The Phenolic Profile of Sweet Cherry Fruits Influenced by Cultivar/Rootstock Combination

The influence of three cultivars (‘Carmen’, ‘Kordia’ and ‘Regina’) grafted on six rootstocks (Mahaleb, ‘Colt’, ‘Oblacinska’, ‘M × M 14′, ‘Gisela 5′ and ‘Gisela 6′) on the phenolic profile of sweet cherry fruits was studied during a two-year period. All the individual phenolic compounds were detected using high-pressure liquid chromatography with diode-array detection coupled with mass spectrometry (HPLC-DAD-MSn). In all the examined samples, 54 compounds were identified and divided into five phenolic classes: anthocyanins (4 compounds), flavonols (7), flavanols (11), flavanones (4), and hydroxycinnamic acids (28). Anthocyanins (58%) and hydroxycinnamic acids (31%) showed the greatest amounts in all the examined fruit samples. PCA analysis revealed that among the cultivars, ‘Kordia’ showed the highest phenolic content. Regarding rootstocks, the lowest values of the most important phenolic compounds were obtained in fruits from trees grafted onto the seedling rootstock Mahaleb. Among the clonal rootstocks, the vigorous ‘Colt’ and dwarf ‘Gisela 5′ promoted the highest values of the evaluated phenolic compounds in the cultivars ‘Kordia’ and ‘Carmen’, while the dwarf ‘Oblacinska’ and semi-vigorous ‘M × M 14′ induced the highest values in the cultivar ‘Regina’. By evaluating the influence of cultivars and rootstocks on the phenolic content in fruit, it has been proven that the cultivar has the most significant influence. However, the rootstock also influences the content of a large number of phenolic compounds. The selection of an adequate cultivar/rootstock combination can also be a powerful tool for improving the phenolic content in fruits, and consequently the nutritional value of sweet cherry fruits.


Introduction
In recent decades, sweet cherry production has seen a significant increase and it continues to spread worldwide with the main trends being to improve growing efficacy and ameliorate premium fruit quality [1]. The impact rate is mainly influenced by the production markets of the United States, Chile, and China [2]. With production of over 2.6 million tonnes per year, sweet cherry is in seventh position in the global production of temperate fruits [3].
Sweet cherry fruits are highly valued on the market due to their main sensory attributes such as firmness, sweetness, sourness, and colouration [4]. Fruit quality is not attributed only by appearance, textural and taste properties, but also by the chemical and nutritional compounds in the fruits [5]. Currently, consumer satisfaction is not only based on the primary pleasure of eating these delicious aromatic fruits, but also on their many human health benefits [6,7].
Phenolics are bioactive compounds naturally occurring in plant-derived foods. They are garnering increasing attention from researchers since they have been proven to play a significant role in improving health. [8,9]. The main dietary phenolic compounds play an Table 1. Phenolic compounds detected in sweet cherry fruit and used abbreviations.
The richest genotype with all the individual and total anthocyanins was 'Kordia', while the lowest levels were detected in the cultivar 'Carmen'. There were no significant differences in the content of all the individual anthocyanins between the cultivars 'Carmen' and 'Regina', while the total anthocyanins were significantly higher in 'Regina'.
All the individual and total anthocyanin contents were lowest in fruits picked from trees grafted onto 'Gisela 6 and Mahaleb rootstocks. On the other hand, the 'Colt' rootstock influenced the highest levels of most individual anthocyanin components and in the amount of the total anthocyanins.
Fruits harvested in the year 2020 had a significantly higher level of all individual anthocyanins compared with the samples harvested in the year 2021.
Among the rootstocks, Mahaleb influenced the lowest level of the total flavanols (33.4 ± 2.7 mg/kg FW), while the highest amount was induced by the 'Oblacinska' rootstock (48.3 ± 3.9 mg/kg FW).
The fruits of the 'Kordia' cultivar showed the highest amount of the total flavanol content (44.3 ± 3.5 mg/kg FW), while 'Carmen' showed the lowest content (40.3 ± 1.7 mg/kg FW). The contents of catechin, epicatechin, and procyanidin dimers were highest in cultivar 'Kordia', while the contents of epicatechin gallate and procyanidin trimers were highest in cultivar 'Regina'.
The amount of the total flavonols is affected by all the examined factors (cultivar, rootstock, year, and cultivar/rootstock interaction). The content of the total flavonols in samples from the combinations 'Regina'/Mahaleb (13.6 ± 1.6 mg/kg FW) and 'Carmen'/Mahaleb (15.1 ± 1.5 mg/kg FW) were half the level of the total flavonol content detected in the combinations 'Kordia'/'Gisela 5 (31.0 ± 2.7 mg/kg FW) and 'Kordia'/'Colt' (31.9 ± 3.8 mg/kg FW). 'Kordia' was the cultivar with the highest, while 'Carmen' was the cultivar with the lowest content of the total flavonols. The lowest level of the total flavonols in cherries was on the Mahaleb rootstock (16.2 ± 2.6 mg/kg FW), while significantly higher levels were found on four rootstocks: 'Gisela 5 , 'M × M 14 , 'Oblacinska', and 'Colt' (≥24.1 mg/kg FW).
The content of the main flavanol quercetin-3-rutinoside (rutin) in the sweet cherry fruit samples varied significantly depending on the cultivar/rootstock interaction. The highest amount of quercetin-3-rutinoside was found in sweet cherry samples of the combination 'Kordia'/'Colt' (20.69 ± 2.49 mg/kg FW), while only about a third of this content was found in the combinations 'Carmen'/Mahaleb (7.66 ± 1.37 mg/kg FW) and 'Regina'/Mahaleb (6.53 ± 0.5 mg/kg FW). Among the cultivars, 'Kordia' (17.37 ± 1.20 mg/kg FW) showed the highest level of quercetin-3-rutinoside, while there was no statistically significant difference in rutin content between 'Carmen' and 'Regina'. Among the rootstocks, 'Colt' induced the highest content of quercetin-3-rutinoside in sweet cherry fruits, while Mahaleb induced the lowest amount of this flavonol.
For some individual flavonol compounds, significant differences for some experimental factors were not found. For isorhamnetin-3-rutinoside and kaempferol-3-glucoside, differences were not significant between rootstocks and years, for kaempferol-3-rutinoside between cultivars, and for quercetin-3-galactoside between cultivars and years.
The cultivar was also a significant factor affecting the amount of both the total and individual flavanones. 'Kordia' influences the highest content of the total flavanones, which was significantly higher than in the cultivar 'Carmen'. Among the rootstocks, Mahaleb showed the lowest amount of total flavanones (3.20 ± 0.19 mg/kg FW). All other rootstocks influenced the significantly higher level of the total flavanones (≥4.23 mg/kg FW).

Content of Hydroxycinnamic Derivatives and Total Hydroxycinnamic Acids
Hydroxycinnamic acids (HCA) were the most diverse phenolic group found in all the studied samples of three sweet cherry cultivars grafted onto different rootstocks. The content of individual and total hydroxycinnamic acids is presented in Table 6. In the group of HCA, derivatives of caffeoylquinic, caffeic, and coumaroylquinic acid were the most expressed and represent more than 91% of the total hydroxycinnamic acids content. Derivatives of ferulic, sinapic, and dicaffeoylquinic acids were detected in levels lower than 1 mg/kg FW of fruit samples.
The cultivar had a significant influence on the content of the individual and total HCA derivatives. The influence of the rootstock was also significant on all HCA compounds, with the exception of the dicaffeoylquinic acids. The interaction of cultivar/rootstock did not show significant differences between the mean values for three compounds (caffeoylquinic acid derivatives, dicaffeoylquinic acids, and feruloylquinic acid derivatives). Similarly, differences between years were not significant for two compounds (dicaffeoylquinic acids and feruloylquinic acid derivatives).
The content of the total hydroxycinnamic acids in the studied samples varied from 75.9 ± 13.2 to 198.4 ± 17.8 mg/kg FW. The highest level was found in samples of the cultivar 'Carmen' grafted onto the dwarf rootstock 'Colt' (198.4 ± 17.8 mg/kg FW). The lowest level was found in the cultivar 'Regina' grafted onto the vigorous seedling rootstock Mahaleb (75.9 ± 13.2 mg/kg FW).

Principal Component Analysis
Since the content of phenolic compounds varied widely among cultivars, rootstocks, years, and their interaction, a principal component analysis (PCA) was performed in order to provide partial visualization of the dataset in a reduced dimension (Figure 1). PCA produced five PCs with eigenvalues greater than 1, explaining 90.7% of the total variability observed. On the basis of the principal component coefficients between the original variables and these five PCs, using an absolute value greater than 0.75 as a criterion for the significance, it was found that these values are present in the first three PCs. The first principal component contributed 46.6%, the second 24.8%, and the third 11.5% of the total variability obtained. Component 1 mainly explained the variability in all anthocyanins, quercetin-3-rutinoside, quercetin-3-glucoside, isorhamnetin-3-rutinoside, naringenin hexoside 2, taxifolin hexoside, procyanidin dimers, and derivatives of caffeoylquinic, caffeic, ferulic and p-coumaric acids. The second factor (PC2) correlated positively with catechin, taxifolin rutinoside, coumaroylquinic acid derivatives and sinapic acid derivatives and negatively with quercetin -7-glucoside-3-rutinoside and dicaffeoylquinic acids ( Figure 1A). Component 1 mainly explained the variability in all anthocyanins, quercetin-3-rutinoside, quercetin-3-glucoside, isorhamnetin-3-rutinoside, naringenin hexoside 2, taxifolin hexoside, procyanidin dimers, and derivatives of caffeoylquinic, caffeic, ferulic and p-coumaric acids. The second factor (PC2) correlated positively with catechin, taxifolin rutinoside, coumaroylquinic acid derivatives and sinapic acid derivatives and negatively with quercetin-7-glucoside-3-rutinoside and dicaffeoylquinic acids ( Figure 1A).
The distribution of cultivar/rootstock combinations along the PC1/PC2 scatter plot ( Figure 1B) showed a split into three main groups. The cultivars 'Regina' and 'Carmen' were negatively linked to the PC1, whereas the cultivar 'Kordia' had positive scores for the same component. This arrangement confirms the results of ANOVA, which determined that the cultivar 'Kordia' contains significantly more phenolic compounds that are significant and positively correlated within PC1. Discrimination between the cultivars 'Regina' and 'Carmen' was highlighted on PC2. 'Carmen' is located on the positive, and 'Regina' is on the negative side of PC2. Furthermore, the distribution of samples within these three main groups indicates a pronounced cultivar/rootstock interaction. In all three cultivars, the worst results in terms of the content of phenolic compounds were obtained from the seedling rootstock Mahaleb. Among the clonal rootstocks, the vigorous 'Colt' and dwarf 'Gisela 5′ promoted the highest values of the evaluated compounds in the cultivars 'Kordia' and 'Carmen', while the dwarf 'Oblacinska' and semi-vigorous 'M × M 14′ induced the highest values in the cultivar 'Regina'. The distribution of cultivar/rootstock combinations along the PC1/PC2 scatter plot ( Figure 1B) showed a split into three main groups. The cultivars 'Regina' and 'Carmen' were negatively linked to the PC1, whereas the cultivar 'Kordia' had positive scores for the same component. This arrangement confirms the results of ANOVA, which determined that the cultivar 'Kordia' contains significantly more phenolic compounds that are significant and positively correlated within PC1. Discrimination between the cultivars 'Regina' and 'Carmen' was highlighted on PC2. 'Carmen' is located on the positive, and 'Regina' is on the negative side of PC2. Furthermore, the distribution of samples within these three main groups indicates a pronounced cultivar/rootstock interaction. In all three cultivars, the worst results in terms of the content of phenolic compounds were obtained from the seedling rootstock Mahaleb. Among the clonal rootstocks, the vigorous 'Colt' and dwarf 'Gisela 5 promoted the highest values of the evaluated compounds in the cultivars 'Kordia' and 'Carmen', while the dwarf 'Oblacinska' and semi-vigorous 'M × M 14 induced the highest values in the cultivar 'Regina'.

Discussion
In the three sweet cherry cultivars grafted onto different rootstocks, 54 individual phenolic compounds were detected and quantified. They were classified into five groups: anthocyanins (4 compounds), flavonols (7), flavanols (11), flavanones (4), and hydroxycinnamic acids (28). Anthocyanins accounted for the highest percentage (62.7%) of all analysed phenolics, while hydroxycinnamic accounted for 26.1% of the total phenolic content. Flavanols, flavonols, and flavanones corresponded to 7.2, 3.2, and 0.7% of total phenolics, respectively. The same components were identified in all the samples of cultivar/rootstock combinations studied. Cyanidin-3-rutinoside was the dominant component in the phenolic profile of all the samples studied (on average 43.87% of the total phenolics).
Anthocyanins and hydroxycinnamic acids represent the main phenolic compounds in sweet cherries, as reported previously [27,28]. In all the analysed fruit samples from different cultivars and rootstocks, anthocyanins are the most dominant components. On the other hand, the hydroxycinnamic acids (HCA) represent the most numerous phenolic compounds. This agrees with the previous findings [14,29]. Martini et al. [14] reported 86 tentatively identified phenolics in six different sweet cherry cultivars, of which 40 belong to the class of hydroxycinnamic acids. Gonçalves et al. [29] reported that the phenolic profile of 23 Portuguese sweet cherry cultivars consists of 46 phenolic compounds: 19 hydroxycinnamic acids, 2 hydroxybenzoic acids, 13 flavonols, 5 flavan-3ols, 2 flavanones, 1 flavanonol and 4 anthocyanins.
The cultivars significantly influenced the total content of all phenolic groups, as well as the content of almost all the detected individual phenolic compounds. The exceptions are only two minor flavanol compounds (kaempferol-3-rutinoside and quercetin-3-galactoside). Our results confirmed the previous finding of a strong genotype influence on the phenolic profile of sweet cherry fruits [29][30][31]. Among the studied cultivars, the highest content of most phenolic compounds was found in 'Kordia', then in 'Regina', while the lowest content was found in 'Carmen'. The higher phenolic content in cultivar 'Kordia' compared to 'Regina' is in agreement with previous research [25].
Fruits of the cultivar 'Kordia' were the richest in major detected anthocyanins, and the values of the total and individual anthocyanins were similar to those found by Milinovic et al. [25] The contents of the total and individual athocyanins were higher than in the 'Black Star', 'Sweetheart', 'Sunburst', 'Summit', and 'Van' sweet cherry cultivars reported by other authors [34,35]. Mozetič et al. reported a higher amount of cyanidin-3-rutinoside during the late phases of maturation [36], which could be explained by different climatic conditions. Cyanidin-3-rutinoside and cyanidin-3-glucoside were the main identified anthocyanins, while the derivatives of caffeoylquinic, caffeic, and coumaroylquinic acids were the leading components among the hydroxycinnamic acids in sweet cherry fruits. The rootstock significantly influenced the amount of detected individual and total anthocyanins. Our results confirm previous findings that anthocyanin content was largely affected by the rootstocks in the cultivars 'Lapins' [16,24] and '0900 Ziraat' [26].
The influence of the year on some phenolic compounds confirmed the results of the previous research [25,37]. During the first year of examination, the level of anthocyanins and flavanols was considerably higher, while the amount of total flavanones and hydroxycinnamic acids were higher during the second year. The significant effects of the harvest year on phenolic content can be explained by different meteorological factors, such as temperature, solar radiation, and the amount of rain (Table S4). The more than doubled content of all the individual and total anthocyanins detected in the year 2020 can be explained as a response to the weather conditions during the first year of examination. During June 2020, the total precipitation was considerably higher (158 mm) than in 2021 (only 34 mm). Also, the average monthly temperatures during the maturation period (May and June) were considerably higher in the second year of study (2021), especially in June. This was the same case for the time of insolation. The water stress during the ripening period is reported to have a positive effect on the biosynthesis of phenolic compounds in the fruits of apricot [38] and peach [39].
Identification concerning the hydroxycinnamic acid profile is in accordance with previous research, which identifies caffeoylquinic acid derivatives, caffeic acid derivatives, and coumaroylquinic acid derivatives as the major hydroxycinnamic components in cherries [40]. According to the previous results, the content of enumerated components of hydroxycinnamic acids varies greatly depending on the cultivar and rootstock [23,25,41]. The dominance of caffeoylquinic acid derivatives in all the samples was the same as in the plum cultivar 'Čačanska lepotica' grafted onto five different rootstocks [42].
The most prominent flavonols detected in our study were quercetin-3-rutinoside and quarcetin-7-glucoside-3-rutinoside. Our values of flavonol glycosides are in accordance with previous studies [28]. It was also found earlier that quercetin-3-rutinoside is the most dominant compound of flavonols and that it is significantly influenced by the rootstock [24].
The detected individual flavanols were reported to be the same as components detected in six ancient sweet cherry cultivars [12]. The obtained results of epicatechin content were in agreement with the report of Kelebek et al. [43] concerning different cultivars, while the content of catechin was half of what was detected in our report. Mikulic-Petkovsek et al. detected higher amounts of procyanidin dimers and procyanidin trimers in wild Prunus species [44] than the values found in our study. It is worth mentioning that the influence of cultivar/rootstock interaction on procyanidin dimers and trimers has not been reported so far.
Flavanones are determined as the smaller group of the phenolic profile representing 1% of the total phenolic compounds. They included four compounds that are in all the examined samples. Two of them are naringenin hexosides, which confirms the results of Gonçalves et al. [29]. The same low content of naringenins was found in raw and frozen sweet cherry fruits, as well as in cherry juice [45].
By PCA, out of the 28 phenolic compounds detected in sweet cherry fruits, 22 showed strong correlations with PCs, namely 14 with PC1, 7 with PC2, and one with PC3 which indicates the high discriminating power of these compounds. These results agree with previous reports, which indicate that it is possible to differentiate between sweet cherry genotypes using PCA based on their phenolic constituents [14,30,34,37,46].
The classification of different cultivar/rootstock combinations into three main groups is primarily a function of the genetic potential of the cultivar. This is in agreement with the results of Radović et al. [22] who studied the chemical composition of three plum cultivars grafted onto four rootstocks and concluded that the chemical composition of the fruits was more cultivar-than rootstock-dependent. In general, the highest phenolic content was found in the cultivar 'Kordia', followed by 'Carmen' and 'Regina'.
The highest variation in phenolic amount due to the rootstock was found in the cultivar 'Kordia', slightly less in the cultivar 'Regina', while the cultivar 'Carmen' showed the greatest stability concerning the phenolic content on different rootstocks. In all three cultivars, the lowest content of phenolic compounds was obtained in fruits from the vigorous seedling rootstock Mahaleb. Among the clonal rootstocks, there was not clear influence of vigour on the content of phenolic compounds in cherry fruit. In the cultivars 'Kordia' and 'Carmen', semi-vigorous 'Colt' and dwarf 'Gisela 5 promoted the highest values of phenolic compounds, while in the cultivar 'Regina' the semi-vigorous 'M × M 14 and the dwarf 'Oblacinska' induced the highest values. The results obtained in our research show that rootstock vigour is not linearly correlated with the content of phenolic compounds in the cherry fruit, which is consistent with data shown by Remorini et al. [47] who reported that the phenolic levels in peaches were influenced by the rootstock, but its vigour did not affect some secondary metabolites. Similarly, Milošević et al. [48] cited that one dwarf (low vigorous) and one vigorous rootstock promoted the best values of the evaluated compounds and antioxidant capacity in the sour cherry cultivar 'Šumadinka'.
Jakobek et al. [27] considered that higher phenolic compound content in sweet cherry fruits probably comes from heterogenic grafting combinations. The differences in the concentration of bioactive compounds in sweet cherry fruit are explained by the effects of the rootstocks on scion physiology [49]. Karakaya et al. [26] found that incompatibility problems can affect the content of individual phenolic compounds in sweet cherry fruits.

Plant Material
The sweet cherry fruits were collected in two consecutive years (2020 and 2021) from the 7-and 8-year-old experimental plantation located at the Fruit Growing Centre "Radmilovac" of the Faculty of Agriculture in Belgrade (44 •

Extraction and Analysis of Phenolic Compounds
Extraction and identification of phenolic compounds in cherries were performed as previously described by Mikulic-Petkovsek et al. [44]. Each cherry sample for extraction of individual phenolic components was weighed with 4 g of mixed fresh fruit, to which 10 mL of extraction solution (methanol/water/formic acid = 70/27/3, v/v/v) was added. Then, the extraction of phenolics was carried out in an ultrasonic bath for 60 min. After that, all extracts were centrifuged at 9000 rpm and the supernatant was filtered through PTFE filters (Macheery Nagel) into vials. Thermo Dionex HPLC system (Thermo Scientific, San Jose, USA) was used in conjunction with a diode array detector (DAD) for the determination of phenolic compounds. The analytical HPLC conditions were the same as previously described by Mikulic-Petkovsek et al. [50]. Phenomenex column (150 × 4.6 mm i.d., 3 µm, Gemini C18) heated to 25 • C was used for separation of phenolic compounds. The studied extract was injected at 20 µL, and the flow rate of the mobile phases was 0.6 mL per minute. The mobile phases were aqueous 0.1% formic acid and 3% acetonitrile in double-distilled water (A), and 0.1% formic acid and 3% distilled water dissolved in acetonitrile (B). Mixing of the mobile phases was performed according to the gradient method described in the study by Mikulic-Petkovsek et al. [51].

Statistical Analysis
All statistical analyses were performed using the "Statistica" (Stat Soft software Inc., Tulsa, OK, USA) program package. Three-way ANOVA was used for the analysis of the effect of cultivar, rootstocks, year, and cultivar/rootstock interaction. Differences between the mean values were estimated with Tukey's test (p < 0.05). Multivariate statistical analysis was conducted in order to interpret the differences between the phenolic compounds in fruits of the cultivars on different rootstocks. For every compound, mean values and standard errors are presented (mean ± SE) and statistical differences among treatments are denoted by different letters.

Conclusions
Evaluation of the results of all the identified individual phenolic compounds in the three sweet cherry cultivars grafted onto six rootstocks revealed that the phenolic content depends mainly on the cultivar, but is also modified by the rootstock, the interaction between the cultivar/rootstock, and the weather conditions during the study years. The dominant phenolic components were cyanidin-3-rutinoside, caffeoylquinic acid derivatives, and caffeic acid derivatives. The highest amounts of the most phenolic compounds were found in fruits of the cultivar 'Kordia'. The significant variability of phenolic compounds was also caused by the rootstock. The seedling rootstock Mahaleb had the lowest content of phenolic compounds. Among the clonal rootstocks, semi-vigorous 'Colt' and dwarf 'Gisela 5 induced the highest values of major phenolic compounds in the cultivars 'Kordia' and 'Carmen', while the dwarfing rootstock 'Oblacinska', and semi-dwarfing rootstock 'M × M 14 induced the highest values in the cultivar 'Regina'. As for the phenolic profile, the best-evaluated cultivar and rootstock combination was 'Kordia' grafted onto the 'Colt' rootstock. In summary, we can conclude that the phenolic content of sweet cherries, as an important element of fruit quality, can be improved not only by the choice of cultivars and rootstocks, but also by the selection of the best combination of cultivar and rootstock. Thus, the selection of an appropriate combination of cultivar and rootstock is an effective tool for improving the nutritional value of sweet cherries.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/plants12010103/s1, Table S1: All 54 identified phenolic compounds in sweet cherry fruits and their abbreviations; Table S2: Content of all individual and total flavanols in fruits of sweet cherry cultivars on different rootstocks; Table S3: Content of all individual and total hydroxycinnamic acids in fruits of sweet cherry cultivars on different rootstocks; Table S4: Meteorological conditions during the first six months in the years of study (2020-2021).