Materials and Methods

Standards.

Commercial carotenoid standards of β-carotene were purchased from Sigma-Aldrich (St. Louis, MO), and standards of lutein, epoxy lutein, violaxanthin, neoxanthin, and β-cryptozanthin were purchased from Carote-Nature (Lupsingen, Switzerland). Six concentrations of carotenoid standards were dissolved in ethanol and diluted in hexane. Standard concentrations were determined spectrophotometrically using their respective molar extinction coefficients (Rodriguez-Amaya, 2001). Commercial flavonoid standards of procyanidins (B1 and B2), catechin, epicatechin, chlorogenic acid, quercetin glycosides, and cyanidin glycosides were purchased from Chromadex (Irvine, CA). Neochlorogenic acid was purchased from Sigma-Aldrich. Five flavonoid standard concentrations were prepared in 50% methanol to generate standard curves. Five microliters of each standard were injected as an external standard followed by the peach extract samples. Methanol, acetonitrile, formic acid, and acetic acid were purchased from Fisher Scientific (Pittsburgh, PA). All solvents were of high-performance liquid chromatography (HPLC) or HPLC–mass spectrometry (MS) grades.

Plant materials.

Tree-ripened fruit (“ready-to-eat” stage) of three peach cultivars, China Pearl, Contender, and Carolina Gold, were collected from established trees at two North Carolina State University (NCSU) agricultural research stations located in Mills River and Jackson Springs, NC, in 2009 and 2010. The Mills River site (lat. 35.41° N, long 79° W) is located in the mountains of western North Carolina, at an elevation of 884 m, with a clay-loam soil type. The Jackson Springs site (lat. 35° N, long. 82° W) is located in the Sandhills region of the southern piedmont of North Carolina. The elevation at this site is 191 m. Soils are predominantly coarse-textured sand demonstrating excellent drainage and low moisture-holding capacity. Trees at the two locations differed by age. All trees at Mills River were planted in 2001, whereas trees at Sandhills were planted in 2001 (‘China Pearl’), 2005 (NC Yellow and NC 97-48), and 2006 (‘Contender’ and ‘Carolina Gold’). All trees were considered “mature” in that they had completed at least five growing seasons by the first year of analysis. Standard cultivation and fertility practices were followed at both sites.

Contender is a high-chilling yellow-fleshed freestone cultivar widely grown in North Carolina and extensively used in cultivar development (Werner et al., 1989). ‘China Pearl’ is a white-fleshed peach resulting from the hybridization of ‘Contender’ (yellow flesh) and a white-fleshed Chinese introduction PI 134401 (Werner and Snelling, 2006). In addition to white flesh, PI 134401 (not included in the study) lacks red coloration in the peel and likely contains a recessive mutation for ANC accumulation in that tissue. ‘Carolina Gold’ is a yellow-fleshed freestone peach developed through the hybridization of ‘Biscoe’ × NC-C5S-067. The pedigree of NC-C5S-067 is ‘Encore’ × Calanda San Miquel 2383 (Werner and Snelling, 2006). Calanda San Miquel 2383 is a yellow-fleshed, Spanish land race that lacks ANC in the fruit and the peel. In addition to the three commercial cultivars, two NCSU advance breeding selections (NC 97-48 and NC Yellow) were also collected from the Jackson Springs site in 2009 and 2010 (not available at second location). Both selections share the same pedigree ‘Contender’ × P I65977 but may have been derived from different F₁ plants. A complete discussion of PI accessions has been previously published (Werner and Okie, 1998). Pedigree information for the peach genotypes evaluated in this study is presented in Figure 1.

View Figure

• Download as Powerpoint
• Download Figure

Fruit of each cultivar were collected from the peak harvest of four pre-labeled trees at each location. Fruit was immediately placed in ice coolers and transported to the Plants for Human Health Institute, NCSU at Kannapolis, NC, where they were processed the same day. Five representative fruit from each tree were selected at random from the peak yield to randomize positional effects. After weighing individual fruit, the fruit was bisected and the peel was separated from flesh. Peel and fruit were individually flash-frozen in liquid nitrogen and stored at –80 °C. Weights of the pit, peel, scar, and flesh were recorded. The total soluble solids were measured using a handheld digital refractometer (PAL-1; Atago, Tokyo, Japan). Frozen fruit peel and flesh were freeze-dried using a freeze dryer (VirTis 24Dx48; SP Scientific, Stone Ridge, NY) and the lyophilized tissue was then ground into fine powder with a coffee grinder and stored at –20 °C in the dark until analyses for carotenoids and flavonoids. Moisture removed from samples after freeze-drying was estimated to be 87.2% ± 2.9% and 85.2% ± 2.1% in the flesh and peel, respectively.

Extraction and analysis of carotenoids.

Under yellow lights, 0.2 g of freeze-dried finely ground tissue was placed in 50-mL tubes and extracted three times with 5 mL hexane for 20 min on a standard laboratory shaker (180 rpm). The extracts were then centrifuged at 2900 g_n for 15 min at 10 °C and supernatants were collected in new 50-mL tubes for carotenoid quantification. After carotenoid extraction, the pellets were completely dried down under a stream of nitrogen gas for flavonoid extraction as described below. The supernatant (hexane containing carotenoids) was saponified by adding methanolic 0.1% KOH at 60 °C for 45 min. A half volume of water was added to remove lipids and KOH and samples were centrifuged where the top layer of hexane was pipetted into a new 15-mL tube. Samples were dried down completely under nitrogen gas and re-suspended in 1.5 mL ethanol. The ethanolic extracts were filtered through 0.2-μm polytetrafluoroethylene (PTFE) filters (Nalgene, Rochester, NY) into amber 2-mL amber vials and stored under nitrogen gas at –80 °C until analysis.

Carotenoids were separated on a 1.8-μm, 2.1 × 100-mm column (HSS T3; Waters, Milford, MA) set at 35 °C using an ultra-performance liquid chromatography (UPLC) unit (ACQUITY; Waters) with a photodiode array detector. The UPLC analysis was conducted following procedures previously described (Guzman et al., 2012). The solvent system consisted of A) 0.05% ammonium acetate and B) 74:19:7 (v/v/v) acetonitrile:methanol:chloroform. The profile included two linear phases (0 to 18 min at 75% B, 18 to 19 min from 75% to 100% B, 19 to 30 min from 100% to 98% B). Flow rate was set at 0.4 mL·min⁻¹ and the injection volume was 5 μL. Spectra were observed for the wavelengths between 240 and 500 nm and carotenoids were detected at 450 nm and identified according to retention time and spectral comparison with commercial standards and published data.

Extraction and analysis of phenolic compounds.

The dried pellets of peel and flesh (after carotenoid extraction) were used to extract flavonoids from the individual fruit. A single sample was used for both extractions as a result of the limited amount of tissue represented by the peel. For flavonoid extraction, 0.5 g of dry pellet was extracted with 5 mL of 47.5% methanol (MeOH), 47.5% water, and 5% formic acid twice (10 mL total) with incubation on ice for 20 min and occasional vortexing. The mixture of sample and solvent was centrifuged (2900 g_n, 10 °C, 20 min) and supernatant was collected with 1.5 mL filtered through a 0.2-μm PTFE syringe filters (Fisher Scientific, Pittsburgh, PA). Samples were stored at –20 °C until further analysis. For each sample, 15 μL was injected into the HPLC attached to a quadratic time of flight (QTOF) MS (6510; Agilent Technologies, Santa Clara, CA) for chromatography separation. The HPLC was an Agilent 1200 system with a temperature-controlled autosampler and column compartment. The system was equipped with reversed-phase C₁₈ column (Zorbax ods, 250 × 4.6 mm 5 μM; Agilent Technologies). The solvents used were A) 5% formic acid and B) 100% MeOH. The gradient was 5% B (5 min), 10% B (10 min), 15% B (10 min), 25% B (5 min), 60% B (15 min), and 5% B (15 min) with a constant flow rate of 1 mL·min⁻¹. Simultaneously, flavan-3-ol, phenolic acids, flavonol, and ANC were separated and quantified using their maximum ultraviolet absorption with this analysis protocol. Flavan-3-ol compounds including the proanthocyanidins (catechin, epicatechin, and procyanidins B1, B2, B3, and B4) as well as other flavan-3-ols were monitored at maximum ultraviolet wavelength absorption of 280 nm. Phenolic acids including chlorogenic, neochlorogenic, and caffeic acids were monitored at ultraviolet 325 nm. Flavonols including quercetin glycosides (glucoside and rutinoside) and other quercetin derivatives were recorded at ultraviolet 360 nm. ANC, including cyanidin glycosides (glucoside and rutinoside), were monitored at ultraviolet 520 nm. Identification of compounds was established by using authentic standards (Chromadex), molecular weights recorded on HPLC-MS-QTOF, and comparison with published data (de Pascual-Teresa et al., 2000; Gu et al., 2004; Haminiuk et al., 2012; Hong et al., 2004; Lavelli et al., 2009; Tomás-Barberán et al., 2001).

Statistical analyses.

Analysis of variance (ANOVA) was performed using the general linear model procedure (PROC GLM) to estimate variation associated with each source of variation using SAS software (Version 9.3; SAS Institute, Cary, NC). The ANOVA was limited to cultivars grown in both environments (‘Contender’, ‘China Pearl’, and ‘Carolina Gold’). For carotenoids and flavonoids quantified in peach peel or flesh, the statistical model used was y_ijkl = μ + Y_i + L_j + C_k + YC_ik + LC_jk + YL_ij + LYC_ijk + T(C)_(k)l + F(T)_(l)m + ɛ_(ijklm), where y = response from the experimental unit, μ = overall mean, Y = year, L = location, C = genotype, T(C) = tree within genotype, F(T) = fruit within tree, and ɛ = experimental random error. Pearson correlation statistics were generated using the PROC CORR module of the same software and was also limited to the three cultivars grown in both locations. Tukey’s honest significant test was used to compare significance differences among cultivars at P < 0.5. The PROC MEANS statement was used to compute cultivar means and sds over the 2 years, 2009 and 2010. Carotenoid concentrations were presented as micrograms per gram dry weight (DW), whereas flavonoids were presented as milligrams per 100 g DW in the peel and flesh of individual peach fruit.

Results

Tandem UPLC and HPLC-MS analysis revealed appreciable levels of 10 carotenoids and 24 phenolic compounds (Fig. 2) in the peel and/or flesh of the three peach cultivars and two advanced selections. Seven carotenoids could be identified on the basis of available standards and included neoxanthin, violaxanthin, lutein epoxide, lutein, zeaxanthin, β-cryptoxanthin, and β-carotene. Three additional minor peaks were identified in both the peel and the flesh but represented relatively minor contributions to overall carotenoid levels. The predominant carotenoid found in both the flesh and peel was violaxanthin, which tended to be two to three times higher than other identified carotenoids. A 3-fold difference was observed in total carotenoids between ‘China Pearl’ and yellow-fleshed peaches. Although all carotenoids were appreciably lower in ‘China Pearl’, the greatest differences were observed with xanthophylls. Violaxanthin accumulation was six times lower in the flesh and 10 times lower in the peel. A reduction in unidentified carotenoids was even more dramatic with an almost 20-fold decrease in both the flesh and peel. Neoxanthin was not detected in either the flesh or peel of ‘China Pearl’. The yellow-fleshed cultivars were not significantly different from each other in the accumulation of total or individual carotenoids in either the peel or flesh.

View Figure

• Download as Powerpoint
• Download Figure

The phenolic content included two major ANCs (cyanidin 3-glucoside and cyanidin 3-rutinoside), three phenolic acids (chlorogenic, neochlorogenic, and caffeic acid), and 14 flavan-3-ols {catechin, epicatchin, procyanidin B1-B4 [B1 = epicatechin-(4β→8)-catechin, B2 = (-)-epicatechin-(4β→8)-(-)-epicatechin, B3 = catechin-(4α→8)-catechin, and B4 = catechin-(4α→8)-epicatechin]}. The unidentified eight flavan-3-ols represented relatively minor peaks in the phenol profile. The flavanol profile included quercetin-3-glucoside, quercetin-3-rutinoside, and three unidentified minor peaks.

ANC concentration in the peel (primarily cyanidin 3-glucoside) ranged from undetectable levels (NC Yellow and NC 97-48) to 178 mg/100 g DW (‘Contender’) (Table 1). ‘Contender’ contained significantly higher levels of cyanidin 3-glucoside than ‘China Pearl’ or ‘Carolina Gold’ (approximately two times), which were not significantly different from each other. ANCs were not detected in the flesh of any genotype (Table 2). Chlorogenic acid was the primary phenolic acid detected in the peel and in general was four to five times higher than neo-chlorogenic acid (Table 1). A 2- to 3-fold variation (52 to 136 mg/100 g DW) among cultivars was observed with ‘China Pearl’, NC Yellow, and NC 97-48 containing significantly higher levels than ‘Contender’ or ‘Carolina Gold’. Chlorogenic acid was present in the flesh at lower concentrations but a similar ranking of the cultivars was apparent with NC Yellow and NC 97-48 containing the highest levels and ‘Contender’ and ‘Carolina Gold’ accumulating the lowest levels.

Table 1.Fruit characteristics and flavonoid and carotenoid concentrations in the peel of ripe peach fruit evaluated in two North Carolina locations over 2 years (2009 and 2010).

View Fullscreen

Table 2.Flavonoid and carotenoid concentrations in the flesh of ripe peach fruit evaluated in two North Carolina locations over 2 years (2009 and 2010).

View Fullscreen

Flavonols constituted a relatively minor contribution to total phenolic concentration in both the peel and the flesh. The predominant flavonols identified in the peel were identified as quercitin-3-glucoside and quercitin-3-rutinoside in approximately equal concentrations. Significant (2-fold) differences in quercetin-3-glucoside accumulation in the peel were noted between ‘China Pearl’ (8 mg/100 g DW) and NC Yellow and NC 97-48 (16 to 19 mg/100 g DW). Flavonols were present in the flesh but at lower concentrations. The flavan-3-ol profile of peach included: catechin, epicatechin, and 12 proanthocyanidin dimers and polymers. The predominant procyanidins were identified as B1 [epicatechin-(4β→8)-catechin], B2 [(-)-epicatechin-(4β→8)- (-)-epicatechin], and B3 [catechin-(4α→8)-catechin]. In general, ‘Carolina Gold’ tended to have lower levels of most of these compounds but these differences were not in general significant.

The ANOVA conducted on cultivars replicated at both locations over 2 years (‘Contender’, ‘China Pearl’, and ‘Carolina Gold’) detected significant genetic variation for all compounds except catechin, procyanidin B4, and the unidentified flavonol derivatives (Tables 3 and 4). Of the five major phenolic compounds found in peach (cyanidin 3-glucoside, chlorogenic acid, catechin, and procyanidins B1 and B3), only catechin was nonsignificant for genetic variation. The genetic variation associated with cyanidin 3-glucoside and chlorogenic acid was 20% and 40%, respectively. The relative importance and significance of location and year effects was compound-specific. Location effects contributed 28% of the variation associated with chlorogenic acid (year nonsignificant), whereas year contributed 35% of the variation associated with cyanidin 3-glucoside (location nonsignificant). Significant two- and three-way interactions were detected for multiple compounds. The relative magnitude of these effects depended on the individual compound. No significant variation could be detected (P < 0.01) for any compounds between fruit from the same tree or trees of the same cultivar within environments.

Table 3.Source of variation in fruit characteristics, flavonoids, and carotenoids measured in the ripe fruit peel of three peach cultivars (‘China Pearl’, ‘Carolina Gold’, and ‘Contender’) harvested at two North Carolina locations (Mills River and Jackson Springs) in 2009 and 2010.

View Fullscreen

Table 4.Source of variation in fruit characteristics, flavonoids, and carotenoids measured in the ripe fruit flesh of three peach cultivars (‘China Pearl’, ‘Carolina Gold’, and ‘Contender’) harvested at two North Carolina locations (Mills River and Jackson Springs) in 2009 and 2010.

View Fullscreen

Moderate to nonsignificant correlations were observed among polyphenolic concentrations of ‘Contender’, ‘Carolina Gold’, and ‘China Pearl’ across locations and years (Table 5). The strongest correlations tended to occur between constituents of the same flavonoid class (neochlorogenic and chlorogenic acid = r = 0.76; cyanidin 3-glucoside and cyanidin 3-rutinoside = r = 0.78). A nonsignificant correlation was observed between the principle ANC (cyanidin 3-glucoside) and phenolic acid (cholorogenic acid) (r = 0.06). Correlations among individual carotenoids tended to be stronger than phenolics and ranged from r = 0.40 to r = 0.87 (Table 6).

Table 5.Pearson’s correlation coefficients for fruit characteristics and phenolic compounds measured in the peel of ripe peach fruit.

View Fullscreen

Table 6.Pearson’s correlation coefficients for fruit characteristics and carotenoid compounds measured in the peel of ripe peach fruit.

View Fullscreen

Discussion

Polyphenolics were found in both the peel and flesh of peach (with the exception of ANC and select flavon-3-ols monomer and dimers found only in the peel) but occurred at significantly higher concentrations in the peel. Carotenoids were also detected in both tissues but were significantly more concentrated in the fruit peel. We will confine the discussion to polyphenolics and carotenoids in the peel, but with the exception of ANC (which were only found in the peel), it should be noted that many of the same observations are relevant to both types of fruit tissues.

Given the nature of the plant material included in the study (white and yellow-fleshed peaches with and without red coloration in the peel), some of the results were not unexpected. Significant differences (3-fold) and genetic variation were observed in total carotenoid accumulation between yellow and white-fleshed cultivars. Our results are consistent with previous reports (Brandi et al., 2011), who observed that xanthophylls were the predominant carotenoids in peach at the fully ripe stage. In the white-fleshed cultivar China Pearl, all carotenoids (with the exception of neoxanthin) could be detected in the peel and the flesh. Although there were significant differences in all carotenoids between white and yellow-fleshed cultivars, the largest differences occurred in the xanthophylls (violaxanthin, neoxanthin, and lutein epoxide) and in the total accumulation of the three unidentified carotenoid peaks. Further examination of these unidentified peaks and comparison with previous spectra data suggest they likely represent isomers of mutatoxanthin, auroxanthin, and neochrome that have been previously described in peach (Brandi et al., 2011).

Yellow flesh color in peach is the result of a recessive mutation at the Y locus on linkage group 1 (Bliss et al., 2002). Recent reports suggest that mutations within the CCD4 carotenoid cleavage dioxygenase gene result in accumulation of yellow-pigmented carotenoids and the subsequent reduction of apocarotenoids that likely contributes to aroma volatiles (Adami et al., 2013; Brandi et al., 2011; Falchi et al., 2013; Fukamatsu et al., 2013). This model is consistent with our observations, but the carotenoid levels observed in ‘China Pearl’ are higher than previously reported in white-fleshed cultivars (Brandi et al., 2011), which could be a function of environment, genetic background, or incomplete dominance at the Y locus. Based on coloration of the peel, we also expected to observe differences in ANC accumulation between the breeding selections lacking red color (NC 97-48 and NC Yellow) and cultivars that display significant red color in the same tissue (‘Contender’, ‘China Pearl’, and ‘Carolina Gold’). The peel-specific ANC mutation in North Carolina peach germplasm has been designated H for Highlighter and differs from both ANC-less and white flower mutations previously described in peach (Beckman et al., 2005). In Rosaceae, MYB transcription factors (MYB10) regulate the transcription of UDP-glucose-flavonoid-3-O-glucosyltransferase (UFGT) and dihydroflavonol 4-reductase (DFR) genes and likely provide specific impact of ANC levels in the peel (Lin-Wang et al., 2010; Ravaglia et al., 2013).

The ANC content of ‘China Pearl’ and ‘Carolina Gold’ represents approximately half of what was observed in ‘Contender’. Given the pedigrees of ‘China Pearl’ and ‘Carolina Gold’ (Fig. 1), this intermediate phenotype (between ‘Contender’ and mutants) could fit a model of additivity at a single locus, because ‘China Pearl’ is a progeny of ‘Contender’ and PI 134401 (peel color-deficient and likely double-recessive). ‘Carolina Gold’ also fits the model because it has an expected genetic contribution of 25% from Calanda San Miquel 2383 (peel color-deficient). Intermediate (pink) phenotypes in peach have been described in flower color mutations (Chaparro et al., 1995), but to our knowledge, this is the first time it has been observed in peel accumulation. Also of interest was the observation that although it represented a relatively minor contribution to overall ANC accumulation, cyanidin-3-rutinoside peel levels were statistically higher (6-fold) in ‘Contender’ and ‘China Pearl’ than they were in ‘Carolina Gold’.

Our results suggest that the mutation for ANC accumulation does not impact the accumulation of flavonoids or phenolic acids. Chlorogenic acid concentration, in particular, appears to be independent of ANC accumulation as evidenced by both the ANOVA and the correlation analysis. Significant differences among cultivars were observed in chlorogenic acid accumulation over multiple years and locations, but these differences did not fit the same pattern observed for ANC accumulation. ‘China Pearl’, NC 97-48, and NC Yellow contained the highest levels of chlorogenic acid (123 to 136 mg/100 g DW), ‘Carolina Gold’ contained the lowest (52 mg/100 g DW), and ‘Contender’ represented an intermediate phenotype (70 mg/100 g DW). ‘Carolina Gold’ also tended to have the lowest levels of overall flavan-3-ols, but these differences were not statistically significant. Quercetin-3-glucoside was highest in ANC mutant advanced breeding lines NC 97-48 and NC Yellow, but this observation should be tempered by the fact that these cultivars were only grown in a single location over 2 years.

The ANOVA detected significant genetic variation among the three cultivars replicated over years and locations for all compounds except for the flavon-3-ol monomer catechin and the procyanidin B4. This suggests that even within commonly grown peach cultivars, significant variation exists for individual constituents of the polyphenol profile. Additionally, because all of the constituents contribute to chemical measurements of antioxidant capacity to varying degrees, selecting on the basis of these measurements could result in multiple phytochemical phenotypes with wide-ranging differences in the amounts of the individual compounds. It might also be a plausible explanation for why some phenolic-rich foods have beneficial impacts on health but these benefits do not directly correlate with antioxidant capacity. The results of the study demonstrate the importance of multiple years and locations for adequately evaluating the phytochemical profile of tree fruit. With respect to chlorogenic acid, 28% of the phenotypic variance was explained by location (year = nonsignificant), whereas 35% of the phenotypic variation of ANC was explained by differences in years (location = nonsignificant). Analyzing fruit from the same environment over 2 years or from two locations in the same year would not have adequately accounted for variation associated with environment. Variation associated with location could be the result of differences in climate, soil fertility, temperature, cultivation, or (potentially) age differences of the trees at the two locations. Differences between locations and multiyear analysis provide for a more reliable estimation of the genetic variation. With respect to phenolic compounds, the percentage of observed genetic variance that we have observed (less than 40%) would suggest that previous non-replicated surveys of these compounds should be treated as preliminary in nature and need to be followed up by more thorough analysis at multiple locations over multiple years.

The detailed phytochemical analysis of peach described in this publication serves multiple purposes. It provides an initial set of peach reference standard cultivars that can be used to identify additional material with unique profiles. Including these commonly grown cultivars in future surveys will provide baseline measurements that will allow meaningful comparisons to unrelated material. Second, the analyses provide evidence for genetic variation among related material that can be used to further investigate the genetics of phytochemical accumulation in peach. It should be noted that most commercially grown peach cultivars in the United States are derived from a narrow genetic base with almost all cultivars derived from an original 1860s introduction called Chinese Cling (Scorza et al., 1985). The variation detected among cultivars and advanced North Carolina breeding lines in this investigation may be a function of the diverse genetic background of the North Carolina peach breeding program, in which germplasm not related to Chinese Cling was often used to broaden the genetic base of the program. The previously described Highlighter gene may itself be relatively unique among peach cultivars. We have noted significant genetic variation among these cultivars for carotenoids, cyanidin-3-glucoside, cyanidin-3-rutinoside, chlorogenic acid, and individual flavan-3-ol monomers and dimers. Inclusion of this material into traditional linkage or association mapping studies should provide insight into how peach accumulates and partitions individual phytochemical compounds in a tissue-specific manner. Finally, because these cultivars differ significantly in individual components with putative health benefits and are widely grown, inclusion of the material in medical studies may provide additional evidence as to the bioactivity of the individual constituents and provide a clearer picture of the optimal phytochemical profile of peach.