Trials were conducted with fruit from commercial apple orchards, located in Chimbarongo, O’Higgins Region, Chile. The cultivars evaluated were Gala ‘Brookfield’, Granny Smith and Fuji ‘Raku Raku’, planted in 2000 on MM106 rootstock. The conventional Granny Smith and Fuji ‘Raku Raku’ fruit was obtained from the orchard “Agrocasan” (34° 40’ S, 70° 53’ W), Gala ‘Brookfield’ fruit from “Los Alamos” (34° 44’ S, 71° 3’ W), and organic fruit of three cultivars was obtained from the orchard “Los Pretiles” (34° 44’ S, 70° 57’ W), in 2011–2012 seasons.

The orchards are located in the same edaphoclimatic zone, at a distance of less than 10 km. The soils in the area are characterized by having a slight slope (less than 3%), being deep (more than 100 cm) and with a clay loam texture. The organic matter content of the orchards was 2.1%, 2.3% and 2.7% for “Agrocasan”, “Los Alamos” and “Los Pretiles”, respectively. The climate of the area is Mediterranean, with dry and warm summers and rainy and mild winters. The average annual rainfall is close to 388 mm. The average annual air temperature is 14.5 °C. Orchard management was according to commercial requirements, with a drip irrigation system.

The fruit of each cultivar was collected at different growth stages: 26, 34, 40, 68, 96 days after full bloom (DAFB) and at commercial harvest (124, 156, and 175 DAFB for Gala ‘Brookfield’, Granny Smith and Fuji ‘Raku Raku’, respectively).

A total of sixteen fruit for each treatment group, in four replicates of four apples each, were randomly selected (two apples per tree). Tissue samples consisted of 2 g of whole fruit (skin plus pulp) for each treatment. Samples were then frozen with liquid nitrogen, pulverized and homogenized in a mortar, and then extracted according to the method described by Coseteng & Lee (1987) with some modifications. The tissue was extracted twice (for 10 and 5 min) at 100 °C, with an 80% ethanol solution (ethanol:water 80:20, v/v), and then filtered. Samples were volumetrically diluted to 10 mL with 80% ethanol and kept at -20 °C until use.

Total phenolic content was determined by the Folin-Ciocalteu method described by Coseteng & Lee (1987). Briefly, 0.1 mL of the extract was mixed with 0.5 mL of the Folin-Ciocalteu phenol reagent (Merck, Darmstadt, Germany). The mixture was incubated for 5 min and then 0.5 mL of Na₂CO₃ (10%, w/v) added and incubated for 15 min at room temperature (20 °C). Absorbance was measured at 640 nm with a spectrophotometer (Pharo 300, Merck KGaA, Darmstadt, Germany). Total phenolic content was expressed as mg of chlorogenic acid equivalents (CAE) per g fresh weight (FW). Moreover, mg of phenolics/ fruit was calculated.

DPPH method: The capture of the free radical 2,2-diphenyl-1-picrylhydrazyl (DPPH; Fluka Chemie, Buchs, Switzerland) was measured colormetry by the method described by von Gadow et al. (1997). Briefly, 0.01 mL of each extract was used, adjusted to a final volume of 0.5 mL with 80% ethanol. They were then mixed with 2 mL of 8×10^-5 M DPPH solution and incubated for 8 min at room temperature. Their absorbance was measured at 515 nm with a spectrophotometer (Pharo 300, Merck KGaA, Darmstadt, Germany). Chlorogenic acid in different concentrations was used as a standard and the capture of the DPPH free radicals was expressed as mg of chlorogenic acid equivalents (CAE) per g FW. Moreover, mg CAE/ fruit was calculated.

Oxygen radical absorbance capacity (ORAC) method: Analysis was performed with a Synergy HT Multi-Detection kit (BioTek Instruments, Inc., Winooski, VT, USA), and a 96-well black-bottom microplate (Nunc, Denmark) was used. Fluorescence was measured from the plate top, wavelength of excitation 485 nm, and emission 528 nm. Equipment was controlled with Gen 5 software. The methodology used was described by Huang et al. (2002). Phosphate buffer 75 mM (pH 7.4) was used to prepare the reagents (Trolox®, AAPH and sodium fluorescein). Samples were diluted with 75 mM phosphate buffer (pH 7.4) until a phenolic content between 0.01 and 0.03 mg CAE/mL or g was obtained. Subsequently, 25 µL of each sample and Trolox® standards (Aldrich, USA) were placed on the plate. Then 150 µL of 0.04 µM sodium fluorescein (Sigma-Aldrich, USA) was added. The mixture was incubated for 30 min at 37 °C. Next, 25 µL of 300 µM AAPH (Aldrich, USA) was added and the equipment shook for 8 s. Kinetics was then started for 60 min, with fluorescence readings every minute. Antioxidant capacity was calculated by software, through interpolation of the Trolox® standard curve (0, 6.25, 12.5, 25, 50 and 100 µM) with the net area under the curve (Net-AUC). Values were expressed as mmol Trolox equivalent (TE) per 100 g FW. Moreover, mmol TE/ fruit was calculated.

The determination of quercetin glycosides in samples was performed using HPLC-DAD Merck Hitachi (LaChrom, Tokyo, Japan) equipment, which consisted of a LaChrom L-7100 pump and a diode array detector, L-7455 LaChrom, and a 100-5 C18 Kromasil column of 259 × 4.6 mm with a pre-column of the same characteristics, maintained at 20 °C. Briefly, 20 µL, previously filtered (0.45 μm filter), were injected. To identify the compounds, different standards of quercetin glycosides were used with the UV-VIS spectra: Quercetin 3-D-galactoside, Quercetin 3-glucoside, Quercetin 3-xyloside, Quercetin 3-arabinoside, Quercetin 3-rhamnoside (Extrasynthese, Genay, France). The chromatogram was monitored at 300 nm. The solvents of the mobile phase were: A: formic acid 1% in H₂O HPLC-grade; B: acetonitrile 40% in H₂O, and C: acetonitrile. The elution time used was: 0-10 min: A (70), B (30), C (0) flow 1 mL/min; 45 min: A (25), B (75), C (0) flow 0.5 mL/min; 52 min: A (0), B (0), C (100) flow 1 mL/min; and 55 min: A (70), B (30), C (0) flow 1 mL/min. Values were expressed as µg equivalents of quercetin glycosides per g FW.

In order to determine fruit skin surface, the nonlinear model presented by Clayton et al. (1995) for apple trees was used. Surface area (A) was determined by the Eq. (1):

where: x = mass (kg) or volume (m³) and d, e = parameters for each cultivar.

Data were analyzed statistically using a truncated multiple regression model (SAS software, SAS Inst., Inc., Cary, NC, USA), which considered management practices (MAN), fruit growth stage (FGS), cultivar (CUL), fruit weight (FW) and skin surface area (SSA) as independent variables (Table 1). The variables evaluated were total phenolic content, antioxidant activity and quercetin glycosides. The statistical model used to calculate each variable was determined by the Eq. (2):

In early developmental stages of apple fruit, total phenolic content reached its highest level and then decreased until harvest (Fig. 1). Conversely, phenolic content per fruit increased steadily by late season (Fig. 2). High initial concentration of polyphenols may be attributed to high levels of enzyme activity (e.g. phenylalanine ammonia-lyase (PAL), chalcone isomerase (CHI) and glycosyltransferase (UFGT)) and biosynthesis of these compounds at cell division (Lister et al., 1996; Treutter, 2001; Yuri et al., 2012). The subsequent decline may be linked to an enzyme activity decrease and dilution due to fruit growth (Awad et al., 2001; Takos et al., 2006; Renard et al., 2007).

Phenolic accumulation rate decreases throughout the fruit growth phases, although it never ceases, possibly explaining the increase in total phenolic content per fruit (Fig. 3). However, even though total phenol concentration showed a strong decreasing tendency, that of specific phenols such as glycosidic quercetins showed a relatively stable concentration that tended to increase late in the season (Fig. 4). Similar results were found by Takos et al. (2006).

According to Awad et al. (2001) changes in the concentration and amount of individual flavonoids and chlorogenic acid during development and ripening were investigated by reversed-phase high performance liquid chromatography (RP-HPLC), quercetin glycoside content tends to increase during fruit growth and development, but with levels and fluctuations mainly attributed to cultivar. In contrast, Yuri et al. (2014) observed that quercetin glycoside concentration in whole fruit from cvs. Granny Smith and Fuji increases up to 32 DAFB, and then decreases until harvest, while the content per fruit increased progressively. However, in another study, Yuri et al. (2012) found that quercetin glycoside levels in three apple cultivars increased during early developmental stages, then decreased and finally increased again. Likewise, Bizjak et al. (2013) reported an increase in quercetin content in cv. Braeburn apples throughout fruit ripening, tending to decrease in the last week before harvest. These results coincide with those observed in the present study, and with those noticed by Takos et al. (2006), who determined that the transcription of the MdFLS gene (flavonol synthetase enzyme), which modulates flavonol synthesis, reached its maximum values at 48 and 147 DAFB in cv. Pink Lady™, so that there would be synthesis peaks after cell division.

The trend for antioxidant activity in all cultivars was similar to that observed for total phenolic content, decreasing from early developmental stages until harvest (Fig. 1), while total antioxidant activity per fruit progressively increased up to the end (Fig. 2). Similar results were found by many investigators (e.g. Awad et al., 2001; Yuri et al., 2012, 2014). Phenolic compounds and antioxidant activity are expected to act similarly during fruit growth and development, as their relationship is well documented in literature (Lee et al., 2003). Antioxidant properties of phenolic compounds are directly linked to their structure, since their aromatic rings are potentially capable of scavenging free radicals, thus forming stable compounds (Rice-Evans et al., 1996; Bors & Michel, 2002).

Concentrations of phenolic compounds and antioxidant activity in fruit from conventional and organic orchards have been of interest to many investigators. The observed reports are inconsistent in determining management practices effects on these compounds concentration (Peck et al., 2009; Roussos & Gasparatos, 2009; Stracke et al., 2009). Valavanidis et al. (2009) found no differences in phenolic content and antioxidant activity according to management in different apple cultivars. Also, no differences in polyphenol levels and antioxidant activity were observed by Peck et al. (2009) in apple cultivars produced in integrated vs. organic system. On the other hand, Stracke et al. (2009) and Petkovsek et al. (2010) found that fruit produced in orchards with conventional or integrated management had lower phenolic content and antioxidant activity than those found in organic orchards, which was associated with the organic orchards being more exposed to biotic and abiotic stress conditions. It is important to mention that the studies cited were carried out taking fruit at harvest, so the results obtained would not necessarily reflect the management practices used in the orchards, while our study was conducted at different season times. According to Weidner et al. (2011), depending on the intensity and duration of stress conditions, changes at cellular level can be reversible or irreversible. Stracke et al. (2009) found that the concentration of phenolic compounds and antioxidant activity varied more between seasons than between organic and conventional management.

Skin surface area and fruit weight were variables incorporated into the models, since literature shows that there is a decrease in phenol concentration over time. The latter seems to be linked to these metabolites dilution due to the increase in fruit size and skin surface area. Yet, results did not show a negative association of fruit weight with phenolic content and antioxidant activity, but rather a positive correlation between both, excluding the antioxidant capacity expressed in ORAC units.

Stopar et al. (2002) reported that a significant decrease in fruit number per plant in cv. Jonagold greatly increased fruit weight and phenol concentration. According to Lux-Endrich et al. (2000), an increase in sucrose and a decrease in macronutrients, in a growing medium with apple shoots, increased accumulation of glycosidic quercetins, catechins, procyanidins and phenolic acids.

Skin surface area showed a negative association with phenolic content and antioxidant activity, although antioxidant activity expressed in ORAC units showed opposite trends. Awad et al. (2001) found similar results, observing that skin phenol concentration in cvs. Elstar and Jonagold decreased as fruit grew and skin surface area increased.

The content of phenolic compounds, glycosidic quercetins and antioxidant activity are variables that are conditioned by several factors. In the study, the cultivar factor was highly determinant in apples, decreasing its content throughout the season, given by a dilution effect as its volume increased. This was shown in the fruit growth stage and skin surface factors. No significant differences were found for fruit weight. Both conventional and organic management practices had a minor relevance in the variables studied, being significant only for quercetin glycoside concentration, which is the main polyphenol in apples.