability of a phenolic compound implies: (i) its release from the food matrix; (ii) gastric and small-intestinal digestion (likely change of phenolic compound structure due to hydrolysis of glycosides and phase I/II metabolism); (iii) cellular uptake of aglycons and some conjugated phenolics by enterocytes; (iv) microbiological fermentation of non-absorbed polyphenols or phenolics re-excreted via bile or the pancreas, to produce additional metabolites; (v) modifications by phase I/II enzymes, upon uptake in the small intestine or in the colon; (vi) transport into the blood stream and redistribution to tissues; (vii) excretion via the kidney or re-excretion into the gut via bile and pancreatic juices (Figure 2).

**Figure 2.** Representation of digestion, absorption and excretion of phenolic compounds and metabolites thereof.

Generally, after the absorption step, phenolics undergo phase I and II metabolic transformation, and metabolites with improved bioactivity or completely inactive can be obtained. As an example, protocatechuic acid, phloroglucinaldehyde, vanillic acid, and ferulic acid are bioactive metabolites obtained by the catabolism of cyanidine-3-*O*-glucoside in the gastrointestinal tract that contribute to maintaining intestinal integrity and function [27]. Hence, the evaluation of polyphenol bioavailability should include not only the determination of native compounds, but also of metabolites thereof.

#### Methods to Evaluate Phenolic Compound Bioaccessibility and Bioavailability

Several approaches have been proposed to evaluate phenolic compound bioaccessibility and bioavailability. They include the use of in vitro methods and in vivo models [23]. In vitro methods comprise simulated gastrointestinal (GI) digestion, artificial membranes, Caco-2 cell cultures and ussing chambers. As regards in vitro digestion models, they can be either static or dynamic [28]. Static models consist of multiple phases, including oral digestion (OD), gastric digestion (GD), intestinal digestion (ID) and dialysate (DIA). Each phase can vary slightly among studies. They can differ in the incubation time and characteristics of the digestive juices, and can also be adjusted for pH on the basis of the specific gut compartment [29]. However, they operate in static mode across the whole process, with prefixed conditions and parameters in terms of concentrations and volumes of digested materials, enzymes and salts, among others. The INFOGEST digestion method is an example of standardised static model [28,30].

Dynamic models include physical and mechanical processes and consider the changes that occur during the digestive process, as well as different physiological conditions. They were developed because static methods do not provide an accurate simulation of the complex dynamic physiological processes occurring under in vivo conditions. A common and very sophisticated gut model is the TIM system, a multi-compartmental dynamic computer-controlled model, used to simulate the human digestive system and to study the bioaccessibility of many compounds, such as vitamins, minerals, as well as phenolics [31].

In addition to the aforementioned methods, gastrointestinal organs in laboratory conditions (ex-vivo models) and intestinal perfusion in animals (in situ model) can be also applied in bioaccessibility/bioavailability determination [23]. In-vivo approaches are based on animal or human studies.

#### **3. Strategies to Modulate Phenolic Compound Content in Pasta**

According to tradition, pasta is seemingly a very simple food, produced with one ingredient, i.e., semolina of durum wheat (*Triticum turgidum* L. var. *durum*), and one "reactant", i.e., water. Pasta final configuration, made of starch granules dispersed within the protein network, is hence obtained upon the biochemical modification of the two main constituents of durum wheat semolina (that is, proteins and starch) prompted by water addition, and mechanical and thermal energy.

Pasta by itself is a healthy food. It is a good source of carbohydrates and energy. However, in recent years, scientists and producers have been striving to develop new formulas, so that pasta can not only provide nutrients and energy, but also beneficially modulate one or more targeted functions in the body, by enhancing a certain physiological response and/or reducing the risk of disease [32]. These new formulations are known as functional pasta products.

The use of functional ingredients, such as whole grain and composite flours, as well as the addition of extracts from plant foods and food wastes, has been increasingly explored as a strategy to improve phenolic content in pasta and gluten-free pasta. In addition, processing technologies have been specifically applied to raw materials or to the pasta-making process in order to increase the content of bioactive components and their bioavailability (Table 2).


**2.**Modulationofphenoliccompoundcontentin

#### *Antioxidants* **2020**, *9*, 343



#### *Antioxidants* **2020**, *9*, 343


4-hydroxy-benzoic

 and salicylic acids.

**Table2.***Cont.*


**Table2.***Cont.*





TFC: Total Flavonoid Content; TPA(s): Total Phenolic Acid(s); TPC: Total Phenolic Content; TPs: Total Phenolics.

#### *3.1. Use of Functional Ingredients in Pasta-Making*

#### 3.1.1. Whole Grain Flours

According to the HEALTHGRAIN Consortium, whole grains (WGs) shall consist of "the intact, ground, cracked or flaked kernel after the removal of inedible parts such as the hull and husk. The principal anatomical components—the starchy endosperm, germ and bran—are present in the same relative proportions as they exist in the intact kernel" [62]. While agreement on the definition of "whole grain" has been reached, there is a lack of consensus on the definition of whole-grain foods, including "whole grain pasta" [63].

In Germany and Italy, "whole grain pasta" is pasta where 100% of the grain component in the final product is whole grain; in Denmark, pasta containing a percentage of whole grain equal or higher than 60% on a dry matter basis, can be classified as "whole grain pasta"; in France and the Netherlands, there are no regulations nor guidelines for whole grain pasta definition [63].

Two main factors explain these different levels of whole grains admitted in whole grain products. On the one hand, foods with high whole grain content are not universally appreciated by consumers, hence manufacturers need to use whole grain ingredients in a level enabling to obtain products with good sensory qualities. On the other hand, the content of whole grain ingredients used for product preparation, must be adequate to guarantee nutritional benefits to consumers.

Cereals included in the whole grain definition are wheat (including spelt, emmer, faro, einkorn, khorasan wheat, durums), rice (including pigmented varieties), barley (including hull-less or naked barley but not pearled), corn, rye, oats (including hull-less or naked oats), millets, sorghum, teff, triticale, Canary seed, Job's tears, fonio, black fonio and Asian millet. Pseudocereals included in the whole grain definition are amaranth, buckwheat and tartar buckwheat, quinoa, and wild rice [62].

In whole grain flours, the outer multi-layered skin (bran) and the germ are retained together with the starchy main part of the grain. The bran is a major source of phenolic acids, dietary fibre (DF) and minerals, while the germ contains vitamins, minerals, fats and some proteins [64]. Phenolic acids, together with DF, are components responsible for many of the health effects associated with whole grain consumption [25]. They have shown to act synergistically and modulate favourably appetite, glucose metabolism, insulin sensitivity, and gut microbiota composition [65], and to have a role in the prevention and treatment of cardiovascular diseases [66,67]. Several studies have evidenced a lower risk from all causes and disease-specific mortality associated with a high intake of WGs [68].

The content of phenolic compounds in whole grain pasta has been, however, poorly investigated. Wójtowicz et al. determined the qualitative and quantitative profile of phenolic compounds in precooked pasta prepared from whole grain wheat and whole grain spelt [33]. Protocatechuic, 4-hydroxybenzoic, vanillic, syringic, *trans*-*p*-coumaric, *cis*-*p*-coumaric, *trans*-ferulic and *cis*-ferulic acids were identified in samples under investigation. *Cis*-ferulic acid was the main phenolic acid in both whole grain wheat and spelt pasta. In whole grain wheat pasta, vanillic acid was the second more abundant phenolic compound, while in whole grain spelt pasta, syringic and vanillic acids were identified as the main phenolics, after *cis*-ferulic acid. Compared to refined flours, the use of whole flours enabled to double the intake of phenolic acids. These data are in keeping with Chen et al. who found ferulic acid as the dominant phenolic compound in six whole grain wheat products, with values ranging between 99.9 and 316.0 μg/g [34]. In whole wheat pasta (41.4% fortification), Total Phenolic Acid (TPA) content was 226.7 μg/g dm.

Hirawan et al. determined the total phenolic content (TPC) in regular and whole grain spaghetti, and found that the former had a TPC level 2-fold lower than the latter [35]. TPC values in whole wheat spaghetti ranged between 1263 and 1423 μg/g Ferulic Acid Equivalents (FAE)/g dm, while in regular spaghetti TPC ranged between 718 and 927 μg/g FAE/g dm. It was also observed that all whole wheat spaghetti samples contained ferulic acid, while this compound was detected only in two out of regular spaghetti samples. However, TPC significantly decreased after cooking (about 40%), both in regular and whole wheat spaghetti. Despite the differences in the TPC, regular and whole grain spaghetti

exhibited the same antioxidant capacity, possibly due to the antioxidant components, such as the Maillard reaction products, formed during pasta drying.

#### 3.1.2. Composite Flours

Composite flours are blends of wheat and varying proportion of legumes, tubers or other cereals, including minor cereals, and pseudocereals. Cassava, maize, rice, sorghum, millets, potato, barley, sweet potato and yam are common ingredients of composite flours [69].

The concept of using composite flours in bread and pasta-making was first elaborated to tackle a low availability of wheat in areas whose climatic conditions are not suitable for wheat production, and to encourage the use of autochthonous crops with economic advantages for local producers and consumers [69]. The concept thus first had an economic value. However, partial or total wheat substitution with composite flours affects also the nutritional profile of the final product. Wheat is, in fact, deficient in essential amino acids, such as lysine and threonine, and, during milling, bioactive compounds and minerals are commonly lost. Hence, the use of composite flours contributes to counteracting these deficiencies. More recently, the concept of "composite flours" has been thus extended to blends of wheat flour/semolina and other flours richer in essential amino acids, minerals, vitamins and phenolic compounds.

Blends of cereal flours with pulse flours have been by far explored in pasta-making. Pulses are an important source of nutrients [69]. They have a low glycemic index and are rich in complex carbohydrates, DF, plant proteins, and micronutrients. They also have high levels of polyphenols with good antioxidant properties, and other plant secondary metabolites and components (i.e., isoflavones, phytosterols, bioactive carbohydrates, alkaloids, and saponins), that are being increasingly recognized for their bioavailability and potential benefits for human health. Among phenolic compounds, phenolic acids, flavonoids and condensed tannins are the most abundant [70].

The use of pulses in pasta-making and their contribution to the content of phenolic compounds have been recently investigated by Turco et al. [36]. They found that, in pasta formulated by wheat semolina and 35% faba bean (*Vicia faba* L.) flour, TPC increased from 63.8 mg Gallic Acid Equivalents (GAE)/100 g dry matter (dm) to 185.3 mg GAE/100 g dm. Cota-Gastélum et al. prepared functional pasta with varying proportions of wheat (*T. durum* L.) semolina (0–100%), chickpea flour (0–90%), and chia flour (0–10%) [37]. In raw samples, the highest phenolic content (approximately 16 mg GAE/g) was observed when durum semolina was totally replaced and a blend of 10% chia flour and 90% chickpea flour was used. This value was approximately 8-fold higher than in durum wheat pasta (2 mg GAE/g) [37]. Carob flour, which is obtained from carob seeds, has been also used in substitution of semolina in pasta-making. Se¸czyk et al. produced pasta by using varying percentages of carob flour (1–5%) [38]. They found that the phenolic content in the produced functional pasta was higher than in the control pasta (3.51 mg GAE/g dm). In pasta with 1% of carob flour, TPC was 5.27 mg GAE/g dm, and it increased to 12.12 mg GAE/g dm in pasta with 5% carob flour [38].

Pseudocereal flours were also used to partially or totally replace semolina in pasta-making, in order to enhance pasta nutritional profile. Pseudocereals are, in fact, characterized by a high nutritional composition, in terms of high content in DF, high-quality protein, essential minerals, vitamins (e.g., folic acid), essential amino acids and unsaturated fatty acids [71,72]. They are also a valuable source of phenolic compounds [73]. Varying levels of amaranth seed flours and dried amaranth leaves (35%, 50%, 55% and 70%) were used as semolina substituents in the preparation of elbow-type pasta [39]. Both grains and leaves are, in fact, rich in bioactive compounds. Grains also show high levels of proteins (15 g/100 g) and are a source of vitamins, such as thiamine, niacin, riboflavin and folate, and minerals, namely iron, calcium, zinc, magnesium, phosphorus, copper, and manganese [74,75]. The study by Cárdenas-Hernández et al. showed that, whichever the substitution levels, amaranth pasta had a TPC higher than 100% semolina pasta (0.98 mg of FAE/g dm), with values ranging from 1.54 to 3.37 mg FAE/g dm [39]. The highest value was observed in pasta with a semolina:amaranth flour/leaves ratio of

65:35. A significant decrease in phenolic content (15–27%) was observed in all amaranth pasta samples, after cooking [39].

Composite flours have been also used to improve the nutritional value of gluten-free (GF) pasta. As a matter of fact, GF pasta is mainly produced with GF flours, such as rice and corn, which are low in micronutrients and bioactive compounds [76]. The use of blue maize in GF pasta-making has been recently explored. Blue maize (*Zea mays* L.), like the red and purple varieties, is rich in anthocyanidins (up to 325 mg/100 g dm), including cyanidin derivatives (75–90%), peonidin derivatives (15–20%) and pelargonidin derivatives (5–10%) [77]. Different percentages of blue maize (25%, 50% and 75%) were added to pasta dough produced with equal amounts of unripe plantain and chickpea flour [40]. It was observed that pasta samples containing 75% of blue maize presented the highest TPC retention after extrusion and cooking. Upon extrusion, TPC in pasta decreased between 20% and 30%, while an additional 10% loss occurred upon cooking. The phenolic compounds, retained after extrusion, were likely bound phenolics, whereas free phenolic species (e.g., free phenolic acids and anthocyanins), not physically trapped in the protein network, were leached into the cooking water.

The fortification of traditional GF flours with sorghum (*Sorghum bicolor* (L.) Moench) flour in pasta-making has been also studied. Sorghum has, in fact, high levels of a diverse array of beneficial bioactive components (e.g., polyphenols, especially flavonoids), and bioactive lipids (such as policosanols and phytosterols) [78–80]. Palavecino et al. produced GF pasta with white and brown sorghum [41]. They compared the two sorghum-based formulations to GF pasta produced with rice, maize and soy flour. Total phenolic compound content was higher in the two sorghum-based pasta samples than in the controls, with a value of 2.41 g GAE kg−<sup>1</sup> and 2.88 g GAE kg−<sup>1</sup> for white and brown sorghum, respectively. Sorghum pasta, after cooking, also showed higher radical scavenging activity and ferric reducing ability than the control samples, without significant differences between sorghum varieties.

#### 3.1.3. Powders and Extracts from Plant Foods and Food By-Products

The use of powders and extracts from plant foods and food by-products in pasta-making is among the strategies recently explored to obtain functional pasta, both gluten-containing and gluten-free.

Functional pasta was prepared by incorporating carrot powder (10%), mango peel powder (5%), moringa leaves powder (3%) and defatted soy flour (15%) in a blend of wheat semolina and pearl-millet [42]. Total flavonoid content (TFC) was determined in order to evaluate the contribution of these ingredients to the phenolic content in pasta. It emerged that, in the control pasta, TFC was 6.30 mg/100 g. The addition of mango peel powder and moringa leaves powder provided the highest values (16.53 and 17.98 mg/100 g, respectively), while carrot powder and defatted soy flour contributed at a lower extent, with values of 7.63 and 8.03 mg/100 g, respectively.

Mushrooms can also contribute to the phenolic dietary intake. The study by Lu et al. investigated the contribution of mushroom powder addition to the phenolic content of spaghetti [43]. Three different powders were used: from white button, from shiitake and from porcini mushrooms. Three different semolina substitution levels were tested: 5 g, 10 g and 15 g/100 g (*w*/*w*). It emerged that all mushroom-powder-supplemented pasta samples showed TPC values significantly higher than semolina pasta, except for 5% and 10% shiitake mushroom pasta. The greatest values were found in porcini mushroom pasta samples (approximately 4–5 mg GAE/g dm), followed by the second button mushroom samples (approximately 2 mg GAE/g dm), and shiitake mushroom pasta.

Plant food industrial processing produces huge amounts of by-products that are a serious disposal issue. However, some by-products have shown to be an abundant source of valuable compounds [81]. Hence, in the domain of circular economy, they have been increasingly turned into functional ingredients. Vegetable wastes, such as peelings, trimmings, stems, seeds, shells, and bran are some by-products from which phenolic compounds can be extracted [82]. Ultrasound-assisted extraction, microwave-assisted extraction, supercritical fluid extraction, pressurized fluid extraction, pulsed electric field extraction and enzyme-assisted extraction are green technologies commonly used in the recovery of phenolic compounds from food wastes [82–84]. The choice of the extraction technique is related to factors including the functional ingredient to extract and the characteristics of the food matrix.

Onion dry skin powder has been used as a functional ingredient to modulate phenolic compound content in pasta [44]. Onion dry skins are by-products generated during industrial peeling and contain bioavailable compounds such as DF, fructo-oligosaccharides and quercetin aglycones. In the study by Michalak-Majewska et al., semolina was replaced by varying amounts of onion powder: 0%, 2.5%, 5% and 7.5%. TPC and TFC were determined both in raw and cooked samples. It was observed that pasta added with onion skin powder showed TPC and TFC higher than the control (100% semolina pasta). The highest TPC was found in pasta with 7.5% substitution level. Moreover, cooked pasta showed TPC not significantly different from the corresponding raw sample, whichever addition level of onion skin powder. Conversely, in the control pasta, TPC decreased after cooking. Hence, the functional pasta ensured a higher intake of phenolic compounds, compared to 100% semolina pasta. As regards TFC, the addition of onion skin powder enabled to obtain pasta with higher level of flavonoids, and after cooking a significant increase was observed.

Durum spaghetti were formulated by the addition of olive paste powder [45]. Olive paste is an industrial by-product of olive oil production, rich in phenolic compounds [45]. Two levels of olive paste powder were added to semolina: 10% and 15%. Phenolic content was determined on spaghetti with 10% addition of olive paste powder, since they showed the best sensory properties. TPC was 82.39 μg/g dm in the control pasta and 245.08 μg/g dm in the enriched spaghetti. In 100% semolina pasta, vanillic acid was the most abundant phenolic compound in free form (0.56 μg/g dm), while ferulic acid was the main bound phenolic compound (67.70 μg/g dm). In spaghetti enriched with olive paste powder, vanillic acid was the main phenolic acid in free form as in the control pasta; however, its content was higher (7.28 μg/g dm) than in the control. HPLC analysis also showed that the addition of olive paste powder increased the content of flavonoids, such as quercetin and luteolin.

Functional spaghetti were also produced by addition of extracts from grape marc, made up of skins, seeds, and stalks [46]. TPC was determined on fresh extruded spaghetti, pasteurized extruded spaghetti and dry spaghetti. It was found that, compared to the control, the addition of grape marc extract increased TPC in all enriched spaghetti samples (approximately 700 mg GAE/100g dm). The pasteurization and drying process did not significantly affect the TPC. Interestingly, after cooking an increase in TPC was observed, with respect to the raw samples.

Bran is the main by-product of cereal milling and is a great source of phenolic compounds and minerals. Despite its functionality, its use in pasta-making is challenging, since it has adverse effects on the quality of the final products, such as an increase of cooking loss, swelling index, and water absorption in pasta [85]. Recently, bran aqueous extract was used in the production of spaghetti [47]. The extract was obtained by ultrasound assisted-extraction at 20 ◦C for 25 min. The ratio between water and bran was 10 L/kg. The bran aqueous extract completely substituted processing water in pasta-making. A significant increase in phenolic compounds was observed in pasta samples due to the bran extract. In detail, TPC was 127 mg FAE/100 g fresh weight (fw) in functional spaghetti and 97 mg FAE/100 g fw in the control pasta.

As regards the formulation of functional GF pasta, different percentages (5% and 10%) of chia (*Salvia hispanica* L.) milled seeds were incorporated into rice flour dough [48]. Chia seed addition allowed increasing phenolic acid content, besides the slowly digestible starch fraction of rice, and protein and DF content. The highest content of TPAs was observed in raw samples of pasta produced with 10% milled chia seeds (164.3 μg/g). TPA content in functional GF pasta did not significantly differ from durum wheat pasta (164.3 vs 149.08 μg/g), but it was by far higher than in pasta produced with commercial GF flour (10.30 μg/g) [48].

After cooking, TPA content was higher in all pasta samples, with an increase of 5.3% in durum wheat pasta, 14.8% in commercial GF pasta, 25.5% in pasta with 5% of milled chia seeds and 13.7% in pasta with 10% of milled chia seeds. The highest content in TPAs was observed in pasta with a 10% milled chia seeds (186.80 μg/g). The increase in TPA content in cooked samples was possibly due to the increased bioaccessibility of bound phenolic acids after boiling [48]. Samples also differed for the content of specific phenolic acids. Addition of milled chia seeds allowed obtaining pasta samples containing chlorogenic acid, which is otherwise absent in commercial GF and durum wheat pasta. Chia seed pasta was also rich in caffeic and vanillic acids, in contrast to durum wheat pasta. The higher was milled chia seed addition, the higher was the content of chlorogenic, caffeic, and vanillic acids.

Oniszczuk et al. investigated the phenolic profile of GF pasta prepared with a blend of rice and field bean flour, enriched with different amounts (2.5%, 5%, 7.5%, 10%, 12.5% and 15%) of pear prickly fruit (*Opuntia ficus indica* (L.) Mill.) [49]. The latter is a source of phenolic compounds and also provides vitamins (C, B1, B2, A, and E), minerals (calcium, potassium, magnesium, iron, and phosphorus), and other bioactive compounds, such as carotenoids and betalains. High-performance liquid chromatography/electrospray ionization tandem mass spectrometry (HPLC-ESI-MS/MS) showed that pasta samples enriched with the different amounts of pear prickly fruit were rich in several phenolic acids: protocatechuic, caffeic, syryngic, 4-OH-benzoic, vanilic, gentisic, *trans*-sinapic, *cis*-sinapic, *p*-coumaric, ferulic, isoferulic, *m*-coumaric, 3,4-dimetoxycinnamic, and salicylic acids. The dominant acid was isoferulic. The higher was the addition of pear prickly fruit, the higher was the content of phenolic acids. Antioxidant activity was also positively correlated with the addition of fruit.

The effect of chestnut fruit (*Castanea sativa* Mill.) addition (10%, 20%, 30%, 40%, and 50%) to the aforesaid blend of rice and field bean flour on pasta phenolic content, was also investigated [50]. Chestnut fruit is rich in phenolic compounds, as well as in proteins, unsaturated fatty acids, DF, vitamins and micronutrients. As regards the content of phenolic compounds, it was observed that the total content of free phenolic acids increased along with the chestnut addition. TPA content was 38.93, 46.98, 51.47, 56.59, and 65.01 μg/g dm in samples with 10%, 20%, 30%, 40% and 50% of chestnut flour, respectively [50]. The content of each phenolic acid also increased at a higher addition of chestnut fruit, with the exception of 4-hydroxy-benzoic and salicylic acids whose level decreased at the increase of chestnut flour addition. This trend might be explained by a content of these two acids higher in the rice and field bean flour blend than in chestnut fruit powder.

#### *3.2. Raw Material Processing, Pasta-Making and Pasta Cooking*

In addition to the use of raw materials naturally rich in phenolic compounds, such as whole grain flour, composite flours, and plant powders and extracts, raw material processing and modulation of pasta-making and pasta cooking parameters have been explored to increase the content of phenolic compounds in pasta.

Debranning, also known as pearling, is a technology based on the gradual removal of the outer bran layers prior to milling process. While in conventional milling the aleurone layer remains attached to the bran, in debranning it remains attached to the endosperm. As a consequence, semolina and flour obtained by debranning are richer in components commonly found in the grain aleurone. The technology also enables to isolate aleurone-rich fractions, which can be used as functional ingredients [86]. Abbasi et al. have recently formulated pasta enriched with a debranning fraction from purple wheat [51]. The debranning fraction (25%) was added to flour and to semolina by dry mixing, and macaroni pasta samples were prepared. Experimental analyses on raw materials showed that phenolic compounds in wheat flour and semolina were negligible compared to the debranning fraction. Despite the debranning technology enabled to obtain raw materials rich in phenolic compounds, pasta samples showed TPC lower than it was expected. This was possibly due to the degradation of phenolics during the pasta-making process, especially in the drying step.

One more study on the formulation of pasta products by using debranning fractions was reported by Zanoletti et al. [52]. Two functional pasta products enriched with a fraction obtained from either the first or the second debranning step of purple wheat were produced. The first fraction corresponded to a debranning level of 3.7% of whole grain, while the second fraction corresponded to 6% of the debranned grain after the first step. The content of anthocyanins, a subclass of phenolic compounds typical of many fruits, vegetables, and cereal grains with red, violet, and blue colour, was determined. The analysis of cooked samples showed that anthocyanin content was 67.9 μg/g dm in pasta enriched with the first debranning fraction and 60.0 μg/g dm in pasta added with the second debranning fraction. These two values were not significantly different. Moreover, anthocyanin content in functional pasta was higher than in pasta added with bran (28 μg/g dm) and used as control sample. In addition, both functional pasta products exhibited an antioxidant activity higher than the control.

Ciccoritti et al. prepared spaghetti enriched with debranning fractions of durum wheat cv Normanno. The addition level of debranning fraction was 30% *w*/*w* [53]. Phenolic acids (PAs) and TPC were determined in raw and cooked samples, in free, esterified and bound forms. It emerged that in raw samples, free PAs content was higher in the control pasta than in functional pasta. As far as conjugated and bound PAs are concerned, values were higher in enriched samples. Conjugated PAs were 59.4 mg/kg dm and bound PAs were 650.0 mg/kg dm in functional pasta, while in control pasta they were 21.6 and 27.2 mg/kg dm, respectively. A similar trend was observed for conjugated and bound TPC. The former was 110.7 mg/kg dm in functional pasta and 31.4 mg/kg dm in control pasta, while the latter was 1308.4 and 156.9 mg/kg dm, respectively. After cooking, it was observed a higher level of PAs, whichever form was considered. Conversely, free and conjugated TPC decreased, and bound TPC increased.

Ciccoritti et al. also explored the use of micronized fractions in pasta-making [53]. Micronization is a mechanical treatment consisting in reducing kernels into a fine powder. For this reason it is also known as ultrafine grinding. It differs from conventional milling because it produces wholegrain flour, without producing by-products such as bran. The treatment damages the fibre matrix, hence the phenolic compounds linked or embedded into the matrix are more bioaccessible. The content of PAs and total phenolics in experimental pasta was determined. Raw pasta prepared from debranned and micronized durum wheat had a higher level of PAs (conjugated and bound) and total phenolics (TPs) than the control. Conjugated PAs were 36.8 mg/kg dm and bound PAs were 357.3 mg/kg dm. The content of conjugated TPs was 75.8 mg/kg dm and bound TPs were 113.3 mg/kg dm. After cooking, the level of free PAs and conjugated PAs increased, while bound PAs decreased. As regards TPs, the content of free forms did not significantly differ from raw samples, while the content of conjugated TPs decreased and the level of bound TPs increased significantly. Data are in keeping with Martini et al. who observed that micronization preserved the content of phenolic acids, while conventional milling determined 89% decrease from seeds to cooked durum wheat pasta [54].

In addition to mechanical treatments, biological processes, such as germination and fermentation, are strategies enabling to increase the phenolic compound content in pasta products.

As regards germination, both cereal grains and pulses can be sprouted. Cereal seed germination may impact on nutritional properties of cereals [87] and possibly cereal-based products. An increase in phenolic compound content ranging from 1.2 to 3.6 folds was reported in wheat, barley, sorghum, rye, oat and brown rice, after germination [87–94]. During sprouting, cell wall-degrading enzymes, such as cellulases, endoxylanases and esterases, are biosynthesized. They can hydrolyze phenolic compounds bound to cell wall constituents, so as to increase the content of free phenolic compounds. Moreover, thanks to the effect of enzymes, bound phenolic compounds are more soluble in extraction solvents and more bioaccessible. Merendino et al. explored the use of sprouted cereals in pasta-making [55]. Spaghetti were formulated by replacing wheat semolina with 30% dry tartary buckwheat sprouts belonging to the Slovenian landrace Ljse. TPC in raw tartary buckwheat spaghetti was 3.7 mg GAE/g, while it was 0.3 mg GAE/g in 100% semolina spaghetti. After cooking, TPC was 2.2 and 0.2 mg GAE/g in tartary buckwheat and in control spaghetti, respectively. Flours from sprouted legumes have been also used in pasta-making, since sprouting increases phenolic compound content in legumes [95]. In detail, Bruno et al. investigated the contribution of sprouting to increasing phenolic content in chickpea pasta [56]. Pasta prepared with sprouted chickpea flour had phenolic content 15% higher than non-sprouted chickpea pasta.

Sourdough-fermented ingredients have been recently proposed to enhance the nutritional and functional properties of pasta [96]. However, to our knowledge, no study investigating the phenolic profile of pasta produced with sourdough-fermented flours has been so far published. Fermentation is indeed one more pre-processing technique that can increase the content of phenolic compounds in pasta ingredients, such as pulses and pseudocereals, and bran. Microorganisms responsible for the fermenting process, produce enzymes which boost the release of insoluble bound phenolic compounds from the food matrix, thus increasing their solubility and bioaccessibility. Rashid et al. reported a content of extractable phenolic compounds, in rice bran fermented by *Aspergillus oryzae*, 3.8-fold higher than in unfermented bran [97]. It was also observed that rice bran solid-state fermentation with *A. oryzae* affected the profile of phenolic acids. In unfermented bran, protocatechuic, coumaric and ferulic acids were found, while in fermented rice bran *p*-coumaric, protocatechuic, ferulic, caffeic and sinapic acids were detected. Moreover, the level of coumaric, ferulic and protocatechuic acids in the fermented bran increased by up to 3.2-fold, 52-fold and 3.2-fold, respectively, compared to its unfermented counterpart. Dey et al. studied the effect of solid state fermentation of wheat by *Rhizopus oryzae* RCK2012 on phenolic content and they found that TPC increased from 5.15 mg GAE/g to 24.55 mg GAE/g [98]. Călinoiu et al. explored the use of solid-state fermentation to improve phenolic content in wheat and oat bran [99]. A 112% increase in TPC was observed on day 3 fermentation in wheat bran and 83% increase on day 4 of fermentation in oat bran, with values reaching 0.84 mg GAE/g and 0.45 mg GAE/g, respectively. Based on these results, it can thus be speculated that fermentation of raw materials may contribute to increasing phenolic content in pasta; however, additional studies on the effect of pasta-making upon TPC are required.

The pasta-making process can also influence the content of phenolic compounds; process parameters and conditions may be thus set in order to limit/avoid phenolic compound degradation and/or increase their bioaccessibility.

Generally speaking, conventional pasta is produced by forcing flour/semolina dough through a die to obtain the required shape, and then drying it. Low shear and heat values (30–40 ◦C) are applied. Gluten-free and precooked pasta are prepared by extrusion-cooking, which is a high-temperature short-time process consisting in a short-term heating of dough, at a high temperature, under high pressure. During extrusion-cooking, raw materials are forced to flow through a die and thermal and shear energies cause structural, chemical, and nutritional transformations, including gelatinization and degradation of starch, denaturalization of proteins, oxidation of lipids, degradation of vitamins and bioactive compounds, and changes in bioavailability of minerals and solubility of dietary fibre [100]. The combination of high temperature, high pressure, and high shearing conditions during extrusion-cooking may affect the content in phenolic compounds. Heat can cause the decomposition of heat-labile phenolics and polymerization of some phenolic compounds, thus decreasing their content. At the same time, heat disrupts cell wall matrices which hinder phenolic molecules to gastrointestinal enzymes, thus promoting their accessibility [101]. Hence, the effect of extrusion-cooking on phenolic content and bioaccessibility depends on which effect prevails.

As regards the effect of extrusion-cooking parameters on phenolic content in pasta, Bouasla et al. observed that the application of higher screw speed (80 rpm) enabled to obtain higher phenolic content in GF precooked rice-yellow pea pasta [57]. As a matter of fact, in cereal grains phenolic acids are mainly found in the bound form and such complexes are difficult to break down at lower screw speeds [57]. Oniszczuk et al. studied the effect of extruder screw speed on free phenolic acid content of GF precooked pasta obtained from roasted buckwheat (*Fagopyrum esculentum* Moench and *F. tataricum* Gaertner) flour [58]. The qualitative and quantitative analysis of FAs in extruded pasta by high-performance liquid chromatography electrospray ionization tandem mass spectrometry (HPLC-ESI-MS/MS) showed that gallic, protocatechuic, gentisic, 4-hydroxybenzoic, vanilic, *trans*-caffeic, *cis*-caffeic, *trans*-*p*-coumaric, *cis*-*p*-coumaric, syryngic, *trans*-ferulic, *cis*-ferulic, salicylic, *trans*-sinapic and *cis*-sinapic acids were present in all pasta samples, regardless of moisture content (30%, 32% and 34%) and screw speed (60, 80, 100 and 120 rpm). However, in pasta samples produced at 100 rpm extruder screw speed and 32%

flour moisture content, benzoic acid derivatives (i.e., gallic, protocatechuic, gentisic, 4-hydroxybenzoic and salicylic acids) were present with the highest amounts. The highest content in cinnamic acid derivatives (i.e., *trans*-caffeic, *trans*-*p*-coumaric, *cis*-*p*-coumaric and *cis*-ferulic) was observed in samples of GF buckwheat pasta produced at 60 rpm extruder screw speed and 30% of flour moisture [58]. Conversely, De Paula et al. reported a significant reduction in total phenolic acid content after pasta extrusion, possibly due to oxidising reactions promoted by water, oxygen and heat [59].

Cooking is a necessary step for pasta consumption, and it may influence the content of phenolic compounds and/or change the ratio between free and bound form of phenolics. De Paula et al. investigated the effect of cooking on phenolic content in barley pasta and observed that TPAs were not greatly affected by this treatment, and both free and bound phenolic compounds were preserved [59]. Conversely, Podio et al. found that cooking promotes the release of bound phenolic compounds, thus increasing the content of the free forms. In addition, pasta-making and cooking produced a change in the phenolic profile with respect to the starting flour [60]. Results are in keeping with Rocchetti et al. who observed that cooking by boiling lowered the bound-to-free ratio of phenolics in GF pasta [61]. They studied six commercially available GF pasta samples (i.e., pasta enriched with black rice, chickpea, red lentil, sorghum, amaranth and quinoa) and observed that in raw GF pasta samples, bound TPC was higher than free TPC, with values ranging from 7.58 mg GAE/100 g (sorghum GF pasta) to 32.68 mg GAE/100 g (quinoa GF pasta). After cooking, the highest free TPC was observed in black rice and quinoa samples, with 27.27 and 19.27 mg GAE 100 g−1, respectively (*p* < 0.01). In conclusion, from a nutritional point of view, understanding the effect of processing on phenolic content and on the ratio between free and bound forms is pivotal. As a matter of fact, the activity of phenolic acids is strictly dependent on the form they reach the gastrointestinal tract. The intake of free forms or soluble conjugated forms has systemic beneficial effects, such as inhibition of LDL cholesterol and liposome oxidation, since they are rapidly absorbed in the stomach and small intestine. Conversely, insoluble bound phenolic compounds reach the colon nearly intact where they are hydrolysed by the esterases and xylanase of colon microorganisms, thus having local activity and protecting against colon cancer [102]. However, the effects of phenolic compounds on human health depend on both the amount consumed and the bioavailability thereof.

#### **4. Bioaccessibility of Phenolic Compounds in Pasta**

As discussed above, several strategies have been explored in order to increase the phenolic content in pasta. Thanks to its low cost and long shelf life, pasta is consumed by people of all ages and all walks of life, hence it is appropriate to be used as a carrier of phenolic compounds, in order to promote health and wellbeing. However, a high dietary phenolic compound intake does not necessarily imply an appropriate bioactivity. As a matter of fact, bioactivity strictly depends on phenolic compound bioaccessibility and bioavailability. Several studies have been published and reviewed on the bioaccessibility of phenolic compounds in bread [29], while bioaccessibility of phenolic compounds in pasta products has been poorly investigated.

Phenolic compound bioaccessibility has been studied in pasta products formulated with whole-wheat flour or composite flours and in pasta samples produced with powders from plant materials or food by-products. Polyphenol bioaccessibility in GF pasta has been investigated, as well (Table 3). Static digestion models have been mainly used.


**3.**Bioaccessibilitystudiesonphenoliccompoundsin



2-Hydroxy-3-

*p*-Coumaroyl-feruloylputrescine;

trans-ferulic acid; TFA: Triferulic acid; TPC: Total Polyphenol Content; Try: Tryptophan.

 QA: Quinic acid; RA: Rosmarinic acid; SA C: Salvianolic acid C; SA E/B/L: Salvianolic acid E/B/L; SA I/H: Salvianolic acid I/H; SF: Salviaflaside;

*O*-β-d-glucopyranosylbenzoic

 acid; ID: intestinal digestion; MeQ:

Methylquercetin;

 MeRA:

Methylrosmarinate;

 OD: oral digestion; *pCoA*:

*p*-coumaric acid; *p*CoFP:

 tFA:

Podio et al. investigated phenolic compound bioaccessibility in whole-wheat pasta by using an experimental model, simulating human gastrointestinal digestion and subsequent absorption [60]. They observed that the conditions found in the intestinal medium (e.g., alkaline pH, pancreatin and bile actions, etc.) are not favourable for the stability of some phenolic compounds, which are changed, among others, by enzymatic, oxidative and other transformations, and by aggregation with food matrix. Generally speaking, they observed that TPC significantly increased after gastric digestion (GD) and intestinal digestion (ID), but in the dialysate (DIA) it was significantly lower than in the GD and ID. As to the polyphenol profile, only 8 out of the 25 compounds identified and quantified in cooked pasta were detected in the four stages of the in vitro digestion. In particular, the analysis of dialysated samples showed that hydroxybenzoic acid diglucoside, hydroxybenzoic acid glucoside, tryptophan and *trans*-ferulic acid content increased with respect to the corresponding intestinal digestion. This is of paramount importance as these compounds represent the bioaccessible and dialyzable fraction of polyphenols, which pass into the blood stream to reach organs or tissues where they would exert their antioxidant action. The authors hypothesized that the alkaline conditions and the action of pancreatin/porcine bile acting during the intestinal phase boosted the release of these phenolic compounds from dietary fibre.

Pigni et al. performed a simulated in vitro gastrointestinal digestion of cooked samples of wheat pasta fortified with 10% of partially-deoiled chia flour (PDCF), to assess the absorption of individual polyphenols through the different stages [103]. Upon oral digestion (OD), a total of 50% of the TPC found in the cooked supplemented pasta was released. Gastric digestion and intestinal digestion determined a higher increase (i.e., 300–500%) indicating that the action of enzymes (pepsin, pancreatin) and pH enables an effective release of polyphenols from the food matrix, including the components of PDCF and wheat. Finally, the DIA samples, representing the fraction absorbed in the intestine, showed an increase of around 50% compared with the values of boiled pasta. As regards the specific phenolic compounds quantified in boiled pasta, only 2 out of 10 were above the limit of detection (LOD) and limit of quantitation (LOQ) in the intestinal samples of pasta with 10% PDCF: rosmarinic acid and salviaflaside. In the DIA samples they were even below the LOQ, however their detection indicates that at least a small fraction is being absorbed at this stage.

Marinelli et al. investigated the bioaccessibility of phenolic compounds in samples of pasta produced with durum wheat semolina and red grape marc, a by-product of winemaking, in combination with transglutaminase [104]. They found that the functional pasta sample showed a significantly higher concentration of bioaccessible total polyphenols than the control sample, formulated only with durum wheat semolina (5.53 vs 4.16 mg GAE/g dm, respectively).

Another study investigated the bioaccessibility and potential bioavailability of phenolics in pasta produced by substituting wheat flour (2.5% and 7.5%) with lyophilised raspberries (*Rubus idaeus* L.), boysenberries (*Rubus idaeus* × *Rubus ulmifolius*), redcurrants (*Ribes rubrum* L.) and blackcurrants (*Ribes nigrum* L.) [105]. It was observed that potentially bioaccessible polyphenols were higher in pasta enriched with fruits from *Rubus* genus than with *Ribes* fruits. Pasta fortified with raspberries and boysenberries showed an increase of 260% in polyphenols, while in samples enriched with red- and blackcurrants, the increase was 360%.

As regards the bioaccessibility of phenolics in GF pasta, Camelo-Méndez et al. investigated samples produced with flours from unripe plantain (*Musa paradisiaca* L.), chickpea and blue maize by using an in vitro model simulating gastrointestinal digestion [106]. During the oral digestion, only free polyphenols were released from the matrix, that is, those compounds not linked to other molecules, such as proteins, lipids and carbohydrates. During the gastric phase, the release of phenolic compounds was higher in samples with a higher amount of blue maize flour (i.e., 50% and 75%). The higher release was likely associated with the breakdown of complexes with proteins, fibre residues and sugars. The low pH and enzymatic activity also favour the release of phenolic compounds, mainly flavonoids, from the food matrix. After ID, the percentage of phenolic compounds released was 40% of the initial value in the samples. In detail, they observed that the bioaccessibility of the phenolic

compounds in pasta was up to 80% and the highest amount was obtained with the pasta manufactured with the highest amounts of blue maize.

Palavecino et al. also studied bioaccessibility of the functional GF pasta they produced with two varieties of sorghum, and found that the white and brown sorghum pasta samples had 2.9- and 2.4-fold higher potentially bioaccessible polyphenol content than in cooked sample, respectively [41]. The antioxidant activity in sorghum pasta did not significantly vary after digestion, and it was approximately 36–48% in DIA samples.

Rocchetti et al. investigated phenolic compound bioaccessibility in six samples of commercially available pasta, formulated with black rice, chickpea, red lentil, sorghum, amaranth and quinoa [107]. They used an in vitro gastrointestinal digestion model comprising two steps: a pre-incubation step with digestive enzymes, and an in vitro large intestine fermentation process. The phenolic profile was investigated at different time points during faecal fermentation. It emerged that GF pasta samples enriched with pseudocereals or legumes were able to deliver phenolics to the large intestine, and this was likely due to the contribution of the food matrix, which acts as a carrier. In addition, once in the large intestine, the main phenolic subclasses (i.e., flavonoids, hydroxycinnamic acids, lignans and stilbenes) degraded, along with a parallel increase in low molecular weight phenolic acids (i.e., hydroxybenzoic acids), alkylphenols, hydroxybenzoketones and tyrosols. As regards phenolic compound bioaccessibility during the large intestine fermentation process, flavonoids reported values lower than 1%, regardless of the time point or matrix considered. Hydroxycinnamic acid bioaccessibility in large intestine ranged from 0.6% to 8.6% at 0 h, from 0.6% to 1.6% at 8 h, and from 0.7% to 5.5% at 24 h. Within lignans, the various classes showed differences in bioaccessibility, with furofurans having very low bioaccessibility, dibenzylbutyrolactones reached the colon in larger amounts (i.e., 2.7–12.2% of bioaccessibility); while tyrosols and alkylresorcinols were the phenolics with the highest bioaccessibility during the in vitro fermentation process.

#### **5. Conclusions**

Phenolic compounds have documented beneficial effects on human health, because of their contribution to preventing chronic diseases. Durum wheat semolina, the main ingredient of pasta, lacks phenolic compounds, since they are lost during conventional milling. Hence, several strategies have been proposed to produce functional pasta whose consumption may contribute to an increased intake of phenolic compounds. Whole grain, legume and composite flours are the main substituents of durum semolina. GF pasta has been functionalized, as well, by using ingredients rich in phenolic compounds. The use of pre-processing technologies on raw materials, such as sprouting, or modulation of extrusion-cooking conditions, may be valuable approaches to increase the phenolic content in pasta. However, a higher intake of phenolic compounds does not necessarily imply a greater bioactivity. Hence, it is pivotal to investigate bioaccessibility and bioavailability of phenolic compounds in functional pasta. Currently, few studies have been performed, and comparing results across different studies is not always reliable due to the diversity of in vitro model conditions and the lack of official methods for the determination of phenolic compound content. Hence, efforts are still needed to evaluate the contribution of functional pasta consumption to maintaining optimal health.

**Author Contributions:** All authors contributed to the conceptualization and design of the study. V.M. and F.M. collected available literature and jointly wrote the paper. V.M. and R.A. contributed expert opinions. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

1. Shahidi, F.; Ambigaipalan, P. Phenolics and polyphenolics in foods, beverages and spices: Antioxidant activity and health effects—A review. *J. Funct. Foods* **2015**, *18*, 820–897. [CrossRef]


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Review*
